rights / license: research collection in copyright - …i abstract fire blight is a bacterial...
TRANSCRIPT
Research Collection
Doctoral Thesis
Molecular breeding for fire blight resistance in apple (Malusspp.)
Author(s): Le Roux, Pierre-Marie F.
Publication Date: 2011
Permanent Link: https://doi.org/10.3929/ethz-a-007303852
Rights / License: In Copyright - Non-Commercial Use Permitted
This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.
ETH Library
DISS. ETH Nr. 20055
MOLECULAR BREEDING FOR FIRE BLIGHT RESISTANCE IN
APPLE (MALUS SPP.)
A dissertation submitted to
ETH ZURICH
For the degree of
Doctor of Sciences
Presented by
PIERRE-MARIE F. LE ROUX
Dipl. Ing. Agr. ENSAR, Rennes, France
Born October 11th
, 1982
Citizen of France
Accepted on the recommendation of
Prof. Dr. Cesare Gessler, examiner
Prof. Dr. Bruce McDonald, co-examiner
Prof. Dr. Magda-Viola Hanke, co-examiner
Dr. Andrea Patocchi, co-examiner
2011
i
Abstract
Fire blight is a bacterial disease caused by Erwinia amylovora that affects a wide variety of
Rosaceae, and primarily Spiraeoideae. Its damage to the production of the cultivated apple
(Malus × domestica Borkhausen) is a major concern since no existing control option has
proven to be completely effective and durable. Selecting pre-breeding genotypes and
subsequently apple cultivars resistant to fire blight by phenotypic evaluation is feasible but it
is expensive, labor-intensive and time-consuming. Selection efficiency may be improved by
the application of molecular markers. For this, a good understanding of the genetic basis of
fire blight resistance in Malus spp., and especially in apple cultivars, is necessary.
Chapter 1 gives an overview of classical breeding, quantitative trait locus (QTL) and gene
mapping, and marker-assisted breeding in apple for diseases resistance, with a focus on fire
blight resistance. The disease, its causative agent E. amylovora and the various control
options are also presented.
Chapter 2 describes a QTL analysis conducted on a F1 progeny of 118 individuals derived
from the cross ‘Florina’ × ‘Nova Easygro’ using two parental linkage maps based on simple
sequence repeats (SSR) and amplified fragment length polymorphism (AFLP) markers. Both
parental cultivars are considered resistant to fire blight. Two significant QTLs for resistance
to E. amylovora strain CFBP 1430 were detected in ‘Florina’; one QTL on linkage group
(LG) 10 explained 15.3 % of the phenotypic variation, and a second QTL on LG 5 explained
10.1 % of the phenotypic variation. Genotyping the plants of ‘Florina’ pedigree with the SSR
markers flanking the QTLs showed that the QTLs on LGs 10 and 5 were inherited from
‘Jonathan’ and ‘Starking’ (a ‘Red Delicious’ sport mutation), respectively. Only putative
QTLs were detected in ‘Nova Easygro’, on LGs 5 and 9.
Chapter 3 reports a QTL analysis performed on a F1 progeny of 92 individuals derived from
a cross between the cultivars ‘Idared’ (fire blight susceptible) and ‘Rewena’ (fire blight
resistant). Two framework parental linkage maps were constructed with diversity arrays
technology (DArT) and SSR markers, covering 44.2 % and 37.1 % of ‘Idared’ and ‘Rewena’
genomes, respectively. One putative QTL for resistance to E. amylovora strain Ea 222 was
detected on the LG 7 of ‘Idared’. Improving genome coverage as well as marker density and
distribution on the linkage maps of ‘Rewena’ and ‘Idared’ may allow to detect additional
QTLs. Together, Chapters 2 and 3 suggest that the resistance to fire blight of the cultivars
ii
‘Florina’, ‘Nova Easygro’ and ‘Rewena’ is quantitative and has an oligo- or polygenic basis,
involving mostly minor-effect additive QTLs and possibly epistatic QTLs.
Quantitative trait loci conferring high levels of fire blight resistance have been identified in
wild and ornamental apple genotypes, such as Malus × robusta ‘Robusta 5’, ‘Evereste’ and
Malus × floribunda clone 821. However, no advanced selections or even pre-breeding
genotypes carrying these QTLs and displaying satisfactory fruit quality and agronomic
performance are yet available. Recently, a breeding technology based on the early flowering
BpMADS4-transgenic line T1190 was developed to accelerate the introgression of genes or
QTLs from wild or ornamental apples into the domesticated apple. Chapter 4 reports the
genetic mapping on LG 4 of the T-DNA integration site in the BpMADS4-transgenic apple
line T1190.
Chapter 5 describes the use of the BpMADS4-transgenic line T1190 to accelerate the
introgression of the highly efficacious fire blight resistance locus Fb_E identified on LG 12 of
the ornamental apple cultivar ‘Evereste’ into a domesticated apple background. The strong
phenotypic effect of the Fb_E locus was confirmed by artificially inoculating a F1 progeny
T1190 × ‘Evereste’ with E. amylovora strain CFBP 1430. Twenty four BC’1 seedlings
carrying the BpMADS4 transgene and the Fb_E locus could be retrieved owing to the
independent segregation of the T-DNA and the locus Fb_E. Among them, two early flowering
BC’1 seedlings estimated to carry less than 15 % of the genome of ‘Evereste’ were identified
using a background selection based on SSR markers regularly distributed over the apple
genome. These two early flowering BC’1 seedlings will be used to further introgress the fire
blight resistance locus Fb_E in pre-breeding genotypes carrying as little as possible genome
from the resistance donor ‘Evereste’.
Finally, Chapter 6 gives a general discussion. Results of QTL analysis in the F1 progenies
‘Florina’ × ‘Nova Easygro’ and ‘Idared’ × ‘Rewena’ are discussed in the context of apple
breeding for fire blight resistance. Alternative QTL mapping strategies in the breeder’s plant
material are examined. Moreover, the potential of the high-speed breeding technology based
on the early flowering BpMADS4-transgenic line T1190 is discussed in a perspective of
introgression of disease resistance genes and QTLs into the domesticated apple.
iii
Zusammenfassung
Feuerbrand ist eine durch Erwinia amylovora verursachte bakterielle Krankheit, die viele
Arten der Rosaceae und besonders der Spiraeoideae befällt. Der durch die Krankheit
verursachte Ernteverlust im Apfelanbau (Malus × domestica Borkhausen) ist ein wesentliches
Problem der Apfelproduktion, da die Krankheit nicht dauerhaft und sicher kontrolliert werden
kann. Die phänotypische Selektion von feuerbrandresistenten Genotypen und Apfelsorten ist
möglich aber teuer sowie arbeits- und zeitaufwändig. Der Aufwand der Selektion könnte
mittels molekularer Marker reduziert werden. Dafür sind Kenntnisse über die genetischen
Grundlagen der Feuerbrandresistenz in Malus spp. und besonders in Apfelkultursorten
wichtig.
Kapitel 1 gibt einen Überblick über die klassische Züchtung, ‘quantitative trait loci’ (QTL)
und die Genkartierung sowie die marker-gestützte Selektion in der Apfelzüchtung für
Krankheitsresistenzen mit einem Schwerpunkt auf der Feuerbrandresistenz. Erwinia
amylovora, der Erreger des Feuerbrands, und verschiedene Möglichkeiten seiner Bekämpfung
werden ebenfalls präsentiert.
Im Kapitel 2 wird die QTL-Analyse mit einer F1 Nachkommenschaft von 118 Individuen der
Kreuzung ‘Florina’ × ‘Nova Easygro’ anhand zweier bereits vorhandener parentaler
genotypischer Karten beschrieben, die auf ‘simple sequence repeats’ (SSRs) und ‘amplified
fragment length polymorphism’-Markern (AFLP) basieren. Beide Eltern waren
feuerbrandresistent. Zwei signifikante QTLs für Resistenz gegen E. amylovora Stamm CFBP
1430 wurden in ‘Florina’ entdeckt; ein QTL lag auf Kopplungsgruppe (LG) 10, der 15.3 %
der phänotypischen Variation erklärt, und ein zweiter QTL auf LG 5, der 10.1 % der
phänotypischen Variation erklärt. Die Genotypisierung der Pflanzen aus dem Stammbaum
von ‘Florina’ mit SSR-Markern, die neben den QTLs lagen, zeigte, dass die QTLs auf LG 10
und 5 von ‘Jonathan’ bzw. ‘Starking’ (eine Mutation von ‘Red Delicious’) vererbt wurden. In
‘Nova Easygro’ konnten nur nicht-signifikante QTLs auf LG 5 und 9 entdeckt werden.
Im Kapitel 3 wird eine QTL-Analyse über eine F1-Nachkommenschaft mit 92 Individuen
einer Kreuzung der Apfelsorten ‘Idared’ (feuerbrandanfällig) und ‘Rewena’
(feuerbrandresistent) beschrieben. Zwei genotypische, minimale Karten wurden mit ‘diversity
arrays technology’ (DArT) und SSR-Markern angefertigt, die 44.2 % bzw. 37.1 % des
Genoms von ‘Idared’ bzw. ‘Rewena’ abdecken. Ein möglicher QTL für Resistenz gegen E.
amylovora Stamm Ea 222 wurde auf LG 7 von ‘Idared’ gefunden. Durch Erhöhung der
iv
Deckung des Genoms mit Markern sowie der Markerdichte und -verteilung auf den
Kopplungsgruppen von ‘Rewena’ und ‘Idared’ könnten weitere QTLs entdeckt werden.
Zusammenfassend beschreiben Kapitel 2 und 3, dass die Resistenz gegen Feuerbrand von
‘Florina’, ‘Nova Easygro’ und ‘Rewena’ quantitativ und mit einer oligo- oder polygenischen
Grundlage ist, was meistens zu schwachen oder möglicherweise epistatischen QTLs führt.
QTLs, die starke Feuerbrandresistenzen beschreiben, wurden in Wild- und
Zierapfelgenotypen wie Malus × robusta ‘Robusta 5’, ‘Evereste’ und Malus × floribunda
Klon 821 gefunden. Dennoch ist weder eine fortgeschrittene Zuchtnummer und noch eine
Zwischenstufe in der Züchtung vorhanden, welcher diese QTLs trägt und gleichzeitig hohe
Frucht- und Anbauqualität zeigt. Vor kurzem wurde basierend auf der transgenen Linie
BpMADS4 T1190 (‘Frühe Blüte’) eine Züchtungstechnik entwickelt, um die erwähnten QTLs
in den domestizierten Apfel einzuführen. Kapitel 4 behandelt die genetische Kartierung der
T-DNA-Integrationsstelle von BpMADS4 in Apfel auf LG 4.
Kapitel 5 beschreibt die Nutzung der BpMADS4-transgenen Linie T1190 für die Einführung
des hoch wirksamen Feuerbrandresistenzlokus Fb_E, der auf LG 12 des Zierapfels ‘Evereste’
gefunden wurde, in den domestizierten Apfel. Der starke phänotypische Effekt des Fb_E-
Lokus wurde an einer F1 Nachkommenschaft von T1190 × ‘Evereste’ mittels künstlicher
Inokulation mit E. amylovora Stamm CFBP 1430 bestätigt. Vierundzwanzig BC’1 Sämlinge,
die infolge der unabhängigen Segregation beider Loci das BpMDAS4-Transgen sowie den
Fb_E-Lokus enthielten, konnten zurückgewonnen werden. Mittels SSR-Markern, die
gleichmässig über das Apfelgenom verteilt waren, konnten zwei frühblühende BC’1 Sämlinge
identifiziert werden, die weniger als 15 % des Genoms von ‘Evereste’ trugen. Diese beiden
frühblühenden BC’1 Sämlinge werden in Folge weiter für die Einführung des
Feuerbrandresistenzlokus Fb_E in ‘pre-breeding genotypes’ verwendet, die so wenig wie
möglich von dem Genom der Resistenz-Donors ‘Evereste’ enthalten sollen.
Schlussendlich gibt es in Kapitel 6 eine generelle Diskussion. Die Resultate der QTL-
Analyse in den F1-Nachkommenschaften von ‘Florina’ × ‘Nova Easygro’ und ‘Idared’ ×
‘Rewena’ werden im Zusammenhang mit der Züchtung für Feuerbrandresistenz diskutiert.
Alternative QTL-Kartierungsstrategien im Pflanzenmaterial des Züchters werden behandelt.
Zusätzlich wird das Potential der Hochgeschwindigkeitszüchtung mittels der frühblühenden
transgenen Linie BpMADS4 T1190 im Hinblick auf die Einführung von Resistenzgenen und
QTLs in den domestizierten Apfel besprochen.
v
Dedicated to my parents and my brother
vi
Table of contents
Abstract ................................................................................................................. i
Zusammenfassung .............................................................................................. iii
Table of contents ................................................................................................. vi
List of Tables ........................................................................................................ x
List of Figures ..................................................................................................... xi
Chapter 1: General introduction ....................................................................... 1
1.1. Introduction to the domesticated apple Malus × domestica (Borkhausen) .................. 2
1.1.1. The genus Malus: a complex nomenclature ......................................................... 2
1.1.2. History of the domesticated apple ........................................................................ 3
1.1.3. Economic importance ........................................................................................... 4
1.1.4. Breeding program organization ............................................................................ 4
1.2. Application of molecular markers to apple breeding ................................................... 8
1.2.1. Genomics, genetics and marker-assisted breeding: definitions ............................ 8
1.2.2. Development of molecular marker–phenotypic trait associations ..................... 10
1.2.2.1. Gene mapping ...................................................................................... 10
1.2.2.2. QTL mapping on a single progeny ...................................................... 11
1.2.2.2.1. Linkage map construction ..................................................... 12
1.2.2.2.2. Statistical and QTL analysis ................................................. 14
1.2.2.3. Pedigree based analysis ....................................................................... 16
1.2.2.4. Candidate gene approach ..................................................................... 17
1.2.3. Implementation of marker-assisted selection ..................................................... 18
1.2.3.1. Challenges for application of molecular markers to breeding ............. 18
1.2.3.2. Marker-assisted parental selection ....................................................... 20
1.2.3.3. Marker-assisted seedling selection ...................................................... 21
1.3. Fire blight, a major apple disease .................................................................................. 23
1.3.1. History of the disease ........................................................................................ 23
1.3.2. Dispersal and host range ..................................................................................... 23
1.3.3. Disease cycle ...................................................................................................... 24
1.3.4. Epidemiology ..................................................................................................... 26
1.3.5. The pathogen, Erwinia amylovora ..................................................................... 28
1.3.5.1. Taxonomy ............................................................................................ 28
1.3.5.2. Diversity .............................................................................................. 28
1.3.6. Non-genetic measures of fire blight management .............................................. 29
1.4. Genetics and breeding of fire blight resistance in apple .............................................. 32
1.4.1. Context and challenges of fire blight resistance breeding .................................. 32
1.4.1.1. Fire blight resistance as breeding objective ......................................... 32
1.4.1.2. Challenges in fire blight resistance breeding ....................................... 33
vii
1.4.2. Conventional breeding for fire blight resistance ................................................ 35
1.4.2.1. Sourcing fire blight resistance in Malus germplasm ........................... 35
1.4.2.2. Apple scion breeding for fire blight resistance .................................... 36
1.4.3. Development of molecular markers: towards MAB for fire blight resistance ... 37
1.4.3.1. QTL mapping and MAS ...................................................................... 37
1.4.3.2. Candidate gene approaches.................................................................. 39
1.4.4. Genetic engineering for fire blight resistance in Malus spp. .............................. 40
1.4.5. Fast breeding strategy: status and prospects ....................................................... 42
1.4.5.1. Juvenile phase and transition to flowering: definitions ....................... 42
1.4.5.2. Shortening the juvenile phase .............................................................. 43
1.4.5.2.1. Agro-technical approaches ................................................... 43
1.4.5.2.2. Transgenic approaches ......................................................... 44
1.4.5.3. Implementation of fast breeding approaches through reduction of the
juvenile phase ........................................................................................................................... 45
1.5. Scope of this Thesis ......................................................................................................... 48
Chapter 2: Mapping of quantitative trait loci for fire blight resistance in the
apple cultivars ‘Florina’ and ‘Nova Easygro’ ................................................ 51
2.1. Abstract ........................................................................................................................... 52
2.2. Introduction ..................................................................................................................... 53
2.3. Material and methods ..................................................................................................... 55
2.3.1. Plant material and evaluation of fire blight resistance ....................................... 55
2.3.2. DNA extraction .................................................................................................. 55
2.3.3. SSR selection ...................................................................................................... 55
2.3.4. Gel and capillary electrophoresis of SSR and AFLP markers ........................... 56
2.3.5. Linkage mapping ................................................................................................ 56
2.3.6. Statistical and QTL analysis ............................................................................... 57
2.3.7. Tracing the significant FLO 10 QTL in the pedigree of ‘Florina’ ..................... 57
2.4. Results .............................................................................................................................. 58
2.4.1. Phenotypic evaluation ........................................................................................ 58
2.4.2. Construction of parental linkage maps ............................................................... 59
2.4.3. Statistical and QTL analysis ............................................................................... 63
2.4.4. Analysis of the pedigree of ‘Florina’.................................................................. 67
2.5. Discussion ......................................................................................................................... 69
2.5.1. Phenotypic screening of the ‘Florina’ × ‘Nova Easygro’ progeny..................... 69
2.5.2. Linkage maps of ‘Florina’ and ‘Nova Easygro’ ................................................. 70
2.5.3. Identification of two new genomic regions involved in fire blight resistance ... 70
2.5.4. Accuracy of QTL mapping in ‘Florina’ × ‘Nova Easygro’ ................................ 72
2.5.5. LG 10 and LG 5 QTLs of ‘Florina’ do not map in homoeologous genomic
regions ...................................................................................................................................... 73
2.5.6. A putative disease resistance cluster on apple linkage group 10 ........................ 73
2.6. Acknowledgements .......................................................................................................... 74
2.7. Addendum ........................................................................................................................ 75
viii
Chapter 3: Quantitative trait loci analysis of fire blight resistance in an
apple F1 progeny ‘Idared’ × ‘Rewena’ ........................................................... 77
3.1. Abstract ............................................................................................................................ 78
3.2. Introduction ..................................................................................................................... 79
3.3. Material and methods ..................................................................................................... 82
3.3.1. Plant material and evaluation of fire blight resistance ....................................... 82
3.3.2. Statistical analysis .............................................................................................. 82
3.3.3. SSR and DArT markers genotyping ................................................................... 83
3.3.4. Genetic map construction ................................................................................... 83
3.3.5. QTL analysis ...................................................................................................... 84
3.4. Results .............................................................................................................................. 85
3.4.1. Evaluation of fire blight resistance and statistical analysis ................................ 85
3.4.2. Linkage map construction .................................................................................. 85
3.4.3. Statistical analysis and QTL detection ............................................................... 93
3.5. Discussion ......................................................................................................................... 94
3.5.1. Impact of DArT markers on linkage map construction ...................................... 94
3.5.2. QTL analysis in ‘Idared’ and ‘Rewena’ ............................................................. 96
3.5.3. Perspectives of molecular breeding towards fire blight resistance using apple
cultivars like ‘Rewena’ ............................................................................................................. 98
3.6. Acknowledgements ........................................................................................................ 100
Chapter 4: Genetic mapping of the T-DNA integration site in the
BpMADS4-transgenic apple line T1190 ........................................................ 103
4.1. Abstract .......................................................................................................................... 104
4.2. Introduction ................................................................................................................... 105
4.3. Materials and methods.................................................................................................. 106
4.4. Results ............................................................................................................................ 107
4.5. Discussion ....................................................................................................................... 108
4.5.1. Relevance of the mapping position of the T-DNA integration site for accelerated
marker-assisted introgression/gene pyramiding ..................................................................... 108
4.5.2. Importance of solid trait-locus-marker associations for marker-assisted gene
pyramiding ............................................................................................................................. 110
4.6. Acknowledgements ........................................................................................................ 111
Chapter 5: Use of a transgenic early flowering approach in apple (Malus ×
domestica Borkh.) to introgress the major quantitative trait locus for fire
blight resistance of the apple genotype ‘Evereste’ ....................................... 113
5.1. Abstract .......................................................................................................................... 114
5.2. Introduction ................................................................................................................... 115
5.3. Materials and methods.................................................................................................. 118
5.3.1. Plant material and introgression scheme .......................................................... 118
5.3.2. Experimental conditions ................................................................................... 120
5.3.3. Genotypic foreground selection of F1 and BC’1 progenies ............................. 120
5.3.4. Phenotypic screening of F1 progeny for fire blight resistance ......................... 121
5.3.5. Genotypic background selection of transgenic BC’1 offspring carrying the fire
blight resistance locus Fb_E ................................................................................................... 123
ix
5.4. Results ............................................................................................................................ 125
5.4.1. Achievement of the first two breeding cycles of the introgression scheme ..... 125
5.4.2. Inheritance of the fire blight resistance locus Fb_E and the BpMADS4 transgene
in F1 and BC’1 progenies ....................................................................................................... 127
5.4.3. Evaluation of the fire blight resistance level in the F1 progeny T1190 ×
‘Evereste’ ............................................................................................................................... 127
5.4.4. Background selection of transgenic BC’1 offspring carrying the fire blight
resistance locus Fb_E ............................................................................................................. 128
5.5. Discussion ....................................................................................................................... 131
5.5.1. Acceleration of the first two breeding cycles of the introgression of the fire
blight resistance locus Fb_E from ‘Evereste’ by inheritance of the BpMADS4 transgene .... 131
5.5.2. Effect of the Fb_E locus from the apple genotype ‘Evereste’ on fire blight
resistance in a T1190 × ‘Evereste’ F1 offspring .................................................................... 133
5.5.3. Efficiency of background selection to assess the proportion of ‘Evereste’
genome in the BC’1 offspring ................................................................................................ 134
5.5.4. Perspectives of enhancing the introgression strategy based on the transgenic
early flowering line T1190 ..................................................................................................... 137
5.6. Acknowledgements ........................................................................................................ 141
Chapter 6: General conclusion....................................................................... 143
6.1. Fire blight resistance in classical apple breeding ....................................................... 144
6.2. Mapping QTLs for fire blight resistance in bi-parental populations ....................... 146
6.3. Alternatives to bi-parental mapping populations ...................................................... 151
6.3.1. Limitations of bi-parental mapping populations .............................................. 151
6.3.2. QTL mapping in multiple related F1 progenies ............................................... 152
6.3.3. Pedigree-based analysis .................................................................................... 152
6.3.4. Association mapping ........................................................................................ 153
6.3.5. Phenotyping issues ........................................................................................... 155
6.4. Accelerating the introgression of fire blight resistance from wild Malus species or
hybrids into the cultivated apple......................................................................................... 156
6.4.1. Considerations on the genetic basis of fire blight resistance in wild Malus
species or hybrids ................................................................................................................... 156
6.4.2. Accelerated introgression of a fire blight resistance locus from a crab apple
based on the transgenic early flowering line T1190 and molecular markers ......................... 157
6.4.3. Towards an optimal use of the high-speed breeding technology in apple pre-
breeding .................................................................................................................................. 158
Appendices ....................................................................................................... 161
References ........................................................................................................ 173
Acknowledgements .......................................................................................... 199
Curriculum Vitae ............................................................................................ 205
x
List of Tables
Table 2.1. Fire blight lesion length as a percentage of shoot length (PLL) of ‘Florina’ (FLO),
‘Nova Easygro’ (NEG), ‘Golden Delicious’ (GD), ‘Idared’ and the F1 progeny ‘Florina’ ×
‘Nova Easygro’ (FLO × NEG).
Table 2.2. Significant and putative QTLs for fire blight resistance detected by the Kruskal-
Wallis test, interval mapping and multiple QTL mapping in the F1 progeny ‘Florina’ × ‘Nova
Easygro’.
Table 3.1. Description of the number and type of molecular markers, average distance
between markers, length and percentage covered with molecular markers for each linkage
group (LG) of the maps of ‘Idared’ (ID) and ‘Rewena’ (RE).
Table 3.2. Characteristics of the 562 polymorphic DArT markers segregating in the F1
progeny ‘Idared’ × ‘Rewena’.
Table 3.3. Genomic regions of the cultivars ‘Idared’ and ‘Rewena’ associated with fire blight
resistance using the Kruskal-Wallis test (p ≤ 0.05).
Table 5.1. Crosses and offspring of the first two introgression cycles of the fire blight
resistance locus Fb_E from the ornamental apple cultivar ‘Evereste’ using the early flowering
transgenic apple line T1190.
Table 5.2. Distribution of the F1 and BC’1 apple seedlings of the fast introgression scheme in
four genotype groups based on the presence/absence of the BpMADS4 transgene and the
resistant allele of SSR marker ChFbE06 co-segregating with the Fb_E locus.
xi
List of Figures
Fig. 1.1. Apple breeding scheme at the Agroscope ACW (Wädenswil, Switzerland).
From Kellerhals et al. (2009b).
Fig. 1.2. Fields of genomics and the related technologies and techniques used for
understanding the genetic control of important agronomic traits with the aim of improving
crops.
From Peace and Norelli (2009).
Fig. 1.3. Disease cycle of fire blight caused by the bacterium Erwinia amylovora.
From Norelli et al. (2003).
Fig. 1.4. Global position of quantitative trait loci (QTL) for fire blight resistance on the apple
genome based on the backbone genetic map developed by Silfverberg-Dilworth et al. (2006).
From Peil et al. (2009).
Fig. 1.5. Schematic illustration of the ontogenic phases of development in Malus spp..
From Hanke et al. (2007).
Fig. 1.6. Scheme of accelerated gene introgression from a small fruited wild apple species
into a domesticated apple genetic background through several pseudo-backcrosses using
transgenic early flowering genotypes.
From Flachowsky et al. (2009).
Fig. 2.1. Distribution of the individuals of the ‘Florina’ × ‘Nova Easygro’ progeny according
to their mean lesion length as a percentage of shoot length (PLL) at 7 (A) and 14 (B) days
after inoculation (PLL1 and PLL2, respectively).
Fig. 2.2. Genetic linkage maps of ‘Florina’ (FLO 1 to FLO 17) and ‘Nova Easygro’ (NEG 1
to NEG 17).
Fig. 2.3. LOD plots for fire blight resistance QTL mapping on FLO 10 (interval mapping, p =
0.05) with traits log10(PLL1) (A) and log10(PLL2) (B).
Fig. 2.4. Analysis of the pedigree of ‘Florina’ with SSR markers flanking the significant
QTLs on FLO 10 and FLO 5.
Fig. 3.1. Distribution of the 92 F1 individuals from the cross ‘Idared’ × ‘Rewena’ (ID × RE)
ranked in ascending order of their mean lesion length in percentage of shoot length (PLL).
Fig. 3.2. Framework parental linkage maps of ‘Idared’ (ID 1 to ID 17) and ‘Rewena’ (RE 2 to
RE 17).
Fig. 3.3. Example of redundant DArT markers on linkage groups 9 and 16 of ‘Idared’ (ID 9
and ID 16) and linkage group 5 of ‘Rewena’ (RE 5).
xii
Fig. 4.1. Genetic mapping position of locus SNP_T1190 on linkage group 4 of ‘Discovery’ as
previously constructed by Silfverberg-Dilworth et al. (2006).
Fig. 4.2. Schematic description of the crossbred-breeding scheme of Flachowsky et al. (2011).
Fig. 5.1. First two breeding cycles of the introgression of the fire blight resistance locus Fb_E
from the ornamental apple cultivar ‘Evereste’ using the BpMADS4-transgenic line T1190.
Modified from Chevreau (2009) and Flachowsky et al. (2009).
Fig. 5.2. Mean fire blight lesion length in percentage of total shoot length (PLL) of the four
genotype groups in the F1 progeny T1190 × ‘Evereste’ in comparison to the controls.
Fig. 5.3. Genome coverage with informative SSR markers and contribution of the
grandparental genome of ‘Evereste’ (both in %) in each BC’1 seedling carrying the
BpMADS4 transgene and the fire blight resistance locus Fb_E.
Fig. 5.4. Hypothetical schemes of introgression and pyramiding of disease resistance (R) loci
from wild apple species or crab apples into pre-breeding genotypes using the high-speed
transgenic introgression strategy.
Fig. 6.1. Global position of significant quantitative trait loci (QTL) with additive effect on fire
blight resistance on the apple (Malus spp.) genome.
Modified from Peil et al. (2009).
xiii
xiv
1
Chapter 1
General introduction
2
1.1. Introduction to the domesticated apple Malus × domestica
(Borkhausen)
1.1.1. The genus Malus: a complex nomenclature
The domesticated (syn. cultivated or eating) apple is a perennial fruit crop which belongs to
the genus Malus. With other genera such as Pyrus, Cydonia, Cotoneaster, Crataegus,
Pyracantha or Sorbus, it is included in the Spiraeoideae subfamily belonging to the Rosaceae
family (Potter et al., 2002; Potter et al., 2007). Depending on the taxonomists and botanists,
the genus Malus would comprise between 8 and 122 species; thus, Forsline et al. (2003)
identified 27 primary Malus species, 5 secondary ones and 11 species hybrids. The difficulty
to determine with certainty the number of species within Malus spp. and their relatedness
comes predominantly from the fact that Malus genotypes are mostly self-incompatible and
tend to interbreed easily. There has been a lot of debate about the exact scientific name of the
domesticated apple. Based first on morphological data (tree and fruit) and later on molecular
data (nuclear ribosomal DNA (rDNA) and chloroplast DNA (cpDNA)), it was found that the
domesticated apple is very closely related to the wild apple from Central Asia, Malus
sieversii. Therefore, it was proposed that M. sieversii would be the major progenitor of the
domesticated apple (Harris et al., 2002; Robinson et al., 2001; Vavilov, 1951); other Malus
species (M. baccata, M. mandshurica, M. orientalis, M. prunifolia, M. sylvestris) would have
contributed to the “domesticated” genetic pool although their degree of parentage is poorly
understood. Together with M. sieversii, the domesticated apple was referred to as M. pumila
Mill. by several authors (Juniper and Mabberley, 2006; Maliepaard et al., 1998; Robinson et
al., 2001). However, the denomination Malus × domestica Borkhausen (Borkh.), referring to
its supposed interspecific origin, has been mostly utilized by the scientific community so far.
The recent completion of a high-quality draft genome sequence of the domesticated apple
cultivar ‘Golden Delicious’ has shed new light on the Malus phylogenetic (Velasco et al.,
2010). This study not only demonstrated the formation of the domesticated apple gene pool
from M. sieversii, but also supports the hypothesis that both are the same species, thus
prolonging the debate on the nomenclature within Malus spp.. As it is not the purpose of this
Thesis to solve the question, we will refer to the domesticated apple and the Central Asian
wild apple as Malus × domestica Borkh. and Malus sieversii, respectively.
3
1.1.2. History of the domesticated apple
It is difficult to determine exactly when the apple was domesticated, as people probably
started planting apple trees unintentionally via garbage disposal or via their domesticated
animals for a long period before conscious selection. A common assumption is that, some
thousands of years ago, M. sieversii was repeatedly dispersed over short and long distances to
the east and west of the Tian Shan range (mountains of Central Asia, now at the border of
Kazakhstan and China) along the so-called Silk Roads, i.e. the caravan routes running from
East China to the Danube. The horse and the donkey, directed by human travellers, would
have carried apple seeds into their guts and unwittingly spread them along the trade routes
(Gladieux et al., 2010). It is thought that apple cultivation then reached the Near East by 3000
years before present (BP), where trees began to be cultivated in more sophisticated ways,
probably using the technique of grafting (Forsline et al., 2003; Harris et al., 2002). From
there, the crop and the grafting technique passed through the Persians and the Greeks to the
Romans who perfected the orchard management and spread the apple throughout Europe
where domestication flourished (Gladieux et al., 2010; Harris et al., 2002).
In the Middle Age, apple cultivation was maintained despite wars in the abbey gardens
throughout Europe and in the Muslim part of the Eastern Mediterranean. Skills of grafting,
training and pruning became there highly developed (Hancock et al., 2008). In the 17th
and
18th
centuries, apple cultivation expanded in Germany, Northern and Eastern Europe due to
the rise of Protestantism, which considered the apple as a special fruit of God. Consequently,
at least 1200 named varieties were acknowledged by the Royal Horticultural Society of
England in 1826, compared to the 120 named varieties described in Western Europe by the
end of the 17th
century (Hancock et al., 2008). At that time also, European colonists started to
move the domesticated apple across the seas: the crop was thus brought to the Americas
starting from the beginning of the 17th
century. European settlers established the first orchards
in New England in the 1620s and 1630s; as settlers were moving westward, they planted
apple seeds throughout Pennsylvania, Ohio, Indiana, and Illinois. This genetic base, which
has been heavily used for breeding new apple varieties in the United States of America
(USA), is known as “the North America gene pool” or “the Johnny Appleseed gene pool”
(Fazio et al., 2009; Janick et al., 1996); the latter refers to the legendary Jonathan Chapman
who devoted 40 years of his live to help settlers establish thousands of apple trees in the Ohio
River drainage (Hancock et al., 2008; Morgan and Richards, 1993). Apple was also
introduced to South Africa, Australia and New Zealand in 1650, 1788 and 1814, respectively
4
(Hancock et al., 2008). Nowadays, apple cultivation occurs in all temperate areas of the world
(Juniper and Mabberley, 2006).
1.1.3. Economic importance
The domesticated apple ranks fourth among the most important fruit crops worldwide (after
Citrus species, grape and banana). Apple is cultivated on about 5 million hectares (ha) all
across the temperate world (35°-50° latitude) with a production exceeding 60 million tons
(World Apple Report 2009). China is the biggest producer worldwide with 30 million tons,
followed by Europe with 14 million and North America with 4.4 million tons. The production
of apple and its consumption as well is expected to increase in the coming years. More
precisely, the production is expected to remain stable in Europe and in North America
whereas it should increase in Southeastern Asia and China. The Asian production is however
predicted to be largely consumed in China itself (http://faostat.fao.org/; Kellerhals, 2009).
Apple is an extremely versatile crop. First, fruits are produced mainly for the fresh market:
they can be eaten directly from the tree or stored up to a year in controlled atmospheres.
Although there are over 6,000 regionally important cultivars (apple varieties, typically given a
name, developed through intentional breeding) and landraces (old cultivated forms of apple
adapted to local growing conditions, but not improved by intentional breeding) worldwide, a
few cultivars dominate the market, representing over 60 % of the world’s production: ‘Golden
Delicious’, ‘Delicious’, ‘Cox’s Orange Pippin’, ‘Rome Beauty’, ‘Granny Smith’, ‘McIntosh’,
‘Jonathan’, ‘Braeburn’ (all chance seedlings, i.e. seedlings grown from open-pollinated
seeds), ‘Fuji’, ‘Gala’ and ‘Jonagold’ (Gardiner et al., 2007; Janick et al., 1996). Second,
apples can be also processed into juice, sauces and slices, three products which represent over
40 % of the apple production in the USA (Pereira-Lorenzo et al., 2009). Another important
product is cider, mainly in France, Spain and the United Kingdom (UK). Crab apples, which
are Malus species, subspecies or hybrids distinct from M. × domestica Borkh. with a small
and bitter pome fruit, are known for their attractive foliage and flowers and are often used as
ornamental plants. Some crab apples are also used as pollinators in commercial orchards.
1.1.4. Breeding program organization
The domesticated apple is grown in commercial orchards as a composite tree with a rootstock
and a fruiting scion (i.e. the cultivar). The main objectives of apple scion breeding programs
worldwide are a high fruit quality, resistance to the most important diseases in orchard,
especially scab and powdery mildew (caused by the fungi Venturia inaequalis and
5
Podosphaera leucotricha, respectively), and a tree architecture enabling high productivity and
regular bearing (Kellerhals, 2009; Laurens, 1999; Lespinasse, 2009; Sansavini et al., 2004).
Nevertheless, M. × domestica Borkh. displays a collection of specific attributes that pose real
challenges to breeders. These attributes are self-incompatibility, high heterozygosity,
perenniality, large plant size, extended juvenility, use of rootstocks, clonal propagation and a
perishable product (Peace and Norelli, 2009). As a consequence, apple breeders cannot afford
long term breeding strategies based on recurrent selection as pursued in field crops, which is a
type of cyclical selection in which the best individual plants of a base population are visually
selected and progeny-tested before being crossed to produce offspring to form the next
generation population; neither can apple breeders afford the time to perform test-crosses or
top-crosses to evaluate the ability of crossing combinations to reach the breeding goals
(Bringhurst, 1983; Gardiner et al., 2007); they have to face huge operational costs for growing
and maintaining inferior seedlings in the field before being able to score them for all traits and
finally discard them (Edge-Garza and Peace, 2010). To sum up, apple breeding is often
described as a long term and labour-intensive approach.
In the past, apple improvement consisted in selecting the best phenotypes from chance
seedlings. Since the early 20th
century, the strategy for apple fruit breeding has been to
generate genetic variability by sexual crossing of two genetically distinct trees (referred to as
parents of the cross), and then select within resulting large F1 (i.e. full-sibs) populations novel
genotypes which are tested in a multistage process for compliance with commercial criteria
(Janick et al., 1996; Johnson, 1999; Magness, 1937); recombination may be continued over a
few cycles in a limited version of recurrent mass selection (Alston and Spiegel-Roy, 1985;
Janick et al., 1996). Very often, the parents are cultivars of known merit or the most
promising breeding material of a program (i.e. advanced selections; (Kellerhals et al.,
2009b)). The parental trees should be chosen carefully by the breeder as the goal is to
combine the good “qualities” of both in some progeny individuals; deciding which “qualities”
should be preferentially combined requires prioritization of the objectives (Janick et al.,
1996). Cummins and Aldwinckle stressed the fact that not every desirable objective can be
achieved, and consequently ranked the objectives of their breeding program (Cornell
University, USA) as “essential”, “important” and “helpful” (Cummins and Aldwinckle,
1995). Similarly, Edge-Garza et al. (2010) listed traits of interest to the Washington breeding
program (USA) under the categories of “highest priority”, “also selected for but of lower
priority”, and “don’t select for but of interest”. Usually, a breeder achieves between 5 and 50
crosses a year, with at least 200–300 seeds per cross (50 to 100 pollinations per cross). Thus,
6
in a conventional apple breeding program, 10,000 to 15,000 seeds can be produced every year
(Durel et al., 2007; Hancock et al., 2008; Kellerhals et al., 2009b).
Evaluation of F1 offspring is a long process which begins as soon as seeds have germinated.
Selection efficiency may be defined as the capacity to discard as early as possible offspring
less prone to become a successful commercial variety (Kellerhals et al., 2009b). Selection for
fruit quality and productivity cannot be made in the first years because of the long juvenile
phase of the apple (period of time during which a seedling cannot be induced to flower; 4 to 5
years on average but up to 10 years for some cultivars like ‘Northern Spy’ (Fischer, 1994;
Hanke et al., 2007; Visser, 1964)). However, the seedling stage is appropriate to select for
resistance to diseases that infects young leaves, such as apple scab or cedar apple rust (incited
by several species of the fungal genus Gymnosporangium). Typically, breeders spray spore
suspensions of these fungi on young seedlings which may show varying levels of infection on
their young leaves after a few weeks. This way, and depending on selection intensity, up to
50–80 % of the progeny can be discarded (Durel et al., 2007; Kellerhals et al., 2009b).
Powdery mildew resistance is usually evaluated in nurseries when seedlings are in their
second or third year of growth, together with traits like plant vigor, growth and architecture
(Kellerhals et al., 2009b). It must be noticed that many programs graft seedlings onto
dwarfing rootstocks quite early, for instance in the second year (e.g. Durel et al., 2007); this
technique allows a more precocious flowering and saves space, but it adds to the cost of the
program (Hancock et al., 2008). Field selection is usually performed in 3 stages
(corresponding to 8 to 12, 10 to 17 and 14 to 20 years after sowing, respectively) and aims at
assessing fruit quality, productivity, tree architecture and resistance to biotic and non biotic
stresses on commercially important rootstocks (Kellerhals et al., 2009b). A schematic
illustration of a classical apple breeding program with multistage selection is given in Fig.
1.1.
7
Fig. 1.1. Apple breeding scheme at the Agroscope ACW (Wädenswil, Switzerland). From
Kellerhals et al. (2009b).
8
1.2. Application of molecular markers to apple breeding
1.2.1. Genomics, genetics and marker-assisted breeding: definitions
So far, classical apple breeding has relied to a large extent on the practical knowledge of
breeders about their breeding germplasm and about how important traits segregate in breeding
populations. Breeding germplasm typically includes founders, i.e. apple cultivars from which
many other cultivars and selections descend, parental cultivars, breeding progenies and
selections. Peace and Norelli (2009) discussed how genomics could be used to decipher the
genetic architecture of important traits and identify the genes that control traits variation in
Rosaceae. This knowledge should, according to them, help breeders to produce optimum
genetic combinations in new cultivars that perform better in orchards. They defined genomics
in comparison to genetics as following: whereas genetics (also called classical genetics) is
based on the principles of Mendelian inheritance of traits and on the theory of quantitative
genetics, genomics is a more holistic approach that includes genetics and eventually aims at
identifying the genetic elements (e.g. genes) responsible of the traits of interest. Three
categories of genomics approaches have been identified by Peace and Norelli (2009), which
require partly common technologies and techniques, namely structural genomics (study of the
genomes’ features), functional genomics (identification of the function of individual genes)
and comparative genomics (evaluation of the commonalities and differences between
genomes of different species) (Fig. 1.2). Practically, genomics is expected to lead to the
development of molecular markers (identifiable location on a chromosome that can reveal
DNA polymorphism), the use of which may aid breeders in different operations of their
breeding program (therefore referred to as marker-assisted breeding (MAB)): assessing the
genetic diversity of the breeding germplasm, fingerprinting cultivars and accessions
(catalogued members of a germplasm collection), verifying parentages or selecting parents
and seedlings (steps referred to as marker-assisted selection (MAS) or more specifically to
marker-assisted parental and seedling selection (MAPS and MASS), respectively) (Bus et al.,
2009; Charcosset and Moreau, 2004).
9
Fig. 1.2. Fields of genomics and the related technologies and techniques used for
understanding the genetic control of important agronomic traits with the aim of improving
crops. From Peace and Norelli (2009).
The earliest genetic (syn. linkage) maps of apple were constructed in the USA (Conner et al.,
1997; Hemmat et al., 1994); shortly after, two European projects addressed the challenge of
developing molecular markers associated to diseases resistance, namely the European Apple
Genome MApping Project (EAGMAP (King et al., 1991)) and the Durable Apple Resistance
in Europe project (DARE (Lespinasse and Durel, 1999)). These two successive projects were
mostly based on the construction of linkage maps using F1 segregating progenies for the
identification of quantitative trait loci (QTL) associated to powdery mildew and scab
resistance. A third European project that took place between 2004 and 2008, named High-
quality Disease Resistant Apples for a Sustainable agriculture (HiDRAS (Gianfranceschi and
Soglio, 2004)), aimed similarly at providing molecular markers associated with fruit quality
and pathogen resistance for application of MAB in apple. Its originality consisted in the
implementation of the pedigree based analysis (PBA (van de Weg et al., 2004)); PBA
explores the identity by descent concept (IBD) and supports integrated QTL analysis of
multiple progenies and multi-generation breeding germplasm including cultivars and breeding
selections (see 1.2.2.3). The pedigree based analysis has been included as a major tool of the
10
US RosBreed project whose aim is the integration of genomics, and not only genetics,
approaches (as defined by Peace and Norelli, 2009) with traditional breeding in Rosaceae
(Iezzoni et al., 2010). More recently, a European Union-funded project (FruitBreedomics) has
been initiated with the similar objective of bridging the gap between genomics research and
traditional breeding in apple and peach (Prunus persica) (Laurens et al., 2011). Similarly to
HiDRAS and RosBreed, Fruit Breedomics will apply PBA in apple and (for the first time) in
peach. New for both species will be the application of association mapping (Oraguzie and
Wilcox, 2007; Rikkerink et al., 2007) in order to find associations between single nucleotide
polymorphism (SNP) markers and agronomically relevant traits. More generally, advances in
technology of marker development and screening is expected to facilitate the transition to
high-throughput MASS, i.e. screening of a large number of seedlings per year (e.g. tens of
thousands) with not only a few markers tagging 2 or 3 traits but with dozens of markers
targeting many traits (Bus et al., 2009).
1.2.2. Development of molecular marker–phenotypic trait associations
A prerequisite to MAS per se is the development of marker-trait associations. Development of
marker-trait associations has been performed in apple using several approaches which are
presented in the following sections; they have been described more in details in several works
(Gardiner et al., 2003; Gardiner et al., 2007; Han and Korban, 2010; Patocchi et al., 2005;
Peace and Norelli, 2009; van de Weg et al., 2004; Xu et al., 2001).
1.2.2.1. Gene mapping
Major genes for diseases and pests resistance are usually responsible for a more or less
qualitative distribution of the resistance phenotype in segregating F1 progenies (i.e. seedlings
can be roughly classified as resistant or susceptible to a given pathogen or pathogen
race/strain); their mapping do not require the construction of complete genetic maps. The bulk
segregant analysis (BSA) technique (Michelmore et al., 1991) was first used in Malus spp. to
map genes for scab (Durham and Korban, 1994; Koller et al., 1994) and powdery mildew
resistance (Gardiner et al., 1999); it implies first to prepare pools of DNA belonging to two
bulks, one from the most susceptible plants and one from the most resistant plants, and then to
identify markers differentiating the two bulks in order to develop partial maps around the
resistance locus (Michelmore et al., 1991). The markers used to screen the bulks were random
amplified polymorphic DNA (RAPD) markers or amplified fragment length polymorphism
(AFLP) markers (Xu and Korban, 2000); for instance, RAPD technology associated with
11
BSA was the most widely employed method to identify Rvi6 (formerly Vf) - linked markers
(Gessler et al., 2006). However, due to the lack of reproducibility and specificity of RAPD
and AFLP markers, researchers usually converted these markers into sequence characterized
amplified region (SCAR) or cleaved amplified polymorphic sequence (CAPS) markers for
final map construction and use for MAS (Gardiner et al., 2003; Gianfranceschi et al., 1996;
Tartarini et al., 1999; Xu et al., 2001). An additional limitation to the BSA-RAPD technology
appeared to be the low success rate (i.e. the low number of RAPD markers linked to the
resistance gene compared to the number of arbitrary primers used to produce these RAPDs)
(Gardiner et al., 1996; Patocchi et al., 2004).
Later on, the genome scanning approach (GSA) was developed with the same purpose of
finding a linkage between a resistance gene and a molecular marker without generating a
complete genetic map. As explained by Patocchi et al. (2005), the method has benefited from
the high number of polymorphic simple sequence repeats (SSR; syn. microsatellite) markers
mapped on the apple genome (Liebhard et al., 2003a). Basically, it consists of testing a
reduced number of resistant (or susceptible) progeny plants with a few selected and well-
spaced SSR markers per linkage group (LG). In the first round of the GSA, the segregation of
the alleles of the resistant parent is tested at each SSR marker against the 1:1 segregation
ratio, which is expected if the SSR marker is not linked to the resistance gene. If a deviation
from the 1:1 segregation is found, this microsatellite is considered as a putative marker for the
resistance gene. The second round of the GSA implies to extend the number of progeny plants
to be screened with additional SSR markers. These markers should be located just above and
below the SSR marker showing skewed allele segregation, in order to finally map the
resistance gene. This approach has been successfully applied to locate the scab resistance
genes Rvi5 and Rvi11 (formerly Vm and Vb, respectively (Erdin et al., 2006; Patocchi et al.,
2005)) and the crown gall (Agrobacterium tumefasciens) resistance gene Cg (Moriya et al.,
2010). Several parameters condition the success of the GSA, the most important being (i) the
number of plants and the number of SSR markers per linkage group to be used, (ii) the
positions of these markers within the linkage group; and (iii) the segregation distortion
occurring in certain regions of the apple genome.
1.2.2.2. QTL mapping on a single progeny
A quantitative trait is a trait for which the observed variation in a segregating progeny is
continuous, a oligo- or poly-genic determinism being most of the time assumed. The genomic
loci controlling such a trait are called quantitative trait loci (QTLs). The approach to genetic
12
mapping of quantitative traits differs from the one applied to major genes. Applying BSA (or
GSA) on extreme phenotypes of a segregating progeny may enable the identification of the
QTL(s) with the strongest effect on the phenotypic variation, but generally will not allow
deciphering the complete genetic architecture of a trait (Collard et al., 2005). For this, a
“high-quality” linkage map is required.
1.2.2.2.1. Linkage map construction
Linkage map construction is basically performed as following: molecular markers are first
screened over a segregating population derived from a cross between two genotypes (a F1
progeny in Malus spp.) and the frequencies of recombination (r) between every pair of
markers are calculated; molecular markers are then assembled together depending if they are
linked (r < 0.5) or not, resulting in so-called linkage groups; in a third step, pairwise
recombination frequencies are converted into genetic distances (in centiMorgan; cM) using
the function of Kosambi and markers are consequently ordered relative to each other within
the groups (i.e. mapped). Linkage groups represent partial segments of chromosomes or
whole chromosomes once enough markers are included to span the gaps between the
segments. Genetic maps have been developed for over 20 pomefruit accessions to date
(Gardiner et al., 2007; Peil et al., 2009). A “high-quality” linkage map should meet several
criteria. First, the linkage map should be saturated, which is the case when every molecular
marker of the map is linked to at least another one, and when the number of linkage groups is
equal to the number of chromosomes of the species. A more pragmatic definition was given
by Liebhard et al. (2003a), for whom the term “saturated” should be regarded as “completely
covered with markers”, thereby emphasizing the necessity to map molecular markers at
regular intervals all over the genome, with as few gaps as possible. Second, the density of
markers on the map (i.e. the average genetic distance between two adjacent molecular
markers) must be taken into account as well; it is usually considered that a genetic distance of
10–20 cM between two adjacent molecular markers is sufficient to identify QTLs explaining
10–20 % of the trait variation (Collard et al., 2005; Darvasi and Soller, 1994; Darvasi et al.,
1993; Piepho, 2000; Van Ooijen, 1992). Third, the linkage map should show as few
inconsistencies as possible, i.e. the markers order on each linkage group should remain stable
whether one marker is added to or removed from the map; this requires beforehand an
accurate location of markers on the map (Liebhard et al., 2003a). Fourth, a linkage map
should comprise if possible “anchor” or “bridge” markers on each linkage group for
alignment with linkage maps from other crosses of the same species (for comparative
13
mapping (Calenge et al., 2004; Liebhard et al., 2003c; Soufflet-Freslon et al., 2008)) or from
closely related species (for syntenic studies (Celton et al., 2009a; Pierantoni et al., 2004;
Yamamoto et al., 2007; Yamamoto et al., 2004b)). Anchor markers offer also the possibility
to integrate the maps of the two parents of the cross, which is especially relevant to QTL
analysis when both parents contribute to the phenotypic variation observed in the segregating
progeny (Durel et al., 2003); anchor markers can finally be used to build a consensus map
from different crosses (N’Diaye et al., 2008).
Liebhard and Gessler (2000) listed three requirements to meet these criteria of linkage map
quality in Malus. The first requirement is the careful selection of the cross and of the mapping
progeny size. The F1 progeny must be segregating for at least one trait of interest. When the
selected F1 progeny is segregating for a range of agronomically relevant traits, the linkage
maps constructed can be re-employed for other traits than the one originally targeted; this is
the case of the linkage maps of the apple cultivars ‘Fiesta’ and ‘Discovery’ (Liebhard et al.,
2003a) that were used to study the following quantitative traits: scab resistance (Liebhard et
al., 2003c), growth habit and fruit quality (Liebhard et al., 2003b), composition of volatile
compounds in ripe fruits (Zini et al., 2005), fire blight resistance (Khan et al., 2006) and pests
resistance (Stoeckli et al., 2009; Stoeckli et al., 2008). The size of the mapping progeny is one
of the most important experimental design factors for QTL analysis (Collard et al., 2005). An
increase in the progeny size increases the power of QTL detection (defined as the probability
to identify “real” QTLs even if they have a small effect, i.e. a small contribution to the
phenotypic variation) and increases the precision for estimating QTLs effects and locations
(Darvasi and Soller, 1994; Tanksley, 1993; Van Ooijen, 1992). Most of the QTL mapping
studies performed in Malus so far were based on 140 to 200 F1 individuals, with some
exceptions (86 (Khan et al., 2006; Zini et al., 2005), 242 (Kenis and Keulemans, 2007) and
251 F1 individuals (Liebhard et al., 2003b)). The second requirement for a linkage map of
good quality derives from an apparently logical statement: a linkage map of good quality is
reliant on the use of molecular markers of good quality (Liebhard and Gessler, 2000). For
genetic studies in Malus spp., molecular markers of good quality should ideally have the
following characteristics: (i) co-dominant inheritance; (ii) reproducibility between
laboratories; (iii) transferability between mapping populations; (iv) cost-effectiveness of
screening via multiplexing (i.e. simultaneous amplification of multiple molecular markers in a
single PCR reaction) and high-throughput genotyping; (v) ease-to-score avoiding genotyping
mistakes which are a major source of experimental errors in QTL mapping studies; (vi)
abundance over the whole genome (Collard et al., 2005). So far, more than 1,100 SSR
14
markers have been developed, mapped and characterized in Malus spp. (Broggini et al., 2009;
Celton et al., 2009b; Cova et al., 2011; Guilford et al., 1997; Han et al., 2011; Hemmat et al.,
2003; Hokanson et al., 1998; Liebhard et al., 2003a; Silfverberg-Dilworth et al., 2006; van
Dyk et al., 2010; Wang et al., 2011). About 80 of them, well-spanned over the genome, were
found to fit the characteristics mentioned above (Evans et al., 2010; Patocchi et al., 2009b).
The third requirement for a good linkage map, still according to Liebhard and Gessler (2000),
is to have a good understanding of the mapping software; the software JoinMap® (e.g. version
4 (Van Ooijen, 2006)), able to deal with linkage mapping in full-sib families, is often handled
for constructing apple linkage maps because it alleviates the grouping and ordering of
markers; however, it does not preclude extensive visual inspection of genotyping data to
verify the presence of putative double recombinants (often due to wrong scoring of markers)
and check suspect genotypes.
1.2.2.2.2. Statistical and QTL analysis
Classically, mapping of QTLs involves determining the degree of association between a
phenotypic trait continuously distributed in a bi-parental mapping progeny and the molecular
markers located on a linkage map. Quantitative trait loci for different components of fruit
quality (size, aroma, firmness, acidity, volatiles), tree architecture and diseases and pests
resistance (e.g. scab, powdery mildew, fire blight and insects) have been identified in Malus
spp. (Gardiner et al., 2007; Gessler et al., 2006; Korban and Tartarini, 2009; Peil et al., 2009).
In this regard, it is important to keep in mind that the percentage of the phenotypic variation
explained (PVE) by a QTL is not an absolute value. When testing for a significant association
with the trait of interest at a molecular marker, one estimates the effect of one parental allele
only relatively to the effect of the other parental allele (Liebhard et al., 2003c). Should each
parental allele have a similar effect on the trait of interest (case of “functional”
homozygosity), no difference could be found between the offspring that inherited either
alleles, and therefore no association with the trait could be detected at this marker (Khan,
2007; Liebhard et al., 2003c).
Several statistical checks have to be performed beforehand to assess the reliability of the QTL
analysis. First, it is always advisable to have a look at the raw phenotypic data collected; the
distribution (continuous, bimodal, normal, …), segregation, transgression and variation of the
phenotyped trait in the mapping population may give first hints on its mode of inheritance and
on the power of the subsequent QTL analysis. Then, based on these observations, one may
check for the presence of outliers and normalize data for analysis of the variance (ANOVA)
15
(Khan et al., 2006; Segura et al., 2007; Stoeckli et al., 2008). The ANOVA allows to estimate
the proportion of the phenotypic variance in the population that is due to the genotype (broad-
sense heritability or h2); for a complete ANOVA model, one may include other explanatory
variables such as the trial (providing that the progeny has been evaluated during several
trials), block, replication or date (e.g. when pathogen inoculation has been performed at
different dates) as well as 2-ways interactions between these variables. Adjustment,
standardization and correction of normalized phenotyping data may be also necessary
(Calenge and Durel, 2006; Dowkiw and Bastien, 2007; Le Clerc et al., 2009).
Once phenotypic data have been processed, the first issue of QTL analysis sensus stricto
consists in identifying the statistically significant QTLs. For this, several tests may be used as
proposed by the software MapMaker, QTL cartographer or MapQTL®, the latter being often
used complimentarily with JoinMap® (Dunemann et al., 2009; Kenis et al., 2008; Khan et al.,
2006; Moriya et al., 2010; Segura et al., 2007; Soufflet-Freslon et al., 2008). The software
MapQTL® provides geneticists with a single marker analysis, which is a non parametric
equivalent of an ANOVA (the Kruskal-Wallis test), and two tests taking advantage of the
linkage map, namely the interval mapping (IM) and the multiple QTL mapping (MQM) tests.
Interval mapping consists in testing for a significant association with the phenotypic trait at
regular intervals on the linkage map by calculating a logarithm of odds (LOD) -likelihood
ratio; the LOD basically compares the likelihood of QTL presence to the likelihood of QTL
absence within one interval bracketed by two markers (also called “window”). Multiple QTL
mapping is similar to interval mapping, except that it tests for the presence of a QTL in a
given interval while taking into account the presence of QTLs in other genomic regions by
selecting cofactors (i.e. molecular markers close to QTLs previously identified). Depending
on the genome coverage and marker density of the linkage map, one, two or all three tests
may be performed; in turn, different thresholds (e.g. LOD threshold) may be utilized to
declare a QTL significant that can be defined arbitrarily, selected from reference tables (Van
Ooijen, 1999) or calculated at different significance levels (95 or 99 %) or scales (e.g.
chromosome or genome scale) by the permutation test (Doerge and Churchill, 1996).
Once significant QTLs have been found, the second issue is to characterize the amplitude of
their effect. Whether a QTL can be considered as of minor, medium or major effect is mostly
determined by the percentage of phenotypic variation it explains (PVE; R2), although there is
no agreement on a R2
threshold for this classification. Thus, Collard et al. (2005) considered a
QTL as “major” when it explains more than 10 % of the phenotypic variation; nonetheless,
QTLs explaining around 20 % of the variation of resistance to Plasmodiophora brassicae in
16
Arabidopsis thaliana were designated as of medium-effect (Jubault et al., 2008); and in Malus
spp., major QTLs were declared only when the PVE was of 40 % or more (Calenge et al.,
2005; Durel et al., 2009; Khan et al., 2006; Peil et al., 2007). However, to fully characterize
the amplitude of a QTL effect, one may better consider the percentage of phenotypic variation
explained by the QTL (R2) in relation with the amplitude of the phenotypic variation in the
mapping progeny; in apple, discrepancies were recently observed between the R2
attributed to
the major fire blight resistance QTL of the cultivar ‘Fiesta’ in original mapping populations
(R2
= 34.3 to 46.6 % (Calenge et al., 2005; Khan et al., 2006)) and the real amplitude of the
effect of this QTL in breeding progenies (Baumgartner et al., 2010; Kellerhals et al., 2011).
Besides its proper (or additive) effect, a QTL may be involved in intra- or inter-loci
interactions (dominance or epistasis, respectively (Calenge et al., 2005; Durel et al., 2003)) as
well, thereby complicating the estimation of its contribution to the phenotypic variation in the
progeny. A global R2 may be eventually calculated using an ANOVA by including all
significant additive, dominant and epistatic QTLs as variables.
The third issue of QTL mapping is to determine as precisely as possible the location of QTLs.
The map position associated to the highest value of a statistical test (e.g. K* for the Kruskal-
Wallis test or LOD score for IM or MQM) may be used, as well as the position of the
molecular markers closest to this highest value. The latter is especially interesting if the
markers are co-dominant and reproducible (e.g. SSR or SNP markers), as they allow to
compare the positions of QTLs detected on linkage maps from several crosses. Co-
localization of QTLs may be also evaluated by the comparison of their confidence intervals
(CI), calculated most of the time by a drop-off of the LOD score at the QTL’s peak.
1.2.2.3. Pedigree based analysis
The “traditional” QTL mapping approach in Malus spp. described above relies on
experimental F1 progenies which are usually created for the specific purpose of identifying
QTLs. A more recent approach called pedigree genotyping or pedigree based analysis (PBA)
avoids the need for such dedicated populations. Indeed, PBA consists in identifying QTLs by
analyzing genotypic and phenotypic data of breeding material itself, which is composed by
multiple pedigree-linked plant populations (which include any combination of crosses,
cultivars, and breeding selections). This approach is therefore well suited to the apple
breeding germplasm (Peace and Norelli, 2009; van de Weg et al., 2004). The pedigree based
analysis is based on two complementary statistical approaches. The first identifies QTL
regions based on Markov chain Monte Carlo simulations and Bayesian statistics. The second
17
refers to the principle of identity by descent (IBD), which expresses the identity of an allele of
a breeding selection in terms of alleles of founding cultivars. Besides the identification of
marker-trait associations and a major reduction in experimental costs since plant material is
already available, the main advantages of the PBA are the following: (i) mining of alleles at
each QTL locus detected in the breeding germplasm; (ii) testing of QTL alleles against a wide
range of genetic backgrounds, i.e. testing of stability; (iii) analysis of intra- as well as inter-
QTL interactions. The PBA approach has been used in the framework of the HiDRAS project
to identify QTLs for apple scab resistance using 13 F1 progenies that were part of an ongoing
breeding program at the Institut National de la Recherche Agronomique (INRA, Angers,
France) (Soufflet-Freslon, 2008).
1.2.2.4. Candidate gene approach
The candidate gene approach sensus largo consists in using the knowledge generated by
structural and functional genomics to identify “candidate” genes with a high likelihood of
playing an important role in the phenotype of a specific trait (Peace and Norelli, 2009). The
candidate gene approach sensus stricto reported by Gardiner et al. (2003) was based on the
mining of an expressed sequence tags (EST) database of Malus containing over 100,000
sequences from 44 cDNA libraries. First, Gardiner and colleagues identified candidate
resistance (R) genes in apple ESTs by bioinformatics comparison with known resistance
genes in EST databases from model plant species. Second, mini-populations of apple
segregating for a number of disease resistance genes, consisting of parents and resistant and
susceptible progeny, were analyzed with the restriction fragment length polymorphism
(RFLP) technique; mini-populations Southern blots were thus set up. Third, candidate R gene
ESTs were screened as RFLP probes over these Southern blots to identify any linkage with
resistance genes. Fourth, putative linkages identified were confirmed in larger progeny sets
and ESTs most closely linked to R genes were converted into PCR-based markers and then
mapped in the entire population. Co-localization of candidate gene markers with either known
major genes or QTLs identified candidates that warrant functional verification. This approach
led to the identification of candidate gene markers for 13 different resistances to apple scab,
powdery mildew and woolly apple aphid (WAA) (Gardiner et al., 2007). Eventually, if further
functional analysis determines a causative role in the trait of interest for one candidate gene, a
molecular marker specific to the functional allele of the candidate gene can be developed that
represents the causative DNA polymorphism underlying the functional differences
(“functional” or “perfect” molecular marker (Peace and Norelli, 2009)). The whole candidate
18
gene approach defined by Gardiner et al. (2003) (from EST-database mining to functional
marker development) has been successfully applied for a component of apple fruit quality, i.e.
the color of the flesh encoded by the Rni locus (Chagné et al., 2007; Espley et al., 2007).
1.2.3. Implementation of marker-assisted selection
1.2.3.1. Challenges for application of molecular markers to breeding
Although molecular markers hold potential to improve breeding efficiency in different ways,
their practical application in non-staple food crops is currently limited. In perennial crops, it
was estimated that only 3 % of the breeding programs worldwide were applying molecular
markers (Byrne, 2010). In Rosaceae, four major factors have to be considered for applying
molecular markers in breeding:
- genotyping costs and investments: according to Bus et al. (2009), the cost of genotyping
several thousands of seedlings (i.e. marker-assisted seedling selection, MASS) is determined
by the number of marker data point (MDP) which is a function of the number of progeny
seedlings times the number of molecular markers applied times the number of generations
required to reach breeding objectives. The authors stressed that the current high costs of the
techniques of DNA extraction, samples genotyping and data analysis explain why MASS
implementation is still limited in breeding programs in Rosaceae, in contrast to major crops
such as rice (Oryza sativa L.) and wheat (Triticum aestivum L.). Byrne (2010) also identified
genotyping costs as the major reason for which molecular markers are not yet applied on a
large scale in perennial plant breeding. Besides the costs of progenies genotyping itself, one
has to consider the investment in equipment, which is often ignored in costs estimates (Bus et
al., 2009). However, nowadays, the use of specific, reproducible and easy-to-use markers
such as SSR markers together with the possibility of multiplexing offers the possibility to
screen large numbers of seedlings in a more cost-effective way (Frey et al., 2004). A further
reduction of costs is expected in the coming years due to the development of high-throughput
genotyping systems that do not require electrophoresis or even PCR (e.g. array-based MASS
or high-throughput SNP genotyping assays (Akhunov et al., 2009; Gupta et al., 2008; Henry
and Edwards, 2009; Mammadov et al., 2011). Last but not least, several techniques have been
developed in recent years to increase the cost-efficiency of identifying molecular markers
associated to a trait of interest (e.g. BSA, GSA, PBA; see 1.2.2); in this regard, the sequence
of the M. × domestica (Borkh.) genome now available will greatly facilitate the identification
of different markers (e.g. SSR, SNP) in specific genomic regions relevant to breeders
(Velasco et al., 2010).
19
- effectiveness of molecular markers: if the targeted trait is highly heritable, simple to
phenotype and controlled by a single gene (e.g. qualitative, monogenic Rvi6 resistance to scab
in apple), genotypic and phenotypic selection may be equally accurate. In this case, the choice
between both of them involves a trade-off between time and money, which is often in favour
of the phenotypic selection (Bliss, 2010). In other words, a breeder has to know the cost
structure for both types of screening (Bus et al., 2009). Molecular markers application can
increase selection efficiency when specialized conditions are necessary for phenotypic
screening (such as quarantine facilities) or for traits difficult to phenotype (e.g. powdery
mildew, woolly aphid or fire blight resistance in apple). More generally, using molecular
markers allows removing environmental constraints from progenies screening. Moreover,
MASS provides the opportunity of early selection for traits expressed in mature plant only
(e.g. fruit quality traits in apple); MASS is also by far the most efficient way of pyramiding
resistance genes showing masking effects that would hinder phenotypic selection (e.g. scab
resistance genes pyramiding; Bus et al., 2009; Soufflet-Freslon, 2008).
- value of the trait for cultivar performance: although the costs of genotyping are decreasing
and will probably continue to decline, the large numbers of seedlings to be genotyped in the
framework of breeding programs will probably force breeders to target the most valuable
traits for MASS; this implies that the choice of traits to be targeted by molecular markers
should be based on the value of the traits (i.e. their contribution to cultivar performance)
rather than on the availability of markers (Bliss, 2010).
- reliability and robustness of molecular markers: molecular markers intended for marker-
assisted breeding (MAB) should be first “reliable”, i.e. should produce neither false positives
nor false negatives. This involves that the markers associated to the trait must display clearly
distinguishable “desirable” (or “favourable”) and “undesirable” (“unfavourable”) alleles.
Reliability also involves a very tight linkage (1 cM or less) between a marker and a locus
involved in the control of the targeted trait, in order to reduce as much as possible the
likelihood that a recombination disrupts the marker-locus linkage (Bus et al., 2009; Gessler et
al., 2006). Gardiner et al. (2007) calculated that if one assumes a recombination rate of 1 %
between a marker and a locus, MASS would result in 3 % of the selected seedlings not
carrying the desired combination of resistances in the case of three pyramided genes, and 5.9
% in the case of six genes; so closely linked markers are yet considered as useful for MASS,
even though they should be soon superseded by markers co-segregating with the genes of
interest developed by map-based cloning or candidate gene approaches (Chagné et al., 2007;
Cova, 2008; Fahrentrapp et al., 2011; Galli et al., 2009; Parravicini et al., 2011; Vinatzer et
20
al., 2004). For the particular case of QTLs with a large confidence interval (CI), use of
flanking molecular markers is considered as especially useful. Second, molecular markers
should be also “robust”, i.e. that they should be transferable across breeding populations,
which is the case of many SSR markers in apple (Evans et al., 2010; N’Diaye et al., 2008;
Patocchi et al., 2009b) but not necessarily of individual SNP markers for instance (Micheletti
et al., 2011; Micheletti et al., 2010). If population-specific markers are needed for different
segregating populations, MASS becomes less efficient and more expensive (Bliss, 2010).
1.2.3.2. Marker-assisted parental selection
Molecular markers are a useful tool to identify among germplasm collections or breeding
material the genotypes carrying alleles of interest at key genes for diverse traits; such
genotypes can be considered as potential parents for future crosses (Bus et al., 2009; Peace
and Norelli, 2009). For instance, Vinatzer and collaborators screened 55 apple cultivars and
selections with two SSR markers (CH-Vf1 and CH-Vf2) tightly linked to the Rvi6 (formerly
Vf) scab resistance gene. This allowed the reliable detection of the Rvi6 gene in the scab-
resistant accessions Malus micromalus SA573-3, ‘Golden Gem’, M. prunifolia 19651 and
MA 16 in which the presence of the gene was only assumed until then; with this work,
additional sources of Rvi6-based scab resistance were provided for breeding programs
(Vinatzer et al., 2004). Two studies reported the phenotypic and molecular screening of open-
pollinated and F1 progenies of M. sieversii, respectively, in an attempt to identify genotypes
carrying known scab-resistance genes (Bus et al., 2005a; Fazio et al., 2009). Bus and
colleagues tested eleven families of M. sieversii from Kazakhstan amplifying the two SCAR
markers flanking the Rvi8 scab-resistance gene (preliminary molecular screening) with the
specific V. inaequalis race 8 that can overcome the Rvi8 resistance (phenotypic screening). In
each family, the scab-resistance was overcome by race 8 and therefore might be traced to the
same gene, presumably Rvi8. Fazio and colleagues screened seven F1 progenies derived from
crosses between seven distinct M. sieversii accessions and ‘Royal Gala’ with molecular
markers linked to known scab-resistance genes (molecular screening) and with V. inaequalis
(phenotypic screening). Correlation of the marker data with the phenotypic data indicated that
some M. sieversii parents may possess known scab-resistance genes with chlorotic or necrotic
symptom reactions. The recent mapping of two scab-resistance loci conditioning stellate
necrotic symptoms in a region of linkage group (LG) 2 known to carry similar scab-resistance
genes (Rvi2, Rvi8, Rvi9 and Rvi11 (Bus et al., 2011)) in the elite M. sieversii accession PI
613988 gives support to this assumption (Wang et al., 2011). However, a strong variability of
21
the segregation of scab resistance in some F1 progenies was observed (10–67 % of resistant
seedlings (Fazio et al., 2009; Forsline and Aldwinckle, 2004)). These observations may
reinforce the necessity to better understand the genetic basis of scab resistance in M. sieversii
to apply more efficiently marker-assisted parental selection (MAPS) among M. sieversii
germplasm. Eventually, elite accessions of M. sieversii carrying scab resistance loci may
contribute to breeding new apple cultivars without the need for extensive pseudo-
backcrossing as some of the elite accessions of M. sieversii are similar to commercial
cultivars for many critical horticultural traits (Forsline et al., 2003).
1.2.3.3. Marker-assisted seedling selection
So far, the vast majority of reports on MASS in Malus spp. have dealt with the pyramiding of
genes for resistance to pests and diseases. Gene pyramiding can be defined as the
accumulation of a number of genes to create an ideotype (Bus et al., 2009; Servin et al.,
2004). Several authors pointed out that natural resistance gene pyramids found in Malus
germplasm seem durable, in the sense that they have not lost their effectiveness over time in
presence of the pathogen, even though the proof of “true” durability would involve an
extensive deployment of these resistance gene pyramids; famous examples are the scab
resistance of Russian apple R12740-7A, M. micromalus or ‘Antonovka’ accessions (Broggini
et al., 2010; Bus et al., 2010; Dayton et al., 1953; Schmidt, 1938). Therefore, the authors
suggested that pyramiding resistance genes may be a more valid strategy to achieve durable
resistance than introgressing one single gene into a cultivar. The New Zealand apple breeding
program has been aiming for many years at developing new apple cultivars with durable pests
and diseases resistances through MASS. Bus et al. (2000) developed segregating populations
to pyramid scab resistance genes as well as other populations to specifically combine
resistances to scab, powdery mildew and/or woolly apple aphids. In particular, when
following the segregation of the genes Rvi4 and Rvi6 using SCAR markers in progenies, Bus
and colleagues realized that the pit resistance symptom of Rvi4 was masking the chlorotic
resistance symptom of Rvi6. Two conclusions could be drawn from this, i.e.: (i) molecular
markers are needed to identify the seedlings pyramiding Rvi4 and Rvi6, as the Rvi6 symptom
is masked in presence of the Rvi4 gene; (ii) it is however still quick and cost-efficient to
perform a traditional phenotypic screening that allows discarding about 25 % of the seedlings
before applying MASS for Rvi4 and Rvi6 on the reduced progenies (Bus et al., 2000).
Recently, Bus et al. (2009) described more in details the segregation of the Rvi2, Rvi4 and
Rvi6 genes in two progenies of MASS, namely A157R08T149 (Rvi6) × A068R03T057 (Rvi2)
22
and A176R02T281 (Rvi2) × A176R03T191 (Rvi4). In particular, the authors examined the
reliability of molecular markers based on their linkage disequilibrium (LD) with the resistance
genes in each progeny, as well as the efficiency of MASS in terms of number of marker data
point (MDP) required for identifying seedlings with pyramided genes. It appeared that the
percentages of recombination between the genes Rvi2 and Rvi4 and their respective markers
(L19SCAR marker or SSR marker CH05e03 for Rvi2; 8283SNP marker for Rvi4) in both
progenies were beyond 5 % (7 % to the least between CH05e03 marker and Rvi2), a rate
which may not be regarded as acceptable by breeders (Bus et al., 2000; Gardiner et al., 2007).
This finding would reinforce the necessity of developing more reliable markers (e.g. co-
segregating with the gene of interest) for optimal MASS efficiency, as it is already the case
with the Rvi6 gene for which co-segregating markers CH-Vf1 and Vfa2SCAR have been
developed (Huaracha et al., 2004; Vinatzer et al., 2004; Xu et al., 2001) and are being used
for MASS (Bus et al., 2009; Kellerhals et al., 2009a; Soufflet-Freslon, 2008). As long as such
markers are not available for Rvi2 and Rvi4 genes, using molecular markers flanking each of
these two genes on either side may be advisable to minimize the risk of selecting false
positives or negatives.
23
1.3. Fire blight, a major apple disease
1.3.1. History of the disease
Fire blight was the first disease shown to be caused by a bacterium, Erwinia amylovora. It
was observed for the first time in the Hudson Valley (New York, USA) in 1780. Its spread
within North America and from there to the rest of the World was previously summarized
(Bonn and van der Zwet, 2000; van der Zwet and Keil, 1979). As early as 1817, fire blight
was recognized as a major problem for apple and pear production in the USA. In the 19th
century, the disease spread with the planting of fruit orchards by settlers from New York to
Ohio, Indiana, Illinois (1840) and then to the Southern and Gulf Coast states and California
(1888). The disease appeared in the Rocky Mountains states in 1905, as well as in Oregon
(1908) and Washington (1915). The first observation of fire blight in Canada and Mexico
were done in 1904 and 1921, respectively. The first two reports of fire blight outside North
America were from Japan and New Zealand in 1903 and 1919, respectively. Australia and
South Korea are the two other countries of the Pacific region where fire blight has been
reported. From North America or New Zealand, fire blight spread to the UK and Northern
Europe in the late 1950s and more widely in Europe and to the Middle East in the 1960s. It
continues now to spread eastward across Europe and the Middle East (Jock et al., 2002). So
far, 46 countries have reported the presence of fire blight on their territory (Malnoy and
Aldwinckle, 2007). Over the last 100 years, several severe epidemics occurred also in North
America as well as outside and were previously summarized (Bonn and van der Zwet, 2000;
van der Zwet and Keil, 1979). More recently, in 2000 and 2007, major epidemics of fire
blight occurred in Switzerland. The epidemic of 2007 led to the removal of over 100 ha of
apple and pear trees and more than 30 million Swiss Francs were spent in the management of
the disease (Holliger et al. 2008; www.feuerbrand.ch).
1.3.2. Dispersal and host range
As stated by Vanneste (2000), fire blight appears to be a remarkable disease to growers and
scientists under various aspects. Although its causative agent E. amylovora has only a
moderate dispersal potential in comparison to other economically important crop pathogens
(such as Blumeria graminis and Puccinia graminis f. sp. tritici on wheat (McDonald and
Linde, 2002)), fire blight has spread from the USA to more than 40 countries including
geographically remote ones such as the UK, Israel, Japan and New Zealand. It is now known
that this dispersal was man-aided and due to the ability of the bacterium to survive in bud
24
woods or trees send by long-distance shipment (Bonn and van der Zwet, 2000). Presence of
endophytic E. amylovora inside apple and pear trees has been detected repeatedly over the
years, and it was shown that the bacterium could survive up to 2 years in healthy buds and be
still virulent on the susceptible pear cultivar ‘Bartlett’ (Thomson, 2000). This is also the
reason why fire blight is currently a contentious trade issue in countries of the Southern
Hemisphere free of the disease.
Erwinia amylovora infects around 200 species, exclusively in the Rosaceae (formerly
Rosaceae) family (van der Zwet and Keil, 1979; Vanneste, 2000). These plant species were
grouped together on the basis of their flower morphology and it was assumed that there must
be one or several common factors that make them susceptible to fire blight (Vanneste, 2000).
Besides the domesticated apple (M. × domestica Borkh.) and the European pear (Pyrus
communis), E. amylovora also infects fruit crops such as quince (Cydonia), loquat
(Eriobotrya japonica), blackberry or raspberry (Rubus spp.), as well as ornamental plants like
mountain ash (Sorbus), hawthorn (Crataegus), firethorn (Pyracantha), Cotoneaster, Photinia
and service berries (Amelanchier). All these ornamental and native forest species are
considered in several countries as important in rural economies, cultural heritage and
landscape ecosystems (Duffy et al., 2005); unfortunately for growers, they represent a major
source of inoculum, sometimes growing wild in the direct vicinity of orchards (Thomson,
2000).
1.3.3. Disease cycle
In orchards, at the beginning of spring, previous year’s cankers are the main source of
primary inocula (Fig. 1.3). From these cankers, bacteria and ooze (viable bacteria embedded
in a polysaccharide matrix) are disseminated by insects, wind or rain to flowers and shoots,
depending on the weather conditions. In flowers, bacteria are first transferred to the stigma.
Erwinia amylovora is not considered to be a good epiphyte under natural conditions, but the
exception seems to be the stigmatic surface in flowers of host plants (Thomson, 2000): in
orchards, the epiphytic phase of the disease cycle of fire blight consists in the multiplication
of the pathogen on the surface of the stigmas. On plants of the Rosaceae family, this surface
represents a nutrient-rich, hydrated habitat on which the E. amylovora populations can grow
to exceed 107 colony forming unit (cfu) per flower. Large epiphytic populations on stigmas
greatly increase the likelihood of pathogen cells sliding down the surface of the style into the
nectary, where infection occurs. The infection is usually first associated with water soaking
and formation of ooze before it reaches the pedicel. It appears that insects (honeybees, flies,
25
aphids, leafhoppers, pear psylla,…) and rain play an important role in secondary
dissemination of the bacterium, especially from flowers to other flowers. Ooze from newly
infected flowers, being viscous and sticky, is indeed suited for subsequent spread. Thompson
(unpublished data mentioned in Vanneste, 2000) reported that the incidence of flowers
colonized by E. amylovora can change from 0 to nearly 100 % in 2 days in an apple orchard if
the weather conditions are favourable and in presence of numerous oozing cankers.
Erwinia amylovora as an endophyte contrasts with most plant-pathogenic bacteria. Indeed,
the causative agent of fire blight induces what was described as an evolutive necrosis
(Vanneste and Eden-Green, 2000): the bacterium can progress internally in host plant tissues,
and if the host is susceptible (as it is the case for popular apple cultivars like ‘Gala’, ‘Fuji’ or
‘Braeburn’ and most pear cultivars), it can move rapidly from the flowers to the pedicel, the
twig and the main branch; it may even reach the trunk and the rootstock, killing the entire tree
within a single season. Erwinia amylovora however does not produce cell-wall degrading
enzymes and does not progress in the plant by dissolving the tissues. This is in contrast to
another necrotroph plant pathogenic bacterium, Pectobacterium carotovorum, responsible of
soft rot and black leg potato diseases. The symptom produced by the progression of the
necrosis in plant tissues is typical, i.e. the formation of the shepherd’s crook (shoot
recurvature) (Fig. 1.3). Other typical symptoms include flower necrosis, immature fruit rot,
profuse bacterial ooze and cankers on woody tissue. All these symptoms correspond to
various infection types, respectively: blossom blight for flower necrosis (infection through
nectarthodes in the nectarial cup of flowers), shoot blight for shepherd’s crook (infection of
young vegetative shoot tips), canker blight for woody tissues (infection of cortical and xylem
parenchyma around overwintering cankers), trauma blight (non-specific infection court;
occurs after traumatic events like hail, high winds or late frost) and rootstock blight (Norelli et
al., 2003; Thomson, 2000).
26
Fig. 1.3. Disease cycle of fire blight caused by the bacterium Erwinia amylovora. From
Norelli et al. (2003).
Dashed lines represent movement of bacteria and spread of disease within the plant, and solid lines represent
movement of bacteria outside the plant.
1.3.4. Epidemiology
Fire blight is often considered by apple and pear growers as a “capricious” disease (Vanneste,
2000). This statement refers to the sporadic occurence of fire blight epidemics, one of the
most striking epidemiological characteristics of the disease. This is in contrast to scab, the
major fungal disease in apple and pear, caused by Venturia inaequalis and Venturia pirina
respectively, which is a regular problem in orchards. Epiphytotics of fire blight occurred in
many countries where the disease was reported and they lead to important, even sometimes
huge crop and financial losses. Actually, the precise economic impact of fire blight is difficult
to estimate, as losses are not determined when they are low (low incidence and/or severity of
flowers infection), and as the removal of infected branches or trees following severe infection
affect the orchards productivity for several years. Weather conditions are of high importance
in the development of fire blight infections. The disease is generally more severe in warm,
humid areas than in cooler and/or dry areas (Bonn and van der Zwet, 2000). In addition, the
27
infection potential is strongly influenced by flower age, temperature and rainfall, which
regulate the size of E. amylovora populations (Pusey, 2000).
Another aspect of fire blight sporadicity is the sudden outbreaks of the disease in the absence
of any obvious source of inocula (Vanneste and Eden-Green, 2000). It was thought that such
unexpected outbreaks were linked to the capacity of E. amylovora to survive symptomless in
host plant tissues, e.g. buds or shoots, as previously mentioned. Contradictory studies have
been published on this topic (recalled by several authors (Billing, 2010; Thomson, 2000;
Vanneste and Eden-Green, 2000)). One hypothesis is that the pathogen could be present in
orchard or nursery healthy trees until it is triggered to cause infection by unknown factors.
However, evidence is still lacking that endophytic bacteria are able to cause outbreaks of fire
blight epidemics. More generally, it has been so far difficult to determine the origin of any
inoculum in fire blight outbreaks; variation in the incidence and severity of fire blight seems
to follow no clear pattern from season to season and orchard to orchard (Biggs et al., 2008).
New molecular tools of forensic plant pathology that are being developed for E. amylovora,
such as clustered regularly interspaced short palindrome repeats (CRISPR) (Rezzonico et al.,
2011) and variable number of tandem repeats (VNTR) (Dreo et al., 2011), may allow genetic
tracking of individual strains in the near future.
Last, but not least, rootstock blight is also a characteristic of fire blight epidemiology. It
occurs in the rootstock of a grafted apple tree, where bark cankers are initiated and can
subsequently girdle and kill the tree (Norelli et al., 2003; Steiner, 2000). The symptoms
associated to rootstock infections before death of the tree differ from the “classical” fire blight
symptoms and consist in delayed bud break, poor growth and early red colour of leaves in
autumn. Among all the types of fire blight infection identified by Steiner, it appears like the
most difficult to predict. Rootstock infections have been reported from the USA as well as
from France and Switzerland, especially in high-density apple orchards planted on susceptible
rootstocks (‘Malling 9’ (‘M.9’) and ‘M.26’ (Holliger et al., 2008; Momol et al., 1998; Steiner,
2000; Vanneste and Eden-Green, 2000)). The cause of rootstock blight is not always clear,
however four mechanisms have been proposed (Fazio et al., 2006; Momol et al., 1998;
Thomson, 2000): (i) infection through rootstock suckers; (ii) washing of bacteria from
infections in the aerial parts of the tree into the grafting point, down the trunk; (iii) internal
migration of bacteria from other infections in the tree; (iv) wounding of the rootstock by
boring insects.
28
1.3.5. The pathogen, Erwinia amylovora
1.3.5.1. Taxonomy
The genus Erwinia belongs to the Enterobacteriaceae family which comprises many species
pathogenic to animals, insects or plants (such as E. amylovora, E. papayae, E.
piriflorinigrans, E. pyrifoliae) or epiphytes (such as E. billingiae, E. tasmaniensis and E.
trachaephila). Erwinia amylovora colonies are white, domed and bright with a dense central
ring after 2 to 3 days of cultivation at 27 °C on King’s B medium. The growth of E.
amylovora under anaerobic conditions is slow, contrary to other enterobacteria. It is also
unable to transform nitrates into nitrites and nicotinic acid is necessary for its growth on
minimal medium (Paulin, 2000).
Based on 16S rRNA gene sequence analysis and hybridization kinetics (Kim et al., 1999) as
well as internal transcribed spacers (ITS) region and plasmid DNA analysis (McGhee et al.,
2002), E. pyrifoliae from South Korea was described as a pathogen closely related but distinct
from E. amylovora; it is primarily pathogenic to Asian or Nashi pear (Pyrus pyrifolia Nakai)
but it seems also able to infect the European pear (P. communis L.) and the domesticated
apple, although with a lower aggressiveness (Kim et al., 2001). It was recently reported that
E. pyrifoliae would be responsible for the bacterial shoot blight of pear (BSBP) observed in
Japan as early as 1981 and first attributed to Erwinia amylovora (Geider et al., 2009).
However, the disease symptoms caused by E. pyrifoliae are essentially indistinguishable from
those of E. amylovora infection, and both species might therefore be referred to as fire blight
pathogens (Smits et al., 2010c).
The comparison of the genomes of E. amylovora, E. pyrifoliae, E. tasmaniensis and E.
billingiae recently sequenced and annotated may shed light on the evolutionary relationships
between these species and may reveal what make them successful plant pathogens (E.
amylovora and E. pyrifoliae) or efficient epiphytes, potential candidates for biological control
(E. tasmaniensis and E. billingiae) (Kube et al., 2008; Kube et al., 2010; Park et al., 2011;
Powney et al., 2011; Sebaihia et al., 2010; Smits et al., 2010a; Smits et al., 2010b; Smits et al.,
2010c).
1.3.5.2. Diversity
Unlike Pseudomonas syringae or Xanthomonas campestris for instance, E. amylovora has
been considered as a homogeneous species based on biochemical, serological, metabolic and
protein electrophoresis characteristics (Paulin, 2000). There is so far no recognized pathovar
within E. amylovora (Paulin, 2000). In the last 15 years, several molecular and genomic
29
approaches were developed that shed light into the diversity of E. amylovora at the genetic
level (Geider et al., 2009; Jock et al., 2002; McGhee et al., 2002; Zhang and Geider, 1997).
To summarize, these studies led to the development of two major groups of E. amylovora
strains, Maloideae (whose taxa have been reclassified into subfamily Spiraeoideae) and
Rubus; the former group ‘Hokkaido’ (Momol and Aldwinckle, 2000) corresponds actually to
E. pyrifoliae (Geider et al., 2009).
The vast majority of E. amylovora strains isolated from nature have a stable, non-
transmissible plasmid called pEA29 (29 kb); its role has not been fully described, but it seems
to be involved in thiamine biosynthesis and it would contribute to the colonization of plant
tissues by bacteria (Mohammadi, 2010). Strains devoid of pEA29 are rare but were found in
different regions of the world with fire blight infections (Llop et al., 2006). Some strains have
additional plasmids of variable size, from 1.7 to 72 kb (Llop et al., 2008; Llop et al., 2006;
Sebaihia et al., 2010). The role of these plasmids is largely unknown. Two RSF1010-based
plasmids were assigned the function of streptomycin resistance (Palmer et al., 1997). Llop et
al. (2008) reported the presence of one particular plasmid, pEI70, in 10 countries in Europe; it
appeared to be widespread among the European E. amylovora strains screened by Llop and
colleagues (Llop et al., in preparation).
1.3.6. Non-genetic measures of fire blight management
In countries like Switzerland where fire blight is present although not in every region, the
general strategy to control fire blight relies heavily on containment (i.e. preventing the entry
of fire blight into disease-free zones) and eradication (when single foci of fire blight are
detected in a commune) (Duffy et al., 2005). In regions where fire blight is established,
management strategies at the orchard scale are based on sanitation which aims at (i) reducing
the number and distribution of primary and secondary inocula; (ii) preventing blossom
infections; and (iii) reducing shoot blight (http://www.caf.wvu.edu/kearneysville/articles/FB-
MANAGE00.html). In the USA, the reduction of the primary inoculum is mostly achieved by
removing overwintering cankers during dormant pruning. Pruning the early-season infections
after bloom also helps to reduce the amount of inoculum available for shoot infection (Norelli
et al., 2003; E. Holliger, personal communication). As the establishment of epiphytic
populations of E. amylovora on the stigmas is critical in blossom infections, prevention of
infection by a series of sprays is essential. Most conventional spray programs make treatments
at regular 4- to 7- days intervals during the bloom period, using copper-based compounds or
antibiotics (Steiner, 2000). The timely application of sprays was improved in the last ten years
30
thanks to the use of forecasting models able to predict the occurrence of fire blight infection
periods during bloom (e.g. Maryblyt™ (Steiner, 1990)). Copper-based compounds show a
bacteriostatic activity that reduces the efficacy of the remaining inoculum; they are efficient
only if applied preventively at green tip or after and can be phytotoxic if applied too late,
leading to fruit russetting (Turechek, 2004). Antibiotics prevent bacterial multiplication by
acting on protein synthesis. Streptomycin is the most efficient antibiotic against E. amylovora:
properly timed applications of streptomycin during bloom can provide over 90 % control
against sensitive strains of the pathogen (Norelli et al., 2003). However, streptomycin
resistance has developed in E. amylovora populations in the Pacific Northwest, California,
and Michigan (USA) (Coyier and Covey, 1975; McGhee et al., 2011; Miller and Schroth,
1972; Russo et al., 2008). This antibiotic is allowed for use in North America, Israel, New
Zealand and the Netherlands and with restrictions in Germany, Austria and Switzerland since
2008. The other antibiotic commonly utilized in the USA, oxytetracyclin, is less efficient than
streptomycin when applied on streptomycin-sensitive E. amylovora strains (Stockwell et al.,
2008). The potential of a third antibiotic, kasugamycin, in fire blight management has been
investigated; however, its application alone may lead to the spontaneous development of
resistant strains (McGhee and Sundin, 2011) and its compatibility with biological control
agents requires further investigation (Johnson et al., 2008). Reducing shoot blight by using
streptomycin is ineffective and copper formulations are potentially phytotoxic
(http://www.caf.wvu.edu/kearneysville/articles/FB-MANAGE00.html). Therefore, reduction
of shoot blight is mostly performed by the application of plant growth regulators such as
Regalis® (BASF, Ludwigshafen, Germany) or Apogee
™ (BASF). Both products are based on
prohexadione-calcium and inhibit gibberellic acid synthesis, thus reducing shoot growth;
controlling shoot growth proved to reduce the incidence and severity of fire blight shoot
infection (Norelli et al., 2003). Inducers of systemic acquired resistance (SAR) are also used
to control shoot blight, e.g. Bion® (Syngenta, Dielsdorf, Switzerland) (Norelli et al., 2003).
Bion® contains the chemical acibenzolar-S-methyl (ASM) that was shown to provide a
significant level of fire blight control under very favorable conditions (Brisset et al., 2000).
The emergence, or the anticipation of the emergence, of streptomycin-resistant strains of E.
amylovora has driven investigations into the effectiveness of microbial antagonists as
biological control (syn. biocontrol) agents targeting the blossom blight phase. In the USA,
three main biocontrol products based on bacterial strains of antagonists to E. amylovora are
commercialized. The first antagonist registered for use was Pseudomonas fluorescens strain
A506 under the product name BlightBan® A506 (Nufarm Americas Inc., Burr Ridge, Illinois,
31
USA); it suppresses the fire blight disease by excluding E. amylovora from the flowers.
Pantoea vagans strain C9-1 is commercialized under the product name BlightBan® C9-1
(Nufarm Americas Inc., Burr Ridge, Illinois, USA) and produces the antibiotics herbicolin O
and I with antibiotic activity against E. amylovora (Ishimaru et al., 1988). Another species of
the genus Pantoea, P. agglomerans strain E325, is the active ingredient in Bloomtime
Biological™
(Northwest Agricultural Products, Wenatchee, Washington, USA) and secretes a
yet unknown compound that inhibits E. amylovora (Pusey et al., 2008). Many other strains of
P. vagans and P. agglomerans have been shown to produce antibiotics in culture that inhibit
growth of E. amylovora, such as P. agglomerans strain Eh252 (Stockwell et al., 2011). In
Europe, three fire blight biocontrol products are registered for use: Serenade® and Biopro
® are
based on Bacillus subtilis strains QST 713 and BD170, respectively (Broggini et al., 2005;
Vanneste, 2011), and Poma Vita™
(distributed under the name Blossom Bless®
in New
Zealand) is based on P. agglomerans strain P10c (Vanneste, 2011). However, several studies
have pointed out the variability of efficacy of existing biological control options (Johnson and
Stockwell, 1998; Sundin et al., 2009; Vanneste, 2011). Therefore, it may be advisable to use a
combination of several antagonistic organisms, possibly coupled with compatible antibiotics
just before predicted blossom infection events (Stockwell et al., 2008; Stockwell et al., 2010;
Stockwell et al., 2011).
On the basis of these observations, it was concluded that an integrated management of fire
blight, combining all adequate control methods, is most likely to succeed in reducing the
damages down to an economically acceptable level (Johnson, 2010; McManus et al., 2002;
Norelli et al., 2003; Psallidas and Tsiantos, 2000).
32
1.4. Genetics and breeding of fire blight resistance in apple
This section deals with the genetics and breeding of fire blight resistance in apple scion
cultivar. The topics of germplasm sourcing, conventional breeding, development of molecular
markers and biotechnological strategies for fire blight resistance have been previously
reviewed (Aldwinckle and Beer, 1979; Khan et al., 2011; Lespinasse and Aldwinckle, 2000;
Malnoy and Aldwinckle, 2007; Peil et al., 2009; van der Zwet and Keil, 1979). Therefore,
only the key points related to these topics will be summarized here .
1.4.1. Context and challenges of fire blight resistance breeding
1.4.1.1. Fire blight resistance as breeding objective
The importance of fire blight resistance as breeding objective is relatively low compared to
fruit quality; in fact, it seems variable, depending on the incidence and severity of the disease
in the region where the breeding program is located. Fire blight resistance is not a high
priority in many international breeding programs, nevertheless some of them have the
development of fire blight resistant cultivars as an objective (e.g. in the USA, Canada, New
Zealand, Germany, Italy, Poland, France and Switzerland) (Laurens, 1999; Peil et al., 2009).
It has to be noticed that the term fire blight “resistance” may lead to confusion. In this Thesis,
it should be understood as the quantitative limitation of bacterial growth in infected plants at
variable degrees (Cooper and Jones, 1983; Poland et al., 2009; Vergne et al., 2010); it will not
be used as a synonym of “tolerance”, which may refer to mild disease symptoms and minor
decreases of plant productivity despite unrestricted bacterial invasion in plant tissues (Cooper
and Jones, 1983).
Unlike pear breeding, the apple scion breeding programs in North America have not aimed
primarily at combining a high fruit quality and productivity with fire blight resistance (van der
Zwet and Keil, 1979); the apple scion breeders have rather included fire blight resistance as
an important but not essential objective, in addition to scab and powdery mildew resistance
considered as more important economically (Lespinasse and Aldwinckle, 2000; van der Zwet
and Keil, 1979). As rootstock blight is a major cause of concern for apple growers in the
Northeastern USA (Norelli et al., 2003; Turechek, 2004), the apple rootstock breeding
program at the New York State Agricultural Experiment Station (Cornell University; and
since 1998 collaboration Cornell University-USDA) in Geneva (New York, USA) has been
aiming since 1968 at selecting apple rootstocks combining excellent orchard performance
33
with diseases resistance, and especially fire blight resistance (Cummins and Aldwinckle,
1995; Johnson, 1999)
In Europe and New Zealand, most apple scion breeders have been considering fire blight
resistance as a helpful objective of apple improvement. Fire blight is considered as of minor
economical importance in England and France (E. Billing and J.-P. Paulin, personal
communications cited in Bonn and van der Zwet, 2000), unlike in the regions Baden-
Württemberg and Rheinland-Pfalz in Germany where it is regarded as an economical problem
((Bonn and van der Zwet, 2000); Moltmann, personal communication cited in Peil et al.,
2009). Some programs have intensified their efforts for breeding fire blight resistant apple
scions following disease epidemics or spread in the country where they are located, e.g. in
France (Lespinasse and Paulin, 1984) and more recently in Switzerland (Kellerhals et al.,
2011). One noticeable program with the explicit essential aim of breeding apple varieties with
high fruit quality and resistance to fire blight, scab and powdery mildew was conducted at the
Institute for Fruit Research at Dresden-Pillnitz (Germany (Fischer and Fischer, 1999)).
1.4.1.2. Challenges in fire blight resistance breeding
When breeding apples for multiple diseases and pests resistance, a balance of resistance and
commercial fruit quality may be difficult to achieve. Adequate levels of resistance to scab,
powdery mildew and fire blight require a large number of seedlings to improve the chances of
meeting all selection criteria, which makes apple breeding more labour intensive (Gardiner et
al., 2007; Hancock et al., 2008). Moreover, numerous sources of fire blight resistance are, as
for scab and powdery mildew, small fruited wild apple species (e.g. M. baccata, M. fusca, M.
× atrosanguinea, M. × robusta (Aldwinckle and van der Zwet, 1979; Peil et al., 2009)),
ornamental crab apples (e.g. ‘Adams’, ‘Adirondack’, ‘David’ and ‘Sentinel’ (Bell et al., 2000;
Bonn and Elfving, 1990; van der Zwet and Keil, 1979)) or landraces (e.g. ‘Heimenhofer’,
‘Ohio Reinette’, ‘Schneiderapfel’ and ‘Waldhöfler’; David Szalatnay, personal
communication) characterized by a poor fruit appearance and/or quality. Several generations
of pseudo-backcrossing of such donor material with high quality recurrent parents are usually
required before seedlings with disease resistance and sufficient fruit quality can be considered
as new potential commercial cultivars (Lespinasse and Aldwinckle, 2000). Norelli et al.
(2003) observed that fire blight resistant apple rootstocks were developed much quicker than
apple scion cultivars partly because fruit quality was precisely not an issue for rootstock
breeding.
34
A second impediment to improve fire blight resistance is the nature of the trait itself. High
levels of fire blight resistance exist in Malus germplasm but no complete immunity to the
disease has been reported (Lespinasse and Aldwinckle, 2000; Peil et al., 2009). Besides, no
major gene for fire blight resistance has been identified yet in Malus germplasm, contrary to
apple scab or powdery mildew for which numerous genes, some involved in gene-for-gene
relationships, have been reported (Bus et al., 2011; Gardiner et al., 2007; Gessler et al., 2006).
Actually, the resistance to fire blight in Malus spp., and especially in M. × domestica
(Borkh.), seems to be mostly partial, continuously distributed in F1 progenies and of oligo- or
polygenic inheritance (Fazio et al., 2008; Gardner, 1976; Hampson and Sholberg, 2008;
Korban et al., 1988; Lespinasse and Paulin, 1990). Such a resistance is difficult to work with
for a breeder as there is no clear threshold to differentiate “resistant” and “susceptible” plants.
Developing markers associated to fire blight resistance for use in marker-assisted breeding
(MAB) may be achieved by a candidate gene approach (Baldo et al., 2010; Norelli et al.,
2009a), or most probably, by a QTL mapping approach (Calenge et al., 2005; Durel et al.,
2009; Khan et al., 2006; Peil et al., 2007). However, a QTL mapping experiment is relatively
expensive and time-consuming, necessitating on the one hand the screening of sufficient
reproducible molecular markers to cover the apple genome for mapping (at least 80 well-
spanned markers on at least 100 F1 progeny plants) and on the other hand the artificial
inoculation of hundreds of plants for fire blight severity scoring, each F1 progeny plant being
replicated 5 to 15 times to guarantee the precision of scoring (Khan et al., 2007; Peil et al.,
2009).
A third major issue in testing fire blight resistance for breeding purposes is the nature of the
bacterial inoculum. Indeed, the response of a genotype, for instance a breeding selection or a
pre-breeding genotype (genotype used as a parent for cultivar breeding) to various bacterial
strains after inoculation may give information on the durability of its resistance. Erwinia
amylovora strains showing differential virulence have been identified in North America but
not in Europe (Lespinasse and Paulin, 1990; Norelli and Aldwinckle, 1986; Norelli et al.,
1987); it was also found that strains vary in aggressiveness (Korban et al., 1988; Norelli et al.,
1987; Paulin and Lespinasse, 1990; Sobiczewski et al., 2008; Taylor et al., 2002; Wang et al.,
2010). Therefore, it has been proposed to use a mixture of strains for screening seedlings for
fire blight resistance, as this would prevent the risk to inoculate only a poorly aggressive
strain (Norelli et al., 1987). The mixture of strains could be modified with time, i.e. strains
showing low aggressiveness could be replaced by more aggressive ones. It has also been
proposed instead to inoculate successively several individual strains to avoid potential
35
interactions resulting from a mixture of all strains (e.g. competition for infection sites or
antagonisms between different strains (Paulin and Lespinasse, 1990)). For inoculation of
field-grown selections, the use of individual, highly aggressive strains showing differential
virulence to specific cultivars has been proposed (Norelli et al., 1987).
1.4.2. Conventional breeding for fire blight resistance
1.4.2.1. Sourcing fire blight resistance in Malus germplasm
As recently pointed out by Peil et al. (2009), several germplasm screening studies have been
published since the review by van der Zwet and Keil (1979). They deal with different plant
material such as cultivars, landraces, M. sieversii as well as other plant material from Central
Asia, ornamental crab apples and small fruited wild apple species. Although no standardized
inoculation and scoring procedure was used, it appears that high levels of fire blight resistance
are found generally in small-fruited wild apple species or in some ornamental crab apples
(Aldwinckle et al., 1999; Bell et al., 2000; Bonn and Elfving, 1990; Norelli et al., 1987;
Paulin and Lespinasse, 1990; Peil et al., 2004; Peil et al., 2009); various levels of resistance
are present in the cultivated apple (Aldwinckle et al., 1999; Kása et al., 2004; Khan et al.,
2007; Korba et al., 2008; Le Lézec et al., 1997; Luby et al., 2002; Szalatnay et al., 2009) and
in its wild relatives M. sieversii and M. orientalis (Fazio et al., 2009; Forsline and
Aldwinckle, 2004; Forsline et al., 2008; Momol et al., 1999; Volk et al., 2008).
These sources of fire blight resistance have been of interest for the apple breeding community
for different reasons. First, the high level of fire blight resistance found in small-fruited wild
apple species (e.g. Malus baccata, M. fusca, M. × atrosanguinea, M. × floribunda, M. ×
prunifolia and M. × robusta) have been very appealing to scion breeders for which high level
of fire blight resistance is a major objective (Baumgartner et al., 2011; Lespinasse and Paulin,
1990; Peil et al., 2009). Second, the existence of various levels of resistance to fire blight in
M. × domestica Borkh. has implied that new cultivars with increased resistance could be
obtained by crossing one already existing tolerant cultivar with another tolerant cultivar or
even with a susceptible cultivar showing outstanding fruit quality (Fischer and Richter, 1999).
Thereby, the long process of pseudo-backcrossing over several generations may not be the
only route to new fire blight resistant apple cultivars. Third, the relatively recent identification
of interesting levels of fire blight resistance in landraces or in M. sieversii (M. orientalis to a
lesser extent) enlarge the pool of resistance donors within Malus spp. that would likely
necessitate few pseudo-backcrosses (Baumgartner et al., 2011).
36
1.4.2.2. Apple scion breeding for fire blight resistance
Apple scion breeding programs aim at improving in-plant resistance to fire blight, although
selection for fire blight avoidance (e.g. absence of secondary flowering) might be a possibility
as demonstrated in pear breeding (Alston, 1994). In the breeding program at the research
station of Cornell University (USA), selection for fire blight resistance used to be done at an
early stage by inoculating 50–90 cm tall seedlings with E. amylovora. It has been more
recently performed at a later stage of the breeding process to be able to make replications for
each F1 individual for more precision (Lespinasse and Aldwinckle, 2000). Moderately
resistant apple cultivars (e.g. ‘Delicious’) were used as sources of resistance (Aldwinckle and
van der Zwet, 1979). In Dresden-Pillnitz (Germany), seedlings were also screened at an early
stage in the field, i.e. in the first year after sowing, with a mixture of three aggressive isolates.
The donors of resistance have been mostly clones of M. × floribunda, M. × robusta as well as
recent “Pi-“ and “Re-cultivars®” released from the breeding program (Fischer and Richter,
1999). At the INRA of Angers (France), the fire blight resistance breeding program was
originally divided into two parts, i.e. a long term and a short term breeding program; the first
one was designed to use as donor of resistance M. × robusta ‘Robusta 5’, the second one scab
resistant cultivars or selections also carrying some level of fire blight resistance. The long-
term breeding program has been abandoned due to the susceptibility of ‘Robusta 5’ to the E.
amylovora strain Ea 266 (Lespinasse and Paulin, 1990). On the contrary, the short-term
programme has been continued, the scab resistant accessions being not more susceptible to Ea
266 than to other strains (Lespinasse and Paulin, 1990). The most promising selections of the
main apple breeding program (not necessarily bred for fire blight resistance in particular) are
generally evaluated for their low level of fire blight susceptibility after selection for
agronomical performance at the end of the selection process. They are challenged with
different strains of E. amylovora under glasshouse conditions (Y. Lespinasse, personal
communication). At the research station Agroscope ACW (Switzerland), phenotypic screening
of promising advanced selections for fire blight resistance is performed in quarantine glasshouse
by syringe inoculation of young shoots (Kellerhals et al., 2008). Considerable differences of
necrotic lesion length were observed among selections (Kellerhals et al., 2008; Kellerhals et
al., 2011). Two of the most resistant advanced selections derived from a cross between two
susceptible cultivars, ‘Topaz’ and ‘Fuji’, meaning that transgressive segregation of the fire
blight resistance trait can be exploited by breeders. Genotypes with low fire blight severity
over several years are preferentially selected for use as parents in the breeding program
(Kellerhals et al., 2011).
37
1.4.3. Development of molecular markers: towards MAB for fire blight resistance
1.4.3.1. QTL mapping and MAS
Calenge et al. (2005) investigated the genetic basis of fire blight resistance in two F1
progenies derived from crosses between the cultivars ‘Fiesta’ and either ‘Discovery’ or
‘Prima’. Both progenies were inoculated in glasshouse with the strain CFBP 1430 of E.
amylovora (107 cfu/ml), and the length of necrosis was scored 7 and 14 days after inoculation.
A major QTL explaining 34.3–46.6 % of the variation of necrosis length in the offspring
(phenotypic variation) was identified on LG 7 of ‘Fiesta’ (F7 QTL) in both progenies at the
same genetic position. Using the Swiss strain Ea 610 for plant inoculation in glasshouse (109
cfu/ml), Khan et al. (2006) calculated the area under the disease progress curve (AUDPC) and
confirmed the F7 QTL in a Swiss ‘Fiesta’ × ‘Discovery’ F1 progeny (37.5–38.6 % of the
phenotypic variation explained). Subsequently, Khan et al. (2007) transformed two RAPD
markers bracketing the QTL into the SCAR markers AE10-375 and GE-8019 and developed a
SSR marker specific for the region. Markers enabled tracking of the F7 QTL allele of ‘Fiesta’
back to ‘Cox’s Orange Pippin’ by pedigree analysis. Stability of the effect of the F7 QTL was
checked by inoculating progeny plants of a cross between ‘Milwa’, a susceptible cultivar, and
‘1217’, a moderately resistant F1 offspring of the Swiss cross ‘Fiesta’ × ‘Discovery’ carrying
both flanking SCAR markers. Calenge et al. (2005) and Khan et al. (2007) both considered
this QTL as promising for MAS, given its stability, its presence in cultivars of good fruit
quality and the availability of reproducible flanking markers. Baumgartner and colleagues
described the use of the F7 QTL in the breeding program of Agroscope ACW (Switzerland)
(Baumgartner et al., 2010). Selected progeny plants of the cross ‘Milwa’ × ‘Enterprise’ were
tested for presence of the SCAR markers AE10-375 and GE-8019 present in ‘Enterprise’
(MASS). The same selected plants were subsequently inoculated with E. amylovora in
quarantine glasshouse for evaluation of their level of fire blight resistance. It appeared that
seedlings of the cross ‘Milwa’ × ‘Enterprise’ carrying the SCAR markers were on average
significantly less susceptible than seedlings lacking the markers. Pyramiding several QTLs
showing each an intermediate phenotypic effect like the F7 QTL may provide a durable level
of fire blight resistance (Peil et al., 2009). In addition to the F7 QTL, Calenge et al. (2005)
detected four minor QTLs on LGs 3, 12 and 13, as well as digenic interactions in the crosses
between the cultivars ‘Fiesta’ and either ‘Discovery’ or ‘Prima’.
Peil et al. (2007) reported a QTL mapping approach for resistance to E. amylovora in a cross
between the susceptible cultivar ‘Idared’ and the highly resistant hybrid species Malus ×
robusta, clone ‘Robusta’ 5 (‘R.5’). Artificial shoot inoculation was performed on a
38
segregating F1 progeny of 146 individuals with the strain Ea 222 (109
cfu/ml). A major QTL
explaining about 80 % of the phenotypic variation was mapped in the proximal part of LG 3,
between the SSR markers CH03g07 and CH03e03. Shortly after, Peil and colleagues
confirmed this QTL by artificial glasshouse inoculation of two F1 progenies, one being the
original F1 progeny ‘Idared’ × ‘R.5’ and the other being a ‘M.9’ × ‘R.5’ family from New
Zealand (Peil et al., 2008). Although the strains inoculated differed between the progenies (Ea
222 for ‘Idared’ × ‘R.5’, 11176 for ‘M.9’ × ‘R.5’), a major QTL was detected at the same
genetic position (proximal part of LG 3), explaining up to 67 % and 83 % of the phenotypic
variation, in both respective progenies. Fazio et al. (2008) studied the inheritance of fire blight
resistance in a F1 progeny derived from a cross between ‘Ottawa 3’ and ‘R.5’. By inoculating
170 F1 individuals with two E. amylovora strains (E2002a and Ea 273), the authors observed
two different distributions of the mean percent lesion length: 147 F1 individuals showed no
necrotic lesion with strain Ea 273 against 47 with strain E2002a. These results have thus re-
opened the debate about the genetic basis of fire blight resistance in Malus × robusta clone
‘R.5’ (Aldwinckle and Beer, 1979; Gardner, 1976; Peil et al., 2007). Further QTL analysis on
the F1 progeny from ‘Ottawa 3’ × ‘R.5’ using the two strains E2002a and Ea 273 may provide
additional information on this (Fazio et al., 2008). Durel et al. (2009) identified by QTL
mapping a genomic region associated with fire blight resistance on the distal part of LG 12 in
the wild apple Malus × floribunda clone 821 and in the ornamental cultivar ‘Evereste’,
explaining 40 % and 70 % of the phenotypic variation, respectively; another minor QTL was
detected in ‘Evereste’ on LG 5. The E. amylovora strain, the inoculation procedure and the
scoring scale used were according to Calenge et al. (2005). Recently, the map position of
‘Evereste’ QTL could be determined more precisely using the BSA technique coupled with
AFLP markers, followed by the development of specific and reproducible SCAR markers
flanking the QTL peak on either side (Parravicini Rusca, 2010); the genomic region of the fire
blight resistance QTL was named Fb_E, according to the phenotypic (resistant/susceptible F1
individual) marker mapping at the distal end of LG 12 (Durel et al., 2009). Baumgartner et al.
(2011) performed a cross between the advanced selection ‘ACW 11303’ (carrying Rvi4 and
Rvi6) and ‘Evereste’ (carrying Rvi6 and Fb_E) with the aim of combining a major QTL for
fire blight resistance with scab resistance genes. In a first step, the progeny plants were
phenotypically tested for scab resistance with a local field inoculum, and 34 out of 38 were
found scab-resistant. In a second step, MASS of scab-resistant seedlings revealed that 2
seedlings were carrying the marker for Fb_E (SCAR M45TA-403c) as well as the markers for
Rvi6 (SSR CH-Vf1) and Rvi4 (SSR CH02c02), one of the two carrying the Rvi6 marker in a
39
homozygous state. Both seedlings were chosen as parents of the next cross to continue the
introgression of the Fb_E locus and simultaneously the pyramiding of the Rvi4 and Rvi6
genes. Several other projects of genetic mapping of fire blight resistance are underway,
involving wild apple species such as M. baccata, M. fusca, M. prunifolia and M. hupehensis
(Andreas Peil, personal communication; (Velasco, 2010)). The positions of QTLs for fire
blight resistance identified so far on the apple genome are summarized on Fig. 1.4.
Fig. 1.4. Global position of quantitative trait loci (QTL) for fire blight resistance on the apple
genome based on the backbone genetic map developed by Silfverberg-Dilworth et al. (2006).
Only the linkage groups that have been shown to carry QTL are presented. From Peil et al.
(2009).
major QTLs (R = ‘Robusta 5’, F = ‘Fiesta’; E/Mf: ‘Evereste’/Malus × floribunda clone 821).
minor QTLs (LG 3: at the top of LG 3 from ‘Fiesta’ and at the bottom of LG 3 from ‘Prima’; LG 12:
from ‘Discovery’; LG 13: from ‘Discovery’).
1.4.3.2. Candidate gene approaches
In addition to the QTL mapping approach, functional genomics approaches called transcript
profiling were also undertaken to develop molecular tools to breed fire blight resistant apple
cultivars. The suppression substractive hybridization (SSH) technique was applied to
characterize the transcriptional response of the cultivar ‘Gale Gala’ (a sport mutation of
‘Gala’) to infection by E. amylovora over 72 hours (Norelli et al., 2009a). Thus, 468 apple
CH03e033
CH03g0724
Hi03d0643
AU22365776
CH03g12y111
LG3
CN4447945
Hi03a10p26
CH04e0535
GE80-1955
Hi05b0965
LG7
CH05d0413
CH04g0429
CH01g1245
CH03c0261
Hi07f0176
CH01d03z89
LG12
CH05h053
Hi04g0518
CH02g0140
NH009b70
CH05f0489
LG13
R
F
E/Mf
40
ESTs were shown to be up- or down-regulated during fire blight challenge. A complementary
cDNA-AFLP analysis was performed to gain insight into the transcriptional response of
resistant (‘M.41’) and susceptible (‘M.26’) Malus genotypes to artificial fire blight infection
(Baldo et al., 2010). A total of 190 ESTs differentially expressed between ‘M.41’ and ‘M.26’
were identified that are involved in recognition, signalling, defense and apoptosis. The
identification of ESTs associated with fire blight response from E. amylovora–challenged
apple leaf tissue was followed by a bioinformatics step, which consisted in ranking the ESTs
for their potential impact on fire blight resistance based on comparison with model systems
(e.g. Arabidopsis thaliana ESTs identified in response to bacterial challenges) (Norelli et al.,
2009b). Then, SSR or SNP markers derived from the ESTs most likely playing a role in fire
blight resistance were mapped on the apple genome using a ‘M.9’ × ‘R.5’ F1 progeny (Celton
et al., 2009b). Among the 28 candidate fire blight resistance genes that were mapped, one (a
secretory class III peroxidase homologous with Arabidopsis thaliana Per53) co-localized with
the major fire blight resistance QTL of ‘R.5’ on LG 3 while another (a putative disease
resistance protein) was mapped to a position corresponding to the location of the F7 QTL.
These markers are in the process of being evaluated for application in MAB (Norelli et al.,
2009b).
1.4.4. Genetic engineering for fire blight resistance in Malus spp.
Genetic engineering or recombinant DNA (rDNA) technology is considered as a promising
strategy to introduce resistance to pathogens in established apple cultivars for two main
reasons: (i) it allows the targeted improvement of an elite cultivar for one trait while
preserving its genetic background and thus the associated agronomical characteristics; (ii) it
avoids some major bottlenecks of traditional breeding such as the long juvenile phase, the
complicated segregation of traits and the presence of unwanted traits carried along the
introgression of a resistance gene/QTL (Chevreau, 2009; Malnoy and Aldwinckle, 2007). Fire
blight resistance was probably the first and most important target of rDNA technology in
apple. Three main strategies have been used so far to improve the resistance to fire blight by
rDNA technology (Gessler and Patocchi, 2007; Malnoy and Aldwinckle, 2007; Peil et al.,
2009):
- production of antimicrobial proteins. Three categories of antimicrobial proteins were
transformed in apple cultivars: first cecropins, which are antibacterial proteins found in the
hemolymph of Hyalophora cecropia pupae (cultivars transformed: ‘Gala’, ‘Royal Gala’ and
‘Galaxy’, the last two being sport mutations of ‘Royal Gala’ (Aldwinckle et al., 1996)); then
41
attacins, which are antimicrobial peptides produced by H. cecropia in response to bacterial
infection (‘Galaxy’ (Hanke et al., 2000)); and also lysozymes, which are bacteriolytic
enzymes characterized from phages, bacteria, fungi, plants, and animals (‘Pinova’, ‘Elstar’
and ‘Galaxy’ mostly (Ko et al., 1998; Ko et al., 2002)). Interestingly, Borejsza-Wysocka and
colleagues reported that the gene attacin E, transformed in the cultivar ‘Galaxy’, showed a
stable expression and conferred a stable, increased fire blight resistance over 12 years in the
field compared to the untransformed cultivar ‘Galaxy’, without affecting fruit characteristics
or tree morphology (Borejsza-Wysocka et al., 2010).
- inhibition of bacterial pathogenicity factors. The studies conducted to inhibit pathogenicity
factors of E. amylovora have aimed at the degradation of the extracellular polysaccharide
amylovoran by expression of depolymerase genes from phages (Borejsza-Wysocka et al.,
2007; Hanke et al., 2002; Hanke et al., 2000; Sule et al., 2002).
- promoting natural plant defense. This was first performed by expressing an elicitor that
induces plant defense mechanisms, e.g. the E. amylovora effector HrpNEa (‘Galaxy’ (Abdul-
Kader et al., 1999)) or MpNPR1, an ortholog of NPR1 in M. × domestica Borkh. that is a key
mediator of systemic acquired resistance (SAR) in A. thaliana (‘Galaxy’ (Malnoy et al.,
2007)). Promoting apple defense to fire blight was also achieved by silencing either the genes
encoding the four DspE-Interacting Proteins of Malus (DIPM genes) (Borejsza-Wysocka et
al., 2004) or the gene encoding the HrpN-Interacting Protein of Malus (HIPM) (Malnoy et al.,
2008). Last but not least, the constitutive over-expression of the TIR-NBS LRR gene mbr4
(Toll interleukin 1 receptor–nucleotide binding site–leucine rich repeat) of M. baccata in the
cultivar ‘Pinova’ also led to an increased resistance to fire blight (Flachowsky et al., 2008).
Recently, the identification and isolation of genes from Malus spp. conferring resistance to
fire blight was undertaken. Parravicini et al. (2011) used the two SCAR markers flanking the
LG 12 QTL Fb_E of the apple genotype ‘Evereste’ as starting points for the positional
cloning of this QTL; after narrowing down the fire blight resistance locus to a region of 78 kb,
they identified two genes as the most probable candidates showing homology with the Pto/Prf
complex involved in the resistance to Pseudomonas syringae pathovar tomato DC3000 in
tomato (Solanum lycopersicum L.) (Chang et al., 2002; Zhou et al., 1997). Similarly,
Fahrentrapp and collaborators reported the fine mapping of the major fire blight resistance
QTL of Malus × robusta ‘R.5’ located at the top of LG 3 (Fahrentrapp et al., 2011). They
determined a small window of less than 1 cM encompassing the resistance locus that is
currently being investigated with the objective of cloning the ‘R.5’ QTL. Together with the
development of “clean vector technologies” (Rommens, 2004), these QTL cloning projects
42
are expected to pave the way to the creation of cisgenic fire blight resistant apple cultivars,
i.e. cultivars which, at the end of the transformation process, contains only genes from Malus
spp. in a sense orientation and controlled by their own promoters and terminators (Joshi,
2010; Schouten et al., 2006; Vanblaere et al., 2011).
1.4.5. Fast breeding strategy: status and prospects
1.4.5.1. Juvenile phase and transition to flowering: definitions
According to Hanke et al. (2007), the life cycle of a plant can be divided into two parts: the
juvenile (syn. vegetative) phase and the adult (syn. generative, reproductive) phase (Fig. 1.5).
The end of the juvenile period is indicated by the attainment of the ability to flower and the
actual production of flowers is the first evidence that a plant is in the adult phase. However,
the end of the juvenile period and the first appearance of flowers may not coincide. Indeed,
the time of transition from one phase to the other is influenced by endogenous (gibberellic
acid, developmental stage) but also environmental (cold treatment, photoperiod) factors which
trigger or repress the change of the shoot meristem from generating leaves to the development
of reproductive organs; even though the seedlings have attained the ability to flower, they
might not flower because of these environmental factors (Hanke et al., 2007). In apple, it was
proposed to name transition phase the intervening period between the juvenile and the adult
phase (Zimmerman, 1973); this period of transition is also defined as the adult vegetative
phase (Poething, 1990) (Fig. 1.5). The adult vegetative phase is an important phase for
breeders, as during this phase most agro-technical floral-inducing methods are applied
successfully.
43
Fig. 1.5. Schematic illustration of the ontogenic phases of development in Malus spp.. From
Hanke et al. (2007).
1.4.5.2. Shortening the juvenile phase
Because the long juvenile phase of the domesticated apple severely limits selection efficiency
and makes very time-consuming the continuation of the breeding process through recurrent
selection or pseudo-backcrosses, shortening the juvenile stage has been heavily investigated
as breeding method, first using chemical or physical methods (Meilan, 1997) and later using
transgenic approaches (Flachowsky et al., 2009; Hanke et al., 2007).
1.4.5.2.1. Agro-technical approaches
In the domesticated apple, early flowering genotypes (i.e. flowering less than 3 years after
sowing) are unavailable. Several agro-technical approaches have been therefore tested with
the aim of reducing the duration of the juvenile phase. A variety of techniques can promote
flowering induction of trees when applied during the adult vegetative phase, including trunk
ringing, bark scoring and inversion, root pruning, light/temperature supplementation and
fertilization, defoliation and placing shoots in a horizontal position (Aldwinckle, 1976;
Flachowsky et al., 2009; Meilan, 1997; Visser, 1964). A frequent method to accelerate the
44
onset of flowering in apple consists in grafting seedlings onto dwarfing rootstocks such as
‘M.9’; this results usually in earlier flowering by 1 or 2 years (Soejima et al., 1998). Recently,
a method aiming at inducing high frequency early flowering of apple seedlings has been
developed (Volz et al., 2009). The method uses optimal conditions in a controlled
environment (CE), including high temperatures and relative humidity (26 °C to 30 °C and 85
%, respectively), high irradiance (> 900 μmol m2s
-1), long photoperiod (18 h photoperiod),
high red:far-red ratio, elevated CO2 concentration (~1800 ppm CO2) and non-limiting
fertigation; in addition, prohexadione-Ca (Regalis®, a gibberellin biosynthesis inhibitor) is
applied to reduce shoot extension without affecting node production. As explained by the
authors, the aim is to grow seedlings “fat, rich and branched”, i.e. with a high vigour that
favours the juvenile-adult transition with extensive branching to maximize sites for floral bud
production. Thus, some progenies displayed seedlings which flowered within 18 months after
sowing (Volz et al., 2009).
1.4.5.2.2. Transgenic approaches
Besides agro-technical methods, alternative transgenic approaches have been tested to shorten
more drastically the juvenile phase and thus the breeding cycle (i.e. the time elapsing between
seed planting and fruiting of the seedling). Penã and collaborators were the first who reported
the reduction of the juvenile phase in a fruit tree species, after transformation of Citrus plants
with the A. thaliana LFY gene (Penã et al., 2001). Since this work, several attempts of genetic
transformation of apple to shorten the juvenile phase have been performed using different
genes involved in the flowering pathway. Thus, the LFY gene itself or its putative apple
homologues AFL1 and AFL2 were over-expressed in apple (Flachowsky et al., 2010; Kotoda
et al., 2003); however, no precocious flowering was observed in both studies. The MdTFL1
gene of apple, homologous to the TERMINAL FLOWER 1 (TFL1) gene of A. thaliana, was
also investigated; TFL1 suppresses the floral meristem identity genes LFY and AP1 and
maintains the inflorescence meristem. It was found that the constitutive over-expression of
MdTFL1 in A. thaliana retards the transition from the vegetative to the reproductive phase as
known for the gene TFL1 (Kotoda and Wada, 2005). The suppression of MdTFL1 by
expressing its antisense RNA in transgenic lines of the apple cultivar ‘Orin’ resulted in a
significantly reduced juvenile period: the first solitary flowers were detected 8 months after
transfer of the transgenic lines to the glasshouse (Kotoda et al., 2003; Kotoda et al., 2006).
Later, Szankowski et al. (2009) also used an RNAi-based approach to induce silencing of the
MdTFL1 gene in apple. The regenerated transgenic plants of the cultivars ‘Holsteiner Cox’
45
and ‘Gala’ started to flower six month after the transformation under in vitro conditions. The
plants were then transferred to the glasshouse where they continued to flower (Szankowski et
al., 2009). Using a similar RNAi-based approach, Flaishman reported on the silencing of the
PcTFL1 gene (homologous to TFL1 of A. thaliana in Pyrus communis L.) in the European
pear cultivar ‘Spadona’; this work resulted in the creation of a transgenic line flowering 4 to 8
months after rooting in the glasshouse (Flaishman, 2009). Flachowsky and colleagues
investigated the effect of over-expressing the FRUITFULL (FUL)-homolog BpMADS4 gene
from silver birch (Betula pendula Roth.) in the apple cultivar ‘Pinova’. It appeared that the
CaMV35::BpMADS4 gene could induce early flowering in 25 transgenic lines of ‘Pinova’
with a dramatic reduction of the juvenile period (Flachowsky et al., 2007): 8 BpMADS4-
transgenic lines showed flowering shoots already in vitro and 2 others produced flowers for
the first time 3 to 4 months after transfer to the glasshouse; these observations were similar to
those made on BpMADS4-transgenic silver birch (Elo et al., 2007). More recently, Kotoda
and colleagues over-expressed the MdFT1 gene using the CaMV35S promoter in apple; they
obtained six transgenic lines, five of which flowered in vitro 8 to 12 months after
transformation (Kotoda et al., 2010). Tränkner and collaborators expressed the MdFT1 gene
(actually MdFT2 (Tränkner et al., 2011)) in apple under the control of the CaMV35S promoter
or the A. thaliana Suc2 promoter. Using the CaMV35S promoter, one transgenic line set up its
first flowers during in vitro cultivation; another transgenic line started flowering immediately
after transfer to the glasshouse (Tränkner et al., 2010). In the European pear, Matsuda et al.
(2009) transformed the cultivars ‘La France’ and ‘Ballade’ with the Citrus FLOWERING
LOCUS T (CiFT) gene; the over-expression of CiFT led to in vitro flowering on transgenic
shoots. The authors could further demonstrate that 5 out of 7 F1 progeny plants from the
CiFT–transgenic line crossed with the cultivar ‘Bartlett’ flowered within 10 months after
transfer to a glasshouse, i.e. inherited the early flowering phenotype (Matsuda et al., 2009).
1.4.5.3. Implementation of fast breeding approaches through reduction of the
juvenile phase
Using CE conditions as described above, Volz et al. (2009) developed a fast breeding
approach to accelerate the introgression of novel fruit traits and diseases resistances from
small fruited wild apple species into elite material. This approach involves the reduction of
both the generation time and the number of generations of pseudo-backcross. This is
respectively achieved by promoting early flowering of seedlings by growing them on their
own roots in CE rooms, and by screening the seedlings with a set of molecular markers well-
46
distributed on the whole genome for selection on three criteria: (i) presence of the desired
gene inherited from the low quality ancestor (i.e. the crab apple or a derived descendant),
referred to as foreground selection; (ii) high proportion of genome inherited from the high
quality grandparent(s); and (iii) low proportion of genome from the low quality ancestor; the
last two criteria being usually referred to as background selection or whole-genome selection
(WGS) (Bus et al., 2009). The assumption underlying selection based on genomic
contributions is that relationship(s) exists between fruit quality and the ancestral genomic
contribution. Practically, Volz and colleagues worked with 330 seedlings of a F1 progeny
from a cross between ‘Royal Gala’ and A689-24 (breeding selection carrying the powdery
mildew resistance gene Pl-2 from A368-12, itself derived from the wild species M. zumi).
Under CE conditions, the F1 seedlings showed precocious flowering less than 2 years after
sowing and 173 of them were screened with 108 SSR markers to perform WGS. Of these 108
markers, 65 covering 9 of the 17 linkage groups were found to be polymorphic in at least two
of the grandparents, and of these 37 were fully informative for all four grandparents. Lack of
informative markers across all LGs did not allow estimating the grand-parental contribution
precisely. Moreover, correlations between the proportion of A368-12 genome and fruit
quality traits such as fruit weight in F1 seedlings were found to be poor. However, the authors
stressed that the potential genetic gain from a selection based on genomic contribution could
be significant when selecting the seedlings containing as less as possible genome from A368-
12. A similar “low input” fast-track breeding method is being applied at Agroscope ACW
(Switzerland) to accelerate the introgression of the Fb_E locus of ‘Evereste’ and combine it
with the scab resistance genes Rvi4 and Rvi6 (Baumgartner et al., 2011).
The feasibility of a fast breeding strategy in apple using early flowering transgenic trees has
recently been demonstrated by Flachowsky and collaborators using the BpMADS4-transgenic
line T1190 (Flachowsky et al., 2007; Flachowsky et al., 2009). The aim is the same as that of
Volz et al. (2009), i.e. using the early flowering plants to accelerate the introgression of a
gene/trait from a small-fruited wild apple species through 4–5 pseudo-backcross generations
(Fig. 1.6). In winter 2005/2006, several plants of the line T1190 were pollinated by the wild
apple species M. fusca. Three fruits with a total of 11 seeds were harvested in fall 2006, and
the seeds were stratified and sown in winter 2006/2007. A total of 7 F1 seedlings were
obtained, 4 of them being transgenic. The transgenic F1 seedlings flowered within a few
weeks and were subsequently pollinated in spring 2007 by the cultivar ‘Topaz’. In October
2007, the resulting fruits were harvested and a total of 41 seeds were obtained, representing
the BC’1 progeny. This fast breeding program is still underway and according to Flachowsky
47
and colleagues, several backcrosses might be manageable within a decade using such a
strategy. At the last pseudo-backcross generation, pre-breeding material can be produced that
carries the gene/trait of interest with reduced proportion of genome from the wild species and
without the CaMV35::BpMADS4 construct with which T1190 was originally transformed
(Fig. 1.6).
Fig. 1.6. Scheme of accelerated gene introgression from a small fruited wild apple species
into a domesticated apple genetic background through several pseudo-backcrosses using
transgenic early flowering genotypes. From Flachowsky et al. (2009).
Brown: genome of a small fruited wild apple species (low fruit quality ancestor); green: genome of a high fruit
quality apple cultivar; P: parental generation; F1: first filial generation, composed of seedlings containing 50 %
genome of each parent; BC’x: generation after x pseudo-backcrosses; t: transgene; goi: gene of interest
introgressed from the wild apple species.
48
1.5. Scope of this Thesis
The major objectives of the work presented in this Thesis were (i) to identify QTLs associated
with fire blight resistance in commercial apple cultivars, using different types of molecular
markers; and (ii) to initiate the accelerated introgression of a strong-effect fire blight
resistance QTL from an ornamental apple cultivar into a domesticated apple background
using molecular markers.
The Chapters 2 and 3 aimed at identifying molecular markers associated with fire blight
resistance in apple cultivars known to be resistant to the disease, namely ‘Florina’, ‘Nova
Easygro’ and ‘Rewena’. Two bi-parental F1 progenies were thus subjected to a classical QTL
mapping approach. The first F1 progeny was derived from a cross between ‘Florina’ and
‘Nova Easygro’; its study had been initiated by Muhammad A. Khan (Diss. ETH Zürich, Nr.
17451, 2007) by the construction of two SSR-based backbone parental linkage maps, the
phenotyping of 118 F1 individuals in glasshouse and a preliminary QTL analysis. The work
described in Chapter 2 aimed at enriching the parental linkage maps with AFLP markers to
perform a more accurate QTL mapping. The second F1 progeny was derived from a cross
between ‘Idared’ and ‘Rewena’, two cultivars showing a contrasting level of resistance to E.
amylovora; the phenotyping data had already been produced by Andreas Peil and Klaus
Richter (JKI, Germany; unpublished data). The objective of the work described in Chapter 3
was to construct two parental linkage maps with diversity arrays technology (DArT) and SSR
markers and to subsequently identify genomic regions associated with fire blight resistance,
especially in ‘Rewena’.
The Chapters 4 and 5 of this Thesis investigated the potential of a high-speed breeding
strategy based on the BpMADS4-transgenic apple line T1190 to accelerate the introgression of
disease resistance genes/QTLs in a domesticated apple background. The Chapter 4 reports the
genetic mapping of the T-DNA integration site in the transgenic line T1190. The Chapter 5
describes the accelerated introgression of the fire blight resistance locus Fb_E of the
ornamental apple cultivar ‘Evereste’ by means of the transgenic line T1190. Marker-assisted
introgression (MAI) and background selection were performed using SSR markers on two (F1
and BC’1) and one (BC’1) breeding generations, respectively, in order to increase the
efficiency of the first two introgression cycles. The ultimate aim of this pre-breeding program
is to produce pre-breeding genotypes carrying the fire blight resistance locus Fb_E of
‘Evereste’ but possessing as less as possible genome of ‘Evereste’.
49
50
51
Chapter 2
Mapping of quantitative trait loci for fire blight resistance
in the apple cultivars ‘Florina’ and ‘Nova Easygro’
Published in the journal Genome as:
Le Roux P.-M., Khan M. A., Broggini G. A. L., Duffy B., Gessler C., Patocchi A. (2010)
Mapping of quantitative trait loci for fire blight resistance in the apple cultivars ‘Florina’ and
‘Nova Easygro’. Genome 53:710-722
P.-M. F. Le Roux and M. A. Khan contributed equally to the work described in Chapter 2.
For the sake of understanding, the parts of the work performed by M. A. Khan were not
removed from this Chapter. They were first described in M. A. Khan’s PhD. dissertation, Nr.
17451 ETH Zürich, and are as follows: (i) evaluation of fire blight resistance of plant
material in a quarantine glasshouse; (ii) SSR markers selection; (iii) gel and capillary
electrophoresis of SSR markers; (iv) pedigree analysis with three SSR markers flanking on
either side the QTL identified on linkage group 10 of the cultivar ‘Florina’. The preliminary
mapping of fire blight resistance QTLs in the cross ‘Florina’ × ‘Nova Easygro’ achieved by
M. A. Khan was repeated and refined.
52
2.1. Abstract
Fire blight is a devastating bacterial disease of rosaceous plants (Rosaceae). Its damage to
apple production is a major concern since no existing control option has proven to be
completely effective. Some commercial apple varieties, such as ‘Florina’ and ‘Nova Easygro’,
exhibit a consistent level of resistance to fire blight. In this study, we used an F1 progeny of
‘Florina’ × ‘Nova Easygro’ to build parental genetic maps and identify quantitative trait loci
(QTL) related to fire blight resistance. Linkage maps were constructed using a set of
microsatellites and enriched with amplified fragment length polymorphism (AFLP) markers.
In parallel, progeny plants were artificially inoculated with Erwinia amylovora strain CFBP
1430 in a quarantine glasshouse. Shoot length measured 7 days after inoculation (DAI) and
lesion length measured 7 and 14 DAI were used to calculate the lesion length as a percentage
of the shoot length (PLL1 and PLL2, respectively). Percent lesion length data were log10-
transformed (log10(PLL)) and used to perform the Kruskal-Wallis test, interval mapping (IM)
and multiple QTL mapping (MQM). Two significant fire blight resistance QTLs were
detected in ‘Florina’. One QTL was mapped on linkage group 10 by IM and MQM; it
explained 17.9 and 15.3 % of the phenotypic variation by MQM with log10(PLL1) and
log10(PLL2) data, respectively. A second QTL was identified on linkage group 5 by MQM
with log10(PLL2) data; it explained 10.1 % of the phenotypic variation. Genotyping the plants
of ‘Florina’ pedigree with the microsatellites flanking the QTLs showed that the QTLs on
linkage groups 5 and 10 were inherited from ‘Starking’ (a ‘Red Delicious’ sport mutation)
and ‘Jonathan’, respectively. Other putative QTLs (defined as QTLs with LOD scores above
the chromosomal threshold and below the genome-wide threshold) were detected by IM on
linkage groups 5 and 9 of ‘Nova Easygro’.
Keywords: Malus × domestica Borkh., Erwinia amylovora, microsatellite, QTL mapping.
53
2.2. Introduction
Fire blight, caused by the bacterium Erwinia amylovora, is a threat to plants of the Rosaceae
family (formerly Rosaceae (Potter et al., 2007)) worldwide, especially apples and pears. The
disease is reported to cause losses of more than $100 million every year in the USA (Norelli
et al., 2003). Fire blight leads to important crop losses in other countries also; for instance in
Italy, one million pear trees were destroyed by fire blight in 1997 and 1998 (Finelli et al.,
2004); and in Switzerland 130 hectares of apple and pear orchards and 10,000 high-stem
meadow trees were destroyed during the epidemic of 2007 (Holliger et al., 2008).
Applications of antibiotics such as streptomycin (and oxytetracyclin to a lesser extent) are so
far the most reliable control strategy to limit the damage due to E. amylovora (Turechek,
2004). They are applied mostly in the USA, Israel and New Zealand and were recently
allowed with restrictions for application in Germany and Switzerland (Duffy and Holliger,
2008). However, E. amylovora strains resistant to streptomycin have been reported
(McManus et al., 2002). Therefore, the use of antibiotics alone is probably not a sustainable
control strategy. Besides regulation measures and chemical or biocontrol treatments, breeding
for host resistance is considered to be an essential component of an integrated fire blight
management strategy (Lespinasse and Aldwinckle, 2000). Variable levels of resistance
against fire blight are present in both wild and cultivated apples. Consequently, several
research stations have been conducting programs for breeding fire blight resistant apple and
pear cultivars, in the USA as well as in Europe (Kellerhals et al., 2009b; Lespinasse and
Aldwinckle, 2000).
Unraveling the genetic determinism of fire blight resistance in apple could help breeders to
introduce quantitative trait loci (QTLs) from sources of resistance into breeding material
using marker-assisted selection. Studying the genetic architecture of a quantitative trait in a
diploid, self-incompatible species like apple requires access to a population segregating for
the trait of interest. A genetic map constructed in the progeny can be used for QTL mapping.
Linkage maps based on both dominant (random amplified polymorphic DNA, RAPD;
amplified fragment length polymorphism, AFLP) and co-dominant (simple sequence repeats,
SSR) molecular markers have been developed in different apple cultivars and have been used
to identify QTLs associated with different quality traits as well as disease resistance (Calenge
and Durel, 2006; Calenge et al., 2004; Calenge et al., 2005; Celton et al., 2009b; Dunemann et
al., 2009; Durel et al., 2009; Durel et al., 2003; Fernández-Fernández et al., 2008; Kenis and
Keulemans, 2007; Liebhard et al., 2003a; Liebhard et al., 2003b; Liebhard et al., 2003c; Peil
54
et al., 2007). The availability of several hundred SSRs mapped in apple (Celton et al., 2009b;
Liebhard et al., 2003a; Silfverberg-Dilworth et al., 2006) now allows the fast construction of
SSR-based linkage maps.
When studying the genetic determinism of fire blight resistance, Calenge et al. (2005) and
Khan et al. (2006) identified a QTL on linkage group (LG) 7 of the cultivar ‘Fiesta’ that
explained about 40 % of the variability in fire blight resistance in two different populations,
i.e. ‘Prima’ × ‘Fiesta’ and ‘Fiesta’ × ‘Discovery’. Stability of the effect of “F7” QTL in a third
genetic background (‘Milwa’ × ‘1217’, ‘1217’ being an F1 individual from ‘Fiesta’ ×
‘Discovery’ carrying the favorable QTL allele) was checked and its applicability in marker-
assisted selection was demonstrated by Khan et al. (2007). Calenge et al. (2005) also
identified four minor QTLs on LG 3 (‘Prima’ and ‘Fiesta’), LG 12 (‘Discovery’) and LG 13
(‘Discovery’), each explaining 4.4–7.9 % of the variation. Using a QTL mapping approach,
Peil et al. (2007) reported a major QTL for fire blight resistance at the proximal end of LG 3
of the wild apple Malus × robusta ‘Robusta 5’ (‘R.5’). This QTL explained up to 80 % of the
phenotypic variability in the progeny ‘R.5’ × ‘Idared’. It was confirmed in two crosses, ‘R.5’
× ‘Idared’ and ‘Malling 9’ (‘M.9’) × ‘R.5’. In both crosses, the fire blight resistance QTL (R2
= 67 % and 83 %) was detected at the same genomic position, i.e. close to the microsatellites
CH03g07 and CH03e03 in the upper part of LG 3 of ‘R.5’ (Peil et al., 2008). Durel et al.
(2009) identified two additional strong-effect QTLs in the ornamental cultivar ‘Evereste’ and
in the wild apple Malus × floribunda clone 821, explaining between 50–70 % and more than
40 % of the phenotypic variation, respectively. The two last studies raised the question of the
existence of “major” genes against fire blight in apple.
QTL analysis for fire blight resistance was undertaken also in pear (Pyrus communis L.)
which belongs, like apple, to the sub-family Spiraeoideae in the family Rosaceae and shows
clear syntenic relationships with apple (Pierantoni et al., 2004; Yamamoto et al., 2007). Using
an F1 progeny derived from a cross between ‘Passe Crassane’ and ‘Harrow Sweet’, four
QTLs were identified on LGs 2A, 2B, 4 and 9, all in the resistant parent ‘Harrow Sweet’, the
first QTL explaining up to 16.4 % of the phenotypic variation (Dondini et al., 2004). In these
studies, the fire blight resistance of the parental cultivars was found to be possibly oligogenic
(Calenge et al., 2005) or polygenic (Dondini et al., 2004).
In this study, we investigated the genetic basis of fire blight resistance in the F1 progeny of
two apple cultivars, ‘Florina’ and ‘Nova Easygro’, upon artificial inoculation under
glasshouse conditions.
55
2.3. Material and methods
2.3.1. Plant material and evaluation of fire blight resistance
Out of 491 F1 progeny plants of the cross ‘Florina’ × ‘Nova Easygro’ grown in an orchard at
Agroscope Changins-Wädenswil (ACW) (Gianfranceschi et al., 1996), 120 were randomly
selected. Ten replications for each of these progeny plants along with 12 replications of each
parent and susceptible controls (‘Idared’ and ‘Golden Delicious’) were bud-grafted to virus-
free ‘M.9’ T337 rootstocks and grown in a glasshouse. After 45 days, 8 well-grown and
healthy plants for each progeny plant and cultivar were selected and moved into the
quarantine glasshouse at ACW in two adjacent cabins (4 replications per cabin). Temperature
and humidity were controlled throughout the experiment (Khan et al., 2006). For the fire
blight resistance test, the reference strain E. amylovora CFBP 1430 was grown on King’s B
medium and the inoculum was prepared as described by Khan et al. (2006). Actively growing
shoots with a minimum length of 13.5 cm were inoculated, and shoot length at 7 days after
inoculation (DAI) as well as lesion length (cm) at 7 and 14 DAI were recorded as described in
Khan et al. (2006).
2.3.2. DNA extraction
The DNA of ‘Florina’ × ‘Nova Easygro’ progeny plants, previously extracted by
Gianfranceschi et al. (1996) and stored at - 20 °C, was used for amplification of microsatellite
(SSR) markers. A fresh DNA extraction from the same F1 progeny plants, necessary for
molecular analysis with AFLP markers, was performed as described by Koller et al. (1994).
The same protocol was used to extract the DNA of the plants of ‘Florina’ pedigree. DNA was
gel-quantified and diluted to 1 ng/µl.
2.3.3. SSR selection
Liebhard et al. (2002) and Silfverberg-Dilworth et al. (2006) developed almost 300 SSR
markers that were mapped on the apple map ‘Fiesta’ × ‘Discovery’. These SSRs were
simultaneously tested on a set of cultivars including ‘Florina’ and ‘Nova Easygro’
(information also available at the home page of the European Union-funded project High-
quality Disease Resistant Apples for a Sustainable agriculture (HiDRAS):
http://www.hidras.unimi.it). Based on this information, SSR markers approximately 15–20
cM distant from each other, polymorphic in ‘Florina’ and ‘Nova Easygro’, and where possible
also polymorphic between the two cultivars, were selected to construct framework parental
56
maps. A few SSR markers developed by Celton et al. (2009b) as well as the SSR markers
TTTC17 (G. Parravicini et al., unpublished data) and NH033b (T. Yamamoto, personal
communication) were also chosen. Regions of the genome where QTLs for fire blight
resistance were identified in previous studies (in apple as well as in pear) were especially
targeted with molecular markers.
2.3.4. Gel and capillary electrophoresis of SSR and AFLP markers
Polymerase Chain Reaction (PCR) amplification of SSRs was performed in a volume of 15 µl
according to Liebhard et al. (2002). Fragment analysis of SSR markers was performed using
one of three detection techniques, depending on the size of the PCR fragments and the
difference in allele size within and between the parents: high-resolution agarose gel, 33
P-
labeling and fluorescent labeling. When high-resolution MetaPhor® agarose (4 %, w/v;
Cambrex Bio Science Inc., East Rutherford, New Jersey, USA) was used, PCR fragments
were separated by running gels stained with ethidium bromide for 1.5 to 3 h and gels were
then photographed. SSR genotyping with 33
P-labeling was performed according to Liebhard
et al. (2002). For the fluorescent labeling technique, 2 to 4 fluorescently labeled PCR products
(1 to 3 μl each) were pooled; the final volume was completed to 15 μl with sterile water.
Denaturation and fragment analysis of pooled PCR products were performed as described by
Patocchi et al. (2009a).
AFLP reactions were performed using the restriction enzymes EcoRI and MseI. A total of 45
EcoRI–MseI primer combinations were tested and screened on the ‘Florina’ × ‘Nova Easygro’
progeny. The restriction digestion, ligation, pre-amplification and selective amplification
steps were performed as previously described (Broggini, 2007; Xu and Korban, 2000) with
the following modification: the EcoRI+3 forward primers used for selective amplification
were not labeled with a radioisotope (33
P) but with a fluorescent dye (6-FAM, HEX, ROX) at
the 5’ end. Products of selective amplification were separated and analyzed as described for
SSR markers.
2.3.5. Linkage mapping
Linkage analysis and map construction were done as described by Liebhard et al. (2003a)
with JoinMap® version 4.0 (Van Ooijen, 2006) using the Kosambi mapping function. A LOD
(logarithm of odds) score of 4 was used to assign markers to linkage groups. For two linkage
groups however (NEG 8 and FLO 16), a decrease of the LOD threshold to 3 was necessary to
57
insert all markers in the map. Drawings of the linkage maps were generated with MapChart
(Voorrips, 2002).
2.3.6. Statistical and QTL analysis
The percent lesion length (PLL; %) was calculated by dividing the lesion length (LL; cm)
recorded at 7 and 14 DAI by the shoot length (SL; cm) recorded at 7 DAI. Shoot length, LL
and PLL measurements for each F1 individual, parent and control plant were averaged and
standard deviation was calculated. Data were checked for interactions, outliers and normal
distribution. The progeny × glasshouse cabin interaction was not significant (p < 0.05; data
not shown) and therefore data for both glasshouse cabins were pooled together. Percent lesion
length values were not normally distributed and were therefore log10-transformed. Random
effect values, variance components and broad-sense heritability (h2) were calculated
according to Calenge et al. (2005) with the log10-transformed PLL data (log10(PLL)).
Log10-transformed PLL data of progeny plants at 7 DAI (log10(PLL1)) and 14 DAI
(log10(PLL2)) were used to perform single marker analysis (Kruskal-Wallis test), interval
mapping (IM) and multiple QTL mapping (MQM) with the software MapQTL®
version 5
(Van Ooijen, 2004). A threshold of p < 0.05 was set to identify markers significantly
associated with fire blight resistance by the Kruskal-Wallis test. For IM and MQM, the
genome-wide LOD threshold value was calculated by the permutation test (Doerge and
Churchill, 1996). A LOD > 2.8 and a LOD > 2.9 were set to declare a QTL significant at the
95 % confidence level at 7 DAI and 14 DAI in ‘Florina’ and ‘Nova Easygro’, respectively.
The confidence interval of QTLs was calculated with a 1 LOD score drop-off from maximum
LOD score. The phenotypic variation explained by a QTL was estimated using MapQTL®
version 5. Multiple QTL mapping was conducted by selecting the AFLP markers E31M41-
351 (FLO 5) and E31M44-368 (FLO 10) as cofactors. Digenic epistatic interactions between
the markers closest to the significant or putative QTLs were checked by a two-way ANOVA
model (SYSTAT, version 12, SPSS, Inc., 2007) with an interaction component as described in
Calenge et al. (2005).
2.3.7. Tracing the significant FLO 10 QTL in the pedigree of ‘Florina’
Available cultivars of ‘Florina’ pedigree were genotyped with three SSR markers (CH02b07,
CH02a10 and CH01f12) flanking the peak of the FLO 10 QTL as well as with SSR CH05e06,
located close to the peak of the FLO 5 QTL. The primers were fluorescently labelled and PCR
products were analyzed as described above.
58
2.4. Results
2.4.1. Phenotypic evaluation
The PLL values of ‘Nova Easygro’ at 7 and 14 DAI were significantly lower than the
respective PLL values for ‘Florina’ using the Mann-Whitney test (p < 0.001; Table 2.1). The
mean PLL of ‘Florina’ × ‘Nova Easygro’ progeny plants (7.7 %) was between the PLLs of
‘Florina’ (11.7 %) and ‘Nova Easygro’ (1.7 %) at 7 DAI, but it was not statistically different
from the PLL of ‘Florina’ at 14 DAI. The cultivars ‘Golden Delicious’ and ‘Idared’
(moderately susceptible and susceptible controls, respectively) were also inoculated. Their
respective mean PLL values were 11.9 % and 29.3 % at 7 DAI, compared to 16 % and 38.2 %
at 14 DAI (Table 2.1). Significant differences among ‘Florina’ × ‘Nova Easygro’ progeny
plants were observed for PLL values (Fig. 2.1) and log10(PLL) values (data not shown) at 7
and 14 DAI using the Kruskal-Wallis test (p < 0.05). Broad-sense heritability (h2) was
estimated to be 71 % for PLL1 (7 DAI) and 74 % for PLL2 (14 DAI).
Table 2.1. Fire blight lesion length as a percentage of shoot length (PLL) of ‘Florina’ (FLO),
‘Nova Easygro’ (NEG), ‘Golden Delicious’ (GD), ‘Idared’ and the F1 progeny ‘Florina’ ×
‘Nova Easygro’ (FLO × NEG).
SD: standard deviation; DAI: days after inoculation.
Trait Statistic FLO NEG FLO × NEG GD ‘Idared’
PLL1 (7 DAI) Mean 11.7 1.7 7.7 11.9 29.3
SD ±4 ±0.6 ±6.3 ±6 ±15.4
PLL2 (14 DAI) Mean 13.4 2.1 10.9 16 38.2
SD ±4 ±1.3 ±8.4 ±5.6 ±20.1
59
Fig. 2.1. Distribution of the individuals of the ‘Florina’ × ‘Nova Easygro’ progeny according
to their mean lesion length as a percentage of shoot length (PLL) at 7 (A) and 14 (B) days
after inoculation (PLL1 and PLL2, respectively).
Individuals are ordered from the lowest to the highest PLL value. FLO: ‘Florina’; NEG: ‘Nova Easygro’; GD:
‘Golden Delicious’.
2.4.2. Construction of parental linkage maps
Out of 120 randomly selected ‘Florina’ × ‘Nova Easygro’ progeny plants, two were identified
as out-breeders and hence discarded. The final map was thus based on data from 118
60
individuals. Among the 98 SSR markers polymorphic between ‘Florina’ and ‘Nova Easygro’,
91 were mapped on 17 linkage groups of ‘Florina’ and ‘Nova Easygro’. The remaining seven
SSRs were discarded because they did not segregate, contrary to what was expected from the
data of Liebhard et al. (2002) and Silfverberg-Dilworth et al. (2006), or because the
amplification profile was not sufficiently clear. Of the 91 mapped SSRs, 52 were common to
both maps, whereas 22 and 17 SSRs mapped only to ‘Florina’ or ‘Nova Easygro’,
respectively (Fig. 2.2). A total of 95 and 135 AFLP markers were mapped on the ‘Florina’
and ‘Nova Easygro’ maps. The average number of polymorphic AFLP bands generated per
EcoRI-MseI combination was 5.1. In summary, the maps of ‘Florina’ and ‘Nova Easygro’
were composed of 169 and 204 molecular markers, spanning 1199 and 1336 cM, respectively,
with an average density of one marker every 7.1 cM for ‘Florina’ and every 6.5 cM for ‘Nova
Easygro’ (Fig. 2.2). The average linkage group length was 70.5 cM for ‘Florina’ and 78.8 cM
for ‘Nova Easygro’. Linkage group 2 of ‘Florina’ remained split in two parts.
61
62
Fig. 2.2. Genetic linkage maps of ‘Florina’ (FLO 1 to FLO 17) and ‘Nova Easygro’ (NEG 1 to NEG 17).
Markers shown in bold are the closest to the peak of the significant (on FLO 5 and FLO 10) or putative (on NEG 5 and NEG 9) QTLs.
63
2.4.3. Statistical and QTL analysis
The Kruskal-Wallis test identified two linkage groups with markers showing significant
association with fire blight resistance in ‘Florina’ (FLO 5 and FLO 10). The AFLP markers
E31M41-351 (FLO 5) and E31M44-368 (FLO 10) were associated with the resistance at 7
and 14 DAI, with significance levels of p < 0.05 (Table 2.2). The Kruskal-Wallis test also
identified two linkage groups of ‘Nova Easygro’ (NEG 5 and NEG 9) with markers associated
with the resistance at 7 and 14 DAI (p < 0.05; Table 2.2). In the case of NEG 5, the marker
with the highest test statistic (K*) was the SSR CH03a09 (14.9 cM) at 7 DAI and the AFLP
marker E34M40-46 (0 cM) at 14 DAI. On NEG 10, the SSR marker Hi03f06 was
significantly associated with fire blight resistance by the Kruskal-Wallis test (p = 0.01) at 7
DAI but not at 14 DAI.
Interval mapping conducted with the traits log10(PLL1) and log10(PLL2) identified the same
region associated with fire blight resistance on FLO 10; the LOD plot was above the genome-
wide threshold (LOD = 2.8) using log10(PLL1) data, but only above the chromosomal
threshold (LOD = 1.8) using log10(PLL2) data. The peak of the LOD plot was located
between the AFLP markers E34M38-121 (15 cM) and E31M44-368 (30.9 cM; LOD = 3.26
and 2.29 with log10(PLL1) and log10(PLL2) data, respectively), the last marker being the
closest to the peak. Beyond these AFLPs, the peak is bracketed by SSR markers, CH02b07
(8.6 cM) on one side and CH01f12 (42.1 cM) and Hi03f06 (46.5 cM) on the other side (Fig.
2.3). The genomic region between AFLP markers E34M38-121 and E31M44-368 explained
15.9 % of the phenotypic variation (PVE) for log10(PLL1) data and 12.5 % of the phenotypic
variation for log10(PLL2) data. Multiple QTL mapping detected a significant association of
the same genomic region between markers E34M38-121 and E31M44-368 on FLO 10 with
log10(PLL1) and log10(PLL2) data. The LOD score associated with the AFLP marker
E31M44-368 was 3.86 with log10(PLL1) data and 2.90 with log10(PLL2) data; the phenotypic
variation explained was 17.9 % and 15.3 %, respectively (Table 2.2).
Interval mapping conducted with the traits log10(PLL1) and log10(PLL2) showed that the
AFLP marker E31M41-351 on FLO 5 was associated with LOD scores of 1.59 and 2.35,
respectively (R2 = 6.1 and 8.8 %), the second value being above the chromosomal LOD
threshold (LOD = 1.6). By MQM, the same AFLP marker, E31M41-351, was found at the
peak of the LOD plot on FLO 5, above the genome-wide threshold, using the log10(PLL2)
data (LOD score = 2.96; R2 = 10.1 %; Table 2.2).
Interval mapping detected two regions in ‘Nova Easygro’ showing a LOD score only above
the chromosomal threshold (LOD = 1.7) on NEG 5 and NEG 9 (Table 2.2). On NEG 5, the
64
peak of the LOD plot was located at the proximal end and co-localized with the SSR marker
CH03a09 (LOD = 2.03, R2
= 7.9 %) with log10(PLL1) data or with the AFLP marker
E34M40-46 (LOD = 1.88, R2
= 7.2 %) with log10(PLL2) data. On NEG 9, the highest LOD
score was found almost in the middle of the linkage group, corresponding to the AFLP marker
E31M33-493 at both time points (LOD=1.91 and 1.73; R2
= 7.3 and 6.6 %; with log10(PLL1)
and log10(PLL2) data, respectively). No digenic epistatic interaction was found involving any
of the genomic regions associated with fire blight resistance by the Kruskal-Wallis test (data
not shown).
65
Table 2.2. Significant and putative QTLs for fire blight resistance detected by the Kruskal-Wallis test, interval mapping and multiple QTL mapping
in the F1 progeny ‘Florina’ × ‘Nova Easygro’.
a: closest marker to the peak of the LOD plot.
b: LOD score associated with the marker closest to the peak of the LOD plot using interval mapping / multiple QTL mapping.
c: percentage of the phenotypic variation explained by the peak of the LOD plot using interval mapping / multiple QTL mapping.
d: genome-wide LOD threshold (95 % of confidence).
e: chromosomal LOD threshold (95 % of confidence).
f: Confidence interval (CI) calculated based on a 1 LOD drop-off from the peak of the LOD plot using interval mapping.
Values are indicated in bold when the LOD score is higher than the genome-wide threshold using interval mapping / multiple QTL mapping.
Values are indicated in italics when the association was found only by the Kruskal-Wallis test.
Trait LG Closest marker a
Distance
(cM) LOD b
R2 (%)
c GW threshold
d Chr. threshold
e CI (cM)
f K* Significance
FLO 10 E31M44-368 30.9 3.26 / 3.86 15.9 / 17.9 2.8 1.8 33.6 11.8 0.001
FLO 5 E31M41-351 34.8 1.59 / 2.19 6.1 / 7.3 2.8 1.6 88.8 6.4 0.05
Log10(PLL1) NEG 5 CH03a09 14.9 2.03 7.9 2.9 1.7 27 8.6 0.005
NEG 9 E31M33-493 37.5 1.91 7.3 2.9 1.7 33.9 5.6 0.05
NEG 10 Hi03f06 56.9 - - 2.9 1.6 - 7.1 0.01
FLO 10 E31M44-368 30.9 2.29 / 2.90 12.5 / 15.3 2.8 1.8 36 8.4 0.005
Log10(PLL2) FLO 5 E31M41-351 34.8 2.35 / 2.96 8.8 / 10.1 2.8 1.6 13.7 9.3 0.005
NEG 5 E34M40-46 0 1.88 7.2 2.9 1.7 48.7 6.7 0.01
NEG 9 E31M33-493 37.5 1.73 6.6 2.9 1.7 33.7 4.5 0.05
66
Fig. 2.3. LOD plots for fire blight resistance QTL mapping on FLO 10 (interval mapping, p =
0.05) with traits log10(PLL1) (A) and log10(PLL2) (B).
The x-axis indicates the linkage map of ‘Florina’ in cM; the y-axis shows the LOD scores. The horizontal lines
represent the significant LOD thresholds (p < 0.05) at the genome scale (GW, smooth line) and the chromosome
scale (Chr., dashed line).
67
2.4.4. Analysis of the pedigree of ‘Florina’
As SSR marker Hi03f06, flanking the peak of the LOD plot on FLO 10, was multilocus, the
marker CH01f12, a single-locus marker located 4.5 cM away from Hi03f06, was used to
analyze the pedigree of ‘Florina’, together with CH02a10 and CH02b07. The following
alleles were associated with an increased fire blight resistance in the parent ‘Florina’: 175 bp,
CH02a10; 152 bp, CH01f12; and 126 bp, CH02b07. Molecular analysis of pedigree plants
showed that these three SSR alleles have been inherited from ‘Jonathan’ (Fig. 2.4A). The
SSR marker CH05e06, mapping close to the FLO 5 QTL, amplified in ‘Florina’ two alleles of
144 and 155 bp, the first being associated with an increased resistance and inherited from
‘Starking’ (Fig. 2.4B).
68
Fig. 2.4. Analysis of the pedigree of ‘Florina’ with SSR markers flanking the significant
QTLs on FLO 10 and FLO 5. (A). Pedigree analysis with the three SSR markers CH02b07,
CH02a10 and CH01f12, flanking on either side the QTL identified on FLO 10. (B). Pedigree
analysis with SSR marker CH05e06 flanking the QTL identified on FLO 5.
Alleles in coupling with the QTLs are shown in bold. “?”: null allele or homozygous.
69
2.5. Discussion
2.5.1. Phenotypic screening of the ‘Florina’ × ‘Nova Easygro’ progeny
The PLL values calculated for the parents ‘Florina’ and ‘Nova Easygro’ and the controls
‘Golden Delicious’ and ‘Idared’ in this study (Table 2.1) are lower than the values obtained
for the same cultivars during previous fire blight resistance tests performed in the same
glasshouse facilities: 17.8 % for ‘Nova Easygro’, 46.4 % for ‘Florina’, 47.5 % for ‘Golden
Delicious’ and 66 % for ‘Idared’ (PLL at 14 DAI; M. A. Khan et al., unpublished data). A
possible explanation might be the difference in temperature and relative humidity between the
experiments. However, as the ranking of these four cultivars is the same in the different
experiments, the phenotyping data collected in this study can be considered suitable for QTL
mapping.
The mean PLL value for ‘Florina’ × ‘Nova Easygro’ progeny plants was in between the mean
PLL values of the parents, which is comparable to the results of Calenge et al. (2005). When
studying fire blight resistance in two crosses involving the cultivars ‘Fiesta’ and either
‘Discovery’ or ‘Prima’, Calenge et al. (2005) observed mean LL score for F1 individuals in
between the LL scores of the respective parents. They concluded that this may indicate an
additive genetic determinism of fire blight resistance, which also seems true for the cross
‘Florina’ × ‘Nova Easygro’. Furthermore, the distribution of the ‘Florina’ × ‘Nova Easygro’
progeny for PLL is continuous, indicating a quantitative resistance to fire blight (Fig. 2.1).
This was also observed in other studies on fire blight resistance QTLs in apple (Calenge et al.,
2005) and pear (Dondini et al., 2004). However, one peculiarity of the distribution of the
‘Florina’ × ‘Nova Easygro’ progeny was not observed by Calenge et al. (2005) in their two F1
progenies, i.e. the absence of any progeny plant showing a lower PLL than the most resistant
parent (‘Nova Easygro’ in this study). On the contrary, in our study some F1 individuals (14
% and 37 % of the progeny at 7 and 14 DAI) were more susceptible than ‘Florina’ (less
resistant parent). To explain the absence of any transgressive F1 individuals more resistant
than ‘Nova Easygro’, one could hypothesize that the relatively high level of fire blight
resistance of ‘Nova Easygro’ might be the result of many small-effect, additive QTLs. These
QTLs would segregate in ‘Florina’ × ‘Nova Easygro’ in such a way that the probability of
finding a transgressive individual cumulating all the additive QTLs of ‘Nova Easygro’ plus an
additive QTL of ‘Florina’ is very small. Our ‘Florina’ × ‘Nova Easygro’ progeny may be too
small and therefore would need to be enlarged to identify such individuals.
70
2.5.2. Linkage maps of ‘Florina’ and ‘Nova Easygro’
The available information on SSR alleles for ‘Florina’ and ‘Nova Easygro’ (HiDRAS:
http://www.hidras.unimi.it) allowed an efficient construction of two parental linkage maps.
Owing to the mapping of 91 SSRs in both parental maps, it was possible to identify and
orientate the corresponding linkage groups with the apple “reference” map ‘Fiesta’ ×
‘Discovery’ (Liebhard et al., 2003a; Silfverberg-Dilworth et al., 2006). Linkage maps were
1199 and 1336 cM long and average linkage group lengths were 70.5 and 78.8 cM for
‘Florina’ and ‘Nova Easygro’, respectively. Marker density was one marker every 7.1 cM in
‘Florina’ and 6.5 cM in ‘Nova Easygro’. These marker densities and map lengths are lower
than those in the “reference” map, ‘Fiesta’ × ‘Discovery’. However, our genetic map and
linkage group lengths and marker densities are similar to those of the QTL mapping
performed by Kenis and Keulemans (2007). In addition, it is thought that a very high marker
density is not necessary for QTL mapping. Using a backcross (BC) population in simulation
studies, Piepho (2000) reported that an increase in marker density beyond 10 cM has only a
very slight effect on the power of QTL detection and the standard errors of genetic effect
estimates. Similarly, in Darvasi et al. (1993), the power of QTL detection in a backcross
simulation experiment was only slightly decreased for a marker spacing of 20 cM compared
with 10 cM.
2.5.3. Identification of two new genomic regions involved in fire blight resistance
The broad-sense heritability estimated for fire blight resistance in the ‘Florina’ × ‘Nova
Easygro’ population was high (71–74 %), albeit lower than the heritabilities estimated by
Calenge et al. (2005) and Khan et al. (2006) in ‘Fiesta’ × ‘Discovery’ (85–88 %), ‘Prima’ ×
‘Fiesta’ (86–88 %) and ‘Fiesta’ × ‘Discovery’-CH populations (90–94 %). The difference in
heritability between these previous studies and this study could be due to environmental
factors such as differences in temperature and humidity between the glasshouse cabins.
Nevertheless, the major proportion of the phenotypic variation within the ‘Florina’ × ‘Nova
Easygro’ progeny could still be attributed to the genetic variability, which allowed a reliable
QTL detection.
The Kruskal-Wallis test identified five genomic regions significantly associated with fire
blight resistance in ‘Florina’ and ‘NovaEasygro’ (p < 0.05). Two of these regions, located on
FLO 10 and FLO 5, can be considered as significant QTLs. Indeed, the LOD plot on FLO 10
exceeds the genome-wide threshold using log10(PLL1) data (by IM and MQM) and
log10(PLL2) data (by MQM; Fig. 2.3 and Table 2.2). On FLO 5, the LOD score associated
71
with the AFLP marker E31M41-351 at the peak of the LOD plot is also above the genome-
wide threshold by MQM using log10(PLL2) data (Table 2.2). It is possible that the QTL on
FLO 5 was not identified as significant with the log10(PLL2) data by IM because its effect
was “masked” by the QTL on FLO10. Selecting two cofactors in MQM (E31M41-351 on
FLO 5 and E31M44-368 on FLO 10) allowed testing for the presence of a QTL on FLO 5
while accounting for the presence of the significant QTL on FLO 10, thereby increasing the
power of detection. However, when log10(PLL1) data were used for IM and MQM, the QTL
peak at the AFLP marker E31M41-351 on FLO5 was not identified as significant. Three
additional genomic regions on NEG 5, NEG 9 and NEG 10 were significantly associated with
fire blight resistance at one or both time points using the Kruskal-Wallis test (p < 0.01; Table
2.2). However, they were not significant at the genome scale using IM or MQM, and thus
cannot be considered as significant QTLs.
None of these five genomic regions co-localizes with QTLs for fire blight resistance identified
in previous studies. The three QTLs with a very strong effect on fire blight resistance detected
recently on LG 3 and LG 12 in wild or ornamental apples (Durel et al., 2009; Peil et al., 2007;
Peil et al., 2008) do not show any homologous counterpart in the cultivars ‘Florina’ or ‘Nova
Easygro’, although it has to be noted that no SSR or AFLP marker was mapped at the distal
end of NEG 12. On LG 7, no association with fire blight resistance was detected at the
position of the ‘Fiesta’ F7 QTL (Calenge et al., 2005; Khan et al., 2006). Neither SSR marker
TTTC17 nor the two sequence characterized amplified region (SCAR) markers AE10-375 and
GE80-19 could be mapped on FLO 7 or NEG 7 because of lack of polymorphism. However,
the AFLP marker E31M53-233 of ‘Florina’ is in the genomic region where the F7 QTL is
located (i.e. the distal part of LG7 ) and it is not associated with fire blight resistance (by
Kruskal-Wallis or IM). In the distal part of NEG 7, SSR markers Hi03a10 and Hi05b09 are
separated by 32.3 cM but neither of them is associated with resistance; thus, the presence of a
strong or moderate-effect QTL between Hi03a10 and Hi05b09 is unlikely. Additional minor-
effect QTLs were identified on LG 3, LG 12 and LG 13 of apple by Calenge et al. (2005), but
again, these regions, although covered by markers, do not display any homologous QTL in
‘Florina’ or ‘Nova Easygro’. In pear, four genomic regions of the cultivar ‘Harrow Sweet’
were found to harbour QTLs related to fire blight resistance (LG 2A and 2B, LG 4 and LG 9;
Dondini et al., 2004). Because there is only one common marker (SSR CH05c07) between the
maps of ‘Nova Easygro’ and ‘Harrow Sweet’ on LG 9, it is so far not possible to conclude
with certainty if the putative QTL of ‘Nova Easygro’ co-localizes with the QTL of ‘Harrow
Sweet’.
72
Pedigree analysis with SSR markers shows that the alleles of ‘Florina’ in coupling with the
fire blight resistance QTLs were inherited from ‘Jonathan’ (FLO 10 QTL) or ‘Starking’ (FLO
5 QTL), but the alleles could not be followed further because of the unknown pedigrees of
‘Jonathan’ and ‘Starking’ (Fig. 2.4). However, as ‘Starking’ is a sport mutation of ‘Red
Delicious’, we can expect that the LG 5 QTL in ‘Starking’ is also present in ‘Red Delicious’,
which is considered resistant to fire blight. The SSR markers CH02a10, CH01f12, CH02b07
(LG 10) and CH05e06 (LG 5) will be useful for breeders to track and confirm the presence of
the QTLs in crosses where ‘Jonathan’, ‘Starking’ and ‘Red Delicious’ are used as a parent.
The alleles of ‘Golden Delicious’ at the corresponding SSR markers (Fig. 2.4) can be used as
standards to identify the alleles in coupling with the fire blight resistance QTLs, no matter
which fragment analysis technique is used, as proposed by Patocchi et al. (2009a).
2.5.4. Accuracy of QTL mapping in ‘Florina’ × ‘Nova Easygro’
The two QTLs that can be truly considered as significant regarding the genome-wide LOD
threshold are contributed by the less resistant parent, ‘Florina’. Contribution of the less
resistant parent has been observed in previous studies on quantitative apple scab resistance
(Liebhard et al., 2003b; Calenge et al., 2004; Soufflet-Freslon et al., 2008). ‘Florina’ would
carry at the QTLs on LG 10 and LG 5 two alleles of dissimilar effect on fire blight resistance,
thus allowing the detection of a significant association with fire blight resistance at both loci.
On the other hand, the apparent absence of a significant QTL in ‘Nova Easygro’ could be due
to the presence of two alleles of similar effect at the same loci, as already hypothesized by
Liebhard et al. (2003b). To address this issue, one could perform a similar QTL analysis on a
progeny derived from an F1 individual of ‘Florina’ × ‘Nova Easygro’ crossed with a
susceptible cultivar such as ‘Gala’. The selected F1 individual ideally should carry all the
favorable alleles at the putative QTLs inherited from ‘Nova Easygro’ (NEG 5, NEG 9 and
NEG 10). This way, the “functionally” homozygous resistance loci from ‘Nova Easygro’
would segregate, making possible the detection of the alleles associated with an increased fire
blight resistance.
There are several other factors that could have hampered the identification of significant
QTLs in ‘Florina’ × ‘Nova Easygro’. First, alleles underlying putative QTLs may individually
have a weak influence on the trait, which makes them difficult to map with accuracy. Second,
the size of the ‘Florina’ × ‘Nova Easygro’ F1 progeny (118 individuals) may not be sufficient
to detect the position and estimate the contribution of such small-effect QTLs. Third, some
genomic regions in ‘Florina’ and ‘Nova Easygro’ have not been completely covered with
73
molecular markers (FLO 2A and 2B, FLO 6, NEG 8 and NEG 17). Increasing the size of the
‘Florina’ × ‘Nova Easygro’ progeny to be phenotyped and genotyped, or simply improving
the maps’ coverage, might lead to the identification of additional, small-effect QTLs in
‘Florina’ or ‘Nova Easygro’.
2.5.5. LG 10 and LG 5 QTLs of ‘Florina’ do not map in homoeologous genomic
regions
Several studies have pointed out the existence of homoeologous regions in the apple genome,
for instance between LG 5 and LG 10 (Celton et al., 2009b; Liebhard et al., 2003a;
Maliepaard et al., 1998). A map of homoeologous regions based on multilocus SSR markers
and restriction fragment length polymorphism (RFLP) markers was proposed by Celton et al.
(2009b). A logical issue is thus whether the QTLs identified on FLO 5 and FLO 10 are
located on homoeologous chromosomal parts. No multilocus SSR marker known to map on
both LGs 5 and 10 were used in this study, which makes direct comparison with the
aforementioned “homoeologous” map difficult. Nevertheless, the question can be addressed
by comparing the linkage map of ‘Florina’ and the “homoeologous” map of Celton et al.
(2009b) through the map ‘Fiesta’ × ‘Discovery’ (Silfverberg-Dilworth et al., 2006). Such a
“consensus” map shows that the FLO 10 QTL, bracketed by SSR markers CH02b07 and
CH02a10, would encompass a genomic region of about 25 cM around the multilocus SSR
CH02a08 at the proximal end of LG 10, which is homoeologous to a chromosome segment at
the distal end of LG 5; whereas the FLO 5 QTL maps close to the SSR marker CH05e06 (less
than 5 cM), i.e. more at the proximal end of LG 5. Thus, it can be concluded that the FLO 5
and FLO 10 QTLs probably do not belong to homoeologous regions of the apple genome.
2.5.6. A putative disease resistance cluster on apple linkage group 10
Liebhard et al. (2003b) found a minor QTL for scab resistance (R2 = 4.4 %) peaking at 2.1 cM
above SSR marker CH02a10 on LG 10 of ‘Fiesta’. In addition, Calenge and Durel (2006)
identified powdery mildew resistance QTLs whose peaks corresponded approximately to the
markers CH02a08 and CH02b07 (scoring 2002) or to a position 10 cM below CH02b07
(scoring 2001), on LG 10 of the integrated map ‘Discovery’ × ‘TN10-8’. In comparison, the
confidence interval of the FLO 10 QTL encompasses a region delimited by the SSR markers
CH02b07 (8.6 cM) and CH02a10 (32.6 cM). Thus, these three QTLs involved in quantitative
resistance to apple diseases seem to cluster in the same region of the apple genome.
Clustering of genetic factors for disease resistance in apple has been already noticed in the
74
bottom part of LG 12 (Durel et al., 2009). Similar results have been found in other crops, for
example in maize (Wisser et al., 2006), where QTLs for resistance to different disease also
co-localize in clusters.
2.6. Acknowledgements
This research was funded by the Swiss Commission for Technology and Innovation (CTI
Grant no. 6502.2 BTS-LS) and the ZUEFOS Project (Züchtung feuerbrandtoleranter
Obstsorten, Federal Office for Agriculture FOAG, Switzerland). Genotyping with AFLP
markers was supported by the Genetic Diversity Centre of ETH Zurich (GDC) and CCES.
The authors gratefully acknowledge laboratory support from Rochina Abbas, Caterina
Matasci, Dr. Davide Gobbin, Paolo Galli (Plant Pathology, ETH Zürich) and Dr. Aria Minder
(Genetic Diversity Center, ETH Zürich). We thank Rolf Blapp (Agroscope Changins-
Wädenswil) for the preparation of grafted material, Regula Bauermeister (Agroscope
Changins-Wädenswil) for glasshouse support, Dr. Hans-Rudolf Roth (Seminar für Statistik,
ETH Zürich) for helping with statistical analysis and Gabriella Parravicini and Dr. Toshiya
Yamamoto for kindly providing the primer sequences of SSR markers TTTC17 and NH033b,
respectively.
75
2.7. Addendum
In Chapter 2, we identified the origin of two significant fire blight resistance QTLs mapped in
the cultivar ‘Florina’ (FLO 10 and FLO 5 QTLs) by SSR-genotyping the selections and
cultivars in ‘Florina’ pedigree. The pedigree of ‘Florina’ we used differs from the one
commonly accepted by apple breeders: we described 9433-2-2 (syn. 9333-2-2) and 9433-2-8,
parents of the F2 descendant 26829-2-2, as being open-pollinated seedlings of ‘Rome Beauty’
and M. × floribunda clone 821, respectively (Fig. 2.4). However, 9433-2-2 and 9433-2-8 are
usually described as two F1 full-sibs of the cross ‘Rome Beauty’ × M. × floribunda 821
(Dayton et al., 1977; Gianfranceschi et al., 1996; Shay and Hough, 1952). This discrepancy
does not affect the origin of the FLO 10 and FLO 5 QTLs, as the QTLs “donors” are the
cultivars ‘Jonathan’ and ‘Starking’, respectively. However, for the sake of clarity, it is
desirable to explain our hypothesis of a different parentage for the genotypes 9433-2-2 and
9433-2-8.
Two studies reported inconsistencies in the F2_26829-2-2 parentage commonly accepted.
Vinatzer et al. (2004) were the first who questioned whether F2_26829-2-2 could really
derive from a cross between two F1 full-sibs of the cross ‘Rome Beauty’ × M. × floribunda
821. Indeed, Vinatzer and colleagues found that F2_26829-2-2 amplified one allele at the SSR
marker CH-Vf1 (137 bp) that was present neither in ‘Rome Beauty’ (163 and 141 bp) nor in
M. × floribunda 821 (159 and 129 bp). More recently, Evans et al. (2010) further addressed
this issue in the frame of the HiDRAS project. The authors screened founders of pedigreed
apple breeding material, including F2_26829-2-2 and its putative grandparents ‘Rome
Beauty’ and M. × floribunda 821, with a genome-covering set of 80 SSR markers. The
percentage of SSR loci of F2_26829-2-2 without any allele from either ‘Rome Beauty’ or M.
× floribunda 821 was found to be 28 %, while many SSR markers displayed an allele in
F2_26829-2-2 that was also present in M. × floribunda 821. Evans and colleagues concluded
that ‘Rome Beauty’ is likely to be incorrect in the way it is indicated in the pedigree
commonly accepted by breeders. Since the genotypes 9433-2-2 and 9433-2-8 no longer exist,
and can therefore not be checked molecularly, we made the hypothesis mentioned above, i.e.
F2_26829-2-2 may have resulted from the cross between two genotypes, one resulting from
the open-pollination of ‘Rome Beauty’ (arbitrarily chosen to be 9433-2-2), the other resulting
from the open-pollination of M. × floribunda 821 (arbitrarily chosen to be 9433-2-8) (see
pedigree of ‘Florina’; Fig. 2.4).
76
77
Chapter 3
Quantitative trait loci analysis of fire blight resistance in
an apple F1 progeny ‘Idared’ × ‘Rewena’
For a better understanding, the parts of the work performed by A. Peil and K. Richter (JKI,
Dresden, Germany) were not removed from Chapter 3. They are as follows: (i) evaluation of
the level of fire blight resistance of 92 F1 individuals from the cross ‘Idared’ × ‘Rewena’ in a
quarantine glasshouse; (ii) capillary electrophoresis of 14 SSR markers on 92 F1 individuals
from the cross ‘Idared’ × ‘Rewena’.
78
3.1. Abstract
Breeding apple (Malus × domestica Borkh.) cultivars resistant to fire blight, a major disease
caused by the bacterium Erwinia amylovora, is a cumbersome task that could be facilitated by
the application of molecular markers. A necessary preliminary to this is a good understanding
of the genetic basis of fire blight resistance in Malus spp., and especially in apple cultivars.
Previous studies identified additive and epistatic quantitative trait loci (QTLs) for fire blight
resistance in the apple cultivars ‘Fiesta’, ‘Discovery’, ‘Prima’, ‘Florina’ and ‘Nova Easygro’
using bi-parental mapping populations. In the present study, we conducted a QTL analysis on
a F1 progeny derived from a cross between the fire blight susceptible cultivar ‘Idared’ (ID)
and the cultivar ‘Rewena’ (RE), one of the most resistant to the disease. One hundred and
fifty one F1 individuals were artificially inoculated with E. amylovora strain Ea 222 and the
length of the necrotic lesion was recorded 28 days after inoculation. Of these, 92 F1
individuals were screened with diversity arrays technology (DArT) and simple sequence
repeats (SSR) markers, and two parental framework genetic maps were constructed thereof.
The genome coverage provided by the combination of DArT and SSR markers was sufficient
to perform a QTL analysis on 44.2 % and 37.1 % of ‘Idared’ and ‘Rewena’ genomes,
respectively; the coverage includes all apple genome regions where significant fire blight
resistance QTLs had been previously identified, except the linkage group (LG) 7 in ‘Rewena’
(RE 7). No significant QTL could be identified by interval mapping (IM) in ‘Idared’ or
‘Rewena’. However, a putative QTL was detected at the distal end of LG 7 of ‘Idared’ (ID 7)
by IM and the Kruskal-Wallis test, corresponding to the DArT marker aPa-526070. It appears
that the high level of fire blight resistance found in the cultivar ‘Rewena’ may have a oligo- or
polygenic basis, possibly involving QTLs alleles with similar effect and therefore
undetectable. Mapping DArT, SSR or single nucleotide polymorphism (SNP) markers at a
medium to high density in genomic regions not yet covered may allow the detection of
additional putative or significant QTLs in ‘Rewena’, and in ‘Idared’ to a lesser extent.
Implications of these results for molecular apple breeding towards fire blight resistance are
discussed.
Key-words: Erwinia amylovora, microsatellite, DArT, linkage map, Malus spp.
79
3.2. Introduction
Apple (Malus × domestica Borkh.) breeding is a long term approach which consists in
identifying the most promising individuals within progenies derived from controlled crosses
between apple cultivars or advanced selections with complementary “qualities” (Kellerhals et
al., 2009b). In general, the major objectives of apple cultivar breeding programs are high fruit
quality, good agronomic performance, and resistance to the fungal pathogens apple scab and
powdery mildew (caused by Venturia inaequalis and Podosphaera leucotricha, respectively)
(Laurens, 1999; Lespinasse, 2009; Sansavini et al., 2004). Although less important
economically than scab and powdery mildew on a worldwide scale, fire blight, caused by the
enterobacterium Erwinia amylovora, can result in devastating epidemics in apple orchards
(Bonn and van der Zwet, 2000; Holliger et al., 2008; Norelli et al., 2003; van der Zwet and
Keil, 1979). Therefore, fire blight resistance is an important additional objective of several
breeding programs in Czech Republic, Germany, Hungary, New Zealand, Poland, Switzerland
and in the USA (Kellerhals et al., 2011; Korba et al., 2008; Peil et al., 2009; Sobiczewski et
al., 2011; Toth et al., 2006). Breeding is considered all the more important, as there is no
completely efficient method to control fire blight, with the exception of streptomycin
application in regions where no streptomycin-resistant E. amylovora strain is present
(Psallidas and Tsiantos, 2000; Sundin et al., 2009).
A prerequisite to the genetic improvement of apple cultivars regarding fire blight resistance is
a better understanding of the genetic architecture of this trait in Malus spp.. This has been
initiated by several studies through the construction of genetic maps and quantitative trait loci
(QTL) analysis in wild apple species, ornamental apples and apple cultivars (reviewed in
Khan et al., 2011; Peil et al., 2009). Several wild apple species and ornamental apples were
found to be highly resistant to fire blight, making them of high interest to geneticists and
breeders (Aldwinckle and van der Zwet, 1979; Bonn and Elfving, 1990; Gardner et al., 1980;
Lespinasse and Paulin, 1984; Norelli and Aldwinckle, 1986; Peil et al., 2004; van der Zwet
and Keil, 1979). Three major QTLs for fire blight resistance were identified in three
genotypes, namely the clone ‘Robusta 5’ of Malus × robusta (QTL on linkage group (LG) 3)
(Peil et al., 2007), the ornamental apple cultivar ‘Evereste’ and the clone 821 of Malus ×
floribunda (both QTLs on LG 12) (Durel et al., 2009). The percentage of the phenotypic
variation explained (PVE) by each QTL varied from 40 % for M. × floribunda 821 to 80 %
for ‘Robusta 5’. Additionally, the phenotypic distribution of the F1 offspring from two
crosses investigated, ‘Idared’ × ‘Robusta 5’ and ‘Malling Merton 106’ (‘MM.106’) ×
80
‘Evereste’, suggested that the major QTLs of ‘Robusta 5’ and ‘Evereste’ might segregate as
major resistance genes; the testing of this hypothesis is underway (Fahrentrapp et al., 2011;
Parravicini et al., 2011). However, the poor fruit quality of ‘Robusta 5’, ‘Evereste’ and M. ×
floribunda 821 is a major hurdle to the immediate use of their strong-effect QTLs in cultivar
breeding via marker-assisted selection (MAS). Besides, such major QTLs may be associated
with a risk of breakdown when present alone in a cultivar, as suggested by the existence or E.
amylovora strains able to overcome the resistance of ‘Robusta 5’ (Fazio et al., 2008; Norelli
and Aldwinckle, 1986).
A large variability in resistance to fire blight seems to exist among apple cultivars, for which a
polygenic determinism has been hypothesized (Korban et al., 1988; Lespinasse and
Aldwinckle, 2000). Calenge et al. (2005) identified in the moderately fire blight resistant
cultivar ‘Fiesta’ a major QTL located on LG 7 explaining 35–40 % of a moderate phenotypic
variation in F1 progenies of crosses with ‘Prima’ and ‘Discovery’; this QTL, called F7 QTL,
was re-identified in another cross between ‘Fiesta’ and ‘Discovery’ (Khan et al., 2006). Khan
et al. (2007) developed two sequence-characterized amplified regions (SCAR) markers
flanking the F7 QTL and validated its effect in a third distinct genetic background derived
from a cross between ‘Milwa’ and ‘1217’, a F1 progeny from the cross ‘Fiesta’ × ‘Discovery’
carrier of the two SCAR markers. Kellerhals and colleagues reported on the first application
of the F7 QTL in an apple breeding program in Switzerland (Baumgartner et al., 2010;
Kellerhals et al., 2011). Four other additive QTLs were found by Calenge et al. (2005) in
‘Fiesta’, ‘Discovery’ and Prima’, but they showed only minor effects (explaining less than 8
% of the phenotypic variation) and were not stable at 7 and 14 days after inoculation with E.
amylovora; 12 digenic interactions were additionally detected in the crosses ‘Fiesta’ ×
‘Discovery’ and Prima’ × ‘Fiesta’. More recently, Le Roux et al. (2010; see Chapter 2)
identified two QTLs for resistance to E. amylovora on LG 10 and LG 5 of the resistant
cultivar ‘Florina’; they respectively explained 15.3 and 10.1 % of the phenotypic variation
recorded 14 days after inoculation in the F1 progeny ‘Florina’ × ‘Nova Easygro’ using
multiple QTL mapping (MQM). No significant QTL could be identified in ‘Nova Easygro’.
Besides ‘Florina’ and ‘Nova Easygro’, several apple cultivars are considered resistant to fire
blight (Khan et al., 2007; Peil et al., 2009; van der Zwet and Keil, 1979). The “Re-cultivars®
”
derived from Malus × floribunda clone 821 and bred in Dresden-Pillnitz (Germany) showed
the highest level of resistance to E. amylovora among a series of cultivars during repeated
artificial inoculations performed with a mixture of three aggressive isolates (Fischer and
Richter, 2004; Fischer and Fischer, 1999; Richter and Fischer, 2002). The genetic basis of fire
81
blight resistance in these “Re-cultivars®” is still unknown, although it would be of high
relevance for application in breeding.
In this study, we describe a QTL analysis of fire blight resistance in a F1 progeny of 92
individuals derived from a cross between ‘Idared’ (fire blight susceptible cultivar) and
‘Rewena’. We applied the diversity arrays technology (DArT) combined with simple
sequence repeats (SSR) markers to build two framework linkage maps of ‘Idared’ and
‘Rewena’, which were subsequently used for QTL analysis.
82
3.3. Material and methods
3.3.1. Plant material and evaluation of fire blight resistance
A F1 progeny of 250 individuals was derived from a cross between the two apple cultivars
‘Idared’ and ‘Rewena’ (ID × RE) and grown in an orchard at the Julius Kühn Institute (JKI)
in Dresden (Germany). In 2005, 151 randomly selected F1 individuals were replicated by
bud-grafting onto ‘Malling 9’ (‘M.9’) rootstock. The 151 F1 individuals were split in two
cabins of a quarantine glasshouse in Quedlinburg (JKI, Germany), all replications of the same
individual being in the same cabin. Each cabin displayed similar conditions: day temperature
between 25 and 27 °C, night temperature of 20 °C and relative humidity of 85 %.
All replications of the 151 ID × RE F1 offspring were inoculated with the bacterial strain Ea
222 of E. amylovora following a previously published protocol (Kleinhempel et al., 1984).
Briefly, scissors were dipped in a bacterial suspension (109 cfu/ml) and were used to incise
the tips of the two youngest leaves of each actively growing shoot (one shoot per replication).
Shoot length and length of shoot necrosis were measured 28 days after inoculation (DAI). The
parental cultivars ‘Idared’ and ‘Rewena’ were not included in the assay. However, their
phenotype (fire blight susceptible and resistant, respectively) had been repeatedly confirmed
in independent glasshouse assays under controlled conditions of temperature and relative
humidity at JKI (Germany) and at Agroscope ACW (Switzerland) (Fischer and Richter, 2004;
Isabelle Baumgartner, unpublished results). The differences in disease resistance assessment
included the inoculum (mixture of three aggressive E. amylovora isolates at JKI vs single
strain ACW 610 at ACW) and the phenotyping scale (semi-quantitative scale at JKI (Fischer
and Richter, 1999; Fischer and Richter, 2004) vs scoring of lesion length at ACW).
Of the 151 ID × RE F1 offspring inoculated, the 92 individuals with the highest number of
replications (average of 8 replications per individual) were randomly selected for linkage map
construction, statistical analysis and QTL analysis of fire blight resistance.
3.3.2. Statistical analysis
Statistical analysis was performed using the software SPSS Statistics 19 (IBM, Germany).
The lesion length in percentage of the shoot length (PLL; %) was calculated by dividing the
length of the necrotic lesion (LL; cm) by the shoot length (SL; cm). A Kruskal-Wallis test was
conducted to investigate a putative “cabin” effect on PLL (significance level 0.05). As no
significant difference was found between the two glasshouse cabins for PLL (64 % and 66 %,
respectively; p = 0.18), the data of both cabins were pooled together and checked for normal
83
distribution. Data of the F1 individuals for PLL were used to perform a one-way ANOVA in
order to calculate the broad sense heritability (h2) as described previously (Calenge et al.,
2005; Khan et al., 2006).
3.3.3. SSR and DArT markers genotyping
DNA of the parents ‘Idared’ and ‘Rewena’ and of the 92 F1 individuals was extracted from
leaf samples using the DNAeasy 96 Plant Kit (Qiagen, Hilden, Germany). DNA was
quantified and checked for integrity (absence of smears) on agarose gel (0.8 %) and finally
diluted to a concentration of 50–100 ng/µl (resp. 10–20 ng/µl) for genotyping with DArT
(resp. SSR) markers. The DArT markers were genotyped by Diversity Arrays Technology Pty
Ltd (Yarralumla, Australia) using the standard genotyping array composed of 14,592 clones
from the PstI/AluI complexity reduction method (Schouten et al., 2011; Wenzl et al., 2004).
The DArT markers with a low polymorphic information content (PIC < 0.10) or polymorphic
in both ‘Idared’ and ‘Rewena’ were excluded from subsequent analysis following Schouten et
al. (2011). Forty-eight SSR markers were selected from published genetic maps of Malus spp.
(Celton et al., 2009b; Silfverberg-Dilworth et al., 2006) to facilitate the identification and
orientation of linkage groups; the genotyping of the SSR markers was performed either on 44
F1 individuals (31 SSR markers) or on 92 F1 individuals (17 SSR markers) for the SSR
markers located in genomic regions where significant QTLs for fire blight resistance had been
previously detected; the latter SSR markers were: CH03e03, CH03g07 and AU223657-SSR
on LG 3 (Calenge et al., 2005; Peil et al., 2007); Hi04a08, CH03a09 and CH05f06 on LG 5
(Le Roux et al., 2010; Peil et al., 2011); CH01c06 and CH01f09 on LG 8 (Lalli et al., 2010);
CH02b07 and CH02a10 on LG 10 (Le Roux et al., 2010); CH03c02 and CH01d03z on LG 12
(Calenge et al., 2005; Durel et al., 2009); CH05h05 and GD147 on LG 13 (Calenge et al.,
2005). PCR amplification, capillary electrophoresis and data analysis were performed
following Patocchi et al. (2009a) for 34 SSR markers. Capillary electrophoresis and data
analysis for the 14 other SSR markers were made using an automatic dual-laser DNA
sequencer LI-COR 4200 (LI-COR, Lincoln, Nebraska, USA) as described in Patocchi et al.
(2009b).
3.3.4. Genetic map construction
Separate genetic maps were constructed for the parents ‘Idared’ and ‘Rewena’ using the
software JoinMap® version 4.0 (Van Ooijen, 2006). A minimal LOD score of 3 was applied
for grouping molecular markers. The marker order within the linkage groups was determined
84
using the exhaustive thresholds for LOD and REC (recombination frequency) (0.01 and
0.499, respectively). Such loose thresholds were applied in order not to discard large but true
recombination frequencies. Linkage groups were identified and aligned to published genetic
maps using SSR (Silfverberg-Dilworth et al., 2006) and DArT (Schouten et al., 2011)
markers. Only linkage groups necessitating one round of calculation (first round linkage
maps) were considered.
3.3.5. QTL analysis
Single marker analysis and QTL mapping were performed using the software MapQTL®
version 5.0 (Van Ooijen, 2004). First, the Kruskal-Wallis test was used to detect an
association between a single marker and the phenotypic traits LL and PLL (significance
threshold p ≤ 0.05). Interval mapping (IM) was then performed with a mapping step size of 1
cM to test for the presence of putative and significant QTLs for LL and PLL. Genome-wide
(resp. chromosomal) logarithm of odds (LOD) threshold value was calculated by the
permutation test (Doerge and Churchill, 1996) to detect significant (resp. putative) QTLs at
the 95 % confidence level. The phenotypic variation explained (PVE) by a QTL was
estimated by MapQTL®
version 5.0. Since the position and effect of the genomic regions
detected were similar for LL and PLL data, only the latter are presented in this Chapter.
85
3.4. Results
3.4.1. Evaluation of fire blight resistance and statistical analysis
The 92 F1 individuals from ID × RE were ranked by ascending order of their mean score of
PLL (Fig. 3.1). The average score of the F1 progeny for PLL was 65.5 %, with individual
mean scores ranging from 36.6 % to 95.2 %. In an independent assay performed at Agroscope
ACW (Switzerland), the mean PLL scores of the cultivars ‘Idared’ and ‘Rewena’ were 3.3 %
and 53.9 %, respectively, compared to 53.6 % for the cultivar ‘Gala Galaxy’ (Isabelle
Baumgartner, unpublished results; see Appendix A). The broad-sense heritability h2 was
0.97.
Fig. 3.1. Distribution of the 92 F1 individuals from the cross ‘Idared’ × ‘Rewena’ (ID × RE)
ranked in ascending order of their mean lesion length in percentage of shoot length (PLL).
3.4.2. Linkage map construction
Sixty seven and 64 polymorphic DArT markers mapped in ‘Idared’ and ‘Rewena’,
respectively, after removal of redundant DArT markers (61 and 37 markers, respectively).
The quality parameters calculated by the DArT software for these 131 DArT markers were as
follows: average P-value of 90.8 % (lowest value 69.4 %), average reproducibility of 99.8 %
(lowest value 97.0 %) and average call rate of 97.7 % (lowest value 79.8 %). Sixty five DArT
86
markers (35 of ‘Idared’, 30 of ‘Rewena’) had been mapped on the integrated linkage maps of
the F1 progenies ‘Prima’ × ‘Fiesta’ or 2010–2012 (Schouten et al., 2011). Out of the 48 SSR
markers selected, 28 mapped on the linkage map of ‘Idared’ and 22 mapped on the linkage
map of ‘Rewena’. Sixteen SSR markers were common to both parental linkage maps (Fig.
3.2).
A total of 95 molecular markers (67 DArT and 28 SSR) were positioned on the linkage map
of ‘Idared’ which reached 646.8 cM of length (Tables 3.1 and 3.2). The seventeen linkage
groups of the apple genome were represented, ID 12 being split in two parts (A and B). The
linkage groups of ‘Idared’ showed a variable coverage with molecular markers compared to
the SSR-based backbone linkage map established during the EU-funded project High-quality
Disease Resistant Apples for a Sustainable Agriculture (HiDRAS) (Patocchi et al., 2009b;
Silfverberg-Dilworth et al., 2006): from 107 % for ID 7 down to 1.3 % for ID 17 with an
average of 44.2 % (Fig. 3.2). The largest gap was 58.5 cM on ID 3, between the SSR markers
CH03g07 and AU223657-SSR.
A total of 86 molecular markers (64 DArT and 22 SSR) were positioned on the linkage map
of ‘Rewena’ which reached 542.9 cM of length (Tables 3.1 and 3.2). Fourteen linkage groups
of the apple genome were identified, linkage groups RE 1, RE11 and RE 16 were missing.
The linkage group RE 15 was split in two parts (A and B) and the part RE 15B could not be
orientated as only one marker (aPa-182445) had been mapped previously on LG 15 of apple
(Schouten et al., 2011). Similarly to ‘Idared’, the 14 linkage groups of the linkage map of
‘Rewena’ displayed a variable molecular marker coverage, from 111.2 % for RE 8 down to
1.5 % for RE 9 with an average of 37.1 % (Fig. 3.2). The largest gap was 41.7 cM on RE
15B, between the DArT markers aPa-443208 and aPa-182445.
87
88
Fig. 3.2. Framework parental linkage maps of ‘Idared’ (ID 1 and ID 17) and ‘Rewena’ (RE 2 to RE 17).
89
← SSR markers in italics allowed linkage groups alignment to the linkage map of the cultivar ‘Discovery’ by
Silfverberg-Dilworth et al. (2006).
DArT markers in bold allowed linkage groups alignment to the integrated linkage maps of the F1 progenies
‘Prima’ × ‘Fiesta’ and 2010–2012 (Schouten et al., 2011). Redundant DArT markers were kept if they enabled
the alignment of the ‘Idared’ and ‘Rewena’ linkage maps to the integrated linkage maps of ‘Prima’ × ‘Fiesta’
and 2010–2012.
Segments of linkage groups not covered with molecular markers were estimated based on the SSR-based
backbone linkage map of Patocchi et al. (2009b) adapted from Silfverberg-Dilworth et al. (2006).
90
Table 3.1. Description of the number and type of molecular markers, average distance between markers, length and percentage covered with
molecular markers for each linkage group (LG) of the maps of ‘Idared’ (ID) and ‘Rewena’ (RE). Linkage group 12 of ‘Idared’ and 15 of ‘Rewena’
are split in two parts, A and B.
ID1 ID2 ID3 ID4 ID5 ID6 ID7 ID8 ID9 ID10 ID11 ID12A ID12B ID13 ID14 ID15 ID16 ID17 All LGs
DArT markers 5 9 4 2 9 2 9 3 3 0 2 1 4 5 2 0 5 2 67
SSR markers 1 4 2 0 3 2 1 2 0 2 0 1 2 2 1 2 3 0 28
Total markers 6 13 6 2 12 4 10 5 3 2 2 2 6 7 3 2 8 2 95
Average distance
between markers
(cM)
6.4 4.2 14.8 12.0 6.8 7.3 7.9 10.0 9.6 6.4 7.1 4.1 2.9 6.5 6.6 5.1 5.3 0.7 6.8
Length (cM) 38.3 54.3 88.8 23.9 82.1 29.1 79.2 50.2 28.9 12.8 14.2 8.2 17.6 45.5 19.8 10.1 42.5 1.3 646.8
Length of LG in
Patocchi et al.
(2009b) (cM)
87 77 111 73 108 78 74 74 73 104 83 89 89 50 116 76 100 1462
Percentage of LG not
spanned with markers
(%)
44.0 70.5 80.0 32.7 76.0 37.3 107.0 67.8 39.6 12.3 17.1 29.0 51.1 39.6 8.7 55.9 1.3 44.2
91
RE1 RE2 RE3 RE4 RE5 RE6 RE7 RE8 RE9 RE10 RE11 RE12 RE13 RE14 RE15A RE15B RE16 RE17 All LGs
DArT markers 0 6 3 1 8 1 5 2 3 7 0 8 2 6 5 2 0 5 64
SSR markers 0 1 3 1 1 2 1 3 0 3 0 2 2 1 1 0 0 1 22
Total markers 0 7 6 2 9 3 6 5 3 10 0 10 4 7 6 2 0 6 86
Average distance
between markers
(cM)
- 7.8 9.7 9.2 6.4 4.3 2.4 16.5 0.4 2.6 - 7.7 5.8 2.9 6.5 20.9 - 2.9 6.3
Length (cM) 0 54.4 58.2 18.3 57.7 12.9 14.2 82.3 1.1 25.6 0 77 23.3 20.1 39 41.7 0 17.1 542.9
Length of LG in
Patocchi et al.
(2009b) (cM)
87 77 111 73 108 78 74 74 73 104 83 89 89 50 116 76 100 1462
Percentage of LG not
spanned with markers
(%)
0.0 70.6 52.4 25.1 53.4 16.5 19.2 111.2 1.5 24.6 0.0 86.5 26.2 40.2 69.6 0.0 17.1 37.1
92
Table 3.2. Characteristics of the 562 polymorphic DArT markers segregating in the F1 progeny ‘Idared’ × ‘Rewena’.
a: DArT markers polymorphic in ‘Idared’ only, ‘Rewena’ only and both parents segregated as following, respectively: < a0 × 00 > (hybridization signal only in ‘Idared’), < 00 ×
c0 > (hybridization signal only in ‘Rewena’) and < a0 × a0 > (hybridization signal in both parents). b: parentage was unclear when the presence of a fragment could not be ascertained in one (or both) parents due to a weak hybridization signal.
c: polymorphism information content; the maximum in a F1 progeny is 0.5, i.e. when the fragment is absent (no hybridization) in 50 % of the F1 individuals and present
(hybridization) in the other 50 % F1 individuals. d: redundant DArT markers displayed the same segregation pattern as other DArT markers in the ID × RE F1 progeny; they mapped at the same locus or within 0.5 cM when
scoring data were missing; they were excluded from mapping data set as they do not provide additional genetic information. e: markers polymorphic in both parents were not considered for mapping as they carry little genetic information (segregation type < a0 × a0 >).
f: ungrouped markers were those showing an insufficient linkage with the other markers or a skewed segregation in the F1 progeny (0.10 < PIC < 0.20).
Mapping stages DArT markers ‘Idared’ ‘Rewena’ Both parents Total
Preparation of
data
Polymorphic a 144 118 243
560 Polymorphic with unclear
parentage b
24 29 2
PIC < 0.10 c 2 0 0 2
Mapping e
Redundant d 61 37 - 98
Ungrouped f 16 17 - 33
Mapped 67 64 - 131
93
3.4.3. Statistical analysis and QTL detection
Two genomic region in ‘Rewena’ (on RE 15A and RE 15B) and four in ‘Idared’ (on ID 7, ID
8, ID 11 and ID 16) were associated with the trait PLL using the Kruskal-Wallis test (p ≤
0.05; Table 3.3). By interval mapping (IM), the LOD score associated with the markers aPa-
526070 on ID 7 was 3.92, above the chromosomal LOD threshold value (3.50) but below the
genome-wide LOD threshold value (4.40) in ‘Idared’. No other genomic region of ‘Idared’ or
‘Rewena’ displayed a LOD score above the chromosomal LOD threshold value by IM.
Table 3.3. Genomic regions of the cultivars ‘Idared’ and ‘Rewena’ associated with fire blight
resistance using the Kruskal-Wallis test (p ≤ 0.05). The molecular marker aPa-526070
associated with a LOD score above the chromosomal LOD threshold value by interval
mapping is highlighted in bold.
LG Marker Distance (cM) K* Significance
ID 7 aPa-526070 79.2 12.37 0.0005
ID 8 aPa-184064 30.6 6.56 0.05
ID 11 aPa-460857 14.2 5.11 0.05
ID 16 aPa-185094 24.6 8.81 0.005
RE 15A aPa-526179 39.0 3.90 0.05
RE 15B aPa-182445 41.7 11.17 0.001
94
3.5. Discussion
3.5.1. Impact of DArT markers on linkage map construction
The basic principle of the diversity arrays technology is to use micro-arrays to identify and
type DNA variation at several hundred or thousand loci over the whole genome of a species
(Jaccoud et al., 2001). It has been applied, solely or in combination with other molecular
markers, to construct medium-density saturated linkage maps for barley (Hordeum vulgare
L.) (Hearnden et al., 2007; Wenzl et al., 2004), wheat (Triticum spp.) (Akbari et al., 2006;
Jing et al., 2009), rye (Secale cereale L.) (Bolibok-Bragoszewska et al., 2009), triticale (×
Triticosecale Wittmack) (Alheit et al., 2011; Tyrka et al., 2011), bananas (Musa spp.)
(Hippolyte et al., 2010) and apple (M. × domestica Borkh.) (Schouten et al., 2011), among
others. All these studies demonstrated (i) the high quality of DArT scoring, including a high
reproducibility (i.e. how repeatable the scoring of fragments is), a high reliability (i.e. how
well the two phases “presence” (hybridization) vs “absence” (no hybridization) of a marker
are separated; measured by the P-value) and a high proportion of F1 individuals without
missing score (call rate); (ii) the generation of several hundreds to several thousands of
polymorphic DArT markers in segregating mapping populations; (iii) the transferability of
DArT markers across mapping populations; and (iv) the cost-efficiency of DArT markers
enabled by automation of scoring and data analysis. On the basis of this knowledge, we
assumed that a DArT genotyping array would be a valuable tool for the construction of
medium-density, well-covered linkage maps of the cultivars ‘Idared’ (ID) and ‘Rewena’ (RE).
Therefore, we genotyped 92 F1 individuals of ID × RE with the standard DArT genotyping
array for apple recently developed by Schouten et al. (2011). This array was composed of
14,592 clones generated from pooled DNA of 44 Malus accessions (founders, modern apple
cultivars and recent selections) by the PstI/AluI complexity reduction method.
Genotyping of the 92 ID × RE F1 individuals with the standard DArT array and SSR markers
resulted only in a partial coverage of the genomes of ‘Idared’ (44.2 %) and ‘Rewena’ (37.1
%) in comparison with the SSR-based backbone linkage map of Patocchi et al. (2009b). In
addition, the average density of DArT and SSR markers, although intermediate (average
genetic distance between adjacent markers of 6.8 and 6.3 cM, respectively), was very variable
across the genomes of the two cultivars (Table 3.1). There seems to be three main
explanations to these unexpected results. First, among the 560 DArT markers segregating in
the ID × RE F1 progeny with a PIC > 0.10, 243 (43.4 %) proved to be heterozygous in both
parents and were therefore unsuitable for mapping (Table 3.2), which is similar to the
95
percentage found in the 2010–2012 F1 progeny (45 %) by Schouten et al. (2011).
Technically, this means that 243 fragments of the array hybridized with DNA of ‘Idared’,
‘Rewena’ and a variable percentage of the F1 individuals (58.7 % to 83.7 %, depending on
the DArT marker). The only logical genetic interpretation of such a hybridization pattern was
that the 243 DArT markers had a < a0 × a0 > mendelian segregation in the ID × RE F1
progeny. As DArT markers can be reliably scored as dominant markers only (Andrzej Kilian,
personal communication), the genotypes “aa” and “a0” cannot be distinguished (both
genotypes leading to hybridization, i.e. “presence”, on the array) and the only identifiable
genotype is “00” (“absence”). Second, among the 229 DArT markers that mapped in ‘Idared’
(128) or ‘Rewena’ (101), 98 (42.9 %) exhibited identical segregation patterns in the ID × RE
F1 progeny compared to other DArT markers (i.e. were redundant), thereby providing no
additional genetic information (Fig. 3.3; Table 3.2). Moreover, non-redundant DArT markers
tended to map in cluster, resulting in linkage groups with uneven DArT markers distribution
(e.g. ID 12 and RE 10; Fig. 3.2). Redundancy and clustering of DArT markers across a plant
genome was already observed in rye (Bolibok-Bragoszewska et al., 2009), Eucalyptus
(Sansaloni et al., 2010) and also apple (Schouten et al., 2011). Third, large regions of the
apple genome were not represented by the 131 DArT markers mapping at unique positions in
‘Idared’ or ‘Rewena’ (Fig. 3.2). Entire linkage groups of ‘Idared’ and ‘Rewena’ are devoid of
DArT markers (ID 10, ID 15, RE 1, RE 11 and RE 16). This was also observed by Schouten
et al. (2011), although at a lower extent.
Fig. 3.3. Example of redundant DArT markers on linkage groups 9 and 16 of ‘Idared’ (ID 9
and ID 16) and linkage group 5 of ‘Rewena’ (RE 5).
96
← SSR markers in italics allowed linkage groups alignment to the linkage map of the apple cultivar ‘Discovery’
by Silfverberg-Dilworth et al. (2006).
DArT markers in bold allowed linkage groups alignment to the integrated linkage maps of the F1 progenies
‘Prima’ × ‘Fiesta’ and 2010–2012 (Schouten et al., 2011).
Segments of linkage groups not covered with molecular markers were estimated based on the SSR-based
backbone linkage map of Patocchi et al. (2009b).
Redundant DArT markers are highlighted in red.
If the standard DArT genotyping array of apple (as of May 2011) does not seem to provide
alone a sufficient marker coverage for genome-wide mapping purposes, it remains attractive
for enriching SSR or SNP-based backbone linkage maps. First, the high P-value,
reproducibility and call rate associated with the 131 DArT markers mapped in ‘Idared’ and
‘Rewena’ after removal of the redundant DArT markers are consistent with the values
reported in wheat, rye, A. thaliana and apple (Akbari et al., 2006; Bolibok-Bragoszewska et
al., 2009; Schouten et al., 2011; Wittenberg, 2007); this emphasize the high quality of this
type of marker in apple, which is essential to construct linkage maps of high quality (Liebhard
and Gessler, 2000). Second, the DArT markers mapping in ‘Idared’ and ‘Rewena’ proved to
be transferable, as 65 of them (35 from ‘Idared’, 30 from ‘Rewena’) were also mapped on the
integrated linkage maps of ‘Prima’ × ‘Fiesta’ and 2010–2012 (Schouten et al., 2011).
Moreover, a strong co-linearity between the linkage maps of the present study (‘Idared’ and
‘Rewena’) and the study of Schouten and colleagues (‘Prima’ × ‘Fiesta’ and 2010–2012) was
observed based on transferable DArT and SSR markers. Transferability and co-linearity of
DArT markers should greatly facilitate comparative mapping between apple cultivars. At last,
only few inconsistencies were noticed at the proximal end of the linkage groups ID 3
(inversion between the DArT markers aPa-187001 and aPa-553952 compared to Schouten et
al., 2011) and on RE 15B (uncertain orientation of RE 15B as the DArT marker aPa-443208
has not been mapped on the apple genome yet).
3.5.2. QTL analysis in ‘Idared’ and ‘Rewena’
Despite the high broad-sense heritability calculated for the trait PLL (0.97), no significant
QTL for fire blight resistance was detected by IM in ‘Idared’ or ‘Rewena’ (i.e. LOD plot was
below the genome-wide LOD threshold value of 4.4 in ‘Idared’ and 4.6 in ‘Rewena’). Only
one putative QTL was identified by IM on ID 7, with a LOD score of 3.92 above the
chromosomal LOD threshold value of 3.5; the QTL corresponds to the DArT marker aPa-
526070 located at the distal end of LG 7 (79.2 cM) and it is therefore clearly distant from the
97
region where the F7 QTL was found in the cultivar ‘Fiesta’ (Calenge et al., 2005; Khan et al.,
2006). The effect of this region on PLL was confirmed by the Kruskal-Wallis test (p ≤ 0.05)
(Table 3.3). Thus, the putative QTL on ID 7 linked with the DArT marker aPa-526070 would
represent one new region of the apple genome involved in the control of fire blight resistance
(Khan et al., 2011; Peil et al., 2009). Three other genomic regions on ID 8, ID 11 and ID 16
showed a significant, small effect on PLL according to the Kruskal-Wallis test (p ≤ 0.05)
(Table 3.3). However they cannot be considered as putative QTLs as the LOD plots in these
genomic regions were below the chromosomal LOD threshold values (between 3.0 and 3.5).
In ‘Rewena’, two DArT markers were significantly associated with the trait PLL by the
Kruskal-Wallis test only, on RE 15A (aPa-526179) and RE 15B (aPa-182445). These two
DArT markers mapped at different positions on the integrated LG 15 of the cross ‘Prima’ ×
‘Fiesta’ (34 cM and 90 cM, respectively) (Schouten et al., 2011). Nevertheless, one can not
exclude that aPa-526179 on RE 15A and aPa-182445 on RE 15B are actually flanking a real,
significant QTL located in between. Saturating the whole linkage group RE 15 with SSR or
SNP markers would allow testing this hypothesis. A similar approach enabled Le Roux et al.
(article submitted) to bridge with SSR markers a gap between the parts A and B of LG 2 in
the European pear cultivar ‘Harrow Sweet’ (Dondini et al., 2004); this allowed herewith to
redefine the position and effect of a major fire blight resistance QTL located in the middle of
LG 2 of ‘Harrow Sweet’.
Interestingly, the SSR marker CH01d03z mapping at the very distal end of RE 12 was not
associated with PLL by the Kruskal-Wallis test. In addition, the most fire blight resistant
individual of the ID × RE F1 progeny showed a relatively high mean PLL (36.6 %) after
inoculation with the E. amylovora strain Ea 222. Both results are in contrast to what would be
expected if a highly-efficacious QTL for resistance to fire blight (such as the QTLs of
‘Robusta 5’, ‘Evereste’ or M. × floribunda clone 821), and especially to the strain Ea 222, was
segregating in the ID × RE F1 progeny. Therefore, it can be concluded that ‘Rewena’
probably did not inherit the major QTL of M. × floribunda clone 821 located at the very distal
end of LG 12 (Durel et al., 2009). This conclusion strengthens the hypothesis that the fire
blight resistance of ‘Rewena’ may be due to multiple additive QTLs of small effects, similarly
to ‘Florina’ or ‘Nova Easygro’ (Le Roux et al., 2010); the genomic regions bearing these
QTLs may not have been identified yet in the apple genome. However, increasing the
coverage of ‘Rewena’ genome will be necessary to perform a comprehensive QTL analysis in
order to test this hypothesis.
98
In this study, the majority of the genomic regions putatively associated with fire blight
resistance (4 out of 6; Table 3.3) were so far identified in the most susceptible parental
cultivar, ‘Idared’. The presence of QTLs in ‘Idared’ may explain why this cultivar does not
display a full susceptibility to fire blight in glasshouse resistance screening, being only
moderately susceptible (PLL = 53.9 %) when the resistant control (e.g. ‘Rewena’) display
very short necrotic lesions (PLL = 3.3 %), i.e. when the disease severity is moderate (Isabelle
Baumgartner, unpublished results; see Appendix A). Furthermore, the presence of QTLs
alleles with similar effects on PLL in ‘Rewena’ may explain the low number of genomic
regions identified in this cultivar, as already postulated for the cultivar ‘Nova Easygro’ in
Chapter 2.
3.5.3. Perspectives of molecular breeding towards fire blight resistance using
resistant apple cultivars like ‘Rewena’
A full understanding of the resistance to fire blight in the ID × RE F1 progeny will require
parental linkage maps completely covered (no chromosome segment missing) with a high
density of markers evenly distributed all over the genome (no redundancy or cluster of
markers). This could be achieved by hybridizing the 92 individuals of the ID × RE F1
progeny on alternative DArT genotyping arrays generated by different, possibly more
efficient, complexity reduction methods. For instance, combining the standard (PstI/AluI) and
one alternative (PstI/EcoRI) DArT genotyping arrays on the 2010–2012 F1 progeny enabled
Schouten and colleagues to construct an integrated medium-density linkage map with 320
uniquely mapped DArT markers (240 and 80 from the two respective genotyping arrays)
(Schouten et al., 2011). Interestingly, the alternative array produced a markedly reduced
number of redundant DArT markers compared to the standard array (23 % vs 52 % ) Another
option would be to genotype the ID × RE F1 progeny with co-dominant specific molecular
markers, like SSR or SNP markers. Numerous SSR and SNP markers have been recently
developed in the cultivated apple (Celton et al., 2009b; Chagné et al., 2008; Han et al., 2011;
Han et al., 2009; Micheletti et al., 2011; van Dyk et al., 2010; Wang et al., 2011). Single
nucleotide polymorphism markers seems superior to SSR markers for construction of high-
density linkage maps due to their abundance and amenability to automation, for instance
using the multiplex genotyping technology SNPlex™
(Applied Biosystems Inc., Foster City,
USA; Micheletti et al., 2011) or the Illumina Infinium apple SNP array (Illumina Inc., San
Diego, California, USA; www.rosaceae.org). A high-density SNP-based linkage map of
‘Rewena’ may allow to detect QTLs with small but nevertheless significant effect by IM. If
99
only few, if any, significant QTLs for fire blight resistance were to be identified with such a
high-density linkage map, then the hypothesis of “functionally homozygous” resistance QTLs
in ‘Rewena’ would prove to be true. One would then need to construct a linkage map from a
cross between a ID × RE F1 seedling combining all the favorable QTL alleles at a
heterozygous state and a fire blight susceptible apple cultivar (e.g. ‘Gala’ or ‘MM.106’); such
a cross would make the favorable QTL alleles of ‘Rewena’ segregate in a F2 progeny, thus
enabling their mapping on the apple genome.
On a breeding point of view, the provisory results of this study contribute to the discussion
about how to use the most efficiently molecular markers for fire blight resistance identified in
apple cultivars. In apple like in many other crops, marker-assisted breeding (MAB) is a
powerful tool to pyramid a reduced number of loci with strong effect on the targeted trait,
such as major genes for apple scab and powdery mildew resistance (Bus et al., 2009;
Kellerhals et al., 2009a). These major genes often originate from wild species or hybrids with
poor agronomic performance and fruit quality. In contrast, MAB has not yet been widely
applied to improve quantitative traits due, in part, to the difficulty to reliably select for small-
effect QTLs. In Malus spp, high levels of quantitative fire blight resistance are found in few
apple cultivars like ‘Rewena’ that have fruits of good quality, unlike wild Malus species or
hybrids. Such quantitative, presumably polygenic resistance to fire blight might be more
efficiently exploited in the future using a variant of MAB called genomic selection (Hamblin
et al., 2011; Meuwissen et al., 2001). This strategy does not focus on selecting for specific
genes or QTLs whose positions and effects were prealably estimated by classical mapping
strategies. Actually, genomic selection can be described in its simplest form as a two-step
approach. First, it utilizes both genome-wide markers data and phenotypic data of a “training
population” (composed of individuals not necessarily related) to establish a “prediction
equation”. Second, a “selection population” (preferably related to the “training population”) is
genotyped with the same set of genome-wide markers and the equation is then applied to
predict the phenotype of each individual of the “selection population” (Meuwissen et al.,
2001). Thanks to the recent development of an Illumina Infinium apple array composed of
9,000 SNP markers (Chagné et al., unpublished), the first results of genomic selection for
quantitative diseases and pests resistance in apple are expected to be released soon (Kumar et
al., 2011).
100
3.6. Acknowledgements
This research was funded by the ZUEFOS Project (Züchtung feuerbrandtoleranter Obstsorten,
Federal Office for Agriculture FOAG, Switzerland). Dr. Andrzej Kilian (Diversity Arrays
Technology, Pty Ltd, Yarralumla, Australia) is gratefully acknowledged for advice in DArT
markers genotyping and mapping. Verena Knorst (Agroscope ACW, Wädenswil,
Switzerland) is also acknowledged for her precious help with SSR markers genotyping.
101
102
103
Chapter 4
Genetic mapping of the T-DNA integration site in the
BpMADS4-transgenic apple line T1190
Published in the journal New Phytologist as:
Flachowsky H, Le Roux P-M, Peil A, Patocchi A, Richter K, Hanke M-V (2011) Application
of a high-speed breeding technology to apple (Malus × domestica) based on transgenic early
flowering plants and marker-assisted selection. New Phytologist: Article first published
online: 8 July 2011.
The work presented in Chapter 4 is the contribution of P.-M. F. Le Roux to the article
published in the journal New Phytologist. The Introduction, Materials and methods and
Results sections were modified for the sake of clarity. The discussion section was enriched
compared to the original article.
104
4.1. Abstract
One characteristic of the life cycle of higher plants is the juvenile phase, which is the period
of time during which young seedlings cannot be induced to flower. In perennial crops like the
domesticated apple (Malus × domestica Borkh.), this juvenile phase hampers long-term
breeding programs, preventing recurrent selection cycles as well as introgressions from
related wild species. Agro-technical approaches involving for instance grafting onto dwarfing
rootstocks or optimal growth conditions were able to shorten the juvenile phase down to 18
months, but not less. Recently, one early flowering transgenic line named T1190 was obtained
by over-expressing the BpMADS4 gene from silver birch (Betula pendula Roth.) in the apple
cultivar ‘Pinova’; the first flowers were produced within several months after transfer to the
glasshouse. It was shown that the transgenic line T1190 carries one single copy of the T-DNA
containing the BpMADS4 transgene. In this chapter, we describe the mapping of the T-DNA
integration site on linkage group 4 of the apple genome using a genetic map of the apple
cultivar ‘Discovery’ previously constructed. The relevance of the knowledge about the T-
DNA integration site is discussed regarding the use of the early flowering transgenic line
T1190 in a high-speed introgression program in apple.
Keywords: Malus × domestica Borkh., early flowering, thermal asymmetric interlaced-PCR,
single nucleotide polymorphism, genetic map.
105
4.2. Introduction
The juvenile phase in higher plants is the period during which young seedlings cannot be
induced to flower (Goldschmidt and Samach, 2004). It can last between four and twelve years
in Malus spp. (Fischer and Richter, 2004; Hanke et al., 2007; Sadamori et al., 1963; Visser,
1964), thus hampering the genetic improvement of the cultivated apple (Malus × domestica
Borkh.) regarding two strategies: (i) classical crossbred-breeding programs (delay of several
years between seed planting and first evaluation of seedlings for fruit quality); (ii) long-term
breeding strategies, such as recurrent selection and introgression of agronomically relevant
traits (e.g. diseases resistance) from wild Malus species or hybrids. Numerous agro-technical
methods have been investigated to reduce the juvenile phase in apple seedlings. Some of them
could initiate a precocious flowering, however it has been impossible to reduce the juvenile
phase to less than 18 months (Flachowsky et al., 2009; Hanke et al., 2007; Volz et al., 2009).
Over the last ten years, evidence has been adduced revealing that the juvenile phase of fruit
trees could be effectively shortened using gene transfer technologies (Elo et al., 2007; Hanke
et al., 2007; Kotoda et al., 2003; Penã et al., 2001). In particular, Flachowsky and
collaborators succeeded in breaking the juvenile phase in apple by over-expressing the
BpMADS4 gene of silver birch (Betula pendula Roth.) in 25 transgenic lines of the apple
cultivar ‘Pinova’ (Flachowsky et al., 2007). Flower induction on transgenic plants occurred
already on shoots in vitro. Glasshouse plants of three BpMADS4-transgenic lines (T1165,
T1187 and T1190) developed flowers within several months after transfer to a glasshouse. In
a following study, the BpMADS4-transgenic lines were further evaluated with regard to
morphological characteristics (Flachowsky et al., 2011). Southern blot analysis were also
performed to identify the number of T-DNA copies inserted in the genome of each
BpMADS4-transgenic line. The transgenic line T1190, showing a single integration site of the
T-DNA, was deemed best suited for application in a high-speed breeding technology because
of the shortness of its juvenile phase, its continuous flower production and its architecture
compatible with fruit bearing (Flachowsky et al., 2011). The present chapter reports on the
genetic mapping of the T-DNA integration site in the BpMADS4-transgenic line T1190. This
work is part of the in-depth characterization of the BpMADS4-transgenic line T1190 described
in entirety in Flachowsky et al. (2011).
106
4.3. Materials and methods
The right and left regions flanking the T-DNA in the BpMADS4-transgenic line T1190 were
isolated by Flachowsky et al. (2011) following a thermal asymmetric interlaced (TAIL)-PCR
protocol previously described (Liu et al., 2005). Primers Right_T1190F (5’-
TACAAGAGCCCTGTTGCAATGC-3’) and Right_T1190R (5’-
CAAGTACAAGGATCTTTATGCA-3’) were designed in the right T-DNA flanking region.
They were tested for amplification in the cultivars ‘Fiesta’, ‘Discovery’, ‘Florina’ and ‘Nova
Easygro’, previously used for construction of the linkage maps ‘Fiesta’ × ‘Discovery’ and
‘Florina’ × ‘Nova Easygro’ (Le Roux et al., 2010; Liebhard et al., 2003a; Silfverberg-
Dilworth et al., 2006). The PCR mix contained 2 μl of genomic DNA (20 ng/μl), 7 μl of
ultrapure water from a Direct-Q® 3 Water Purification System (Millipore, Zug, Switzerland),
10 μl of Reaction Mix from the HotStar HiFidelity Polymerase Kit (Qiagen, Magden,
Switzerland) and 0.4 μl of each primer (10 μM). PCR amplification conditions were as
follows: 15 min of initial denaturation at 94 °C, 35 cycles at 94 °C for 60 s, 55 °C for 30 s and
72 °C for 60 s, and 5 min of final extension at 72 °C. The presence of amplified fragments
was checked on a 1 % agarose gel. Sequencing was performed as previously described
(Rezzonico et al., 2009).
Presence of single nucleotide polymorphisms (SNPs) in the amplicons differentiating ‘Fiesta’
and ‘Discovery’, and ‘Florina’ and ‘Nova Easygro’, was investigated using the Software
Sequencher 4.9 (Gene Codes Corporation, Ann Arbor, USA). Segregation of a SNP detected
in ‘Discovery’ was followed in 54 plants of the segregating F1 progeny ‘Fiesta’ ×
‘Discovery’. The position of the locus SNP_T1190 on the linkage map ‘Discovery’ was
calculated using the Kosambi mapping function with the software JoinMap®
4 (Van Ooijen,
2006) by adding the segregation data of SNP_T1190 to those used to generate the linkage
maps of Silfverberg-Dilworth et al. (2006).
107
4.4. Results
The primers Right_T1190F and Right_T1190R amplified one single band in the cultivars
‘Fiesta’, ‘Discovery’, ‘Florina’ and ‘Nova Easygro’. After sequencing of the corresponding
fragment, one SNP named SNP_T1190 was identified in the cultivar ‘Discovery’ whereas
‘Fiesta’, ‘Florina’ and ‘Nova Easygro’ did not display any polymorphism. The alleles of
‘Discovery’ at this locus segregated in the F1 progeny ‘Fiesta’ × ‘Discovery’ with a slight but
significant distortion (χ2 = 3.63) at p = 0.5. Nevertheless, the integration site of the T-DNA
could be mapped at the locus SNP_T1190 on linkage group (LG) 4 of ‘Discovery’ at 50.3 cM
using a logarithm of odds (LOD) of 5.0 (Fig. 4.1).
Fig. 4.1. Genetic mapping position of locus SNP_T1190 on linkage group 4 of ‘Discovery’ as
previously constructed by Silfverberg-Dilworth et al. (2006).
Genetic distances are indicated on the left side of ‘Discovery’ LG 4 in centiMorgan (cM).
108
4.5. Discussion
4.5.1. Relevance of the mapping position of the T-DNA integration site for
accelerated marker-assisted introgression/pyramiding
Knowledge about the site of T-DNA integration in the apple genome is essential in order to
efficiently use the BpMADS4-transgenic line T1190 in high-speed marker-assisted
introgression programs. The introgression of a locus of interest primarily requires
hybridization between a wild “donor” of the locus of interest and the early flowering
transgenic line. The T-DNA and the locus of interest will segregate independently in the first
generation (F1 progeny) as they are inherited separately from the two parents, no matter
where they are located in the genome. The ratio of F1 individuals carrying both loci is
therefore expected to be approximately 25 % according to Mendelian segregation. However,
the map position of the T-DNA is relevant starting at the second generation (BC’1, first
pseudo-backcross). Indeed, if the T-DNA and the locus of interest are located on homologous
chromosomes, then both loci will be inherited separately in a BC’1 progeny; BC’1 individuals
inheriting the two loci are only expected in cases where a crossing over between the loci has
taken place during meiosis in the parental transgenic F1 seedling. The shorter the genetic
distance between the loci, then the less recombination will occur between them during
meiosis and the less BC’1 individuals will inherit both loci. The “worst case scenario” occurs
if the T-DNA and the locus of interest are located in the same genomic region of two
homologous chromosomes, leading to a T-DNA segregating in the BC’1 progeny in repulsion
to the locus of interest, and vice versa. In contrast, if the T-DNA and the locus of interest are
located on two non-homologous chromosomes, then they will segregate independently in the
BC’1 progeny just as they did in the F1 progeny. In this case, a 25 % ratio of carriers of both
loci will be expected.
The accelerated introgression program based on the BpMADS4-transgenic line T1190
described by Flachowsky et al. (2011) is focused on (i) the introgression of the resistance to
fire blight (caused by Erwinia amylovora) from the accession 76 of the wild apple Malus
fusca (Julius Kühn Institute, Dresden, Germany); (ii) the pyramiding of this trait with several
genes/quantitative trait loci (QTLs) for resistance to apple scab, powdery mildew (incited by
Venturia inaequalis and Podosphaera leucotricha, respectively) and fire blight (Fig. 4.2).
Malus fusca was chosen by Flachowsky et al. (2011) as source of fire blight resistance
because several studies had reported a high level of disease resistance in some genotypes of
this species (Aldwinckle et al., 1999; Gardner, 1976; Peil et al., 2009; van der Zwet and Keil,
109
1979). However, no major gene or QTL for fire blight resistance has been mapped so far in
M. fusca. Therefore, transgenic F1 seedlings from the cross T1190 × M. fusca were evaluated
for fire blight resistance by phenotypic screening with the E. amylovora strain Ea 222.
Marker-assisted seedling selection (MASS) was initiated by Flachowsky and collaborators at
the BC’1 generation of the introgression program, when the apple scab and powdery mildew
resistant parents ‘Regia’ and 98/6-10 were used for the first pseudo-backcrosses ((T1190 × M.
fusca) × ‘Regia’ and (T1190 × M. fusca) × 98/6-10; Fig. 4.2). The apple scab resistance genes
Rvi2 and Rvi4 (both on LG 2) and the fire blight resistance QTL F7 (LG 7) originating from
the cultivar ‘Regia’ are not located on LG 4 where the T-DNA is located (Fig. 4.1).
Consequently, it was possible to obtain three transgenic BC’1 seedlings derived from a fire
blight resistant transgenic F1 seedling and carrying the favourable alleles at all molecular
markers of Rvi2 and Rvi4 (one simple sequence repeat (SSR) marker and one sequence
characterized amplified region (SCAR) marker each); two of the three transgenic BC’1
seedlings are carrying additionally the two SCAR markers flanking the F7 QTL. Similarly,
the powdery mildew resistance genes Pl-1 and Pl-2, originating from the advanced selection
98/6-10, are located on LG 12 and 11, respectively. The favourable allele at the marker AT20-
SCAR flanking the Pl-1 gene and the PCR fragment for the Pl-2 gene could be detected in
two transgenic BC’1 seedlings, one of them derived from a fire blight resistant transgenic F1
seedling. The fire blight resistant BC’1 seedlings cumulating either the genes Rvi2, Rvi4 and
the F7 QTL, or the genes Pl-1 and Pl-2, will be pseudo-backcrossed with apple cultivars that
have good fruit quality (e.g. ‘Golden Delicious’) to remove all unwanted traits of M. fusca
step by step (Fig. 4.2).
110
Fig. 4.2. Schematic description of the crossbred-breeding scheme of Flachowsky et al. (2011).
The designation of the different crosses is given in orange boxes. TSA … TSE: transgenic seedling of cross A
… E; FB-F7: fire blight resistance QTL; Rvi2 and Rvi4: scab resistance genes; Pl-1 and Pl-2: powdery mildew
resistance genes; F1: first filial generation; BC’1/2: pseudo-backcross generation 1/2.
4.5.2. Importance of solid trait-locus-marker associations for marker-assisted
gene pyramiding
In marker-assisted gene pyramiding, a breeder should always consider the solidity of the trait-
locus-marker (TLM) association (Edge-Garza and Peace, 2010), i.e.: (i) how much of the
phenotypic variation is explained by each gene/QTL of interest; and (ii) how close are the
molecular marker(s) to the gene or QTL of interest. In the BC’1 progenies derived from the
crosses (T1190 × M. fusca) × ‘Regia’ and (T1190 × M. fusca) × 98/6-10, the segregating
genes Rvi2, Rvi4, Pl-1 and Pl-2 are all considered as major genes, which means that they were
found to confer individually a high level of scab or powdery mildew resistance when present
in a genotype (Bus et al., 2009; Bus et al., 2005b; Knight and Alston, 1968). Pyramiding the
Rvi2 and Rvi4 genes is relevant to apple breeders because both genes have a large and
complementary spectrum of resistance to the currently known scab races (www.vinquest.ch);
in particular, they both show an incompatible interaction with the so far-tested scab isolates of
races 6 and 7 that overcome the Rvi6 (formerly Vf) resistance present in many modern apple
cultivars (Parisi et al., 1993; Parisi et al., 2004). Combining the Pl-2 gene with another gene
such as Pl-1 seems all the more necessary for durable resistance to P. leucotricha in apple as
Pl-2 has been recently overcome by a virulent population of the pathogen (Caffier and
111
Laurens, 2005). Unlike the Pl-2 gene, the Pl-1, Rvi2 and Rvi4 genes could not be selected
with very tightly linked molecular markers. The development of very tightly linked, and if
possible co-segregating, markers for the Pl-1, Rvi2 and Rvi4 genes by using additional
mapping populations and the apple genome sequence (Velasco et al., 2010) would be
advisable to optimize TLM associations. Moreover, pyramiding more than two major genes in
a single cultivar might be necessary to achieve durable resistance to a particular disease, as
observed in the Malus-V. inaequalis pathosystem (Bus, 2006; Bus et al., 2005b; Gessler et al.,
2006). A reason for this is that fungal pathogens can apparently afford the loss of many
avirulence genes without loss of fitness (Bus, 2006; Parlevliet, 2002). Contrary to Rvi2, Rvi4,
Pl-1 and Pl-2, the locus F7 from ‘Regia’ is not a major gene but a QTL whose presence is not
sufficient to confer alone a high level of fire blight resistance. Nevertheless, the F7 QTL is of
interest for MASS purposes because it appears stable in different genetic backgrounds
(Baumgartner et al., 2011; Calenge et al., 2005; Khan et al., 2006; Khan et al., 2007) and also
because of the availability of two flanking, easy-to-use SCAR markers (Khan et al., 2007).
More generally, numerous major genes/QTLs for apple scab, powdery mildew and fire blight
resistance have been mapped in the apple genome for which molecular markers suitable for
MASS are available or in development (Bus et al., 2011; Gardiner et al., 2007; Gessler et al.,
2006; Khan et al., 2011; Peil et al., 2009). As none of them, except one (apple scab resistance
gene Rvi3; Bus et al., 2011), map on LG 4 where the T-DNA of the transgenic line T1190 is
inserted, they represent ideal targets for accelerated introgression and pyramiding in a
genotype using the high-speed breeding technology developed by Flachowsky et al. (2011).
4.6. Acknowledgements
This research was funded by the ZUEFOS Project (Züchtung feuerbrandtoleranter Obstsorten,
Federal Office for Agriculture FOAG, Switzerland).
112
113
Chapter 5
Use of a transgenic early flowering approach in apple
(Malus × domestica Borkh.) to introgress the major
quantitative trait locus for fire blight resistance of the
apple genotype ‘Evereste’
Published in the journal Molecular Breeding as:
Le Roux P-M F, Flachowsky H, Hanke M-V, Gessler C, Patocchi A (2011) Use of a
transgenic early flowering approach in apple (Malus × domestica Borkh.) to introgress the
major quantitative trait locus for fire blight resistance of the apple genotype ‘Evereste’.
Molecular Breeding: Article first published online: 22 November 2011.
114
5.1. Abstract
The long juvenile phase of Malus spp. has always been a major drawback for the
introgression of agronomically relevant traits (e.g. diseases resistance) from wild apples, crab
apples or ornamental genotypes into domestic apple cultivars (M. × domestica Borkh.). In the
past, several agro-technical approaches were investigated but none was able to cut the
juvenile phase down to less than 18 months. Recently, one early flowering transgenic line
named T1190 was obtained by over-expressing the BpMADS4 gene from silver birch (Betula
pendula Roth.) in the apple cultivar ‘Pinova’. In this study, we report on the acceleration of
the first two introgression cycles (F1 and BC’1) of the highly efficacious fire blight resistance
locus Fb_E from the ornamental apple cultivar ‘Evereste’, using the BpMADS4-transgenic
line T1190. A background selection based on simple sequence repeats (SSR) markers
regularly distributed over the apple genome was applied on the 24 BC’1 seedlings carrying
the BpMADS4 transgene and the Fb_E locus. Two early flowering BC’1 seedlings estimated
to carry less than 15 % of the genome of ‘Evereste’ were identified. They are currently (July
2011) being used in reciprocal crosses with the apple cultivar ‘Royal Gala’ to continue the
introgression of the Fb_E locus. Additionally, the strong phenotypic effect of the Fb_E locus
from ‘Evereste’ was confirmed by artificially inoculating a F1 progeny T1190 × ‘Evereste’
with the causal agent of fire blight, Erwinia amylovora. Possible ways of enhancing the fast
introgression of disease resistance genes in domestic apple using the transgenic line T1190
are discussed.
Keywords: high-speed breeding technology, marker-assisted selection, microsatellite,
Erwinia amylovora.
115
5.2. Introduction
The production of the cultivated apple (Malus × domestica Borkh.), equivalent to 70 million
metric tons a year (http://faostat.fao.org/), is today dominated by a handful of successful
cultivars, namely ‘Golden Delicious’, ‘Red Delicious’, ‘Cox’s Orange Pippin’, ‘Granny
Smith’, ‘McIntosh’, ‘Jonathan’ and ‘Braeburn’, mostly selected from chance seedlings over
100 years ago, as well as ‘Fuji’, ‘Gala’ and ‘Jonagold’ obtained from crossbred breeding
programs (Gardiner et al., 2007). The trend to genetic uniformity of commercial apple
orchards was accentuated by the traditional apple breeding strategy in the 20th
century which
consisted in crossing among the few popular commercial cultivars and their derived
selections. The cultivars ‘Golden Delicious’, ‘Red Delicious’, ‘Jonathan’, ‘McIntosh’ and
‘Cox Orange Pippin’ were thus found in the parentage of numerous modern cultivars (Noiton
and Alspach, 1996). It was suggested that this traditional breeding strategy would not be
sustainable in the long term (Kumar et al., 2010; Noiton et al., 1998). A greater genetic
diversity is considered as necessary to develop innovative cultivars meeting the future market
needs (such as increased health benefits), being more productive and more resistant to
diseases and pests, reducing handling costs as well as the risk of inbreeding depression
(Kumar et al., 2010).
Different approaches were proposed to enlarge the genetic basis of apple breeding programs
or to introduce specific novel traits in domestic apple cultivars. In Europe, several national
and regional projects of inventory and characterization of local apple varieties (syn.
landraces) were implemented (Kellerhals and Egger, 2004; Lateur et al., 1998; Laurens et al.,
2004a; Laurens et al., 2004b; Pereira-Lorenzo et al., 2007; Szalatnay et al., 2009); a detailed
evaluation of the landraces at the molecular and phenotypic level (fruit quality traits, diseases
resistance, …) may provide breeders with additional genotypes to be used as parents in their
programs. The United States Department of Agriculture (USDA) has been characterizing for
many years at the Plant Genetic Resources Unit (PGRU) located in Geneva (New York, USA)
1,582 Malus sieversii and 822 Malus orientalis trees for a broad range of phenotypic traits
using 30 descriptors (Forsline and Aldwinckle, 2004); these evaluations allowed already
pomologists and geneticists to verify that M. sieversii is very diverse and has all the qualities
present in M. × domestica (Borkh.). Some M. sieversii genotypes proved to be very similar to
commercial cultivars for many critical horticultural traits and may be thereby directly
valuable to breeders (Fazio et al., 2009; Forsline and Aldwinckle, 2004; Luby et al., 2006).
The Institute of Plant and Food Research (PFR) in New Zealand has developed a recurrent
116
breeding program based on a highly diverse Apple Breeding Population, established by
collecting open-pollinated seeds of about 520 genotypes of commercial cultivars, breeding
selections and wild apple species worldwide (Noiton and Shelbourne, 1992); this long-term
recurrent breeding strategy, based on general combining ability (RS-GCA), is expected to
enable the continual improvement of the population to finally provide good genotypes for use
as parents (Kumar et al., 2010). A fourth approach is to use wild apple species, crab apples
(Malus species, subspecies or hybrids distinct from M. × domestica Borkh. with a small and
bitter pome fruit) or ornamental apples as sources of agronomically relevant traits (e.g.
diseases resistance) to be introgressed in a domestic apple genetic background by a succession
of pseudo-backcrosses. For autogamous annual crops (e.g. wheat, rice, soybean), the
introgression of major genes or quantitative trait loci (QTL) from wild germplasm into
commercial varieties is a relatively standard breeding strategy. It is not as simple in non-
autogamous perennial crops like the cultivated apple. First, a gametophytic self-
incompatibility mechanism guarantees out-breeding in the Spiraeoideae subfamily (Rosaceae
family) and thus prevents “true” backcrossing (de Nettancourt, 2001; Dreesen et al., 2010).
Second, and most importantly, the juvenile phase (period elapsing between seed germination
and the attainment of the ability to flower) can last up to 12 years in Malus spp. (Fischer,
1994; Hanke et al., 2007; Visser, 1964), thereby considerably slowing down the introgression
of a trait of interest. For instance, ‘Prima’, the first cultivar carrying the Rvi6/Vf apple scab
resistance from the crab apple M. × floribunda clone 821, was released in the USA in 1970,
i.e. 56 years after C. S. Crandall (in 1914) performed the first hybridizations with M. ×
floribunda clone 821 (Crandall, 1926; Janick, 2006); more precisely, almost 30 years and two
additional pseudo-backcrosses were necessary to the Purdue-Rutgers-Illinois (PRI) breeding
program to introgress consciously the Rvi6/Vf scab resistance from the F2 descendant 26829-
2-2 of M. × floribunda clone 821 (originating from the work of C. S. Crandall and identified
by F. Hough in 1943) into the first bred scab resistant apple cultivar ‘Prima’ (Bus, 2006;
Crosby et al., 1992; Hough, 1944; Janick, 2006).
In apple, several agro-technical approaches have been investigated to shorten the juvenile
phase and thus the breeding cycle (time between seed planting and fruiting of the seedling),
such as keeping seedlings in a continuous state of growth (Aldwinckle, 1976) or budding
seedlings onto dwarfing rootstocks (Soejima et al., 1998). Recently, optimal conditions in a
controlled environment were utilized to accelerate the onset of flowering in apple, including
high temperatures and relative humidity (26 °C to 30 °C and 85 %, respectively), high
irradiance (> 900 μmol m-2
s-1
), long photoperiod (18 h photoperiod), high-red:far-red ratio,
117
elevated CO2 concentration (~1800 ppm CO2) and non-limiting fertigation (Volz et al., 2009).
None of these technical approaches was able to shorten the juvenile phase to less than 18
months (Hanke et al., 2007; Meilan, 1997; Volz et al., 2009). Alternative transgenic
approaches have been tested to shorten more drastically the juvenile phase in apple. In
particular, the effect of over-expressing the FRUITFULL (FUL)-homolog BpMADS4 gene
from silver birch (Betula pendula Roth.) in the apple cultivar ‘Pinova’ was investigated
(Flachowsky et al., 2007). It appeared that the CaMV35::BpMADS4 gene could induce a
dramatic reduction of the juvenile period in 25 transgenic lines of ‘Pinova’; eight transgenic
lines produced solitary flowers on in vitro shoots and two other lines produced first flowers
three to four months after transfer to the glasshouse (Flachowsky et al., 2007). Shortly after,
Flachowsky and colleagues provided the proof of concept of the use of transgenic early
flowering trees to shorten breeding cycles in apple: one of the 25 transgenic lines, T1190, was
crossed with the wild apple Malus fusca; four transgenic F1 seedlings resulted from this cross
that flowered within few weeks after seed planting; in turn, these transgenic F1 seedlings
were pollinated with the cultivar ‘Topaz’ and 41 seeds from this first pseudo-backcross
(BC’1) germinated, meaning that one crossbred generation per year is feasible (Flachowsky et
al., 2009). Recently, the in-depth characterization of the BpMADS4-transgenic line T1190 at
the molecular and phenotypic level was described, as well as its use in a fast breeding
program for pyramiding sources of resistance to scab, powdery mildew and fire blight (incited
by Venturia inaequalis, Podosphaera leucotricha and Erwinia amylovora, respectively)
(Flachowsky et al., 2011). In this study, we describe the use of the transgenic line T1190 to
accelerate the marker-assisted introgression of the fire blight resistance locus Fb_E from the
ornamental apple ‘Evereste’ (Durel et al., 2009; Parravicini et al., 2011) over two breeding
generations (F1 and BC’1); we also report the selection of BC’1 seedlings for the lowest
possible genomic contribution from ‘Evereste’ using a set of simple sequence repeats (SSR)
markers spanned over the apple genome. Moreover, possible enhancements of the transgenic
early flowering-based introgression strategy are discussed.
118
5.3. Materials and methods
5.3.1. Plant material and introgression scheme
The first crosses were performed reciprocally between the BpMADS4-transgenic apple line
T1190, originating from the apple cultivar ‘Pinova’ (Flachowsky et al., 2007), and the
ornamental apple cultivar ‘Evereste’ carrying the major QTL for fire blight resistance Fb_E
(Durel et al., 2009). Among the resulting F1 offspring of the cross T1190 × ‘Evereste’
(hereafter designated as T × E), two seedlings carrying the BpMADS4 transgene and the Fb_E
locus, F1_74 and F1_81, were used to pollinate 3-years old potted trees of the apple cultivars
‘Topaz’ and ‘Maloni Sally®
’ grafted on ‘Malling 9’ (‘M.9’) rootstock (pseudo-backcrosses,
BC’1). Reciprocally, the own-rooted transgenic F1 seedling number 5 (hereafter referred to as
F1_5), derived from the cross ‘Evereste’ × T1190 (hereafter designated as E × T) and also
carrier of the Fb_E locus, was fertilized with pollen of the apple cultivar ‘Milwa’ (‘Diwa®
’).
Among the resulting BC’1 offspring, nine seedlings carrying the BpMADS4 transgene and the
Fb_E locus were pollinated with the apple cultivar ‘Royal Gala’ and are currently being used
as pollinators of ‘Royal Gala’ trees grafted on ‘M.9’ rootstock (as of July 2011) (Table 5.1).
119
Table 5.1. Crosses and offspring of the first two introgression cycles of the fire blight resistance locus Fb_E from the ornamental apple cultivar
‘Evereste’ using the early flowering transgenic apple line T1190.
Breeding
cycle Progeny
N° of maternal
trees pollinated
N° of flowers
pollinated N° of fruits N° of seeds N° of seedlings
N° of BpMADS4-
transgenic seedlings
Time from seed planting to first
flowering of BpMADS4-
transgenic seedlings (in weeks)
F1 T1190 × ‘Evereste’ 15 54 28 134 56 35 15 -> 43
‘Evereste’ × T1190 3 58 3 6 6 3 44 -> 50 a
BC’1 c
‘Topaz’ × F1_74 b 1 62 5 16 10 6 14 -> …
‘Topaz’ × F1_81 b 1 67 8 49 35 20 14 -> …
‘Maloni Sally®
’ ×
F1_74 1 50 7 33 26 11 22 -> …
‘Maloni Sally®
’ ×
F1_81 2 135 0 0 0 0 0
F1_5 × ‘Milwa’ d 1 30 9 39 19 12 28 -> …
BC’2 e
BC’1_s × ‘Royal Gala’ 9 27 9 - - - -
‘Royal Gala’ × BC’1_s 1 21 - - - - -
a One of the three BpMADS4-transgenic F1 seedlings derived from the cross ‘Evereste’ × T1190 suffered from hydric stress; its development was impaired and its time to first
flowering was thus not recorded. b Seedlings F1_74 and F1_81 were selected as pollinators for BC’1 crosses among all BpMADS4-transgenic F1 seedlings carrying the fire blight resistance locus Fb_E because
they flowered the most precociously. c For reasons of space limitation in the glasshouse cabins, the BpMADS4-transgenic BC’1 seedlings not carrying the Fb_E locus were not kept after molecular selection; therefore
the time to first flowering applies only to BC’1 seedlings carrying the BpMADS4 transgene and the Fb_E locus. d The BpMADS4-transgenic seedling F1_5, carrier of the Fb_E locus, was selected as maternal tree to be pollinated by the apple cultivar ‘Milwa’ because of an architecture
compatible with fruit bearing (main shoot of 150 cm long). e Ten BpMADS4-transgenic BC’1 seedlings carrying the Fb_E locus and derived from the four BC'1 crosses ‘Topaz’ × F1_74, ‘Topaz’ × F1_81, ‘Maloni Sally
®’ × F1_74 and
F1_5 × ‘Milwa’ have flowered so far (as of July 2011) and nine have been pollinated with the apple cultivar ‘Royal Gala’. A reciprocal cross between ‘Royal Gala’ and BC’1
seedlings (so far BC’1_7, 19, 21, 75 and 85) is underway.
120
5.3.2. Experimental conditions
Trees of the transgenic line T1190 and of apple cultivars, as well as F1 and BC’1 seedlings,
were grown under long day conditions (16 h light and 8 h darkness) at 24 °C during the day
and 20 °C during the night; relative humidity was maintained between 35 and 50 %
throughout the experiment. Transgenic and non-transgenic pollen was collected, stored and
used for hybridization as described previously (Flachowsky et al., 2007). Seeds derived from
F1 and BC’1 crosses were extracted from fully ripe fruits and stratified according to
Flachowsky et al. (2011). Seeds were subsequently sown in plastic seed trays containing
white peatmoss substrate with clay (Floragard, Oldenburg, Germany). Six to eight weeks after
seed planting, the transgenic F1 (respectively BC’1) seedlings were transferred to 9 cm
diameter plastic pots containing the same substrate. Three to four months after seed planting,
they were transferred to 2.5 L plastic pots containing a mixture of Floragard Pflanzsubstrat
(40 L), Floradur B fine (10 L) and Osmocote (150 g) (Floragard, Oldenburg, Germany).
Sulphur vaporizer was utilized for powdery mildew treatment when necessary. Insecticides
Confidor® (Bayer CropScience Deutschland GmbH, Germany) and Pegasus
® (Syngenta Agro
AG, Dielsdorf, Switzerland) were used to combat white flies, aphids and spider mites. Trees
of the apple cultivars ‘Topaz’, ‘Maloni Sally®
’ and ‘Royal Gala’ were stored in a dark cold
room (+ 4 °C) for a period of three months to synchronize flowering with early flowering
transgenic seedlings in order to perform simultaneous reciprocal crosses at the F1 and BC’1
generations.
All crossing experiments were performed in quarantine glasshouse cabins either at the Swiss
Federal Institute of Technology (Zürich, Switzerland) or at the Federal Research Station
Agroscope ACW (Wädenswil, Switzerland).
5.3.3. Genotypic foreground selection of F1 and BC’1 progenies
DNA was extracted from young leaf samples and quantified as described by Patocchi et al.
(2009a). Inheritance of the BpMADS4 transgene from the transgenic line T1190 to the F1 and
BC’1 progenies was followed using specific primers named BpMADS4_F and BpMADS4_R,
with the same PCR conditions as described previously (Flachowsky et al., 2007). Inheritance
of the fire blight resistance locus Fb_E from the ornamental apple cultivar ‘Evereste’ was
followed using the SSR marker ChFbE06 that co-segregates with Fb_E, with the same PCR
conditions as described previously (Parravicini et al., 2011). This SSR marker amplifies an
allele of 273 bp in coupling with fire blight resistance that is specific to the donor ‘Evereste’
and an allele of 256 bp that is common to ‘Evereste’ and few other so far tested domestic
121
apple cultivars (i.e. ‘Gala’, ‘Enterprise’, ‘Florina’, ‘Nova Easygro, ‘Golden Delicious’,
‘Topaz’, ‘Liberty’; G. Parravicini Rusca, personal communication; Le Roux et al.,
unpublished).
5.3.4. Phenotypic screening of F1 progeny for fire blight resistance
The F1 progeny derived from the cross T × E was divided into four genotype groups based on
the presence/absence of the BpMADS4 transgene and the Fb_E locus (Table 5.2). Green shoot
material of five to nine F1 seedlings per genotype group was grafted on ‘M.9’ rootstock, with
a number of replications comprised between five and thirteen for each seedling. Inoculation of
the grafted F1 seedlings with E. amylovora was performed as described previously (Khan et
al., 2007; Le Roux et al., 2010) with some modifications, including the inoculum
concentration (3 × 108 cfu/ml) and a single scoring date of the percent lesion length 21 days
after inoculation (PLL 21 DAI). The transgenic seedlings F1_74, F1_81, and F1_5 (derived
from the reciprocal F1 cross E × T), used as parents of the first pseudo-backcrosses, were
bud-grafted for fire blight resistance screening and vegetative propagation as well. The
transgenic seedling F1_74 displayed only two growing clones after grafting and was therefore
not included in the phenotypic screening. The cultivars ‘Enterprise’ and ‘Evereste’ were
included as resistant controls (4 and 6 replications, respectively) whereas the transgenic line
T1190 (female parent of the F1 cross, generated from the fire blight susceptible cultivar
‘Pinova’) was included as susceptible control (14 replications). Significant pair-wise
differences of PLL 21 DAI within the genotype groups and between the genotype groups and
the controls were investigated by the Mann-Whitney U-test (significance level 0.05) using the
software SPSS Statistics 19 (IBM, Germany).
122
Table 5.2. Distribution of the F1 and BC’1 apple seedlings of the fast introgression scheme in four genotype groups based on the presence/absence
of the BpMADS4 transgene and the resistant allele of SSR marker ChFbE06 co-segregating with the Fb_E locus.
Crossing
generation Progeny
Genotype groups
Total
BpMADS4 a Fb_E
b
Fb_E +
BpMADS4 none
F1 T1190 × ‘Evereste’ 17 12 18 9 56
‘Evereste’ x T1190 2 1 1 2 6
Total F1 19 13 19 11 62
BC’1
‘Topaz’ × F1_74 3 4 3 0 10
‘Topaz’ × F1_81 9 11 11 4 35
‘Maloni Sally®
’ × F1_74 9 7 2 8 26
F1_5 × ‘Milwa’ 3 5 9 c 2 19
Total BC’1 24 27 25 c 14 90
a Presence of the BpMADS4 transgene in each seedling was assessed by using the specific primers BpMADS4_F and BpMADS4_R (Flachowsky et al., 2007).
b The SSR marker ChFbE06 co-segregates with the fire blight resistance locus Fb_E from ‘Evereste’; the length of the resistant allele is 273 bp (Parravicini et al., 2011).
c All BC’1 seedlings carrying the BpMADS4 transgene and the resistant allele of SSR marker ChFbE06 were selected for subsequent background selection, except one seedling
from the cross F1_5 × ‘Milwa’.
123
5.3.5. Genotypic background selection of transgenic BC’1 offspring carrying the
fire blight resistance locus Fb_E
A background selection was performed on the BC’1 seedlings that inherited the BpMADS4
transgene and the fire blight resistance locus Fb_E to estimate the proportion of genome
inherited from the ornamental grandparent ‘Evereste’ in each seedling. Sixty-one SSR
markers amplifying 63 loci on the 17 linkage groups of the apple genome were selected from
a core set of genome-covering SSR markers, reported to be often polymorphic across
domestic apple varieties during the EU-funded project High-quality Disease-Resistant Apples
for a Sustainable agriculture (HiDRAS) (Patocchi et al., 2009b). Sixteen additional SSR
markers were selected from other studies on apple or pear (Pyrus spp.) in order to (i) fill gaps
existing in the core set of HiDRAS SSR markers, or to (ii) find alternatives to SSR markers of
the core set that amplified multiple loci or were insufficiently polymorphic, difficult to
amplify reliably in multiplex PCR or too complex to score: NH033b (Yamamoto et al.,
2004a), TsuENH045 and TsuENH062 (Nishitani et al., 2009) on LG 2, NB141b (Terakami et
al., 2009), AT000420-SSR and Hi07b02 (Silfverberg-Dilworth et al., 2006) on LG 4, HB09-
SSR (Broggini et al., 2009) and NZ23g4 (Guilford et al., 1997) on LG 6, CH-Sd1, CH-F7-
Fb1 (Khan et al., 2007), Hi12f04 and Hi05b09 (Silfverberg-Dilworth et al., 2006) on LG 7,
CH02g09 (Liebhard et al., 2003a) on LG 8, Hi04a05 (Silfverberg-Dilworth et al., 2006) on
LG 9, Hi02d04 (Silfverberg-Dilworth et al., 2006) on LG 10 and NZmsPal51 (Celton et al.,
2009b) on LG 14. These 79 SSR markers were organized in 25 multiplex PCRs; the forward
primers were either directly labeled with a fluorophore (6FAM, NED, PET or VIC) or were
extended at the 5’ end with the M13 sequence 5’-GACTGCGTACCAATTCAAA-3’ to
permit concurrent fluorescent labeling of PCR products by a third primer (M13) labeled with
the fluorophore 6FAM (Schuelke, 2000). PCR amplification of SSR markers from genomic
DNA of the BC’1 seedlings, parental F1 seedlings (F1_5, F1_74 and F1_81), parental apple
cultivars (‘Topaz’, ‘Maloni Sally®’ and ‘Milwa’) and grandparental genotypes T1190 and
‘Evereste’ was performed as described previously (Patocchi et al., 2009a).
The proportion of each BC’1 seedling’s genome contributed by the ornamental grandparent
‘Evereste’ was estimated by deriving allele content information from the order and genetic
distances of the SSR markers selected, following previous estimations (Patocchi et al.,
2009b). Three criteria were considered: (i) the informative SSR markers were those for which
it was possible to distinguish unambiguously between the alleles transmitted by ‘Evereste’
and those transmitted by the transgenic line T1190 through the parental F1 seedlings (F1_5,
F1_74 or F1_81); (ii) at least 3 SSR loci per linkage group were used to infer the inheritance
124
of chromosomal intervals in BC’1 seedlings from either ‘Evereste’ or T1190 (Volz et al.,
2009); (iii) a maximum genetic distance of 40 cM between two SSR loci was considered to
infer the inheritance of the chromosomal interval in-between; 40 cM corresponds to a
probability of 16 % that a double recombination event occurs within this interval between the
homologous chromosomes of a parental F1 seedling (one chromosome from T1190, the other
from ‘Evereste’) without being detected. Five SSR markers (CH05d04, CH04g04, CH03c02,
CH01d03z and ChFbE06) informative on LG 12 for all BC’1 seedlings were applied for
recombinant selection in the surrounding region of the Fb_E locus.
125
5.4. Results
5.4.1. Achievement of the first two breeding cycles of the introgression scheme
Three trees of the ornamental apple cultivar ‘Evereste’ were pollinated with pollen of the
BpMADS4-transgenic line T1190 in July 2008 (E × T cross, Fig. 5.1), whereas the pollination
of 15 T1190 trees with pollen of ‘Evereste’ (T × E cross) was performed between November
2008 and January 2009 (reciprocal cross). Six (resp. 134) seeds were obtained from the E × T
(resp. T × E) cross and were planted in April (resp. October) 2009 (Table 5.1). The first
flowers in the E × T and T × E F1 progenies were observed simultaneously in February 2010,
44 and 15 weeks after seed planting of the respective progenies. Five pseudo-backcrosses
(BC’1) were then achieved in March and April 2010 involving three apple cultivars and three
early flowering F1 seedlings (F1_5, F1_74 and F1_81) that proved to carry the fire blight
resistance locus Fb_E from ‘Evereste’ in addition to the transgene BpMADS4 from T1190:
‘Topaz’ × F1_74, ‘Topaz’ × F1_81, ‘Maloni Sally®
’ × F1_74, ‘Maloni Sally®
’ × F1_81 and
F1_5 × ‘Milwa’. Of these BC’1 crosses, one (‘Maloni Sally®
’ × F1_81) did not produce any
fruit; the four others produced in total 137 seeds that were stratified and sown in early
November 2010. The first flowers in two BC’1 progenies (BC’1_74 and 81) were observed in
February 2011, i.e. 14 weeks after seed planting; the first flowers in the two others BC’1
progenies were observed 22 (BC’1_7) and 28 (BC’1_33) weeks after seed planting (Table
5.1). Since March 2011, nine early flowering BC’1 seedlings carrying the Fb_E locus
(BC’1_7, 16, 19, 21, 33, 74, 75, 81 and 85) have been pollinated with pollen of the apple
cultivar ‘Royal Gala’. A reciprocal cross involving trees of ‘Royal Gala’ and the pollen of
five BC’1 seedlings carrying the locus Fb_E (BC’1_7, 19, 21, 75 and 85) has been performed
in June 2011.
126
Fig. 5.1. First two breeding cycles of the introgression of the fire blight resistance locus Fb_E from the ornamental apple cultivar ‘Evereste’ using
the BpMADS4-transgenic line T1190. Modified from Chevreau (2009) and Flachowsky et al. (2009).
.
127
← 1 For the sake of clarity, the timeline of the reciprocal F1 cross ‘Evereste’ × T1190 was not described on the
scheme; it started earlier, i.e. the pollination of the ornamental apple cultivar ‘Evereste’ with the BpMADS4-
transgenic line T1190 took place in July 2008; three fruits were harvested in January 2009 from which six seeds
were extracted and stratified for two months. The six ‘Evereste’ × T1190 F1 seeds were planted in April 2009. 2 Harvest of fruits on the T1190 trees was spaced out over four months as the pollination of T1190 trees was
made itself over three months. 3 The first flower of the ‘Evereste’ × T1190 F1_5 seedling, carrying the BpMADS4 transgene and the fire blight
resistance locus Fb_E, was observed in February 2010. Thus, the BC’1 crosses could be performed
simultaneously in both directions starting from March 2010, the seedling F1_5 having been used as female
parent due to its architecture compatible with fruit bearing.
5.4.2. Inheritance of the fire blight resistance locus Fb_E and the BpMADS4
transgene in F1 and BC’1 progenies
About three weeks after seed germination (i.e. at the 3 or 4 leaves stage), 62 F1 seedlings
derived from 140 seeds (resp. 90 BC’1 seedlings derived from 137 seeds) were screened for
the presence of the BpMADS4 transgene from the transgenic line T1190 and the fire blight
resistance locus Fb_E from ‘Evereste’ using molecular markers developed in the transgene’s
sequence or tightly linked to the resistance locus, respectively (Table 5.2). Thirty-eight F1
seedlings inherited the BpMADS4 transgene, among which 19 (50 %) inherited also the Fb_E
locus; forty-nine BC’1 seedlings inherited the BpMADS4 transgene, among which 25 (51 %)
inherited also the Fb_E locus.
5.4.3. Evaluation of the fire blight resistance level in the F1 progeny T1190 ×
‘Evereste’
The T × E F1 progeny that inherited the fire blight resistance locus Fb_E only (genotype
group “Fb_E”; n = 5 F1 individuals) displayed a lesion length of 7 % 21 days after
inoculation (PLL 21 DAI) (Fig. 5.2). The T × E F1 progeny that inherited the fire blight
resistance locus Fb_E and the BpMADS4 transgene (genotype group “Fb_E + BpMADS4”; n
= 9 F1 individuals) were significantly more susceptible in a pair-wise comparison using the
Mann-Whitney U-test, with a PLL of 21 % 21 DAI (p < 0.001). The T × E F1 offspring that
inherited either the BpMADS4 transgene only or none of the two loci (genotype groups
“BpMADS4” and “none”; n = 6 and 7 F1 individuals, respectively) were significantly more
susceptible to fire blight than the two previous groups (p < 0.001), but they did not show any
significant difference in a pair-wise comparison (PLL 21 DAI = 52 % and 53 %, respectively;
p = 0.808). The resistant controls ‘Enterprise’ and ‘Evereste’ (male parent) did not display
any necrotic lesion extending on the shoot from the syringe-inoculation point. The transgenic
line T1190 (female parent) was significantly more susceptible to fire blight than any of the
four genotype groups or the resistant controls (PLL 21 DAI = 96 %; p < 0.001) (Fig. 5.2).
128
Fig. 5.2. Mean fire blight lesion length in percentage of total shoot length (PLL) of the four
genotype groups in the F1 progeny T1190 × ‘Evereste’ in comparison to the controls.
The x-axis indicates the four genotype groups of the F1 progeny T1190 × ‘Evereste’ (“Fb_E”, “Fb_E +
BpMADS4”, “BpMADS4”, “none”), the parents T1190 and ‘Evereste’ and the additional control ‘Enterprise’; n
refers to the number of F1 individuals per genotype group (the seedling F1_5 from the cross ‘Evereste’ × T1190
being included in the “Fb_E + BpMADS4” group), r refers to the total number of replications per genotype or
genotype group.
The y-axis indicates the mean percent lesion length recorded 21 days after inoculation (PLL 21 DAI) with the E.
amylovora strain CFBP 1430 (3 × 108 cfu/ml).
The vertical bars indicate the standard errors.
5.4.4. Background selection of transgenic BC’1 offspring carrying the fire blight
resistance locus Fb_E
Among the 79 SSR loci that were selected to estimate the proportion of ‘Evereste’ genome in
the BC’1 seedlings carrying the BpMADS4 transgene and the Fb_E locus, 40 proved to be
informative (see Materials and methods for definition) in the four BC’1 progenies, 24 were
informative in one to three BC’1 progenies and 15 were not informative in any BC’1 progeny.
The linkage groups 8 and 10 could not be included in the background selection of the ‘Topaz’
× F1_81 BC’1 seedlings as only one informative SSR marker was scored on LG 8 (CH01c06)
and as the informative SSR markers Hi02d04 and CH02b03b were separated by a distance
beyond the threshold of 40 cM (67 cM) on LG 10. The linkage groups 4 and 11 could not be
included in the background selection of other BC’1 seedlings for similar reasons: Hi07b02
129
was the only informative SSR marker on LG 4 in the BC’1 cross F1_5 × ‘Milwa’, and the
only two informative SSR markers on LG 11 in the BC’1 crosses ‘Topaz’ × F1_74, ‘Maloni
Sally®
’ × F1_74 and F1_5 × ‘Milwa’ were further apart from each other (43 cM between
CH02d08 and CH04g07). Reasons of the lack of informativeness of the SSR markers selected
can be found in the Appendix D. On the whole, each BC’1 seedling could be screened with at
least 52 informative SSR loci, corresponding to a coverage ranging from 52 % to 72 % of the
SSR-based genome coverage defined during the HiDRAS project (Patocchi et al., 2009b)
(Fig. 5.3). The average genetic distance between informative SSR loci was 14.7 cM. The
proportion of BC’1 seedling’s genome inherited from the ornamental grandparent ‘Evereste’
was 24.3 % on average, ranging from 13.8 to 38.8 % (Fig. 5.3).
130
Fig. 5.3. Genome coverage with informative SSR markers and contribution of the
grandparental genome of ‘Evereste’ (both in %) in each BC’1 seedling carrying the
BpMADS4 transgene and the fire blight resistance locus Fb_E.
The x-axis indicates the BC’1 seedlings genotyped with the set of informative, genome-covering SSR markers;
they are ranked by their estimated proportion of grandparental genome from ‘Evereste’ within each progeny; a:
F1_5 × ‘Milwa’; b: ‘Maloni Sally®
’ × F1_74; c: ‘Topaz’ × F1_74; d: ‘Topaz’ × F1_81.
The y-axis indicates (i) the estimated proportion (%) of each BC’1 seedling’s genome covered with informative
SSR markers (light grey); (ii) the estimated proportion of grandparental genome from ‘Evereste’ in each BC’1
seedling relative to the genome covered with informative SSR markers (black).
BC’1 seedlings that already produced first flowers (as of July 2011) are indicated by an asterisk (*).
131
5.5. Discussion
5.5.1. Acceleration of the first two breeding cycles of the introgression of the fire
blight resistance locus Fb_E from ‘Evereste’ by inheritance of the BpMADS4 transgene
Applying the BpMADS4-transgenic line T1190 led to a dramatic acceleration of the
introgression of the fire blight resistance locus Fb_E from the genotype ‘Evereste’ over the
first two breeding generations (F1 and BC’1). All transgenic F1 seedlings derived from the
reciprocal crosses between ‘Evereste’ and T1190 produced first flowers within 15 to 50
weeks from seed planting. Moreover, ten transgenic BC’1 seedlings derived from the four
pseudo-backcrosses ‘Topaz’ × F1_74, ‘Topaz’ × F1_81, ‘Maloni Sally®’ × F1_74 and F1_5 ×
‘Milwa’ produced first flowers within 14 to 28 weeks from seed planting (as of July 2011)
(Table 5.1). These results are in agreement with Flachowsky et al. (2011), who initiated the
first crossbred-breeding program in apple using the BpMADS4-transgenic line T1190. Indeed,
after crossing T1190 with the fire blight resistant wild species Malus fusca, Flachowsky and
colleagues obtained transgenic F1 seedlings flowering within 15 to 40 weeks from seed
planting; at the next breeding cycle, transgenic BC’1 seedlings were obtained that flowered
within a year after seed planting. As our study took place in different facilities compared to
Flachowsky et al. (2011), it represents an independent confirmation that one crossbred
generation is feasible within approximately one year (12 to 15 months, see Fig. 5.1) in an
apple crossbred-breeding program based on the transgenic line T1190; this reproducibility
may be of valuable interest for apple breeders in general.
The growth and development of the BpMADS4-transgenic seedlings, relative to the non-
transgenic seedlings, was in agreement with the observations of Flachowsky et al. (2009 and
2011) who described the BpMADS4-transgenic F1 seedlings as being small and multi-
branched and continuously flowering under summer-like conditions. The continuous
flowering of the transgenic F1 and BC’1 seedlings after the appearance of the first flower
permitted us to pollinate one transgenic F1 seedling (F1_5) and nine transgenic BC’1
seedlings carriers of the Fb_E locus with the apple cultivars ‘Milwa’ and ‘Royal Gala’,
respectively. It proved to be possible to pollinate 30 flowers on the transgenic seedling F1_5,
which is significantly more than the 3 to 5 flowers per transgenic plant recommended by
Flachowsky et al. (2007) (Table 5.1). This was made possible by the fact that the transgenic
seedling F1_5 flowered relatively late (44 weeks after seed planting) compared to other
transgenic F1 seedlings (as few as 15 weeks after seed planting), thus having a taller
132
architecture: the main shoot of the transgenic seedling F1_5 reached 150 cm, compared to 50
cm for the most precociously flowering transgenic F1 seedlings. The fruit yield of the
transgenic seedling F1_5 was high (9 fruits obtained from 30 flowers pollinated, 30 %) and
satisfying in so far as it allowed to get 39 seeds from one plant. Upcoming observations on
fruit development in the transgenic BC’1 seedlings pollinated with ‘Royal Gala’ may give
further insight on the optimal number of flowers to be pollinated per transgenic plant, in
relation with their small architecture and the risk of fruit drop. In addition, grafting of green
shoot material of the early flowering transgenic F1 seedlings F1_5, F1_74 and F1_81 on
‘M.9’ rootstock proved to be successful: the grafted clones were all able to grow, except for
F1_74 for which only two grafted clones out of 10 grew, and continued to produce flowers
after grafting. This implies that the number of fruits born by an interesting early flowering
transgenic seedling could be increased by pollination of its clones derived from grafted green
shoots (without need of vernalization).
In parallel, we performed reciprocal crosses, using the pollen of transgenic F1 and BC’1
seedlings carrying the Fb_E locus and the BpMADS4 transgene to pollinate 3-years old,
‘M.9’-grafted flowering trees of the apple cultivars ‘Topaz’, ‘Maloni Sally®
’ and ‘Royal
Gala’, in order to ensure a minimum number of fruits/seeds and therefore seedlings at the next
breeding cycle. The fruit yield of the ‘Topaz’ and ‘Maloni Sally®’ trees pollinated with the
transgenic seedlings F1_74 and F1_81 turned out to be lower than the fruit yield on the
transgenic seedling F1_5 pollinated with ‘Milwa’, as only 20 fruits developed from 314
flowers pollinated (6.4 %) (Table 5.1). Relative low fruit yield of apple trees pollinated under
glasshouse conditions would seem to be quite common (A. Peil, personal communication).
The absence of fruit produced by the cross ‘Maloni Sally®
’ × F1_81 could be also due to a
gametophytic self-incompatibility mechanism between the pollen of the F1_81 seedling and
the cultivar ‘Maloni Sally®
’, which was not investigated in our study. On the whole, it appears
that the fruit bearing capacity of the most interesting, i.e. the most early flowering,
BpMADS4-transgenic seedlings carrying Fb_E (i.e. flowering within 15 to 28 weeks after
seed planting) is limited because of their weak tree architecture, whether they are own-rooted
or grafted; this observation justifies the use of non-transgenic flowering apple trees as female
parents in the introgression scheme, the number of which may depend on the space available
in glasshouse cabins and the number of traits that are selected (see below).
As the two loci of interest for the foreground selection, namely the fire blight resistance locus
Fb_E introgressed from the genotype ‘Evereste’ and the BpMADS4 transgene from the
transgenic line T1190, are located on two different chromosomes (chromosome 4 for the
133
BpMADS4 transgene (Flachowsky et al. 2011) and chromosome 12 for the Fb_E locus (Durel
et al. 2009)), an average of 25 % of the offspring at each breeding generation was expected to
carry both loci. In fact, among the 62 F1 seedlings and 90 BC’1 seedlings obtained through
the various crosses performed, 38 (61 %) and 49 (54 %) BpMADS4-transgenic seedlings were
obtained, among which 19 (30 %) and 25 (28 %) seedlings were carrying the two loci of
interest, respectively (genotype group “Fb_E + BpMADS4”; Table 5.2). The slightly higher
than expected proportion of F1 and BC’1 seedlings carrying Fb_E and BpMADS4 may be a
bias associated with the relatively small size of progenies, itself resulting from a low
germination rate due to seed rot. A minimum number of seedlings carrying Fb_E and
BpMADS4 at each breeding cycle is all the more necessary as the time to flowering of
BpMADS4-transgenic seedlings is variable. Disposing of 19 and 25 seedlings carrying Fb_E
and BpMADS4 at the F1 and BC’1 generations, respectively, allowed us to select for the most
early flowering genotypes and thus significantly accelerated the breeding cycles (gain of 43 -
15 = 28 weeks for the F1 cross T × E; Table 5.1). Presently, the reasons of such a variable
time to flowering in BpMADS4-transgenic seedlings are still unclear.
5.5.2. Effect of the Fb_E locus from the apple genotype ‘Evereste’ on fire blight
resistance in a T1190 × ‘Evereste’ F1 offspring
Several wild Malus species, crab apples or ornamental apple cultivars have been described to
be highly resistant to fire blight (Aldwinckle and van der Zwet, 1979; Aldwinckle et al., 1999;
Gardner et al., 1980; Norelli and Aldwinckle, 1986; Peil et al., 2009). Among them, the
genotype ‘Evereste’ has been well-studied for several years (Dugé de Bernonville, 2009;
Pontais et al., 2008; Venisse et al., 2002). The genetic basis of its resistance to E. amylovora
was recently deciphered through the identification of a strong-effect QTL located on LG 12,
named Fb_E, explaining from 50 to 70 % of the phenotypic variation in a ‘MM.106’ ×
‘Evereste’ F1 progeny (Durel et al., 2009). At the onset of our study, the QTL region was
being encompassed with molecular markers (Parravicini Rusca, 2010) and SSR markers co-
segregating with the resistance locus Fb_E were being developed (Parravicini et al., 2011).
For these reasons, the fire blight resistance locus Fb_E from ‘Evereste’ appeared to us as a
prime candidate for marker-assisted introgression in a domestic apple genetic background.
In the present study, the locus Fb_E showed a strong effect on the level of fire blight
resistance of F1 seedlings derived from the T × E cross. The F1 seedlings not carrying the
Fb_E locus showed necrotic lesions with a mean length of about 50 % of the inoculated
shoots, whether they were transgenic or not (Fig. 5.2). By contrast, the non-transgenic F1
134
seedlings carrying the Fb_E locus displayed much shorter necrotic lesions (PLL 21 DAI = 7
%). These observations are in agreement with the results of Parravicini Rusca (2010) who
noticed a significant difference of average necrotic lesion length in two F1 progenies
‘MM.106’ × ‘Evereste’ between the individuals not carrying the Fb_E locus and those which
did carry it (11.66 cm vs. 0.46 cm 14 DAI, respectively). Despite differences in the
inoculation technique (syringe vs. scissors), the tissue inoculated (actively growing shoot tips
vs. youngest well-expanded leaves) and the scoring date (21 DAI vs. 14 DAI), our experiment
confirms that the effect of the fire blight resistance locus Fb_E from the genotype ‘Evereste’
is stable in a different genetic background and therefore stresses the relevance of its
introgression for breeding more fire blight resistant apple cultivars. Interestingly, the
transgenic T × E F1 seedlings that inherited the Fb_E locus showed a significantly longer
necrotic lesion compared to the non-transgenic F1 seedlings carrying Fb_E (PLL 21 DAI = 21
% vs. 7 %, respectively). This higher susceptibility manifested itself in the development of
necrotic lesions of several centimeters three weeks after inoculation (data not shown). We
hypothesize that the active growth and the slender phenotype promoted by the BpMADS4
transgene in transgenic seedlings may favor the invasion of the shoots by the bacteria and
thereby render the Fb_E locus less efficacious. If true, this hypothesis would indicate (i) the
risk of insufficient reliability of fire blight resistance phenotypic screening of BpMADS4-
transgenic seedlings, which is all the more important as the effect of the QTL introgressed is
low; (ii) the relevance of using molecular marker(s) co-segregating with or tightly flanking
the fire blight resistance locus and displaying specific favorable allele(s). Nevertheless, our
hypothesis needs to be further studied, as the increase of fire blight susceptibility assigned to
the presence of the BpMADS4 transgene is not observed when comparing the genotype groups
“BpMADS4” and “none” in the T × E F1 progeny (Fig. 5.2).
5.5.3. Efficiency of background selection to assess the proportion of ‘Evereste’
genome in the BC’1 offspring
The use of wild germplasm as a source of disease resistance is a common strategy in crops
breeding, e.g. in cereals (Feuillet et al., 2008) and potato (Solanum tuberosum L.) (Park et al.,
2009; Tan et al., 2010). One issue of introgressing a resistance gene from a crossable wild
species is to minimize the contribution of the wild genome (often carrying agronomical
undesirable genes) in the advanced selections by (pseudo-)backcrossing. This issue includes
the well-known phenomenon of linkage drag, which is the linkage of the gene of interest to
undesirable genes on the same chromosome that are introduced along with it during (pseudo-
135
)backcrossing (Feuillet et al., 2008; Jacobsen and Schouten, 2007). The ornamental apple
cultivar ‘Evereste’ bears small-sized fruits, a trait probably inherited from its male
(presumable wild) parent (Durel et al., 2009). Therefore, we assumed that genes associated
with poor fruit quality and appearance may be carried along in the process of Fb_E
introgression. Consequently, we used SSR markers to assess the amount of ‘Evereste’ genome
in each BC’1 seedling carrying the BpMADS4 transgene and the Fb_E locus.
The proportion of SSR loci whose alleles could be traced back unambiguously to one of the
two grandparents, either ‘Evereste’ or T1190, was of 51 % (40 out of 79), which is close to
the proportion found by Volz et al. (2009) in their SSR-based whole genome selection of the
F1 progeny ‘Royal Gala’ × A689-24 (65 out of 108, 60 %). However, this percentage is lower
than we expected, considering that the majority of the SSR loci used here were reported to be
polymorphic (63 out of 79) or even highly polymorphic (i.e. more than 10 alleles amplified;
50 out of 79) across domestic apple varieties (Patocchi et al., 2009b). By contrast, a subset of
the SSR markers tested in the present study proved to be almost all informative in a second
BC’1 progeny ((T1190 × M. fusca) × ‘Antonovka’ accession CG01 from the Julius Kühn
Institute, Dresden, Germany) on which a background selection was also performed (55 out of
56 single locus SSR markers, 98 %; Le Roux et al., unpublished data). Ideally, additional SSR
markers should be tested to increase the coverage of the background selection, which
presently ranges within 52 % to 72 % of the SSR-based apple genome coverage (Patocchi et
al., 2009b) (Fig. 5.3). For this purpose, few SSR markers from the core set of the HiDRAS
project may be considered, although their level of polymorphism is expected to be lower.
More likely, one may capitalize on new sources of SSR markers spanning the apple genome
and possibly filling gaps in the core set identified by Patocchi et al. (2009b), namely the about
740 expressed sequence tag (EST)- or bacterial artificial chromosome (BAC)-based SSR
markers recently developed and mapped in Malus spp. (Celton et al., 2009b; Cova et al.,
2011; Han et al., 2011; van Dyk et al., 2010; Wang et al., 2011), the published complete
genomic sequence of the cultivar ‘Golden Delicious’ (Velasco et al., 2010) and the SSR
markers recently developed and mapped in the closely-related Pyrus spp. (Celton et al.,
2009a; Nishitani et al., 2009; Yamamoto et al., 2007). Their relevance to whole genome
selection purposes will be definitely established once (i) they are all more precisely mapped
relative to the HiDRAS core set of SSR markers; (ii) their level of polymorphism in M. ×
domestica Borkh. is accurately assessed, which requires to test them on at least ten unrelated
apple cultivars (Liebhard et al., 2003a; Moriya et al., 2011; Silfverberg-Dilworth et al., 2006);
(iii) their ease-to-use in multiplex PCRs is confirmed.
136
As far as the genome covered with informative SSR markers is concerned, a strong variability
in the contribution of the ornamental grandparent ‘Evereste’ was observed among the BC’1
seedlings (Fig. 5.3). On the whole, among the ten most precociously flowering BC’1
seedlings carrying the BpMADS4 transgene and the Fb_E locus, all except two (BC’1_21 and
BC’1_74) inherited a proportion of the ‘Evereste’ genome below the theoretical average of 25
% after one pseudo-backcross. Two of them, BC’1_16 and BC’1_19, are especially
interesting as parents for the next breeding cycle due to their low genomic contribution from
‘Evereste’ (13.8 and 14.4 %, respectively). Fewer breeding cycles will be necessary with
BC’1_16 and BC’1_19 rather than with BC’1_74 (38.8 % of ‘Evereste’ genome) for instance,
providing that a sufficient amount of seedlings is obtained at each breeding cycle of the
introgression program. Impediments to the background selection were the double-
recombination events likely to occur in parental F1 seedlings within large intervals separating
adjacent SSR loci, and remaining undetected in BC’1 seedlings. Such double-recombination
events probably led us to over- or under-estimate the proportion of ‘Evereste’ genome at the
chromosome scale. However, we assume that these undetected exchanges of chromosomal
segments balanced each other at the genome scale, thus making our estimations of ‘Evereste’
genomic contribution still sound. In the following breeding cycle, SSR markers located in the
midpoint of these intervals may be used to identify the actual double-recombination events.
Furthermore, among the five SSR markers that were informative on LG 12 for all BC’1
seedlings (CH05d04, CH04g04, CH03c02, CH01d03z and ChFbE06), only the two closest to
the Fb_E locus (CH01d03z and ChFbE06) displayed the alleles from ‘Evereste’ in the
BC’1_19 seedling; the chromosomal segment from ‘Evereste’ that was removed by
recombination represents about 85 % of the whole chromosome 12, based on the estimated
genetic distances between the aforementioned five markers. Hence, the recombinant selection
on LG 12 proved to be very efficient in selecting simultaneously for the presence of the Fb_E
locus and against the linked chromosomal segment from ‘Evereste’ that may carry
undesirable genes (linkage drag). More than BC’1_16, which retained all the chromosome 12
from ‘Evereste’, the seedling BC’1_19 appears to be the prime candidate to continue the
introgression of ‘Evereste’ fire blight resistance locus Fb_E while reducing the proportion of
‘Evereste’ genome carried along. The seedling BC’1_19 has already been pollinated and is
currently bearing three fruits in development; its pollen has been used for fertilization of a
‘Royal Gala’ tree (July 2011).
137
5.5.4. Perspectives of enhancing the introgression strategy based on the
transgenic early flowering line T1190
By combining the high-speed breeding method based on the early flowering transgenic line
T1190 (Flachowsky et al., 2009; Flachowsky et al., 2011) with a marker-assisted
introgression approach composed of foreground, recombinant and background selections, it
was possible to accelerate the first two introgression cycles of the Fb_E locus from the apple
genotype ‘Evereste’ up to the selection of promising BC’1 seedlings carrying little of the
‘Evereste’ genome (BC’1_19 and BC’1_16). More generally, with the ongoing advances in
genetic mapping, gene cloning and genomic studies in Malus spp. (Bus et al., 2010; Bus et al.,
2011; Fahrentrapp et al., 2011; Galli et al., 2010; Patocchi et al., 2009a; Velasco et al., 2010),
an increasing number of diseases resistance genes and QTLs are being identified in wild
Malus germplasm or crab apples that are suited for such a fast introgression strategy.
It should be noticed that our strategy does not aim at selecting outstanding apple genotypes
for use as candidates to become a new variety directly at the last breeding cycle (e.g. BC’4 or
BC’5), when it is expected that less than 5 % of the wild or crab apple’s genome is still
present, while the BpMADS4 transgene is still segregating in the progeny. For this purpose, a
genetic basis of as many as 15,000 to 30,000 non-transgenic seedlings would be necessary
(Durel et al., 2007; Kellerhals et al., 2009b), which is difficult to produce for two major
reasons: (i) given that the segregation of the BpMADS4 transgene in the last breeding
generation would halve the proportion of seedlings usable for breeding purposes, twice as
many seedlings in total should be produced to get 15,000 to 30,000 non-transgenic seedlings
from which a classical selection process could be initiated; (ii) obtaining such a high amount
of seedlings at the last breeding generation (i.e. 30,000 to 60,000) would imply pollinating
several dozen of non-transgenic flowering apple trees, which appears barely feasible under
strict quarantine requirements. Instead, our strategy is intended to make available to apple
breeders non-transgenic pre-breeding genotypes, i.e. potential parents of crosses, carrying loci
of interest and pre-eminently diseases resistance genes/QTLs (R loci) originating from wild
Malus germplasm or crab apples, but with the smallest as possible proportion of wild or crab
apple genome. Pre-breeding material carrying one R locus (e.g. R1 from a hypothetical wild
species Malus 1) may be obtained as early as the BC’4 generation, when the proportion of
wild genome is on average of 3.12 % and practically even lower in some seedlings, especially
if a background selection could be applied (Fig. 5.4, scheme a). When it comes to pyramide
two R loci in a pre-breeding genotype (e.g. R1 and R2), one approach may consist in
introgressing independently the two R loci over four breeding cycles to reduce first the
138
proportion of wild genome before pyramiding them in a BC’5 offspring (Fig. 5.4, scheme a).
Another approach may consist in introgressing the two R loci simultaneously, i.e. by
pyramiding them as early as the BC’1 generation and then performing pseudo-backcrosses
over four breeding cycles to introgress them together while reducing the proportion of wild
genome (Fig. 5.4, scheme b). In theory, both approaches of pyramiding seem equivalent in
terms of number of generations (until BC’5), whilst the total number of offspring needed
might be higher for the latter approach due to the necessity to remove a higher proportion of
wild genome from two wild Malus donors. However, considering that (i) apple production is
confronted to multiple pathogens and pests in orchards (the most important being probably
apple scab, powdery mildew, fire blight and woolly apple aphid (Eriosoma lanigerum)); (ii)
no single R loci seems durably resistant to a particular disease per se (as observed in the
Malus - V. inaequalis pathosystem (Bus, 2006; Bus et al., 2005b; Gessler et al., 2006); (iii)
the most efficacious combinations of R loci may vary from a geographical area to another,
depending on the virulences of local pathogen populations (e.g. in V. inaequalis (Beckerman
et al., 2009; Parisi et al., 1993; Parisi et al., 2004; Patocchi et al., 2009a), P. leucotricha
(Caffier and Laurens, 2005; Caffier and Parisi, 2007), E. amylovora (Norelli and Aldwinckle,
1986) and E. lanigerum (Bus et al., 2008)), the optimal use of the high-speed introgression
strategy may involve three steps. First, R loci can be independently introgressed in transgenic
BC’4 seedlings with a domestic apple background (Fig. 5.4, scheme a). Second, combinations
of two to four R loci to one or more diseases can be achieved in (non-) transgenic BC’5-6
seedlings after one or two successive crosses (e.g. R1/R2 in BC’5 or R1/R2/R3/R4 in BC’6;
Fig. 5.4, scheme c). Third, non-transgenic pre-breeding genotypes carrying various
combinations of more than four R loci to the main apple diseases can be produced starting
from the BC’7 generation. Such non-transgenic pre-breeding genotypes would then appear
ideal donors of R loci to be used in large breeding crosses with high fruit quality apple
varieties or elite selections (Kellerhals et al., 2009b; Peil et al., 2009).
139
Fig. 5.4. Hypothetical schemes of introgression and pyramiding of disease resistance (R) loci
from wild apple species or crab apples into pre-breeding genotypes using the high-speed
transgenic introgression strategy.
140
← The hypothetical loci R1, R2, R3 and R4 are introgressed from four hypothetical wild apple species/crab
apples donors named Malus 1, 2, 3 and 4, respectively. These four loci are assumed to segregate independently.
Introgression scheme a: introgression of R1 into a domestic apple background, followed by pyramiding with R2
at the BC’5 generation (R2 having been independently but contemporaneously introgressed).
Introgression scheme b: simultaneous introgression of R1 and R2 in a domestic apple background after
pyramiding both loci at the BC’1 generation.
Introgression scheme c: pyramiding of R1, R2, R3 and R4 at the BC’6 generation after independent but
contemporaneous introgressions of the four R loci into a domestic apple background (BC’4), followed by their
pyramiding in pairs (BC’5 via scheme a).
Percentages in colored boxes indicate the decreasing proportion of the wild apple(s) or crab apple(s) genome(s)
in the offspring at each breeding generation.
Percentages in bracket indicate the proportion of non-transgenic or transgenic seedlings of interest at each
breeding generation. The BpMADS4-transgenic genotypes are underlined.
1: the commercial apple cultivars to be used as parents of pseudo-backcrosses in the introgression schemes may
be selected based on (i) the absence of known relatedness among them and (ii) the difference of S-alleles
genotypes, if known (e.g. S2S
5 for ‘Gala’, here represented by its sport ‘Royal Gala’, S
3S
23 for ‘Granny Smith’,
S9S
24 for ‘Braeburn’ (Dreesen et al., 2010)), in order to decrease the risk of self-incompatibility.
2: the genotypes “BC’4 R2” (scheme a) and “BC’5 R3/R4” (scheme c) are assumed to be derived from separate
introgression schemes similar to that of “BC’4 R1” and “BC’5 R1/R2”, respectively, but involving different
pseudo-backcrossing parents to decrease the risk of self-incompatibility.
3: transgenic BC’5 and BC’6 seedlings of interest carrying only one copy of the BpMADS4 transgene.
Despite a moderate genome coverage, the SSR-based background selection of transgenic
BC’1 seedlings performed in this study proved to be efficient to identify the individuals with
a reduced proportion of ‘Evereste’ genome. In comparison, a phenotypic (positive) selection
of transgenic seedlings, based on fruit quality evaluation, would probably not be relevant
considering that (i) the effect of the BpMADS4 transgene on plant physiology, and thereby on
fruit development, is largely unknown (Hoenicka et al., 2007); (ii) the transgenic seedlings
can apparently not bear more than 10 relatively small fruits due to their weak architecture;
(iii) the transgenic seedlings have to be grown under artificial glasshouse conditions that may
not adequately mimic the conditions of fruit development in the outside environment. In the
future, the development and validation of molecular markers associated with fruit quality
parameters such as firmness, texture, sugar content, acidity and levels of fruit volatiles (Costa
et al., 2008; Dunemann et al., 2009; Kenis et al., 2008; Laurens et al., 2011; Peace and
Norelli, 2009; Rowan et al., 2009; Zhu and Barritt, 2008) may allow selecting for specific
genomic regions brought by parents of pseudo-backcrosses, even though this will require the
production of larger progenies.
141
The legal status of the pre-breeding genotypes produced by the high-speed transgenic
introgression strategy is not yet defined, although it would be of high relevance to apple
breeders. It is clear that the T-DNA carrying the transgene BpMADS4 segregates and can be
inherited or not by seedlings at the final breeding cycle. Absence of the transgenes BpMADS4
and nptII (selectable marker gene) as well as of the whole T-DNA can be demonstrated by
PCR reactions, sequencing and Southern blot analysis (Flachowsky et al., 2011). If plants
devoid of any such foreign DNA can be considered as non-transgenic, it does not necessarily
imply that they can be defined by law as not genetically modified.
5.6. Acknowledgements
This research was funded by the ZUEFOS Project (Züchtung feuerbrandtoleranter Obstsorten)
of the Federal Office for Agriculture (FOAG, Switzerland). The authors gratefully
acknowledge Dr. Giovanni Broggini, Dr. Gabriella Parravicini Rusca (Swiss Federal Institute
of Technology Zürich, Switzerland), and Dr. Charles-Eric Durel (Institut National de la
Recherche Agronomique, INRA Angers, France) for kindly providing the primer sequences
of SSR marker ChFbE06 before publication. The authors are also grateful to Rolf Blapp,
Juergen Krauss, Reto Leumann, Isabelle Baumgartner (Agroscope Changins-Wädenswil
ACW, Wädenswil, Switzerland), Maja Frei, Sabine Klarer, Andre Imboden and Urs Moser
(Swiss Federal Institute of Technology Zürich, Switzerland) for the preparation of grafted
material and help in plant maintenance.
142
143
Chapter 6
General conclusion
144
6.1. Fire blight resistance in classical apple breeding
The cultivated apple (Malus × domestica Borkh.) is one of the most important fruit crop in the
world, with a production reaching 68 million metric tons in 2009 (World Apple Report 2009).
Apple cultivation dates back at least 3,000 years ago (Hancock et al., 2008); though apple
breeding, involving well-planned crossing schemes and stepwise selection of seedlings before
release of a new cultivar, started in the early 20th
century (Crandall, 1926; Magness, 1937; van
der Zwet and Keil, 1979). Nowadays, it is considered that apple breeding programs should
contribute to (i) produce fruits meeting consumer’s expectation of pleasurable consumption
and associated with increased health benefits; (ii) reduce tree growing and fruit handling
costs; and (iii) decrease the use of pesticides potentially harmful to human health and the
environment (Gardiner et al., 2007; Kumar et al., 2010; Laurens et al., 2011). For this, new
apple varieties with a fruit quality similar to the most popular varieties (e.g. ‘Golden
Delicious’, ‘Red Delicious’, ‘Granny Smith’, ‘McIntosh’, ‘Jonathan’, ‘Braeburn’, ‘Fuji’ and
‘Gala’), but able to differentiate themselves based on appearance, texture and nutritional
values and being more resistant to biotic and a-biotic stresses are needed. If the fungal
diseases scab (Venturia inaequalis) and powdery mildew (Podosphaera leucotricha) are the
two main biotic factors affecting production in apple orchards at a global scale, fire blight,
incited by Erwinia amylovora, causes regularly epidemics resulting in severe crop and
financial losses (Bonn and van der Zwet, 2000). Consequently, several apple breeding
programs in the USA, in Europe and in New Zealand have decided to take fire blight into
consideration for selecting new apple cultivars (Peil et al., 2009).
In most of these programs, breeding for fire blight resistance consists in selecting the least
susceptible advanced selections after artificial bacterial inoculation of grafted replications,
generally under quarantine glasshouse conditions. As already mentioned in Chapter 1, such a
phenotypic selection is feasible but has drawbacks. First, it is delayed of at least four years
after crossing, i.e. until collection of sufficient graft-woods from a sufficiently reduced
number of breeding selections is possible (e.g. from 6 % of the total breeding progenies; Fig.
1.1). Second, it is expensive, due to the necessity to make per genotype at least 10 replications
to assess with enough precision its level of resistance. Third, it is labour intensive, not only
because of the total number of replications to handle but also because of the necessity to
perform phenotypic selection under strict quarantine conditions in many countries. Fourth,
phenotypic glasshouse screenings often need to be repeated multiple times to establish the
145
reproducibility of the results, due to the strong influence of parameters such as temperature
and relative humidity on disease severity.
146
6.2. Mapping QTLs for fire blight resistance in bi-parental populations
The application of molecular markers has the potential to increase the efficiency of selection
for fire blight resistance while decreasing the costs and making the selection process less
labour intensive (Kellerhals et al., 2009b; Khan et al., 2007). As a prerequisite, a good
knowledge of the genetic basis of resistance to E. amylovora in Malus spp is necessary;
molecular markers associated with various fire blight resistance sources need to be identified
that are reliable (i.e. as tightly linked as possible to a QTL or gene controlling the resistance),
robust (i.e. usable across as many as possible breeding populations) and easy to use (involving
reproducibility across laboratories and ability to be multiplexed). In this context, Chapters 2
and 3 are a contribution to a better understanding of the quantitative resistance to fire blight
found in apple cultivars (Korban et al., 1988; Lespinasse and Aldwinckle, 2000). In the past
years, standard QTL mapping in bi-parental populations (i.e. F1 progenies) proved to be
successful in identifying 8 additive QTLs associated with fire blight resistance in Malus spp
(Khan et al., 2011; Peil et al., 2009). Consequently, this strategy was chosen to study the
genetic basis of this trait in two F1 progenies; one progeny of 118 individuals was derived
from a cross between two resistant apple cultivars, namely ‘Florina’ (FLO) and ‘Nova
Easygro’ (NEG) (Chapter 2); while another progeny of 92 individuals was derived from a
cross between a susceptible cultivar and a resistant cultivar, respectively ‘Idared’ (ID) and
‘Rewena’ (RE) (Chapter 3).
In both Chapters 2 and 3, the approach of QTL mapping in bi-parental populations required
(i) genotyping the F1 individuals with molecular markers distributed over the apple genome,
i.e. simple sequence repeats (SSR) in Chapter 2 and 3, coupled with amplified fragment
length polymorphism (AFLP) markers in Chapter 2 and diversity arrays technology (DArT)
markers in Chapter 3; (ii) phenotyping the same F1 individuals after artificial inoculation with
one strain of E. amylovora (CFBP 1430 in Chapter 2 and Ea 222 in Chapter 3). In Chapter 2,
two significant fire blight resistance QTLs were detected in the cultivar ‘Florina’ using log10-
transformed data of lesion length in percentage of the shoot length (PLL), collected one week
(PLL1) and two weeks (PLL2) after inoculation. One QTL was mapped on linkage group
(LG) 10 by interval mapping (IM) and multiple QTL mapping (MQM); by MQM, it
explained 17.9 % and 15.3 % of the phenotypic variation in the FLO × NEG F1 progeny with
log10(PLL1) and log10(PLL2) data, respectively. The second QTL was identified on LG 5; it
explained 10.1 % of the phenotypic variation by MQM with log10(PLL2) data. Other putative
147
QTLs were detected by IM on LGs 5 and 9 of the cultivar ‘Nova Easygro’. In Chapter 3, no
significant fire blight resistance QTL was detected in the cultivars ‘Idared’ or ‘Rewena’ by
IM or MQM with the PLL data collected 4 weeks after inoculation. However, one putative
QTL was detected close to the DArT marker aPa-526070 at the distal end of LG 7 of ‘Idared’
(LOD score of 3.92 above the chromosomal LOD threshold of 3.5). Collectively, studies
described in Chapters 2 and 3 identified 5 significant or putative QTLs involved in
quantitative resistance to fire blight. This makes a total of seven significant QTLs with an
additive effect on fire blight resistance that have been identified in the cultivated apple as of
August 2011 (Fig. 6.1). Thus, results of Chapters 2 and 3 give support to the hypothesis of an
oligo- or polygenic determinism of fire blight resistance in apple cultivars (Calenge et al.,
2005; Khan et al., 2006). All additive, significant QTLs identified so far on the genome of
Malus spp. are depicted on Fig. 6.1.
148
Fig. 6.1. Global position of significant quantitative trait loci (QTL) with additive effect on fire
blight resistance on the apple (Malus spp.) genome.
149
← This SSR-based backbone map of the apple genome is an update (as of August 2011) of the backbone map
made by Peil et al. (2009), itself based on the core-set of SSR markers proposed by Silfverberg-Dilworth et al.
(2006) and Patocchi et al. (2009b). Only linkage groups that have been shown to carry significant, additive QTLs
for fire blight resistance by interval mapping (IM) or multiple QTL mapping (MQM) are presented. Please note
that vertical bars do not represent the precise confidence intervals of the QTLs but are approximations of the
genomic regions of the QTLs. The name of the QTL along the vertical bar is indicated first and is followed by
the name of the Erwinia amylovora strain(s) used for phenotyping (CFBP 1430, Ea 222, Ea 273, Ea 610, Ea
3049 and NZ11176). Black vertical bars represent highly efficacious QTLs explaining 40 % or more of a large
phenotypic variation in a F1 progeny and for which phenotypic markers have been mapped at the peak of the
QTLs: Fb_E (‘Evereste’ (Durel et al., 2009; Parravicini et al., 2011)), Fb_Mf (Malus × floribunda clone 821
(Durel et al., 2009; Parravicini Rusca, 2010)) and Fb_R5 (Malus × robusta ‘Robusta 5’ (Fahrentrapp et al., 2011;
Peil et al., 2007; Peil et al., 2008)). Grey vertical bar represents a major effect QTL: F7 (‘Fiesta’ (Calenge et al.,
2005; Khan et al., 2006)); it explained 34.3 to 46.6 % of the phenotypic variation in three F1 progenies but
showed less efficacy in decreasing the extent of fire blight lesions than the three aforementioned QTLs
(Baumgartner et al., 2010; Khan et al., 2007). White vertical bar represent minor- to medium-effect QTLs: F3
(‘Fiesta’), P3 (‘Prima’), D12 and D13 (‘Discovery’) (Calenge et al., 2005); E5 (‘Evereste’ (Durel et al., 2009));
FLO5 and FLO10 (‘Florina’ (Le Roux et al., 2010)); R5-5 (‘Robusta 5’ (Peil et al., 2011)); MsPI613981-8 and
MsPI613981-10 (Malus sieversii accession PI613981 (Lalli et al., 2010)).
On a more technical viewpoint, Chapters 2 and 3 confirm the usefulness of SSR markers (i) to
build backbone linkage maps of quality (e.g. for ‘Florina’ and Nova Easygro’ in Chapter 2);
(ii) to enrich genomic regions of interest for comparative mapping across populations of
Malus spp. genotyped in independent studies (e.g. Chapter 2 and 3 vs previous QTL mapping
studies for fire blight resistance). After Schouten et al. (2011), Chapter 3 further demonstrates
that DArT markers are a valuable alternative to AFLP markers used in Chapter 2 to saturate a
SSR-based backbone linkage map, based on four observations: (i) DArT markers are more
reproducible and robust than AFLP markers and therefore allows alignment of linkage maps
for comparative mapping, like SSR markers; (ii) several hundreds of polymorphic DArT
markers can be produced in parallel by a single hybridization assay on a DArT array, whereas
a single AFLP primer combination yielded a maximum of 34 polymorphic markers in Malus
spp. (Xu and Korban, 2000); (iii) DArT markers are probably more cost-efficient than AFLP
markers: a period of time of three months appears sufficient to extract DNA from a
population of 100–200 individuals, check DNA quantity and integrity, outsource the micro-
array hybridization to Diversity Arrays Technology Pty Ltd (DArT P/L, Yarralumla,
Australia) and receive the hybridization scores corresponding to several hundreds of
polymorphic DArT markers; in comparison, the AFLP technique requires at best 4 months to
produce and score a similar amount of polymorphic markers, assuming a maximum yield of
150
AFLP primer combinations (D. Socquet-Juglard, personal communication); more realistically,
the AFLP technique would require even more time considering that an AFLP primer
combination yields on average only 4 to 10 polymorphic markers in Malus spp. (Kenis and
Keulemans, 2005; Le Roux et al., 2010; Liebhard et al., 2003a); and (iv) DArT markers
analysis through the DArT software (i.e. conversion of hybridization signals for each
fragment on the array into a binary score, “1” (present) or “0” (absence)) proved to be very
reliable, thus allowing straightforward integration of DArT and SSR markers on a linkage
map. Further improvements of the DArT technology would be however desirable, such as the
reduction of DArT markers redundancy using alternative complexity reduction methods
(Schouten et al., 2011).
151
6.3. Alternatives to bi-parental mapping populations
6.3.1. Limitations of bi-parental mapping populations
Standard QTL analysis in bi-parental populations like the F1 progenies from FLO × NEG and
ID × RE have limitations (Danan, 2009; Hamblin et al., 2011; Mackay and Powell, 2007;
Stich et al., 2008b). First, only a maximum of four alleles per locus can be studied in a F1
progeny, which represents a small proportion of the alleles available to geneticists and
breeders in the breeding germplasm. Investigating the individual effect of each QTL allele is
rendered even more challenging in Malus spp. by the high level of heterozygosity: the effect
estimated for one QTL allele is only relative to the effects of the other alleles at the same QTL
(Liebhard et al., 2003c); the “worst case scenario” occurs when the 2 alleles of one parent, or
even the 4 alleles of both parents, have a similar effect (“functional homozygosity” as
hypothesized in Chapter 2 and 3). This issue can only be addressed by generating a mapping
population in which the favorable QTLs alleles would segregate; in the case of the FLO ×
NEG (resp. ID × RE) F1 progeny, this would imply to produce a F2 progeny from the cross
between a FLO × NEG (resp. ID × RE) F1 individual cumulating all favorable QTLs alleles at
a heterozygous state and a fire blight susceptible apple cultivar. Second, the resolution of bi-
parental mapping populations is low due to the fact that only one generation of meiosis (and
therefore one possibility of recombination events between genomic loci) occurs between the
parents and F1 individuals. Third, bi-parental populations have a low power to detect small-
effect QTLs, which may be the reason why several QTLs in ‘Nova Easygro’ and ‘Idared’
were considered as putative only and not significant; they also tend to overestimate the effect
of significant QTLs (Beavis, 1998), especially when the progeny size is reduced like in the
case of FLO × NEG (118 F1 individuals) and ID × RE (92 F1 individuals). Fourth, mapping
diseases resistance QTLs in a bi-parental population does not allow to determine if the QTLs
are stable across various genetic backgrounds; for this, other F1 progenies must be
specifically produced, if possible by crossing the resistant cultivar carrying the QTL(s)
identified with a susceptible cultivar, like ‘Gala’, ‘Idared’ or ‘MM.106’ for fire blight. These
additional F1 progenies should be maintained in orchards, genotyped with molecular markers
flanking the QTL(s) and evaluated for disease resistance, which requires space, labor and
additional funds.
152
6.3.2. QTL mapping in multiple related F1 progenies
In the past years, several strategies have been developed that would ideally complement the
use of standard bi-parental populations to identify associations between molecular markers
and fire blight resistance in Malus spp., and especially in the cultivated apple. One strategy
would consist in performing QTL mapping on a composite linkage map derived from multiple
related F1 progenies, for instance 4 to 8 F1 progenies linked with common parental cultivars
(N’Diaye et al., 2008). Quantitative trait loci could be mapped with more precision thanks to
the numerous recombination events in the multiple F1 progenies. In addition, the stability of a
QTL detected in a parental cultivar could be assessed over several genetic backgrounds, thus
allowing its identification and validation in a single mapping experiment. Also, multiple
connected F1 progenies would increase the power of QTL detection (Danan, 2009; N’Diaye
et al., 2008). This strategy is likely to become more and more popular in the coming years
with the advent of moderate to high throughput DArT and single nucleotide polymorphism
(SNP) genotyping platforms that can provide dozens to hundreds of reproducible markers
“bridging” individual linkage maps in a short time (Micheletti et al., 2011; Schouten et al.,
2011).
6.3.3. Pedigree-based analysis
A second strategy is the pedigree based analysis (PBA) outlined in Chapter 1 (van de Weg et
al., 2004). It consists in detecting QTLs in the breeding material itself, which is composed of
many progenies covering multiple breeding generations (up to 7 in some apple breeding
programs; François Laurens, personal communication) and linked by common ancestors
(founders). In such a pedigreed population, multi-allelic markers such as SSR markers are
applied to connect founders, cultivars, progenies and breeding selections at the molecular
marker level by monitoring specific chromosomal segments along the breeding lines (van de
Weg et al., 2004). Thereby, a QTL allele of an advanced selection can be traced back to the
founder cultivar (identity by descent (IBD) concept); conversely, the transmission of this
founder allele to other advanced selections through alternative breeding lines can be
monitored as well. Applying PBA to fire blight resistance in apple breeding material will be
addressed in the recently funded ZUEFOS II project (Federal Office for Agriculture, FOAG,
Switzerland); it is especially promising for three reasons. First, three of the most important
founders of apple breeding programs worldwide may carry already identified QTLs
associated with fire blight resistance: based on genotyping results with flanking molecular
markers, ‘Cox’s Orange Pippin’ would carry the major QTL identified originally on the LG 7
153
of ‘Fiesta’ (Calenge et al., 2005; Khan et al., 2006; Khan et al., 2007); ‘Jonathan’ might carry
the significant QTL identified on LG 10 of ‘Florina’ (Chapter 2); and ‘Starking’, a sport of
‘Red Delicious’, would carry the significant QTL identified on LG 5 of ‘Florina’ (Chapter 2);
these QTLs could be validated and detected in other cultivars or advanced selections
connected with the founders ‘Cox’s Orange Pippin’, ‘Jonathan’ and ‘Red Delicious’ using the
IBD concept. Besides, additional QTLs are probably present in ‘Cox’s Orange Pippin’,
‘Jonathan’ and ‘Red Delicious’, but also in ‘McIntosh’ and ‘James Grieve’ that were shown
to carry some level of quantitative resistance to fire blight (Aldwinckle et al., 1999; Khan et
al., 2007). Second, the recent development and mapping of hundreds of SSR and SNP
markers on the apple genome (Celton et al., 2009b; Chagné et al., 2008; Han et al., 2011; van
Dyk et al., 2010; Velasco, 2010; Wang et al., 2011) provide now geneticists and breeders with
a wealth of genome-covering, multi-allelic SSR markers and bi-allelic SNP markers that
complement the already existing core-set of SSR markers for PBA (Evans et al., 2010;
Patocchi et al., 2009b). Third, PBA would give the possibility to test the stability of QTLs
against a range of genetic backgrounds (like with multiple related F1 progenies, but on a
bigger scale); these may include progenies derived from the most fire blight susceptible apple
cultivars carrying few, if any, favorable QTLs alleles and that are part of ongoing breeding
programs (which is not always the case of multiple related F1 progenies) (Bink et al., 2008;
van de Weg et al., 2004).
6.3.4. Association mapping
A third strategy that would complement the mapping of QTLs in bi-parental populations is
association mapping (AM), also termed linkage disequilibrium (LD) mapping or genome
wide association (GWA) mapping (Mackay and Powell, 2007; Oraguzie and Wilcox, 2007).
The basic principle is the same as in all aforementioned mapping strategies, i.e. the search for
statistically significant associations between alleles at molecular markers and quantitative
traits of interest (e.g. fire blight resistance in apple). The major difference is that AM is not a
family-based approach; instead, statistical analysis are performed on a large population of
loosely related individuals, which can be a natural population or a collection of accessions
(cataloged member of a gene-bank), landraces (old cultivated varieties adapted to local
growing conditions) or breeding material (founders, cultivars or selections of ongoing
breeding programs). Because the AM approach takes into account all the historic
recombination events by which a diverse population is removed from its progenitors, the
resolution of QTL mapping can be much higher than in family-based populations. Besides,
154
AM offers the possibility to study a larger number of alleles at the species or at breeding-
programs level, depending on the population selected (Hamblin et al., 2011); in the latter case,
the QTLs identified can be more immediately used in cultivar improvement through marker-
assisted breeding (MAB). However, prior knowledge of the population structure (distribution
of genotypes among individuals within a population), family kinship (genetic relationships
between individuals), linkage disequilibrium extent (non-random association of alleles at
distinct loci) and allele frequencies is necessary in order to perform a genome-wide
association mapping; correlatively, several thousands of polymorphic molecular makers
covering the entire genome, and the adequate high-throughput genotyping technologies, are
also required (Mackay and Powell, 2007; Oraguzie et al., 2007). In Malus spp., information
on populations structure, linkage disequilibrium and allele frequencies is still scarce. Malus
spp being out-bred, it has been assumed that the LD level is low, and that the density of
molecular markers therefore needed to cover the whole genome is high (Plomion and Durel,
1996; Rikkerink et al., 2007); preliminary results of LD estimation in a collection of 200
apple accessions support this hypothesis (Micheletti et al., 2010; Velasco, 2010). Accurate
genome-wide association mapping has clearly the potential to fully exploit the variability of
fire blight resistance in the cultivated apple germplasm, following the model of annual crops
such as rice (Oryza sativa L.) (Huang et al., 2010). The current advances in apple structural
genomics and SNP markers development, e.g. in the framework of the FruitBreedomics project
(Laurens et al., 2011), might make it feasible in the coming years.
In the meantime, it may be reasonably envisaged to use AM in the cultivated apple as a
validation method. Molecular markers linked to fire blight resistance QTLs previously
detected in apple cultivars by classical mapping strategies could be screened in an
unstructured population of founders, cultivars and breeding selections for QTL validation. It
may allow for instance to confirm the effect of genomic regions identified in ‘Florina’, ‘Nova
Easygro’, ‘Idared’ and ‘Rewena’ (Chapters 2 and 3), and thereby may establish their
usefulness in MAB. Such a validation approach was already successfully applied to confirm
QTLs controlling agronomically relevant traits in the tetraploid oilseed rape (Brassica napus
L. (Jestin et al., 2011)) and in the diploid sugar beet (Beta vulgaris L. (Stich et al., 2008a)),
using 70 and 26 SSR markers, respectively.
Also in the short term, AM for fire blight resistance in the cultivated apple may be exploited
using another approach called joint linkage association mapping (Yu et al., 2008). With this
method, associations between a quantitative trait and molecular markers distributed over the
whole genome are investigated in multiple related segregating breeding populations. Joint
155
linkage analysis combines the high resolution of association mapping (by densely genotyping
the diverse founders and parents of the populations) with the high power to detect QTLs of
classical linkage mapping (by genotyping the founders, parents and progenies with a smaller
number of molecular markers allowing to monitor the inheritance of chromosome segments).
Joint linkage association mapping was conducted in sugar beet under the name of association
mapping in multiple segregating populations (AMMSP (Reif et al., 2010; Stich et al., 2008b)).
In particular, Stich and colleagues genotyped 768 Fn individuals derived from multiple
related crosses of sugar beet breeding programs with 49 SSR markers to detect associations
with nine quantitative agronomical traits. The existence of similar pedigreed populations in
apple breeding programs (Bannier, 2011; Evans et al., 2010; Kouassi et al., 2009; Noiton and
Alspach, 1996), together with the amount of Malus SSR and SNP markers now available, lead
to the conclusion that the AMMSP approach is feasible with apple breeding material. It
remains to be seen to what extent the AMMSP approach could complement the PBA
approach.
6.3.5. Phenotyping issues
With the development of high-throughput genotyping/sequencing technologies (Akhunov et
al., 2009; Gupta et al., 2008; Mammadov et al., 2011; Rothberg and Leamon, 2008; Shendure
and Ji, 2008), the major bottleneck for mapping QTLs in major crops is not the genotyping
anymore, but rather the phenotyping. In apple, this may be true for a trait like fire blight
resistance whose precise evaluation requires many replications per genotype and (often)
quarantine facilities. Alternative QTL mapping strategies described above (composite linkage
mapping, PBA, AM) would require genotyping several hundreds of individuals, implying
about a ten-fold higher number of plants to be inoculated with E. amylovora for disease
resistance screening. Most probably, this would be possible only in the frame of national or
transnational collaborations between research and breeding institutes. On a practical
viewpoint, homogeneous procedures for inoculating plants with E. amylovora and scoring fire
blight lesions would be required, as already advocated by Peil et al. (2009); careful
experimental designs would be also needed to allow comparing the phenotyping data
collected in different places, e.g. the use of common resistant and susceptible controls
(Sobiczewski et al., 2011) and sufficient “bridge” genotypes across places for standardization
of the results (Dowkiw and Bastien, 2007; Durel et al., 2009).
156
6.4. Accelerating the introgression of fire blight resistance from wild Malus
species or hybrids into the cultivated apple
6.4.1. Considerations on the genetic basis of fire blight resistance in wild Malus
species or hybrids
Several accessions of small-fruited wild Malus species, hybrid species or ornamentals (all
being designated as crab apples for their small-sized and bitter fruits) have long been shown
to carry higher levels of fire blight resistance than apple cultivars (Aldwinckle and van der
Zwet, 1979; Aldwinckle et al., 1999; Gardner, 1976; Shaw, 1934; van der Zwet and Keil,
1979). Recently, the evidence has been adduced that the genetic basis of the high levels of
resistance in crab apples also differs from apple cultivars: by classical interval mapping,
QTLs explaining more than 40 % of large, continuous phenotypic variations (0–100 % of
lesion length in percentage of the shoot length, PLL) were identified in F1 progenies derived
from crosses between susceptible apple genotypes and resistant crab apples such as Malus ×
robusta ‘Robusta 5’ (Peil et al., 2007; Peil et al., 2008), the ornamental genotype ‘Evereste’
(Durel et al., 2009) and the clone 821 of Malus × floribunda (Durel et al., 2009). Despite the
absence of visible, qualitative resistance phenotypes as for scab resistance genes (Gessler et
al., 2006), phenotypic markers of fire blight resistance were established following
mendelization of the fire blight resistance trait in the F1 progenies ‘MM.106’ × ‘Evereste’ and
‘Idared’ × ‘Robusta 5’; interestingly, these phenotypic markers mapped at the peak of both
major QTLs of ‘Evereste’ and ‘Robusta 5’ (Durel et al., 2009; Fahrentrapp et al., 2011).
Moreover, following fine-mapping and map-based cloning approaches, candidate genes for
fire blight resistance could be identified in ‘Evereste’ (Parravicini et al., 2011); the most
promising candidates showed homology with the Pto/Prf complex of genes that confers
resistance to the bacterium Pseudomonas syringae pathovar tomato DC3000 in tomato
(Solanum lycopersicum L.) (Chang et al., 2002; Zhou et al., 1997). All in all, these findings
suggest that the major fire blight resistance QTLs of ‘Evereste’ and ‘Robusta 5’, and possibly
the additional major QTLs on the way to be discovered in wild apples (Andreas Peil, personal
communication), blur the classical distinction between monogenic, qualitative resistance and
polygenic, quantitative resistance (Poland et al., 2009). As a consequence, the genomic region
associated with fire blight resistance at the very distal end of ‘Evereste’ LG 12 was designated
in this Thesis in two ways: either as a major-effect, strong-effect or highly efficacious QTL,
referring to its actual, quantitative phenotypic effect in segregating F1 progenies (Durel et al.,
157
2009); or as a fire blight resistance locus called Fb_E, referring to the phenotypic marker
Fb_E co-segregating with the Pto/Prf complex of genes found on a single bacterial artificial
chromosome (BAC) of ‘Evereste’ (Parravicini et al., 2011).
6.4.2. Accelerated introgression of a fire blight resistance locus from a crab apple
based on the transgenic early flowering line T1190 and molecular markers
Flachowsky and colleagues developed a high-speed breeding technology in apple based on a
transgenic early flowering line originating from the cultivar ‘Pinova’, named T1190
(Flachowsky et al., 2007; Flachowsky et al., 2009; Flachowsky et al., 2011). This transgenic
line was transformed with the BpMADS4 gene from silver birch (Betula pendula Roth.) that
promoted early flowering during the first season of glasshouse cultivation (Flachowsky et al.,
2007). Chapter 4 describes the linkage mapping of the T-DNA integration site on LG 4 of the
apple genome using a SNP marker (SNP_T1190) designed on the right side of the integration
site; it was reported in Flachowsky et al. (2011). As explained in Chapter 4, this knowledge
allows to anticipate the segregation of the early flowering trait in comparison with the
resistance gene(s)/QTL(s) introgressed; it thus helps to plan the size of the crossbred
progenies needed to get sufficient transgenic seedlings cumulating the expected number of
resistance gene(s)/QTL(s).
In Chapter 5, the transgenic early flowering line T1190 was applied to accelerate the first two
generations of introgression of the fire blight resistance locus Fb_E from the apple ornamental
cultivar ‘Evereste’. One F1 and two BC’1 transgenic seedlings were able to flower very
precociously (15 and 14 weeks after seed planting, respectively), confirming the ability of this
technology to dramatically reduce generation time in an apple pseudo-backcross scheme. The
use of a SSR marker co-segregating with the fire blight resistance locus Fb_E (Parravicini et
al., 2011) enabled a very reliable marker-assisted introgression of this locus; the strong effect
of the locus Fb_E on fire blight resistance could be also confirmed by a phenotypic test
carried out on 27 F1 seedlings from T1190 × ‘Evereste’.
A background selection was performed on 24 transgenic BC’1 seedlings carrying the Fb_E
locus using a minimum of 52 SSR markers distributed over all LGs of the apple genome. The
background selection aimed at estimating the proportion of ‘Evereste’ genome linked (linkage
drag) or unlinked to the Fb_E locus introgressed from the grandparent ‘Evereste’. The set of
SSR markers used did not cover the whole apple genome, and therefore did not enable a very
precise estimation of the proportion of ‘Evereste’ genome left on every chromosome of every
BC’1 seedling. Double-recombination events were probably overlooked and led us to over- or
158
under-estimate the proportion of ‘Evereste’ genome at the chromosome scale. However, these
undetected exchanges of chromosomal segments likely balanced each other at the genome
scale, making our SSR-based estimations representative of the actual ‘Evereste’ genomic
contribution. In particular, two BC’1 seedlings were found to carry less than 15 % of
‘Evereste’ genome, being thus prime candidates to continue the introgression of the Fb_E
locus; this result highlights the possibility to reduce the number of breeding cycles necessary
for the introgression of a trait from small-fruited wild Malus (hybrid) species, as previously
observed (Volz et al., 2009). All in all, it appears that the high-speed breeding technology
based on the transgenic early flowering line T1190 and molecular markers is now well-
established for apple pre-breeding, i.e. for the selection of parental genotypes with a
domesticated genetic background and donors of resistance for cultivar breeding; it makes the
wild germplasm more accessible and therefore more relevant to apple pre-breeders who will
be able to introgress at an unprecedented pace additional strong-effect QTLs for fire blight
resistance (possibly from M. × robusta, M. baccata, M. fusca, M. prunifolia and M.
hupehensis; Andreas Peil, personal communication; Fahrentrapp et al., 2011; Velasco, 2010).
6.4.3. Towards an optimal use of the high-speed breeding technology in apple
pre-breeding
The high-speed breeding technology based on the transgenic line T1190 has proven its worth
in accelerating pseudo-backcross introgression schemes in apple. However, the legal status of
the pre-breeding genotypes derived from such an introgression strategy is still unclear and
needs to be determined in order to benefit to the apple pre-breeders and breeders (see Chapter
5). Besides the regulatory aspect which is not the scope of this Thesis, the efficiency of the
high-speed breeding technology can be further improved by considering some technical and
strategical aspects.
On a technical point of view, the completion of the genome sequence of the apple cultivar
‘Golden Delicious’ has made more than 3 million predicted SNP markers available to apple
geneticists and breeders, which will facilitate the mapping of resistance loci in crab apples,
and the development of reliable and robust markers suited for their introgression (Gardiner et
al., 2010; Micheletti et al., 2011; Velasco et al., 2010). Besides, the wealth of BAC- and EST-
based SSR markers recently mapped in the apple genome should quickly enable a full SSR-
based genome coverage for exhaustive background selection. Additionally, the development
of two DArT genotyping arrays whose fragments can be readily sequenced and mapped to the
‘Golden Delicious’ genome may provide a valuable alternative to perform a background
159
selection against the genome of a crab apple resistance donor (Chapter 3; see also Schouten et
al., 2011).
On a strategic point of view, the number of disease resistance loci to be pyramided in a pre-
breeding genotype is a major issue. It is considered that cumulating 3 to 5 major resistance
genes may provide durable resistance to apple scab, following the example of natural gene
pyramids found in M. micromalus, Russian apple R12740-7A and some ‘Antonovka’
genotypes (Bus et al., 2000; Dayton et al., 1953; Kellerhals et al., 2009a; Patocchi et al.,
2009a; Schmidt, 1938). As research on the genetic basis of fire blight resistance is less
advanced, no “ideotype” of durable resistance to fire blight can be defined yet with certainty.
Moreover, how E. amylovora would respond to the presence of highly-efficacious resistance
loci in new apple cultivars can only be hypothesized (Smits et al., 2010a). Historical
breakdowns of major resistance genes introgressed individually in crop varieties have been
reported, e.g. for the phoma stem canker resistance gene Rlm6 in oilseed rape (Brassica napus
L.) (Brun et al., 2000), the stripe rust resistance gene Sr24 in wheat (Triticum spp.) (Bhardwaj
et al., 1990) and the scab resistance gene Rvi6/Vf in the cultivated apple (Parisi et al., 1993).
On this basis, it appears wise not to introgress highly efficacious fire blight resistance loci
individually in pre-breeding genotypes; the rationale being to avoid a quick loss of efficiency
of individual loci in the advanced selections derived from the pre-breeding genotypes.
Instead, various pyramids of fire blight resistance loci should be constructed in pre-breeding
genotypes; they could simultaneously be used as parents for cultivar breeding and tested for
their efficiency and putative durability by artificial inoculations of multiple E. amylovora
strains in quarantine facilities. Ideally, 2 to 3 fire blight resistance loci should be combined in
pre-breeding genotypes with other disease resistance genes/QTLs., in order to allow the
breeding of new apple cultivars resistant to multiple diseases (Bus et al., 2009; Kellerhals et
al., 2008). These new fire blight resistant cultivars would be an essential component of an
integrated disease management strategy for sustainable apple production in areas where fire
blight is present or even established.
160
161
Appendices
The data presented in the Appendices were produced either by P.-M. F. Le Roux or by
colleagues of the Federal Research Station Agroscope Changins-Wädenswil ACW
(Wädenswil, Switzerland) or of the Swiss Federal Institute of Technology Zürich (Zürich,
Switzerland). They complement the results presented in Chapters 2, 3 and 5.
162
Appendix A. Results of fire blight resistance tests on various apple
genotypes
Fire blight resistance tests were conducted in a quarantine glasshouse at the Federal research
Station Agroscope Changins-Wädenswil ACW (Wädenswil) as follows:
- Strain of E. amylovora: ACW 610
- Inoculum concentration: 109 cfu/ml
- Inoculation technique: syringe inserted at the tip of actively growing shoots (shoot
length > 15 cm)
- Temperature range: 18–24 °C
- Disease scoring: the mean fire blight lesion length in percent of total shoot length was
scored 21 days after inoculation (PLL 21 DAI)
Fig. A.1. Mean fire blight lesion length in percentage of total shoot length of apple cultivars
and genotypes in the pedigree of ‘Florina’ and ‘Nova Easygro’ (2007).
The x-axis indicates the apple cultivars and genotypes tested; the ornamental apple cultivar ‘Evereste’ was
included as resistant control. The y-axis indicates the mean percent lesion length recorded 21 days after
inoculation (PLL 21 DAI). The vertical bars indicate the standard errors. The figures above the bars refer to the
number of replications per cultivar
The inoculation and disease scoring were performed by Dr. Muhammad A. Khan (ETHZ) in the old quarantine
glasshouse of Agroscope ACW in Wädenswil (Switzerland).
6 6
10 10
7
10 9
10 10 10
10
10
0
10
20
30
40
50
60
70
80
90
100
Mea
n P
LL
21
DA
I
163
Fig. A.2. Mean fire blight lesion length in percentage of total shoot length of the apple
cultivars ‘Rewena’, ‘Gala Galaxy’ and ‘Idared’ (2009).
The x-axis indicates the apple cultivars tested; the apple cultivar ‘Gala Galaxy’ is a sport mutation of ‘Gala’ used
classically as susceptible control. The y-axis indicates the mean percent lesion length recorded 21 days after
inoculation (PLL 21 DAI). The vertical bars indicate the standard errors. The figures above the bars refer to the
number of replications per cultivar.
The inoculation and disease scoring were performed by Isabelle Baumgartner (ACW) in the new quarantine
glasshouse of Agroscope ACW in Wädenswil (Switzerland).
Fig. A.3. Mean fire blight lesion length in percentage of total shoot length of the apple
cultivars ‘Rewena’, ‘Gala Galaxy’ and ‘Pinova’ (2009).
The x-axis indicates the apple cultivars tested; the apple cultivar ‘Gala Galaxy’ is a sport mutation of ‘Gala’ used
classically as susceptible control. The y-axis indicates the mean percent lesion length recorded 21 days after
inoculation (PLL 21 DAI). The vertical bars indicate the standard errors. The figures above the bars refer to the
number of replications per cultivar.
8
10 10
0
10
20
30
40
50
60
70
Rewena Gala Galaxy Idared
Mea
n P
LL
21
DA
I
10
10
9
0
20
40
60
80
100
120
Rewena Gala Galaxy Pinova
Mea
n P
LL
21
DA
I
164
The inoculation and disease scoring were performed by Gabriella Silvestri (ACW) in the new quarantine
glasshouse of Agroscope ACW in Wädenswil (Switzerland).
165
Appendix B. Overview of the diversity arrays technology (DArT) procedure
From Wittenberg (2007)
Array development: genomic DNA of genotypes representative of the diversity within a
given species (e.g. M. × domestica Borkh.) are pooled (meta-genome), cut with a chosen
combination of restriction enzymes and ligated with adapters; the genome complexity is then
reduced by PCR using primers with selective overhangs, leading to a defined genomic
representation (i.e. a representative subset of genomic fragments) of the meta-genome. The
amplicons from the representation are cloned into a vector that is introduced in Escherichia
166
coli to form a library; the inserts (or fragments) are amplified using vector-specific primers,
purified and spotted onto glass slides to form a genotyping microarray.
Genotyping: the genomic DNA sample for analysis (e.g. individual of a mapping population,
accession of a genebank, …) is converted to a representation as in the microarray
development and labeled with a fluorescent dye. The DNA sample representation is then
hybridized to the genotyping array together with a reference DNA (e.g. poly-linker region of
the cloning vector) labeled with another dye that quantifies the amount of DNA spotted on the
microarray. The relative hybridization intensity, proportional to the relative abundance of a
fragment in the genomic representation, is measured for each fragment by confocal laser
scans. The DArT markers are scored in a dominant manner, i.e. presence (“1”) or absence
(“0”) of individual fragments in the DNA sample representation (Gupta et al., 2008; Jaccoud
et al., 2001; Wenzl et al., 2004; Wittenberg, 2007).
167
Appendix C. Genotyping of apple cultivars and wild Malus species and
hybrids with SSR markers co-segregating with (ChFbE02) or tightly linked
to (ChFbE09) the fire blight resistance locus Fb_E of the apple genotype
‘Evereste’
Malus genotype ChFbE02 5 ChFbE09
5
‘Evereste’ 229 249
M × E F1 R 1 229 249
M × E F1 S 2 -
6 -
‘Enterprise’ - -
‘Rewena’ - -
‘Florina’ - -
‘Nova Easygro’ - -
‘Priscilla’ - -
Gala - -
‘Golden Delicious’ - -
F2 26830-2 3 - -
F2 26829-2-2 3 - -
M.× floribunda clone 821 4 229 249
M baccata jackii 4 - -
Hansen’s baccata #2 - -
M. fusca 4 - -
M. × robusta ‘Robusta 5’ 4 - -
1 F1 individual derived from the cross ‘MM.106’ × ‘Evereste’ that inherited the fire blight resistance locus Fb_E
(G. Parraviccini Rusca, personal communication).
2 F1 individual derived from the cross ‘MM.106’ × ‘Evereste’ that did not inherit the fire blight resistance locus
Fb_E (G. Parravicini Rusca, personal communication).
3 F2 descendants of the cross ‘Rome Beauty’ × (Malus × floribunda clone 821) (Dayton et al., 1977; Shay and
Hough, 1952).
4 Wild Malus species or hybrids known to display a high level of fire blight resistance.
5 The alleles of SSR markers ChFbE02 and ChFbE09 displayed (in bp) are in coupling with the fire blight
resistance of ‘Evereste’; they are shorter than the alleles originally reported by Parravicini et al. (2011) who
extended the 5’ end of the forward SSR primers with the M13 sequence 5’-GACTGCGTACCAATTCAAA-3’
(Schuelke, 2000); SSR marker ChFbE09 is separated from the locus Fb_E by 5 recombination events within
2703 individuals derived from the crosses ‘MM.106’ × ‘Evereste’ and ‘Evereste’ × ‘MM.106’.
6 “-” means that no fragment was amplified by PCR for SSR markers ChFbE02 or ChFbE09.
168
Appendix D. Screening results of the SSR markers used for the background selection of BC’1 seedlings carrying the
BpMADS4 transgene from the transgenic line T1190 and the fire blight resistance locus Fb_E from the apple genotype
‘Evereste’ (see Chapter 5)
SSR marker 1 Linkage group Informativeness
2 Reason for lack of informativeness
CH03g12z 1 - Amplification in multiplex PCR was stable but assigning each allele to a locus was uncertain
CH05g08 1 a, b, c, d
CH-Vf1 1 a, c, d ‘Topaz’ × F1_81 showed a partially inconclusive < ab × ab > segregation
HB11-SSR 1 - Stable amplification in multiplex PCR but presumably multi-locus; also, alleles segregation too complicated to be
determined with certainty in a small progeny
Hi02c07 1 a, b, c, d
CH02f06 2 b F1_5 and F1_74 inherited two alleles of the same size from T1190 and ‘Evereste’
CH03d01 2 a, b, c, d
CH05e03 2 b F1_5 and F1_74 inherited two alleles of the same size from T1190 and ‘Evereste’
NH033b 2 a, b, c, d
TsuENH045 2 - Only one allele amplified in T1190, ‘Evereste’ and the seedlings F1_5, F1_74 and F1_81
TsuENH062 2 - Only one allele amplified in T1190, ‘Evereste’ and the seedlings F1_5, F1_74 and F1_81
AU223657-SSR 3 a, b, c F1_5 inherited two alleles of the same size from T1190 and ‘Evereste’
CH03e03 3 a, b, c, d
CH03g07 3 a, b, c F1_5 inherited two alleles of the same size from T1190 and ‘Evereste’
CH03g12y 3 - Amplification in multiplex PCR was stable but assigning each allele to a locus was uncertain
Hi03d06 3 b, c, d ‘Topaz’ × F1_74 showed a partially inconclusive < ab × ab > segregation
AT000420-SSR 4 a, b, c F1_5 inherited two alleles of the same size from T1190 and ‘Evereste’
CH01d03y 4 - Only the alleles of the locus z (LG 12) could be identified due to higher amplification efficiency
CH04e02 4 a, c F1_5 and F1_81 inherited two alleles of the same size from T1190 and ‘Evereste’
Hi07b02 4 a, b, c, d
NB141b 4 -
F1_5, F1_74 and F1_81 inherited two alleles of the same size from T1190 and ‘Evereste’
169
CH03a09 5 a, b, c, d
CH04e03 5 a, c, d F1_81 inherited two alleles of the same size from T1190 and ‘Evereste’
GD103 5 d F1_74 and F1_81 inherited two alleles of the same size from T1190 and ‘Evereste’
Hi04a08 5 b, d F1_74 inherited two alleles of the same size from T1190 and ‘Evereste’
Hi04d02 5 a, b, c, d
CH03d07 6 - F1_5, F1_74 and F1_81 inherited two alleles of the same size from T1190 and ‘Evereste’
CH03d12 6 - Only one allele amplified in T1190, ‘Evereste’, the seedlings F1_5, F1_74 and F1_81, and ‘Topaz’
CH05a05 6 a, b, c, d
HB09-SSR 6 d ‘Topaz’ × F1_74, ‘Topaz’ × F1_81 and ‘Maloni Sally
®’ × F1_74 showed a partially inconclusive < ab × ab >
segregation
NZ23g04 6 a, b, c, d
CH04e05 7 a, b, c, d
CH-F7-Fb1 7 b F1_5 and F1_74 inherited two alleles of the same size from T1190 and ‘Evereste’
CH-Sd1 7 a, b, c, d
Hi05b09 7 a, c, d ‘Topaz’ × F1_81 showed a partially inconclusive < ab × ab > segregation
Hi12f04 7 c ‘Topaz’ × F1_74 and ‘Topaz’ × F1_81 showed partially inconclusive < ab × ab > and < a0 × ab > segregations,
respectively; F1_5 × ‘Milwa’ showed an inconclusive < a0 × aa > segregation
CH01c06 8 a, b, c, d
CH01f09 8 a, d F1_81 inherited two alleles of the same size from T1190 and ‘Evereste’; ‘Maloni Sally
®’ × F1_74 showed a
partially inconclusive < ab × ab > segregation
CH01h10 8 d F1_74 and F1_81 inherited two alleles of the same size from T1190 and ‘Evereste’
CH02g09 8 a, c, d F1_81 inherited two alleles of the same size from T1190 and ‘Evereste’
CH01f03b 9 a, b, c, d
CH01h02 9 a, b, c, d
CN444542-SSR 9 a, b, c, d
Hi04a05 9 a, b, c, d
Hi05e07 9 a, b, c, d
CH02b03b 10 a, b, c, d
CH02b07 10 a, b, c, d
170
CH02c11 10 a, c, d F1_81 inherited two alleles of the same size from T1190 and ‘Evereste’
Hi02d04 10 a, b, c, d
MS06g03 10 - Multiple alleles that could not be assigned to a specific locus were amplified
CH02d08 11 a, b, c, d
CH04g07 11 a, b, c, d
Hi16d02 11 b F1_5 and F1_74 inherited two alleles of the same size from T1190 and ‘Evereste’
CH01d03z 12 a, b, c F1_5 × ‘Milwa’ showed a partially inconclusive < ab × ab > segregation
CH03c02 12 a, b, c, d
CH04g04 12 a, b, c, d
CH05d04 3 12 a, b, c, d
CH05f04 13 c, d ‘Topaz’ × F1_81 and ‘Topaz’ × F1_74, showed a partially inconclusive < ab × b0> segregation
CH05h05 13 - Three to four alleles that could not be assigned to a specific locus were amplified in every sample, although
CH05h05 was described as single locus (Patocchi et al. 2009b)
GD147 13 a, b, c, d
Hi04g05 13 a, b, c, d
CH01g05 14 a, b, c, d
CH04c07 14 b, c ‘Topaz’ × F1_74 and F1_5 × ‘Milwa’ showed an inconclusive < ab × b0 > segregation
MDAJ761-SSR 14 a, b, c, d
NZmsPal51 14 - T1190 × ‘Evereste’ showed a partially inconclusive < ab × ab > segregation
CH01d08 15 a, b, c, d
CH02c09 15 a, b, c, d
CH02d11 15 a, b, c, d
Hi03g06 15 a, b, c, d
CH02a03 16 a, b, c, d
CH04f10 16 a, b, c, d
CH05a04 16 a, b, c, d
Hi04e04 16 - F1_5 and F1_81 inherited two alleles of the same size from T1190 and ‘Evereste’; also, alleles segregation was
too complicated to be determined with certainty in a small progeny
AT000174-SSR 17 a, b, c, d
CH01h01 17 a, b, c, d
171
Hi02f12 17 - F1_5 and F1_74 inherited two alleles of the same size from T1190 and ‘Evereste’; ‘Topaz’ × F1_81 showed a
partially inconclusive < ab × b0 > segregation
Hi03c05 17 a, b, c, d
Hi07h02 17 a, b, c F1_5 inherited two alleles of the same size from T1190 and ‘Evereste’
1: the SSR markers underlined were not present in the core set of genome-covering SSR markers defined by Patocchi et al. (2009b); their reference is indicated in the text (see
Chapter 5). The SSR markers in italic are highly polymorphic in the sense of Patocchi et al. (2009b) (more than 10 alleles amplified). 2: informativeness refers here to the possibility to determine unambiguously for each seedling of the four BC’1 crosses (designated as a, b, c and d) which allele originated
from the ornamental apple cultivar ‘Evereste’; a: ‘Topaz’ × F1_74; b: ‘Topaz’ × F1_81; c: ‘Maloni Sally®’ × F1_74; d: F1_5 × ‘Milwa’.
3: the allele of the LG 12 SSR marker ChFbE06 in coupling with the fire blight resistance (273 bp) was unambiguously identified in each BC’1 seedling tested and could be
traced back to the ornamental apple cultivar ‘Evereste’.
172
173
References
174
Abdul-Kader, A.M., D.W. Bauer, S.V. Beer, J.L. Norelli, and H.S. Aldwinckle. 1999.
Evaluation of the hrpN gene for increasing resistance to fire blight in transgenic apple.
Acta Hortic 489:247-250.
Akbari, M., P. Wenzl, V. Caig, J. Carling, L. Xia, S. Yang, G. Uszynski, V. Mohler, A.
Lehmensiek, H. Kuchel, M. Hayden, N. Howes, P. Sharp, P. Vaughan, B. Rathmell, E.
Huttner, and A. Kilian. 2006. Diversity arrays technology (DArT) for high-throughput
profiling of the hexaploid wheat genome. Theoretical and Applied Genetics 113:1409-
1420.
Akhunov, E., C. Nicolet, and J. Dvorak. 2009. Single nucleotide polymorphism genotyping in
polyploid wheat with the Illumina GoldenGate assay. Theoretical and Applied
Genetics 119:507-517.
Aldwinckle, H.S. 1976. Early flowering of cultivated apple seedlings forced in the
greenhouse. Acta Hortic 56:201-203.
Aldwinckle, H.S., and S.V. Beer. 1979. Fire Blight and Its Control. Horticultural Reviews
1:423-474.
Aldwinckle, H.S., and T. van der Zwet. 1979. Recent Progress in Breeding for Fireblight
Resistance in Apples and Pears in North America. OEPP/EPPO Bulletin 9:27-34.
Aldwinckle, H.S., H.L. Gustafson, and P.L. Forsline. 1999. Evaluation of the core subset of
the USDA apple germplasm collection for resistance to fire blight. Acta Hortic
489:269-272.
Aldwinckle, H.S., J.L. Norelli, J.Z. Mills, and S.K. Brown. 1996. Transformation of Gala
apple with lytic protein genes. Acta Hortic 411:411-412.
Alheit, K.V., J.C. Reif, H.P. Maurer, V. Hahn, E.A. Weissmann, T. Miedaner, and T.
Würschum. 2011. Detection of segregation distortion loci in triticale (× Triticosecale
Wittmack) based on a high-density DArT marker consensus genetic linkage map.
BMC Genomics 12:380.
Alston, F.H. 1994. The position of fire blight (Erwinia amylovora) avoidance and resistance
amongst pear breeding priorities. Norwegian Journal of Agricultural Sciences Suppl.
17:331-336.
Alston, F.H., and P. Spiegel-Roy. 1985. Fruit Tree Breeding: Strategies, Achievements and
Constraints, p. 49-67, In M. G. R. Cannel and J. E. Jackson, eds. Attributes of Trees as
Crop Plants. Institute of Terrestrial Ecology Huntington, UK.
Baldo, A., J. Norelli, R. Farrell, C. Bassett, H. Aldwinckle, and M. Malnoy. 2010.
Identification of genes differentially expressed during interaction of resistant and
susceptible apple cultivars (Malus × domestica) with Erwinia amylovora. BMC Plant
Biology 10:1.
Bannier, H.-J. 2011. Moderne Apfelzüchtung: Genetische Verarmung und Tendenzen zur
Inzucht. Erwerbs-Obstbau 52:85-110.
Baumgartner, I., L. Franck, G. Silvestri, A. Patocchi, B. Duffy, J. Frey, and M. Kellerhals.
2010. Advanced strategies for breeding fire blight resistant high quality apples, pp. 31-
37. 14th
International conference on organic fruit-growing Foerdergemeinschaft
Oekologischer Obstbau e.V. (FOEFO), Weinsberg, Hohenheim, Germany.
Baumgartner, I.O., L. Franck, G.A.L. Broggini, A. Patocchi, and M. Kellerhals. 2011. Fire
Blight Resistance from ‘Evereste’ and Malus sieversii Used in Breeding for New High
Quality Apple Cultivars: Strategies and Results. Acta Hortic 896:391-398.
Beavis, W.D. 1998. QTL analyses: power, precision, and accuracy, p. 145-162, In A. H.
Paterson, ed. Molecular Dissection of Complex Traits. CRC Press, New York, USA.
Beckerman, J., J. Chatfield, and E. Draper. 2009. A 33-year Evaluation of Resistance and
Pathogenicity in the Apple Scab-crabapples Pathosystem. Hortscience 44:599-608.
175
Bell, A.C., T.G. Ranney, T.A. Eaker, and T.B. Sutton. 2000. Evaluating Fire Blight
Resistance among Flowering Crabapples (Malus spp.) and Pears (Pyrus spp.) 11th
Metropolitan Tree Improvement Alliance (METRIA) Conference. Landscape Plant
Development Center, Gresham, Oregon, USA.
Bhardwaj, S.C., S.K. Nayar, M. Prashar, J. Kumar, M.K. Menon, and S.B. Singh. 1990. A
pathotype of Puccinia graminis f. sp. tritici on Sr24 in India. Cereal Rusts Powdery
Mildews Bulletin 18:35-37.
Biggs, A.R., W.W. Turechek, and T.R. Gottwald. 2008. Analysis of Fire Blight Shoot
Infection Epidemics on Apple. Plant Disease 92:1349-1356.
Billing, E. 2010. Fire blight. Why do views on host invasion by Erwinia amylovora differ?
Plant Pathology 60:178-189.
Bink, M., A. Anderson, W. van de Weg, and E. Thompson. 2008. Comparison of marker-
based pairwise relatedness estimators on a pedigreed plant population. Theoretical and
Applied Genetics 117:843-855.
Bliss, F.A. 2010. Marker-Assisted Breeding in Horticultural Crops. Acta Hortic 859:339-350.
Bolibok-Bragoszewska, H., K. Heller-Uszynska, P. Wenzl, G. Uszynski, A. Kilian, and M.
Rakoczy-Trojanowska. 2009. DArT markers for the rye genome - genetic diversity
and mapping. BMC Genomics 10:578.
Bonn, W.G., and D.C. Elfving. 1990. Evaluation of crabapple cultivars and selections for
resistance to fire blight. Acta Hortic 273:311-317.
Bonn, W.G., and T. van der Zwet. 2000. Distribution and economic importance of fire blight,
p. 37-55, In J. L. Vanneste, ed. Fire Blight: the disease and its causative agent,
Erwinia amylovora. CABI Publishing, Wallingford, UK.
Borejsza-Wysocka, E., J. Norelli, H. Aldwinckle, and M. Malnoy. 2010. Stable expression
and phenotypic impact of attacin E transgene in orchard grown apple trees over a 12
year period. BMC Biotechnology 10:41.
Borejsza-Wysocka, E.E., M. Malnoy, W.S. Kim, K. Geider, S. Beer, and H.S. Aldwinckle.
2007. Expression of Phi-Ea1h phage depolymerase gene with constitutive and
inducible promoters, translation enhancer and signal sequence in transgenic apple
plants increases resistance to fire blight. Acta Hortic 738:273-276.
Borejsza-Wysocka, E.E., M. Malnoy, X. Meng, J.M. Bonasera, R.M. Nissinen, J.F. Kim, S.V.
Beer, and H.S. Aldwinckle. 2004. Silencing of apple proteins that interact with DspE,
a pathogenicity effector from Erwinia amylovora, as a strategy to increase resistance
to fire blight. Acta Hortic 663:469-474.
Bringhurst, R.S. 1983. Breeding strategy, In J. N. Moore and J. Janick, eds. Methods in Fruit
Breeding. Purdue University Press, West Lafayette, Indiana, USA
Brisset, M.-N., S. Cesbron, S.V. Thomson, and J.-P. Paulin. 2000. Acibenzolar-s-methyl
induces the accumulation of defense-related enzymes in apple and protects from fire
blight. European Journal of Plant Pathology 106:529-536.
Broggini, G.A.L. 2007. Identification of apple scab avirulence gene AvrVg candidates. PhD
thesis, ETHZ, Zürich, Switzerland.
Broggini, G.A.L., B. Duffy, E. Holliger, H.-J. Schärer, C. Gessler, and A. Patocchi. 2005.
Detection of the fire blight biocontrol agent Bacillus subtilis BD170 (Biopro) in a
Swiss apple orchard. European Journal of Plant Pathology 111:93-100.
Broggini, G.A.L., P. Galli, G. Parravicini, L. Gianfranceschi, C. Gessler, and A. Patocchi.
2009. HcrVf paralogs are present on linkage groups 1 and 6 of Malus. Genome
52:129-138.
Broggini, G.A.L., V.G.M. Bus, G. Parravicini, S. Kumar, R. Groenwold, and C. Gessler.
2010. Genetic mapping of 14 avirulence genes in an EU-B04 × 1639 progeny of
Venturia inaequalis. Fungal Genetics and Biology 48:166-176.
176
Brun, H., S. Levivier, I. Somda, D. Ruer, M. Renard, and A.M. Chèvre. 2000. A field method
for evaluating the potential durability of new resistance sources: application to the
Leptosphaeria maculans-Brassica napus pathosystem. Phytopathology 90:961-966.
Bus, V. 2006. Differential Host-Pathogen Interactions of Venturia inaequalis and Malus. PhD
thesis, University of Auckland, Auckland, New Zealand.
Bus, V., C. Ranatunga, S. Gardiner, H. Bassett, and E. Rikkerink. 2000. Marker assisted
selection for pest and disease resistance in the New Zealand apple breeding program.
Acta Hortic 538:541-547.
Bus, V., H. Bassett, D. Bowatte, D. Chagné, C. Ranatunga, D. Ulluwishewa, C. Wiedow, and
S. Gardiner. 2010. Genome mapping of an apple scab, a powdery mildew and a
woolly apple aphid resistance gene from open-pollinated Mildew Immune Selection.
Tree Genetics & Genomes 6:477-487.
Bus, V., D. Chagné, H. Bassett, D. Bowatte, F. Calenge, J.M. Celton, C.E. Durel, M. Malone,
A. Patocchi, A. Ranatunga, E. Rikkerink, D. Tustin, J. Zhou, and S. Gardiner. 2008.
Genome mapping of three major resistance genes to woolly apple aphid (Eriosoma
lanigerum Hausm.). Tree Genetics & Genomes 4:223-236.
Bus, V.G.M., D. Esmenjaud, E. Buck, and F. Laurens. 2009. Application of Genetic Markers
in Rosaceous Crops, p. 563-599, In K. M. Folta and S. E. Gardiner, eds. Genetics and
Genomics of Rosaceae, Vol. 6. Springer Science+Business Media, New York, USA.
Bus, V.G.M., E.H.A. Rikkerink, V. Caffier, C.-E. Durel, and K.M. Plummer. 2011. Revision
of the Nomenclature of the Differential Host-Pathogen Interactions of Venturia
inaequalis and Malus. Annual Review of Phytopathology 49:391-413
Bus, V.G.M., F.N.D. Laurens, E.W. van de Weg, R.L. Rusholme, E.H.A. Rikkerink, S.E.
Gardiner, H.C.M. Bassett, L.P. Kodde, and K.M. Plummer. 2005a. The Vh8 locus of a
new gene-for-gene interaction between Venturia inaequalis and the wild apple Malus
sieversii is closely linked to the Vh2 locus in Malus pumila R12740-7A. New
Phytologist 166:1035-1049.
Bus, V.G.M., E.H.A. Rikkerink, W.E. van de Weg, R.L. Rusholme, S.E. Gardiner, H.C.M.
Bassett, L.P. Kodde, L. Parisi, F.N.D. Laurens, E.J. Meulenbroek, and K.M. Plummer.
2005b. The Vh2 and Vh4 scab resistance genes in two differential hosts derived from
Russian apple R12740-7A map to the same linkage group of apple. Molecular
Breeding 15:103-116.
Caffier, V., and F. Laurens. 2005. Breakdown of Pl2, a major gene of resistance to apple
powdery mildew, in a French experimental orchard. Plant Pathology 54:116-124.
Caffier, V., and L. Parisi. 2007. Development of apple powdery mildew on sources of
resistance to Podosphaera leucotricha, exposed to an inoculum virulent against the
major resistance gene Pl-2. Plant Breeding 126:319-322.
Calenge, F., and C.-E. Durel. 2006. Both stable and unstable QTLs for resistance to powdery
mildew are detected in apple after four years of field assessments. Molecular Breeding
17:329-339.
Calenge, F., A. Faure, M. Goerre, C. Gebhardt, W.E. Van de Weg, L. Parisi, and C.E. Durel.
2004. Quantitative Trait Loci (QTL) Analysis Reveals Both Broad-Spectrum and
Isolate-Specific QTL for Scab Resistance in an Apple Progeny Challenged with Eight
Isolates of Venturia inaequalis. Phytopathology 94:370-379.
Calenge, F., D. Drouet, C. Denancé, E.W. Van de Weg, M.N. Brisset, J.P. Paulin, and C.E.
Durel. 2005. Identification of a major QTL together with several minor additive or
epistatic QTLs for resistance to fire blight in apple in two related progenies.
Theoretical and Applied Genetics 111:128-125.
177
Celton, J.-M., D. Chagne, S. Tustin, S. Terakami, C. Nishitani, T. Yamamoto, and S.
Gardiner. 2009a. Update on comparative genome mapping between Malus and Pyrus.
BMC Research Notes 2:182.
Celton, J.M., D. Tustin, D. Chagné, and S. Gardiner. 2009b. Construction of a dense genetic
linkage map for apple rootstocks using SSRs developed from Malus ESTs and Pyrus
genomic sequences. Tree Genetics & Genomes 5:93-107.
Chagné, D., K. Gasic, R.N. Crowhurst, Y. Han, H.C. Bassett, D.R. Bowatte, T.J. Lawrence,
E.H.A. Rikkerink, S.E. Gardiner, and S.S. Korban. 2008. Development of a set of SNP
markers present in expressed genes of the apple. Genomics 92:353-358.
Chagné, D., C.M. Carlisle, C. Blond, R.K. Volz, C.J. Whitworth, N.C. Oraguzie, R.N.
Crowhurst, A.C. Allan, R.V. Espley, R.P. Hellens, and S.E. Gardiner. 2007. Mapping
a candidate gene (MdMYB10) for red flesh and foliage colour in apple. BMC
Genomics 8:212-223.
Chang, J.H., Y.S. Tai, A.J. Bernal, D.T. Lavelle, B.J. Staskawicz, and R.W. Michelmore.
2002. Functional analyses of the Pto resistance gene family in tomato and the
identification of a minor resistance determinant in a susceptible haplotype. Molecular
Plant-Microbe Interactions 15:281-291.
Charcosset, A., and L. Moreau. 2004. Use of molecular markers for the development of new
cultivars and the evaluation of genetic diversity. Euphytica 137:81-94.
Chevreau, E. 2009. La transgenèse pour l’innovation variétale fruitière: état des lieux et
perspectives. Innovations Agronomiques 7:153-163.
Collard, B.C.Y., M.Z.Z. Jahufer, J.B. Brouwer, and E.C.K. Pang. 2005. An introduction to
markers, quantitative trait loci (QTL) mapping and marker-assisted selection for crop
improvement: The basic concepts. Euphytica 143:169-196.
Conner, P.J., S.K. Brown, and N.F. Weeden. 1997. Randomly amplified polymorphic DNA-
based genetic linkage maps of three apple cultivars. Journal of the American Society
of Horticultural Science 122 350-359.
Cooper, J.I., and A.T. Jones. 1983. Responses of plants to viruses - proposals for the use of
terms. Phytopathology 73:127-128.
Costa, F., W.E. Van de Weg, S. Stella, L. Dondini, D. Pratesi, S. Musacchi, and S. Sansavini.
2008. Map position and functional allelic diversity of Md-Exp7, a new putative
expansin gene associated with fruit softening in apple (Malus × domestica Borkh.) and
pear (Pyrus communis). Tree Genetics & Genomes 4:575-586.
Cova. 2008. Isolamento di marcatori microsatelliti strettamente associati al gene di resistenza
a tricchiolatura Vm in melo. PhD thesis, University of Bologna, Bologna, Italy.
Cova, V., D. Perini, V. Soglio, M. Komjanc, E. van de Weg, C. Gessler, and L.
Gianfranceschi. 2011. Exploiting expressed sequence tag databases for mapping
markers associated with fruit development and fruit quality in apple. Molecular
Breeding:DOI: 10.1007/s11032-011-9584-8.
Coyier, D.L., and R.P. Covey. 1975. Tolerance of Erwinia amylovora to streptomycin sulfate
in Oregon and Washington. Plant Disease Reporter 59:849-852
Crandall, C.S. 1926. Apple breeding at the University of Illinois. University of Illinois
(Urbana) Agricultural Experiment Station Bulletin 275.
Crosby, J.A., J. Janick, P.C. Pecknold, S.S. Korban, P.A. O'Connon, S.M. Ries, J. Goffreda,
and V. A. 1992. Breeding Apples for Scab Resistance: 1945-1990. Fruit Varieties
Journal 46:145-166.
Cummins, J.N., and H.S. Aldwinckle. 1995. Breeding rootstocks for tree fruit crops. New
Zealand Journal of Crop and Horticultural Science 23:395-402.
178
Danan, S. 2009. Diversité structurale des locus de résistance à Phytophtora infestans chez la
pomme de terre et synténie chez les Solanacées. PhD thesis, SupAgro Montpellier,
Montpellier, France.
Darvasi, A., and M. Soller. 1994. Optimum spacing of genetic markers for determining
linkage between marker loci and quantitative trait loci. Theoretical and Applied
Genetics 89:351-357.
Darvasi, A., A. Weinreb, V. Minke, J.I. Weller, and M. Soller. 1993. Detecting Marker-QTL
Linkage and Estimating QTL Gene Effect and Map Location Using a Saturated
Genetic Map. Genetics 134:943-951.
Dayton, D.F., J.R. Shay, and L.F. Hough. 1953. Apple scab resistance from R12740-7A, a
Russian apple. Proceedings of the American Society for Horticultural Science 62:334-
340.
Dayton, D.F., E.B. Williams, J. Janick, F.H. Emerson, L.F. Hough, and C.H. Bailey. 1977.
Co-op 19, 20, 21, and 22: Four Scab-Resistant Apple Selections Released for
Advanced Testing. University of Illinois (Urbana) Agricultural Experiment Station
Bulletin 755.
de Nettancourt, D. 2001. Incompatibility and Incongruity in Wild and Cultivated Plants
Springer Science+Business Media.
Doerge, R.W., and G.A. Churchill. 1996. Permutation tests for multiple loci affecting a
quantitative character. Genetics 142:285-294.
Dondini, L., L. Pierantoni, F. Gaiotti, R. Chiodini, S. Tartarini, C. Bazzi, and S. Sansavini.
2004. Identifying QTLs for fire-blight resistance via a European pear (Pyrus
communis L.) genetic linkage map. Molecular Breeding 14:407-418.
Dowkiw, A., and C. Bastien. 2007. Presence of defeated qualitative resistance genes
frequently has major impact on quantitative resistance to Melampsora larici-populina
leaf rust in P. × interamericana hybrid poplars. Tree Genetics & Genomes 3:261-274.
Dreesen, R.S.G., B.T.M. Vanholme, K. Luyten, L. Van Wynsberghe, G. Fazio, I. Roldan-
Ruiz, and J. Keulemans. 2010. Analysis of Malus S-RNase gene diversity based on a
comparative study of old and modern apple cultivars and European wild apple.
Molecular Breeding 26:693-709.
Dreo, T., M. Ravnikar, J.E. Frey, T.H.M. Smits, and B. Duffy. 2011. In Silico Analysis of
Variable Number of Tandem Repeats in Erwinia amylovora Genome. Acta Hortic
896:115-118.
Duffy, B., and E. Holliger. 2008. Abschlussbericht Streptomycin-Monitoring-Modul 2008
[Streptomycin-Monitoring-Programme Report 2008]. Forschungsanstalt Agroscope
Changins-Wädenswil ACW, Wädenswil, Switzerland.
Duffy, B., H.J. Schärer, M. Bünter, A. Klay, and E. Holliger. 2005. Regulatory measures
against Erwinia amylovora in Switzerland. OEPP/EPPO Bulletin 35:239-244.
Dugé de Bernonville, T. 2009. Caractérisations histologique, moléculaire et biochimique des
interactions compatible et incompatible entre Erwinia amylovora, agent du feu
bactérien, et le pommier (Malus × domestica), PhD thesis, Université d'Angers,
Angers, France.
Dunemann, F., D. Ulrich, A. Boudichevskaia, C. Grafe, and W.E. Weber. 2009. QTL
mapping of aroma compounds analysed by headspace solid-phase microextraction gas
chromatography in the apple progeny ‘Discovery’ × ‘Prima’. Molecular Breeding
23:501-521.
Durel, C.-E., C. Denancé, and M.-N. Brisset. 2009. Two distinct major QTL for resistance to
fire blight co-localize on linkage group 12 in apple genotypes ‘Evereste’ and Malus
floribunda clone 821. Genome 52:139-147.
179
Durel, C.-E., F. Laurens, V. Caffier, B. Le Cam, and N. Sapoukhina. 2007. Les apports de
l'innovation variétale: Recherches menées pour améliorer la résistance du pommier à
la tavelure. Innovations Agronomiques 1:47-61.
Durel, C.-E., L. Parisi, F. Laurens, W.E. Van de Weg, R. Liebhard, and M.F. Jourjon. 2003.
Genetic dissection of partial resistance to race 6 of Venturia inaequalis in apple.
Genome 46:224-234.
Durham, R.E., and S.S. Korban. 1994. Evidence of gene introgression in apple using RAPD
markers. Euphytica 79:109-114.
Edge-Garza, D.A., and C.P. Peace. 2010. Enabling Marker-Assisted Seedling Selection in the
Washington Apple Breeding Program. Acta Hortic 859:369-374.
Elo, A., J. Lemmetyinen, A. Novak, K. Keinonen, I. Porali, M. Hassinen, and T. Sopanen.
2007. BpMADS4 has a central role in inflorescence initiation in silver birch (Betula
pendula). Physiologia Plantarum 131:149-158.
Erdin, N., S. Tartarini, G.A.L. Broggini, F. Gennari, S. Sansavini, C. Gessler, and A.
Patocchi. 2006. Mapping of the apple scab-resistance gene Vb. Genome 49:1238-
1245.
Espley, R.V., R.P. Hellens, J. Putterill, D.E. Stevenson, S. Kutty-Amma, and A.C. Allan.
2007. Red colouration in apple fruit is due to the activity of the MYB transcription
factor, MdMYB10. The Plant Journal 49:414-427.
Evans, K., A. Patocchi, F. Rezzonico, F. Mathis, C. Durel, F. Fernández-Fernández, A.
Boudichevskaia, F. Dunemann, M. Stankiewicz-Kosyl, L. Gianfranceschi, M.
Komjanc, M. Lateur, M. Madduri, Y. Noordijk, and W. van de Weg. 2010.
Genotyping of pedigreed apple breeding material with a genome-covering set of
SSRs: trueness-to-type of cultivars and their parentages. Molecular Breeding:DOI:
10.1007/s11032-010-9502-5.
Fahrentrapp, J., G.A.L. Broggini, A. Peil, M. Kellerhals, M. Malnoy, K. Richter, and C.
Gessler. 2011. Fine Mapping of Fire Blight Resistance Locus in Malus × robusta 5 on
Linkage Group 3. Acta Hortic 896:243-244.
Fazio, G., R.P. McQuinn, T.L. Robinson, and H.S. Aldwinckle. 2006. Differential
Susceptibility to Fire Blight in Commercial and Experimental Apple Rootstock
Cultivars. Acta Hortic 704:527-530.
Fazio, G., Y. Wan, N.L. Russo, and K.S. Aldwinckle. 2008. Investigation on the Inheritance
of Strain Specific Resistance to Erwinia amylovora in an Apple Rootstock
Segregating Population. Acta Hortic 793:331-335.
Fazio, G., H.S. Aldwinckle, G.M. Volk, C.M. Richards, W.J. Janisiewicz, and P.L. Forsline.
2009. Progress in Evaluating Malus sieversii for Disease Resistance and Horticultural
Traits. Acta Hortic 814:59-66.
Fernández-Fernández, F., K.M. Evans, J.B. Clarke, C.L. Govan, C.M. James, S. Maric, and
K.R. Tobutt. 2008. Development of an STS map of an interspecific progeny of Malus.
Tree Genetics & Genomes 7:469-479.
Feuillet, C., P. Langridge, and R. Waugh. 2008. Cereal breeding takes a walk on the wild
side. Trends in Genetics 24:27-32.
Finelli, F., A. Contessi, A. Calzolari, A. Saccardi, G. Zecchin, and M. Vettorazzo. 2004.
Fireblight in the Po valley (Italy): the monitoring programmes of Emilia-Romagna and
Veneto regions. OEPP/EPPO Bulletin 34:331-334.
Fischer, C. 1994. Shortening of the juvenile period in apple breeding, p. 161-164, In H.
Schmidt and M. Kellerhals, eds. Progress in Temperate Fruit Breeding. Kluwer
Academic Publishers, Dordrecht, Boston, London.
Fischer, C., and K. Richter. 1999. Results on fire blight resistance in the Pillnitz apple
breeding programme. Acta Hortic 489:279-285.
180
Fischer, C., and K. Richter. 2004. Fire Blight Resistant Apple Cultivars Produced by
Conventional Breeding. Acta Hortic 663:721-724.
Fischer, M., and C. Fischer. 1999. Evaluation of Malus species and cultivars at the Fruit
Genebank Dresden-Pillnitz and its use for apple resistance breeding. Genetic
Resources and Crop Evolution 46:235-241.
Flachowsky, H., A. Peil, T. Sopanen, A. Elo, and M.V. Hanke. 2007. Overexpression of
BpMADS4 from silver birch (Betula pendula Roth.) induces early-flowering in apple
(Malus × domestica Borkh.). Plant Breeding 126:137-145.
Flachowsky, H., M.V. Hanke, A. Peil, S.H. Strauss, and M. Fladung. 2009. A review on
transgenic approaches to accelerate breeding of woody plants. Plant Breeding
128:217-226.
Flachowsky, H., C. Hättasch, M. Höfer, A. Peil, and M.-V. Hanke. 2010. Overexpression of
LEAFY in apple leads to a columnar phenotype with shorter internodes. Planta
231:251-263.
Flachowsky, H., A. Peil, J. Rollins, M.-V. Hanke, K. Richter, and D.-H. Lee. 2008. Improved
Fire Blight Resistance in Transgenic Apple Lines by Constitutive Overexpression of
the mbr4 Gene of Malus baccata. Acta Hortic 793:287-292.
Flachowsky, H., P.-M. Le Roux, A. Peil, A. Patocchi, K. Richter, and M.-V. Hanke. 2011.
Application of a high-speed breeding technology to apple (Malus × domestica) based
on transgenic early flowering plants and marker-assisted selection. New
Phytologist:DOI:10.1111/j.1469-8137.2011.03813.x.
Flaishman, M.A. 2009. Fast breeding system in pear by the use of juvenile free transgenic
plants COST Action 864 PomeFruitHealth, Working Groups 1-4 Valencia, Spain.
Forsline, P.L., and H.S. Aldwinckle. 2004. Evaluation of Malus sieversii Seedling
Populations for Disease Resistance and Horticultural Traits. Acta Hortic 663:529-534.
Forsline, P.L., H.S. Aldwinckle, and J.J. Luby. 2008. Fire Blight Incidence on Malus sieversii
Grown in New York and Minnesota. Acta Hortic 793:345-350.
Forsline, P.L., H.S. Aldwinckle, E.E. Dickson, J.J. Luby, and S.C. Hokanson. 2003.
Collection, Maintenance, Characterization, and Utilization of Wild Apples of Central
Asia. Horticultural Reviews 29:1-61.
Frey, J.E., B. Frey, C. Sauer, and M. Kellerhals. 2004. Efficient low-cost DNA extraction and
multiplex fluorescent PCR method for marker-assisted selection in breeding. Plant
Breeding 123:554-557.
Galli, P., A. Patocchi, G.A.L. Broggini, and C. Gessler. 2010. The Rvi15 (Vr2) Apple Scab
Resistance Locus Contains Three TIR-NBS-LRR Genes. Molecular Plant-Microbe
Interactions 23:608-617.
Galli, P., G.A.L. Broggini, M. Kellerhals, C. Gessler, and A. Patocchi. 2009. High-resolution
genetic map of the Rvi15 (Vr2) apple scab resistance locus. Molecular Breeding 26
561-572.
Gardiner, S., V. Bus, H. Bassett, M. Goodman, L. Greer, C. Ranatunga, E. Rikkerink, and R.
Forster. 1999. Identification of molecular markers linked to major resistances to apple
scab, powdery mildew and woolly apple aphid in apple Plant and Animal Genome VII
Conference, San Diego, California, USA.
Gardiner, S., J. Murdoch, S. Meech, R. Rusholme, H. Bassett, M. Cook, E. Rikkerink, A.
Gleave, R. Crowhurst, G. Ross, V. Bus, and I. Warrington. 2003. Candidate
Resistance Genes from an EST Database Prove a Rich Source of Markers for Major
Genes Conferring Resistance to Important Apple Pests and Diseases. Acta Hortic
622:141-151.
181
Gardiner, S.E., V.G.M. Bus, R.L. Rusholme, D. Chagné, and E.H.A. Rikkerink. 2007. Apple,
p. 1-62, In C. Kole, ed. Genome Mapping and Molecular Breeding in Plants Vol.
Fruits and Nuts. Springer-Verlag Berlin Heidelberg, Berlin, Germany.
Gardiner, S.E., H.C.M. Bassett, D.A.M. Noiton, V.G.M. Bus, M.E. Hofstee, A.G. White, R.D.
Ball, R.L.S. Forster, and E.H.A. Rikkerink. 1996. A detailed linkage map around an
apple scab resistance gene demonstrates that two disease resistance classes both carry
the Vf gene Theoretical and Applied Genetics 93:485-493.
Gardiner, S.E., V.G.M. Bus, D. Bowatte, H.C.M. Bassett, C. Ranatunga, D. Chagne, M.
Troggio, M. Pindo, R. Velasco, and A. Zarkikh. 2010. Application Of Apple Whole
Genome Sequence To Enhance Mapping Of Resistance Gene Loci Plant & Animal
Genomes XVIII Conference, San Diego, California, USA.
Gardner, R.G. 1976. Breeding apple rootstocks resistant to fire blight. PhD thesis, Cornell
University, Geneva, New York, USA.
Gardner, R.G., J.N. Cummins, and H.S. Aldwinckle. 1980. Fire blight resistance in the
Geneva apple rootstock breeding program. Journal of the American Society for
Horticultural Science 105:907-912.
Geider, K., G. Auling, V. Jakovljevic, and B. Völksch. 2009. A polyphasic approach assigns
the pathogenic Erwinia strains from diseased pear trees in Japan to Erwinia pyrifoliae.
Letters in Applied Microbiology 48:324-330.
Gessler, C., and A. Patocchi. 2007. Recombinant DNA Technology in Apple. Advances in
Biochemical Engineering and Biotechnology 107:113-132.
Gessler, C., A. Patocchi, S. Sansavini, S. Tartarini, and L. Gianfranceschi. 2006. Venturia
inaequalis Resistance in Apple. Critical Reviews in Plant Sciences 25:473-503.
Gianfranceschi, L., and V. Soglio. 2004. The European project HiDRAS: Innovative and
Multidisciplinary approach to breed high quality disease resistant apples. Acta Hortic
663:327-330.
Gianfranceschi, L., B. Koller, N. Seglias, M. Kellerhals, and C. Gessler. 1996. Molecular
selection in apple for resistance to scab caused by Venturia inaequalis. Theoretical
and Applied Genetics 93:199-204.
Gladieux, P., X.-G. Zhang, I. Roldán-Ruiz, V. Caffier, T. Leroy, M. Devaux, S. Van Glabeke,
E. Coart, and B. Le Cam. 2010. Evolution of the population structure of Venturia
inaequalis, the apple scab fungus, associated with the domestication of its host.
Molecular Ecology 19:658-674.
Goldschmidt, E.E., and A. Samach. 2004. Aspects of flowering in fruit trees. Acta Hortic
653:23-27.
Guilford, P., S. Prakash, J.M. Zhu, E. Rikkerink, S. Gardiner, H. Bassett, and R. Forster.
1997. Microsatellites in Malus × domestica (apple): abundance, polymorphism and
cultivar identification. Theoretical and Applied Genetics 94:249-254.
Gupta, P.K., S. Rustgi, and R.R. Mir. 2008. Array-based high-throughput DNA markers for
crop improvement. Heredity 101:5-18.
Hamblin, M.T., E.S. Buckler, and J.-L. Jannink. 2011. Population genetics of genomics-based
crop improvement methods. Trends in Genetics 27:98-106.
Hampson, C.R., and P.L. Sholberg. 2008. Estimating Combining Ability for Fire Blight
Resistance in Apple Progenies. Acta Hortic 793:337-344.
Han, Y., and S.S. Korban. 2010. Strategies for Map-Based Cloning in Apple. Critical
Reviews in Plant Sciences 29:265 - 284.
Han, Y., D. Zheng, S. Vimolmangkang, M.A. Khan, J.E. Beever, and S.S. Korban. 2011.
Integration of physical and genetic maps in apple confirms whole-genome and
segmental duplications in the apple genome. Journal of Experimental Botany
DOI:10.1093/jxb/err215.
182
Han, Y., D. Chagné, K. Gasic, E.H.A. Rikkerink, J.E. Beever, S.E. Gardiner, and S.S.
Korban. 2009. BAC-end sequence-based SNPs and Bin mapping for rapid integration
of physical and genetic maps in apple. Genomics 93:282-288.
Hancock, J.F., J.J. Luby, S.K. Brown, and G.A. Lobos. 2008. Apples, p. 1-38 In J. F.
Hancock, ed. Temperate Fruit Crop Breeding, Vol. 1. Springer Science+Business
Media B.V.
Hanke, M.-V., H. Flachowsky, A. Peil, and C. Hättasch. 2007. No Flower no Fruit - Genetic
Potentials to Trigger Flowering in Fruit Trees. Genes, Genomes and Genomics 1:1-20.
Hanke, V., W.S. Kim, and K. Geider. 2002. Plant transformation for induction of fire blight
resistance: transgenic apples expressing viral EPS depolymerase. Acta Hortic
590:393-395.
Hanke, V., I. Hiller, G. Klotzsche, K. Richter, J.L. Norelli, and H.S. Aldwinckle. 2000.
Transformation in apple for increased disease resistance. Acta Hortic 538:611-616.
Harris, S.A., J.P. Robinson, and B.E. Juniper. 2002. Genetic clues to the origin of the apple.
Trends in Genetics 18:426-430.
Hearnden, P., P. Eckermann, G. McMichael, M. Hayden, J. Eglinton, and K. Chalmers. 2007.
A genetic map of 1,000 SSR and DArT markers in a wide barley cross. Theoretical
and Applied Genetics 115:383-391.
Hemmat, M., N.F. Weeden, and S.K. Brown. 2003. Mapping and evaluation of Malus ×
domestica microsatellites in apple and pear. Journal of the American Society for
Horticultural Science 128:515-520.
Hemmat, M., N.F. Weeden, A.G. Manganaris, and D.M. Lawson. 1994. Molecular marker
linkage map for apple. Journal of Heredity 85:4-11.
Henry, R., and K. Edwards. 2009. New tools for single nucleotide polymorphism (SNP)
discovery and analysis accelerating plant biotechnology. Plant Biotechnology Journal
7:311-311.
Hippolyte, I., F. Bakry, M. Seguin, L. Gardes, R. Rivallan, A.-M. Risterucci, C. Jenny, X.
Perrier, F. Carreel, X. Argout, P. Piffanelli, I. Khan, R. Miller, G. Pappas, D.
Mbeguie-A-Mbeguie, T. Matsumoto, V. De Bernardinis, E. Huttner, A. Kilian, F.-C.
Baurens, A. D'Hont, F. Cote, B. Courtois, and J.-C. Glaszmann. 2010. A saturated
SSR/DArT linkage map of Musa acuminata addressing genome rearrangements
among bananas. BMC Plant Biology 10:65.
Hoenicka, H., O. Nowitzki, D. Hanelt, and M. Fladung. 2007. Heterologous overexpression of
the birch FRUITFULL-like MADS-box gene BpMADS4 prevents normal senescence
and winter dormancy in Populus tremula L. Planta 227:1001-1011.
Hokanson, S.C., A.K. Szewc-McFadden, W.F. Lamboy, and J.R. McFerson. 1998.
Microsatellite (SSR) markers reveal genetic identities, genetic diversity and
relationships in a Malus × domestica Borkh. core subset collection. Theoretical and
Applied Genetics 97:671-683.
Holliger, E., J. Vogelsanger, B. Schoch, B. Duffy, L. Lussi, and M. Bünter. 2008. Das
Feuerbrandjahr 2007. Schweizerische Zeitschrift für Obst- und Weinbau 27:9-12.
Hough, L.F. 1944. A survey of the scab resistance of the foliage on seedlings in selected apple
progenies. Proceedings of the American Society of Horticultural Science 4:260-272.
Huang, X., X. Wei, T. Sang, Q. Zhao, Q. Feng, Y. Zhao, C. Li, C. Zhu, T. Lu, Z. Zhang, M.
Li, D. Fan, Y. Guo, A. Wang, L. Wang, L. Deng, W. Li, Y. Lu, Q. Weng, K. Liu, T.
Huang, T. Zhou, Y. Jing, W. Li, Z. Lin, E.S. Buckler, Q. Qian, Q.-F. Zhang, J. Li, and
B. Han. 2010. Genome-wide association studies of 14 agronomic traits in rice
landraces. Nature Genetics 42:961-967.
183
Huaracha, E., M. Xu, and S.S. Korban. 2004. Narrowing down the region of the Vf locus for
scab resistance in apple using AFLP-derived SCARs. Theoretical and Applied
Genetics 108:274-279.
Ishimaru, C.A., E.J. Klos, and R.R. Brubaker. 1988. Multiple antibiotic production by
Erwinia herbicola. Phytopathology 78:746-750.
Jaccoud, D., K. Peng, D. Feinstein, and A. Kilian. 2001. Diversity arrays: a solid state
technology for sequence information independent genotyping. Nucleic Acids Research
29:e25.
Jacobsen, E., and H.J. Schouten. 2007. Cisgenesis strongly improves introgression breeding
and induced translocation breeding of plants. Trends in Biotechnology 25:219-223.
Janick, J. 2006. The PRI Apple Breeding Program. HortScience 41:8-10.
Janick, J., J.N. Cummins, S.K. Brown, and M. Hemmat. 1996. Apples, p. 1-77, In J. Janick
and J. N. Moore, eds. Fruit Breeding, Vol. 1. Wiley, New York, USA.
Jestin, C., M. Lode, P. Vallee, C. Domin, C. Falentin, R. Horvais, S. Coedel, M.J.
Manzanares-Dauleux, and R. Delourme. 2011. Association mapping of quantitative
resistance for Leptosphaeria maculans in oilseed rape (Brassica napus L.). Molecular
Breeding 27:271-287.
Jing, H.-C., C. Bayon, K. Kanyuka, S. Berry, P. Wenzl, E. Huttner, A. Kilian, and K. E
Hammond-Kosack. 2009. DArT markers: diversity analyses, genomes comparison,
mapping and integration with SSR markers in Triticum monococcum. BMC Genomics
10:458.
Jock, S., V. Donat, M.M. López, C. Bazzi, and K. Geider. 2002. Following spread of fire
blight in Western, Central and Southern Europe by molecular differentiation of
Erwinia amylovora strains with PFGE analysis. Environmental Microbiology 4:106-
114.
Johnson, K.B. 2010. Pathogen Refuge: A Key to Understanding Biological Control. Annual
Review of Phytopathology 48:141-160.
Johnson, K.B., and V.O. Stockwell. 1998. Management of fire blight: A case study in
microbial ecology. Annual Review of Phytopathology 36:227-248.
Johnson, K.B., V.O. Stockwell, and T.N. Temple. 2008. Evaluation of Kasugamycin as a
Component of an Integrated Biological and Chemical Strategy for Suppression of
Blossom Blight. Acta Hortic 793:439-444.
Johnson, W.C.B. 1999. The USDA-ARS/Cornell University Apple Rootstock Breeding and
Evaluation Program 42nd
Annual IDFTA Conference, Hamilton, Ontario, Canada.
Joshi, S.G. 2010. Towards durable resistance to apple scab using cisgenes. PhD Thesis,
Wageningen University, Wageningen, The Netherlands.
Jubault, M., C. Lariagon, M. Simon, R. Delourme, and M.J. Manzanares-Dauleux. 2008.
Identification of quantitative trait loci controlling partial clubroot resistance in new
mapping populations of Arabidopsis thaliana. Theoretical and Applied Genetics
117:191-202.
Juniper, B.E., and D.J. Mabberley. 2006. The Story of the Apple Timber Press, Portland,
Oregon, USA.
Kása, K., M. Hevesi, and M. Tóth. 2004. Evaluation of Traditional Hungarian Apple
Cultivars as Sources of Resistance to Fire Blight. Acta Hortic 663:225-228.
Kellerhals, K., A. Patocchi, B. Duffy, and J. Frey. 2008. Modern approaches for breeding
high quality apples with durable resistance to scab, powdery mildew and fire blight,
pp. 226-231, In M. Boos, (ed.) Ecofruit-13th
International Conference on Cultivation
Technique and Phytopathological Problems in Organic Fruit-Growing, Weinsberg,
Germany.
184
Kellerhals, M. 2009. Introduction to apple (Malus × domestica), p. 73-84, In K. M. Folta and
S. Gardiner, eds. Genetics and Genomics of Rosaceae Springer Science+Business
Media, New York, USA.
Kellerhals, M., and S. Egger. 2004. Inventory on Fruit Genetic Resources in Switzerland.
Acta Hortic 663:585-588.
Kellerhals, M., T. Székely, C. Sauer, J. Frey, and A. Patocchi. 2009a. Pyramidisieren von
Schorfresistenzen in der Apfelzüchtung. Erwerbs-Obstbau 51:21-28.
Kellerhals, M., M. Spuhler, B. Duffy, A. Patocchi, and J.E. Frey. 2009b. Selection Efficiency
in Apple Breeding. Acta Hortic 814:177-184.
Kellerhals, M., L. Franck, I. Baumgartner, A. Patocchi, and J. Frey. 2011. Breeding for Fire
Blight resistance in Apple. Acta Hortic 896:385-390.
Kenis, K., and J. Keulemans. 2005. Genetic linkage maps of two apple cultivars (Malus ×
domestica Borkh.) based on AFLP and microsatellite markers. Molecular Breeding
15:205-219.
Kenis, K., and J. Keulemans. 2007. Study of tree architecture of apple (Malus × domestica
Borkh.) by QTL analysis of growth traits. Molecular Breeding 19:193-208.
Kenis, K., J. Keulemans, and M.W. Davey. 2008. Identification and stability of QTLs for fruit
quality traits in apple. Tree Genetics & Genomes 4:647-661.
Khan, M., B. Duffy, C. Gessler, and A. Patocchi. 2006. QTL mapping of fire blight resistance
in apple. Molecular Breeding 17:299-306.
Khan, M.A. 2007. Genetic Basis of Fire Blight Resistance in Apples PhD thesis, ETHZ,
Zürich, Switzerland.
Khan, M.A., Y. Zhao, and S.S. Korban. 2011. Molecular Mechanisms of Pathogenesis and
Resistance to the Bacterial Pathogen Erwinia amylovora, Causal Agent of Fire Blight
Disease in Rosaceae. Plant Molecular Biology Reporter:DOI:10.1007/s11105-011-
0334-1.
Khan, M.A., C.-E. Durel, B. Duffy, D. Drouet, M. Kellerhals, C. Gessler, and A. Patocchi.
2007. Development of molecular markers linked to the ‘Fiesta’ linkage group 7 major
QTL for fire blight resistance and their application for marker-assisted selection.
Genome 50:568-577.
Kim, W.-S., L. Gardan, S.-L. Rhim, and K. Geider. 1999. Erwinia pyrifoliae sp. nov., a novel
pathogen that affects Asian pear trees (Pyrus pyrifolia Nakai). International Journal of
Systematic and Evolutionary Bacteriology 49:899-906.
Kim, W.-S., M. Hildebrand, S. Jock, and K. Geider. 2001. Molecular comparison of
pathogenic bacteria from pear trees in Japan and the fire blight pathogen Erwinia
amylovora. Microbiology 147:2951-2959.
King, G.J., F.H. Alston, I. Battle, E. Chevreau, C. Gessler, J. Janse, P. Lindhout, A.G.
Manganaris, S. Sansavini, H. Schmidt, and K. Tobutt. 1991. The ‘European Apple
Genome Mapping Project’-developing a strategy for mapping genes coding for
agronomic characters in tree species. Euphytica 56:89-94.
Kleinhempel, H., H. Kegler, W. Ficke, and H.-J. Schaefer. 1984. Methods of testing apples
for resistance to fire blight. Acta Hortic 151:261-265.
Knight, R.L., and F.H. Alston. 1968. Sources of field immunity to mildew (Podosphaera
leucotricha) in apple. Canadian Journal of Genetics and Cytology 10:294-298.
Ko, K., A. Grethel, M.T. Momol, J.L. Norelli, and H.S. Aldwinckle. 1998. Enhanced
resistance to fire blight (Erwinia amylovora) of Gala apple lines transgenic for lytic
protein genes, In A. Altman, et al., eds. Plant Biotechnology and In Vitro Biology for
the 21st Century, Vol. 36. Springer Science+Business Media, Jerusalem, Israel
185
Ko, K.S., J.L. Norelli, J.P. Reynoird, H.S. Aldwinckle, and S.K. Brown. 2002. T4 lysozyme
and attacin genes enhance resistance of transgenic ‘Galaxy’ apple against Erwinia
amylovora. Journal of the American Society for Horticultural Science 127:515-519.
Koller, B., L. Gianfranceschi, N. Seglias, J. McDermott, and C. Gessler. 1994. DNA markers
linked to Malus floribunda 821 scab resistance. Plant Molecular Biology 26:597-602.
Korba, J., J. Šillerová, and V. Kůdela. 2008. Resistance of apple varieties and selections to
Erwinia amylovora in the Czech Republic. Plant Protection Science 44:91-96.
Korban, S.S., and S. Tartarini. 2009. Apple Structural Genomics, In K. M. Folta and S. E.
Gardiner, eds. Genetics and Genomics of Rosaceae, Vol. 6. Springer
Science+Business Media, New York, USA.
Korban, S.S., S.M. Ries, M.J. Klopmeyer, J.F. Morrisey, and D.R. Hattermann. 1988.
Genotypic responses of scab-resistant apple cultivars/selections to two strains of
Erwinia amylovora and the inheritance of resistance to fire blight. Annals of Applied
Biology 113:101-105.
Kotoda, N., and M. Wada. 2005. MdTFL1, a TFL1-like gene of apple, retards the transition
from the vegetative to reproductive phase in transgenic Arabidopsis. Plant Science
168:95-104.
Kotoda, N., M. Wada, T. Masuda, and J. Soejima. 2003. The Break-through in the Reduction
of Juvenile Phase in Apple Using Transgenic Approaches. Acta Hortic 625:337-342.
Kotoda, N., H. Iwanami, S. Takahashi, and K. Abe. 2006. Antisense expression of MdTFL1, a
TFL1-like gene, reduces the juvenile phase in apple. Journal of the American Society
for Horticultural Science 131:74-81.
Kotoda, N., H. Hayashi, M. Suzuki, M. Igarashi, Y. Hatsuyama, S. Kidou, T. Igasaki, M.
Nishiguchi, K. Yano, T. Shimizu, S. Takahashi, H. Iwanami, S. Moriya, and K. Abe.
2010. Molecular Characterization of FLOWERING LOCUS T-Like Genes of Apple
(Malus × domestica Borkh.). Plant and Cell Physiology 51:561-575.
Kouassi, A., C.-E. Durel, F. Costa, S. Tartarini, E. van de Weg, K. Evans, F. Fernandez-
Fernandez, C. Govan, A. Boudichevskaja, F. Dunemann, A. Antofie, M. Lateur, M.
Stankiewicz-Kosyl, A. Soska, K. Tomala, M. Lewandowski, K. Rutkovski, E.
Zurawicz, W. Guerra, and F. Laurens. 2009. Estimation of genetic parameters and
prediction of breeding values for apple fruit-quality traits using pedigreed plant
material in Europe. Tree Genetics & Genomes 5:659-672.
Kube, M., A.M. Migdoll, I. Müller, H. Kuhl, A. Beck, R. Reinhardt, and K. Geider. 2008. The
genome of Erwinia tasmaniensis strain Et1/99, a non-pathogenic bacterium in the
genus Erwinia. Environmental Microbiology 10:2211-2222.
Kube, M., A.M. Migdoll, I. Gehring, K. Heitmann, Y. Mayer, H. Kuhl, F. Knaust, K. Geider,
and R. Reinhardt. 2010. Genome comparison of the epiphytic bacteria Erwinia
billingiae and E. tasmaniensis with the pear pathogen E. pyrifoliae. BMC Genomics
11:393.
Kumar, S., R. Volz, P. Alspach, and V. Bus. 2010. Development of a recurrent apple breeding
programme in New Zealand: a synthesis of results, and a proposed revised breeding
strategy. Euphytica 173:207-222.
Kumar, S., M. Bink, R. Volz, V. Bus, and D. Chagné. 2011. Towards genomic selection in
apple (Malus × domestica Borkh.) breeding programmes: Prospects, challenges and
strategies. Tree Genetics & Genomes:1-14.
Lalli, D., J. Norelli, M. Malnoy, M. Pindo, G. Fazio, P.L. Forsline, H. Aldwinckle, S.
Gardiner, D. Chagné, and M. Wisniewski. 2010. Determination of fire blight
resistance and construction of a molecular genetic map in a bi-parental Malus sieversii
population 12th
International Workshop on Fire Blight - Program and abstracts,
Warsaw, Poland.
186
Lateur, M., C. Wagemans, and C. Populer. 1998. Evaluation of fruit tree genetic resources as
sources of polygenic scab resistance in an apple breeding programme. Acta Hortic
484:35-42.
Laurens, F. 1999. Review of the current apple breeding programs in the world: objectives for
scion cultivar improvement. Acta Hortic 484:163–170
Laurens, F., C.-E. Durel, and M. Lascostes. 2004a. Molecular Characterization of French
Local Apple Cultivars Using SSRs. Acta Hortic 663:639-642.
Laurens, F., C.E. Durel, A. Patocchi, A. Peil, S. Salvi, S. Tartarini, R. Velasco, and E. van de
Weg. 2011. Review on apple genetics and breeding programmes and presentation of a
new European initiative to increase fruit breeding efficiency. Journal of fruit science
27:102-107.
Laurens, F., M. Chevalier, E. Dolega, F. Gennari, M. Goerre, C. Fischer, M. Kellerhals, M.
Lateur, B. Lefrancq, L. Parisi, H.J. Schouten, and S. Tartarini. 2004b. Local European
Cultivars as Sources of Durable Scab Resistance in Apple. Acta Hortic 663:115-121.
Le Clerc, V., A. Pawelec, C. Birolleau-Touchard, A. Suel, and M. Briard. 2009. Genetic
architecture of factors underlying partial resistance to Alternaria leaf blight in carrot.
Theoretical and Applied Genetics 118:1251-1259.
Le Lézec, M., P. Lecomte, F. Laurens, and J.-C. Michelesi. 1997. Sensibilité variétale au feu
bactérien (1re
partie). L' Arboriculture fruitière 503:57-62.
Le Roux, P.-M., M.A. Khan, G.A.L. Broggini, B. Buffy, C. Gessler, and A. Patocchi. 2010.
Mapping of quantitative trait loci for fire blight resistance in the apple cultivars
‘Florina’ and ‘Nova Easygro’. Genome 53:710-722.
Lespinasse, Y. 2009. Review of Pome Fruit Breeding in Europe: Which Strategies for the
Near Future? Acta Hortic 814:865-871.
Lespinasse, Y., and J.P. Paulin. 1984. Apple breeding programme for fireblight resistance.
Acta Hortic 151:301-306.
Lespinasse, Y., and J.-P. Paulin. 1990. Apple breeding programme for fire blight resistance:
strategy used and first results. Acta Hortic 273:285-295.
Lespinasse, Y., and C.-E. Durel. 1999. D.A.R.E. - Durable Apple Resistance in Europe: a new
European project on apple scab and powdery-mildew disease resistance based on a
close collaboration between geneticists, pathologists, breeders and nurserymen Plant
and Animal Genome VII Conf, San Diego, California, USA.
Lespinasse, Y., and H.S. Aldwinckle. 2000. Breeding for resistance to fire blight, p. 253-265,
In J. L. Vanneste, ed. Fire blight: the disease and its causative agent, Erwinia
amylovora. CAB International, Wallingford, UK.
Liebhard, R., and C. Gessler. 2000. Possible errors in genome mapping. IOBC wprs Bulletin
23:127-135.
Liebhard, R., B. Koller, L. Gianfranceschi, and C. Gessler. 2003a. Creating a saturated
reference map for the apple (Malus × domestica Borkh.) genome. Theoretical and
Applied Genetics 106:1497-1508.
Liebhard, R., M. Kellerhals, W. Pfammatter, M. Jermini, and C. Gessler. 2003b. Mapping
quantitative physiological traits in apple (Malus × domestica Borkh.). Plant Molecular
Biology 52:511-526.
Liebhard, R., B. Koller, A. Patocchi, M. Kellerhals, W. Pfammatter, M. Jermini, and C.
Gessler. 2003c. Mapping Quantitative Field Resistance Against Apple Scab in a
‘Fiesta’ × ‘Discovery’ Progeny. Phytopathology 93:493-501.
Liu, Y.G., Y. Chen, and Q. Zhang. 2005. Amplification of genomic sequences flanking T-
DNA insertions by thermal asymmetric interlaced polymerase chain reaction. Methods
in Molecular Biology 286:341-348.
187
Llop, P., M.M. López, J. Cabrefiga, L. Ruz, and E. Montesinos. 2008. Study of the virulence
in wild-type strains of Erwinia amylovora devoid of the plasmid pEa29. Acta Hortic.
793:145-148.
Llop, P., V. Donat, M. Rodríguez, J. Cabrefiga, L. Ruz, J.L. Palomo, E. Montesinos, and
M.M. López. 2006. An indigenous virulent strain of Erwinia amylovora lacking the
ubiquitous plasmid pEA29. Phytopathology 96:900-907.
Luby, J.J., P.A. Alspach, V.G.M. Bus, and N.C. Oraguzie. 2002. Field Resistance to Fire
Blight in a Diverse Apple (Malus sp.) Germplasm Collection. Journal of the American
Society for Horticultural Science 127:245-253.
Luby, J.J., K. Hokanson, P.L. Forsline, H.S. Aldwinckle, S. Gardiner, and V. Bus. 2006.
Evaluation of horticulturally elite Malus sieversii germplasm for apple scab resistance
genes using phenotypic and marker-based screenings 3rd
International Rosaceae
Genomics Conference, Napier, New Zealand.
Mackay, I., and W. Powell. 2007. Methods for linkage disequilibrium mapping in crops.
Trends in Plant Science 12:57-63.
Magness, J.R. 1937. Progress in apple improvement, p. 575-614 Yearbook U.S. Department
of Agriculture.
Maliepaard, C., F.H. Alston, G. van Arkel, L.M. Brown, E. Chevreau, F. Dunemann, K.M.
Evans, S. Gardiner, P. Guilford, A.W. van Heusden, J. Janse, F. Laurens, J.R. Lynn,
A.G. Manganaris, A.P.M. den Nijs, N. Periam, E. Rikkerink, P. Roche, C. Ryder, S.
Sansavini, H. Schmidt, S. Tartarini, J.J. Verhaegh, M. Vrielink-van Ginkel, and G.J.
King. 1998. Aligning male and female linkage maps of apple (Malus pumila Mill.)
using multi-allelic markers. Theoretical and Applied Genetics 97:60-73.
Malnoy, M., and H.S. Aldwinckle. 2007. Development of Fire Blight Resistance by
Recombinant DNA Technology. Plant Breeding Reviews 29:315-358.
Malnoy, M., Q. Jin, E.E. Borejsza-Wysocka, S.Y. He, and H.S. Aldwinckle. 2007.
Overexpression of the Apple MpNPR1 Gene Confers Increased Disease Resistance in
Malus × domestica. Molecular Plant-Microbe Interactions 20:1568-1580.
Malnoy, M., E.E. Borejsza-Wysocka, L. Pascal-Omenaca, H.S. Aldwinckle, C.S. Oh, and
S.V. Beer. 2008. Silencing of HIPM, the apple protein that interacts with HrpN of
Erwinia amylovora. Acta Hortic 793:261-264.
Mammadov, J., W. Chen, J. Mingus, S. Thompson, and S. Kumpatla. 2011. Development of
versatile gene-based SNP assays in maize (Zea mays L.). Molecular Breeding:DOI:
10.1007/s11032-011-9589-3.
Matsuda, N., K. Ikeda, M. Kurosaka, T. Takashina, K. Isuzugawa, T. Endo, and M. Omura.
2009. Early Flowering Phenotype in Transgenic Pears (Pyrus communis L.)
Expressing the CiFT Gene. Journal of the Japanese Society for Horticultural Science
78:410-416.
McDonald, B.A., and C. Linde. 2002. Pathogen population genetics, evolutionary potential,
and durable resistance. Annual Review of Phytopathology 40:349-379.
McGhee, G.C., and G.W. Sundin. 2011. Evaluation of Kasugamycin for Fire Blight
Management, Effect on Nontarget Bacteria, and Assessment of Kasugamycin
Resistance Potential in Erwinia amylovora. Phytopathology 101:192-204.
McGhee, G.C., E.L. Schnabel, K. Maxson-Stein, B. Jones, V.K. Stromberg, G.H. Lacy, and
A. Jones. 2002. Relatedness of chromosomal and plasmid DNAs of Erwinia pyrifoliae
and Erwinia amylovora. Applied Environmental Microbiology 68:6182-6192.
McGhee, G.C., J. Guasco, L.M. Bellomo, S.E. Blumer-Schuette, W.W. Shane, A. Irish-
Brown, and G.W. Sundin. 2011. Genetic Analysis of Streptomycin-Resistant (SmR)
Strains of Erwinia amylovora Suggests that Dissemination of Two Genotypes Is
188
Responsible for the Current Distribution of SmR E. amylovora in Michigan.
Phytopathology 101:182-191.
McManus, P.S., V.O. Stockwell, G.W. Sundin, and A.L. Jones. 2002. Antibiotic use in plant
agriculture. Annual Review of Phytopathology 40:443-465.
Meilan, R. 1997. Floral induction in woody angiosperms. New Forests 14:179-202.
Meuwissen, T.H.E., B.J. Hayes, and M.E. Goddard. 2001. Prediction of total genetic value
using genome-wide dense marker maps. Genetics 157:1819-1829.
Micheletti, D., M. Troggio, A. Zharkikh, F. Costa, M. Malnoy, R. Velasco, and S. Salvi.
2011. Genetic diversity of the genus Malus and implications for linkage mapping with
SNPs. Tree Genetics & Genomes 7:857-868.
Micheletti, D., M. Troggio, P. Baldi, F. Costa, M. Pindo, M. Komjanc, M. Malnoy, A.
Zarkikh, P. Magnago, R. Velasco, and S. Salvi. 2010. LD Estimation And Analysis Of
Diversity In Apple Plant & Animal Genomes XVIII Conference, San Diego,
California, USA.
Michelmore, R.W., I. Paran, and R.V. Kesseli. 1991. Identification of markers linked to
disease resistance genes by bulked segregant analysis: A rapid method to detect
markers in specific genomic regions by using segregating populations. Proceedings of
the National Academy of Sciences USA 88:9828-9832.
Miller, T.D., and M.N. Schroth. 1972. Monitoring the epiphytic populations of Erwinia
amylovora on pear with a selective medium. Phytopathology 62:1175-1182.
Mohammadi, M. 2010. Enhanced colonization and pathogenicity of Erwinia amylovora
strains transformed with the near-ubiquitous pEA29 plasmid on pear and apple. Plant
Pathology 59:252-261.
Momol, M.T., and H.S. Aldwinckle. 2000. Genetic diversity and host range of Erwinia
amylovora, p. 55-72, In J. L. Vanneste, ed. Fire blight: the disease and its causative
agent, Erwinia amylovora. CABI Publishing, Wallingford, UK.
Momol, M.T., H.S. Aldwinckle, P.L. Forsline, and W.F. Lamboy. 1999. Fire blight resistance
and horticultural evaluation of wild Malus populations from Central Asia. Acta Hortic
489:229-233.
Momol, M.T., J.L. Norelli, D.E. Piccioni, E.A. Momol, H.L. Gustafson, J.N. Cummins, and
H.S. Aldwinckle. 1998. Internal Movement of Erwinia amylovora Through
Symptomless Apple Scion Tissues into the Rootstock. Plant Disease 82:646-650.
Morgan, J., and A. Richards. 1993. The Book of Apples Ebury Press, London, UK.
Moriya, S., H. Iwanami, K. Okada, T. Yamamoto, and K. Abe. 2011. A practical method for
apple cultivar identification and parent-offspring analysis using simple sequence
repeat markers. Euphytica 177:135-150.
Moriya, S., H. Iwanami, S. Takahashi, N. Kotoda, K. Suzaki, T. Yamamoto, and K. Abe.
2010. Genetic mapping of the crown gall resistance gene of the wild apple Malus
sieboldii. Tree Genetics & Genomes 6:195-203.
N’Diaye, A., W. Van de Weg, L. Kodde, B. Koller, F. Dunemann, M. Thiermann, S.
Tartarini, F. Gennari, and C. Durel. 2008. Construction of an integrated consensus
map of the apple genome based on four mapping populations. Tree Genetics &
Genomes 4:727-743.
Nishitani, C., S. Terakami, Y. Sawamura, N. Takada, and T. Yamamoto. 2009. Development
of novel EST-SSR markers derived from Japanese pear (Pyrus pyrifolia). Breeding
Science 59:391-400.
Noiton, D., and C.J.A. Shelbourne. 1992. Quantitative genetics in an apple breeding strategy.
Euphytica 60:213-219.
Noiton, D., M. Hofstee, P. Alspach, L. Brewer, and C. Howard. 1998. Increasing genetic
diversity for apple breeding: A preliminary report. Acta Hortic 184:105-107.
189
Noiton, D.A.M., and P.A. Alspach. 1996. Founding clones, inbreeding, coancestry, and status
number of modern apple cultivars. Journal of the American Society for Horticultural
Science 121:773-782.
Norelli, J.L., and H.S. Aldwinckle. 1986. Differential susceptibility of Malus spp. cultivars
Robusta 5, Novole, and Ottawa 523 to Erwinia amylovora. Plant Disease 70:1017-
1019.
Norelli, J.L., A.L. Jones, and H.S. Aldwinckle. 2003. Fire blight management in the twenty-
first century, using new technologies that enhance host resistance in apple. Plant
Disease 87:756-765.
Norelli, J.L., H.S. Aldwinckle, S.V. Beer, and R.C. Lamb. 1987. The effects of virulence of
Erwinia amylovora on the evaluation of fire blight resistance in Malus.
Phytopathology 77:1551-1555.
Norelli, J.L., J.R.E. Farrell, C.L. Bassett, A.M. Baldo, D.A. Lalli, H.S. Aldwinckle, and M.E.
Wisniewski. 2009a. Rapid transcriptional response of apple to fire blight disease
revealed by cDNA suppression subtractive hybridization analysis. Tree Genetics &
Genomes 5:27-40.
Norelli, J.L., D.A. Lalli, C.L. Bassett, M.E. Wisniewski, S.E. Gardiner, J.M. Celton, D.R.
Bowatte, C.M. Carlisle, M. Malnoy, H.S. Aldwinckle, J. Farrell, R. E. , A.M. Baldo,
M.B. Horner, and V.G.M. Bus. 2009b. Using Functional Genomics to Identify
Molecular Markers for Fire Blight Resistance (Erwinia amylovora) in Apple (Malus).
Acta Hortic 839:415-420.
Oraguzie, N.C., and P.L. Wilcox. 2007. An Overview of Association Mapping, p. 1-9, In N.
C. Oraguzie, et al., eds. Association Mapping in Plants. Springer New York, USA.
Oraguzie, N.C., E.H.A. Rikkerink, S.E. Gardiner, H.N. Silva, J. Batley, and D. Edwards.
2007. SNP Applications in Plants, p. 95-102 Association Mapping in Plants. Springer
New York, USA.
Palmer, E.L., B.L. Teviotdale, and A.L. Jones. 1997. A relative of the broad-host-range
plasmid RSF1010 detected in Erwinia amylovora. Applied Environmental
Microbiology 63:4604-4607.
Parisi, L., Y. Lespinasse, J. Guillaumes, and J. Krüger. 1993. A new race of Venturia
inaequalis virulent to apples with resistance due to the Vf gene. Phytopathology
83:533-537.
Parisi, L., V. Fouillet, H. Schouten, R. Groenwold, F. Laurens, F. Didelot, K. Evans, C.
Fischer, F. Gennari, H. Kemp, M. Lateur, A. Patocchi, J. Thissen, and C. Tsipouridis.
2004. Variability of the Pathogenicity of Venturia inaequalis in Europe. Acta Hortic
663:107-114.
Park, D.H., S.P. Thapa, B.-S. Choi, W.-S. Kim, J.H. Hur, J.M. Cho, J.-S. Lim, I.-Y. Choi, and
C.K. Lim. 2011. Complete Genome Sequence of Japanese Erwinia Strain Ejp617, a
Bacterial Shoot Blight Pathogen of Pear. Journal of Bacteriology 193:586-587.
Park, T.H., V. Vleeshouwers, E. Jacobsen, E. Vossen, and R.G.F. Visser. 2009. Molecular
breeding for resistance to Phytophthora infestans (Mont.) de Bary in potato (Solanum
tuberosum L.): a perspective of cisgenesis. Plant Breeding 128:109-117.
Parlevliet, J.E. 2002. Durability of resistance against fungal, bacterial and viral pathogens:
present situation Euphytica 124:147-156.
Parravicini, G., C. Gessler, C. Denancé, P. Lasserre-Zuber, E. Vergne, M.-N. Brisset, A.
Patocchi, C.-E. Durel, and G.A.L. Broggini. 2011. Identification of serine/threonine
kinase and nucleotide-binding site-leucine-rich repeat (NBS-LRR) genes in the fire
blight resistance quantitative trait locus of apple cultivar ‘Evereste’. Molecular Plant
Pathology 12:493-505.
190
Parravicini Rusca, G. 2010. Candidate genes for fire blight resistance in apple cultivar
‘Evereste’, PhD thesis, ETHZ, Zürich, Switzerland.
Patocchi, A., A. Frei, J. Frey, and M. Kellerhals. 2009a. Towards improvement of marker
assisted selection of apple scab resistant cultivars: Venturia inaequalis virulence
surveys and standardization of molecular marker alleles associated with resistance
genes. Molecular Breeding 24:337-347.
Patocchi, A., B. Bigler, B. Koller, M. Kellerhals, and C. Gessler. 2004. Vr2: a new apple scab
resistance gene. Theoretical and Applied Genetics 109:1087-1092.
Patocchi, A., M. Walser, S. Tartarini, G.A.L. Broggini, F. Gennari, S. Sansavini, and C.
Gessler. 2005. Identification by genome scanning approach (GSA) of a microsatellite
tightly associated with the apple scab resistance gene Vm. Genome 48:630-636.
Patocchi, A., F. Fernández-Fernández, K. Evans, D. Gobbin, F. Rezzonico, A.
Boudichevskaia, F. Dunemann, M. Stankiewicz-Kosyl, F. Mathis-Jeanneteau, C.
Durel, L. Gianfranceschi, F. Costa, C. Toller, V. Cova, D. Mott, M. Komjanc, E.
Barbaro, L. Kodde, E. Rikkerink, C. Gessler, and W. van de Weg. 2009b.
Development and test of 21 multiplex PCRs composed of SSRs spanning most of the
apple genome. Tree Genetics & Genomes 5:211-223.
Paulin, J.-P. 2000. Erwinia amylovora: general characteristics, biochemistry and serology, p.
87-116, In J. L. Vanneste, ed. Fire blight: the disease and its causative agent, Erwinia
amylovora. CABI Publishing, Wallingford, UK.
Paulin, J.-P., and Y. Lespinasse. 1990. Pathogenicity of strains of Erwinia amylovora to some
apple cultivars in the greenhouse. Acta Hortic 273:319-326.
Peace, C., and J.L. Norelli. 2009. Genomics Approaches to Crop Improvement in the
Rosaceae, p. 19-53, In K. M. Folta and S. E. Gardiner, eds. Genetics and Genomics of
Rosaceae. Springer Science+Business Media, New York, USA.
Peil, A., K. Richter, M. Höfer, and V. Hanke. 2004. Beschreibung des Feuerbrandbefalls am
Institut für Obstzüchtung in Dresden-Pillnitz im Jahr 2003. Erwerbs-Obstbau 46:141-
148.
Peil, A., H. Flachoswky, M.-V. Hanke, K. Richter, and J. Rode. 2011. Inoculation of Malus ×
robusta 5 Progeny with a Strain Breaking Resistance to Fire Blight Reveals a Minor
QTL on LG5. Acta Hortic 896:357-362.
Peil, A., V.G.M. Bus, K. Geider, K. Richter, H. Flachowsky, and M.V. Hanke. 2009.
Improvement of Fire Blight Resistance in Apple and Pear. International Journal of
Plant Breeding 3:1-27
Peil, A., T. Garcia-Libreros, K. Richter, F.C. Trognitz, B. Trognitz, M.V. Hanke, and H.
Flachowsky. 2007. Strong evidence for a fire blight resistance gene of Malus robusta
located on linkage group 3. Plant Breeding 126:470-475.
Peil, A., K. Richter, T. Garcia-Libreros, V. Hanke, H. Flachowsky, J.-M. Celton, M. Horner,
S. Gardiner, and V. Bus. 2008. Confirmation of the Fire Blight QTL of Malus ×
robusta 5 on Linkage Group 3. Acta Hortic 793:297-303.
Penã, L., M. Martin-Trillo, J. Juarez, J.A. Pina, L. Navarro, and J.M. Martinez-Zapater. 2001.
Constitutive expression of Arabidopsis LEAFY or APETALA1 genes in citrus reduces
their generation time. Nature Biotechnology 19:263-267
Pereira-Lorenzo, S., A. Ramos-Cabrer, and M. Díaz-Hernández. 2007. Evaluation of genetic
identity and variation of local apple cultivars (Malus × domestica Borkh.) from Spain
using microsatellite markers. Genetic Resources and Crop Evolution 54:405-420.
Pereira-Lorenzo, S., A.M. Ramos-Cabrer, and M. Fischer. 2009. Breeding Apple (Malus ×
domestica Borkh.), p. 1-49, In S. M. Jain and P. M. Priyadarshan, eds. Breeding
Plantation Tree Crops: Temperate Species. Springer New York, USA.
191
Piepho, H.-P. 2000. Optimal marker density for interval mapping in a backcross population.
Heredity 84:437-440.
Pierantoni, L., K.H. Cho, I.S. Shin, R. Chiodini, S. Tartarini, L. Dondini, S.J. Kang, and S.
Sansavini. 2004. Characterisation and transferability of apple SSRs to two European
pear F1 populations. Theoretical and Applied Genetics 109:1519-1524.
Plomion, C., and C.-E. Durel. 1996. Estimation of the average effects of specific alleles
detected by the pseudo-testcross QTL mapping strategy. Genetics, Selection,
Evolution 28:223-235.
Poething, R.S. 1990. Phase change and the regulation of shoot morphogenesis in plants.
Science 250:923-930.
Poland, J.A., P.J. Balint-Kurti, R.J. Wisser, R.C. Pratt, and R.J. Nelson. 2009. Shades of gray:
the world of quantitative disease resistance. Trends in Plant Science 14:21-29.
Pontais, I., J.P. Paulin, M.N. Brisset, and D. Treutter. 2008. Are phenolic compounds
involved in the differential responses of apple genotypes to Erwinia amylovora? Acta
Hortic 793:247-248.
Potter, D., F. Gao, P.E. Bortiri, S.H. Oh, and S. Baggett. 2002. Phylogenetic relationships in
Rosaceae inferred from chloroplast matK and trnL-trnF nucleotide sequence data.
Plant Systematics and Evolution 231:77-89.
Potter, D., T. Eriksson, R.C. Evans, S. Oh, J.E.E. Smedmark, D.R. Morgan, M. Kerr, K.R.
Robertson, M. Arsenault, T.A. Dickinson, and C.S. Campbell. 2007. Phylogeny and
classification of Rosaceae. Plant systematics and evolution 266:5-43.
Powney, R., T.H.M. Smits, T. Sawbridge, B. Frey, J. Blom, J.E. Frey, K.M. Plummer, S.V.
Beer, J. Luck, B. Duffy, and B. Rodoni. 2011. Genome Sequence of an Erwinia
amylovora Strain with Pathogenicity Restricted to Rubus Plants. Journal of
Bacteriology 193:785-786.
Psallidas, P.B., and J. Tsiantos. 2000. Chemical control of fire blight, p. 99–234, In J. L.
Vanneste, ed. Fire Blight: the disease and its causative agent, Erwinia amylovora.
CABI Publishing, Wallingford, UK.
Pusey, P.L. 2000. The role of water in epiphytic colonization and infection of pomaceous
flowers by Erwinia amylovora. Phytopathology 90:1352-1357.
Pusey, P.L., V.O. Stockwell, and D.R. Rudell. 2008. Antibiosis and acidification by Pantoea
agglomerans strain E325 may contribute to suppression of Erwinia amylovora.
Phytopathology 98:1136-1143.
Reif, J., W. Liu, M. Gowda, H. Maurer, J. Möhring, S. Fischer, A. Schechert, and T.
Würschum. 2010. Genetic basis of agronomically important traits in sugar beet (Beta
vulgaris L.) investigated with joint linkage association mapping. Theoretical and
Applied Genetics 121:1489-1499.
Rezzonico, F., T.H.M. Smits, and B. Duffy. 2011. Diversity, Evolution and Functionality of
CRISPR Regions in the Fire Blight Pathogen Erwinia amylovora. Applied
Environmental Microbiology:DOI:10.1128/AEM.00177-11.
Rezzonico, F., T.H.M. Smits, E. Montesinos, J.E. Frey, and B. Duffy. 2009. Genotypic
comparison of Pantoea agglomerans plant and clinical strains. BMC Microbiology
9:204.
Richter, K., and C. Fischer. 2002. Stability of Fire Blight Resistance in Apple. Acta Hortic
590:381-384.
Rikkerink, E., N. Oraguzie, and S. Gardiner. 2007. Prospects of Association Mapping in
Perennial Horticultural Crops, p. 249-269, In N. C. Oraguzie, et al., eds. Association
Mapping in Plants. Springer, New York, USA.
192
Robinson, J.P., S.A. Harris, and B.E. Juniper. 2001. Taxonomy of the genus Malus Mill.
(Rosaceae) with emphasis on the cultivated apple, Malus domestica Borkh. Plant
systematics and evolution 226:35-58.
Rommens, C.M. 2004. All-native DNA transformation: a new approach to plant genetic
engineering. Trends in Plant Science 9:457-464.
Rothberg, J.M., and J.H. Leamon. 2008. The development and impact of 454 sequencing.
Nature Biotechnology 26:1117-1124.
Rowan, D.D., M.B. Hunt, A. Dimouro, P.A. Alspach, R. Weskett, R.K. Volz, S.E. Gardiner,
and D. Chagné. 2009. Profiling Fruit Volatiles in the Progeny of a ‘Royal Gala’ ×
‘Granny Smith’ Apple (Malus × domestica) Cross. Journal of Agricultural and Food
Chemistry 57:7953-7961.
Russo, N.L., T.J. Burr, D.I. Breth, and H.S. Aldwinckle. 2008. Isolation of streptomycin-
resistant isolates of Erwinia amylovora in New York. Plant Disease 92:714-718.
Sadamori, S., Y. Yoshida, H. Murakami, and S. Ishizuka. 1963. New apple variety ‘Fuji’
Bulletin of the Horticultural Research Station, Japan 1:1-6.
Sansaloni, C., C. Petroli, J. Carling, C. Hudson, D. Steane, A. Myburg, D. Grattapaglia, R.
Vaillancourt, and A. Kilian. 2010. A high-density Diversity Arrays Technology
(DArT) microarray for genome-wide genotyping in Eucalyptus. Plant Methods 6:1-11.
Sansavini, S., F. Donati, F. Costa, and S. Tartarini. 2004. Advances in apple breeding for
enhanced fruit quality and resistance to biotic stresses: new varieties for the European
market. Journal of Fruit and Ornamental Plant Research 12:13-52.
Schmidt, M. 1938. Venturia inaequalis, Cooke. Aderhold. VIII. Weitere Untersuchungen zur
Züchtung schorfwiderstandsfähiger Apfelsorten. Züchter 10:280-291.
Schouten, H., W. van de Weg, J. Carling, S. Khan, S. McKay, M. van Kaauwen, A.
Wittenberg, H. Koehorst-van Putten, Y. Noordijk, Z. Gao, D. Rees, M. Van Dyk, D.
Jaccoud, M. Considine, and A. Kilian. 2011. Diversity arrays technology (DArT)
markers in apple for genetic linkage maps. Molecular Breeding:DOI: 10.1007/s11032-
011-9579-5.
Schouten, H.J., F.A. Krens, and E. Jacobsen. 2006. Cisgenic plants are similar to traditionally
bred plants. EMBO reports 7:750-753.
Schuelke, M. 2000. An economic method for the fluorescent labeling of PCR fragments.
Nature Biotechnology 18:233-234.
Sebaihia, M., A.M. Bocsanczy, B.S. Biehl, M.A. Quail, N.T. Perna, J.D. Glasner, G.A.
DeClerck, S. Cartinhour, D.J. Schneider, S.D. Bentley, J. Parkhill, and S.V. Beer.
2010. Complete Genome Sequence of the Plant Pathogen Erwinia amylovora Strain
ATCC 49946. Journal of Bacteriology 192:2020-2021.
Segura, V., C. Denancé, C.-E. Durel, and E. Costes. 2007. Wide range QTL analysis for
complex architectural traits in a 1-year-old apple progeny. Genome 50:159-171.
Servin, B., O.C. Martin, M. Mézard, and F. Hospital. 2004. Toward a theory of marker-
assisted gene pyramiding. Genetics 168:513-523.
Shaw, L. 1934. Studies on resistance of apple and other rosaceous plants to fire blight. Journal
of Agricultural Research 49:283-313.
Shay, J.R., and L.F. Hough. 1952. Evaluation of Apple Scab Resistance in Selections of
Malus. American Journal of Botany 39:288-297.
Shendure, J., and H. Ji. 2008. Next-generation DNA sequencing. Nature Biotechnology
26:1135-1145.
Silfverberg-Dilworth, E., C.L. Matasci, W.E. Weg, M.P.W. Van Kaauwen, M. Walser, L.P.
Kodde, V. Soglio, L. Gianfranceschi, C.E. Durel, and F. Costa. 2006. Microsatellite
markers spanning the apple (Malus × domestica Borkh.) genome. Tree Genetics &
Genomes 2:202-224.
193
Smits, T.H.M., F. Rezzonico, and B. Duffy. 2010a. Evolutionary insights from Erwinia
amylovora genomics. Journal of Biotechnology 155:34-39.
Smits, T.H.M., S. Jaenicke, F. Rezzonico, T. Kamber, A. Goesmann, J.E. Frey, and B. Duffy.
2010b. Complete genome sequence of the fire blight pathogen Erwinia pyrifoliae
DSM 12163T and comparative genomic insights into plant pathogenicity. BMC
Genomics 11:2.
Smits, T.H.M., F. Rezzonico, T. Kamber, J. Blom, A. Goesmann, J.E. Frey, and B. Duffy.
2010c. Complete Genome Sequence of the Fire Blight Pathogen Erwinia amylovora
CFBP 1430 and Comparison to Other Erwinia spp. Molecular Plant-Microbe
Interactions 23:384-393.
Sobiczewski, P., E. Zurawicz, S. Berczyński, A. Mikiciński, and M. Lewandowski. 2008. The
importance of the type of Erwinia amylovora inoculum in screening of apple
genotypes susceptibility to fire blight. Journal of Fruit and Ornamental Plant Research
16:305-313.
Sobiczewski, P., A. Mikiciński, M. Lewandowski, M. Żurawicz, A. Peil, K. Richter, and M.
Kellerhals. 2011. Selection for Fire Blight Resistance of Apple Genotypes Originating
from European Genetic Resources and Breeding Programs. Acta Hortic 896:399-406.
Soejima, J., H. Bessho, S. Tsuchiya, S. Komori, K. Abe, and N. Kotoda. 1998. Breeding of
Fuji apples and performance on JM rootstocks. Compact Fruit Tree 31:22-24
Soufflet-Freslon, V. 2008. Résistance du pommier à la tavelure (Venturia inaequalis) :
Recherche de nouveaux loci et construction de génotypes "prototypes" en vue d’une
gestion durable de la résistance. PhD thesis, Université d'Angers, Angers, France.
Soufflet-Freslon, V., L. Gianfranceschi, A. Patocchi, and C.E. Durel. 2008. Inheritance
studies of apple scab resistance and identification of Rvi14, a new major gene that acts
together with other broad-spectrum QTL. Genome 51:657-667.
Steiner, P.W. 1990. Predicting Apple Blossom Infections by Erwinia amylovora Using the
Maryblyt Model. Acta Hortic. 273:139-148.
Steiner, P.W. 2000. Integrated orchard and nursery management for the control of fire blight,
p. 339, In J. L. Vanneste, ed. Fire blight: the disease and its causative agent, Erwinia
amylovora. CABI Publishing, Wallingford, UK.
Stich, B., H.-P. Piepho, B. Schulz, and A. Melchinger. 2008a. Multi-trait association mapping
in sugar beet (Beta vulgaris L.). Theoretical and Applied Genetics 117:947-954.
Stich, B., A. Melchinger, M. Heckenberger, J. Möhring, A. Schechert, and H.-P. Piepho.
2008b. Association mapping in multiple segregating populations of sugar beet (Beta
vulgaris L.). Theoretical and Applied Genetics 117:1167-1179.
Stockwell, V.O., T.N. Temple, K.B. Johnson, and J.E. Loper. 2008. Integrated Control of Fire
Blight with Antagonists and Oxytetracycline. Acta Hortic 783:383-390.
Stockwell, V.O., K.B. Johnson, D. Sugar, and J.E. Loper. 2010. Control of Fire Blight by
Pseudomonas fluorescens A506 and Pantoea vagans C9-1 Applied as Single Strains
and Mixed Inocula. Phytopathology 100:1330-1339.
Stockwell, V.O., K.B. Johnson, D. Sugar, and J.E. Loper. 2011. Mechanistically Compatible
Mixtures of Bacterial Antagonists Improve Biological Control of Fire Blight of Pear.
Phytopathology 101:113-123.
Stoeckli, S., K. Mody, A. Patocchi, M. Kellerhals, and S. Dorn. 2009. Rust mite resistance in
apple assessed by quantitative trait loci analysis. Tree Genetics & Genomes 5:257-
267.
Stoeckli, S., K. Mody, C. Gessler, A. Patocchi, M. Jermini, and S. Dorn. 2008. QTL analysis
for aphid resistance and growth traits in apple. Tree Genetics & Genomes 4:833-847.
194
Sule, S., W.S. Kim, and K. Geider. 2002. Transformation of SR1 tobacco and JTE-H apple
rootstock with the EPS depolymerase Gene from Erwinia amylovora phage. Acta
Hortic 590:407-409.
Sundin, G.W., N.A. Werner, K.S. Yoder, and H.S. Aldwinckle. 2009. Field Evaluation of
Biological Control of Fire Blight in the Eastern United States. Plant Disease 93:386-
394.
Szalatnay, D., R. Eder-Bauermeister, B. Duffy, and M. Kellerhals. 2009. Characterization of
Fruit Genetic Resources in Switzerland. Acta Hortic 814:143-148.
Szankowski, I., S. Waidmann, A. El-Din Saad Omar, H. Flachowsky, C. Hättasch, and M.-V.
Hanke. 2009. RNAi-Silencing of MdTFL1 Induces Early Flowering in Apple. Acta
Hortic 839:633-636.
Tan, M.Y.T., R.C.B. Hutten, R.G.F. Visser, and H.J. van Eck. 2010. The effect of pyramiding
Phytophthora infestans resistance genes RPi-mcd1 and RPi-ber in potato. Theoretical
and Applied Genetics 121:117-125.
Tanksley, S.D. 1993. Mapping polygenes. Annual Review of Genetics 27:205-233.
Tartarini, S., L. Gianfranceschi, S. Sansavini, and C. Gessler. 1999. Development of reliable
PCR markers for the selection of the Vf gene conferring scab resistance in apple. Plant
Breeding 118:183-186.
Taylor, R.K., C. Ranantunga, V. Bus, and P.A. Alspach. 2002. Host-Pathogen Interactions of
Erwinia amylovora on Apple and Pear in New Zealand. Acta Hortic 590:363-367.
Terakami, S., T. Kimura, C. Nishitani, Y. Sawamura, T. Saito, T. Hirabayashi, and T.
Yamamoto. 2009. Genetic Linkage Map of the Japanese Pear Housui Identifying
Three Homozygous Genomic Regions. Journal of the Japanese Society for
Horticultural Science 78:417-424.
Thomson, S.V. 2000. Epidemiology of fire blight, p. 9-36, In J. L. Vanneste, ed. Fire blight:
the disease and its causative agent, Erwinia amylovora. CABI Publishing,
Wallingford, UK.
Toth, M., K. Kasa, M. Gondor, K. Honty, and M. Hevesi. 2006. First results of fire blight
resistance screening in a Hungarian breeding programme. Acta Hortic 704:545-549.
Tränkner, C., S. Lehmann, H. Hoenicka, M.-V. Hanke, M. Fladung, D. Lenhardt, F.
Dunemann, A. Gau, K. Schlangen, M. Malnoy, and H. Flachowsky. 2010. Over-
expression of an FT homologous gene of apple induces early flowering in annual and
perennial plants. Planta 232:1309-1324.
Tränkner, C., S. Lehmann, H. Hoenicka, M.-V. Hanke, M. Fladung, D. Lenhardt, F.
Dunemann, A. Gau, K. Schlangen, M. Malnoy, and H. Flachowsky. 2011. Note added
in proof to: Over-expression of an FT-homologous gene of apple induces early
flowering in annual and perennial plants. Planta 233:217-218.
Turechek, W.W. 2004. Apple Diseases and their Management, p. 1-108, In S. A. M. H.
Naqvi, ed. Diseases of Fruits and Vegetables, Vol. I. Kluwer Academic Publishers,
The Netherlands.
Tyrka, M., P.T. Bednarek, A. Kilian, M. Wedzony, T. Hura, and E. Bauer. 2011. Genetic map
of triticale compiling DArT, SSR, and AFLP markers. Genome 54:391-401.
van de Weg, W.E., R.E. Voorrips, R. Finkers, L.P. Kodde, J. Jansen, and M.C.A.M. Bink.
2004. Pedigree Genotyping: A New Pedigree-based Approach of QTL Identification
and Allele Mining. Acta Hortic 663:45-50.
van der Zwet, T., and H.L. Keil. 1979. Fire blight: a bacterial disease of rosaceous plants
United States Department of Agriculture, Washington D.C., USA.
van Dyk, M., M. Soeker, I. Labuschagne, and D. Rees. 2010. Identification of a major QTL
for time of initial vegetative budbreak in apple (Malus × domestica Borkh.). Tree
Genetics & Genomes 6:489-502.
195
Van Ooijen, J.W. 1992. Accuracy of mapping quantitative trait loci in autogamous species.
Theoretical and Applied Genetics 84:803-811.
Van Ooijen, J.W. 1999. LOD significance thresholds for QTL analysis in experimental
populations of diploid species. Heredity 83:613-624.
Van Ooijen, J.W. 2004. MapQTL® 5, software for the mapping of quantitative trait loci in
experimental populations. Kyazma B.V., Wageningen, Netherlands.
Van Ooijen, J.W. 2006. JoinMap® 4, software for the calculation of genetic linkage maps in
experimental populations. Kyazma B.V., Wageningen, Netherlands.
Vanblaere, T., I. Szankowski, J. Schaart, H. Schouten, H. Flachowsky, G.A.L. Broggini, and
C. Gessler. 2011. The development of a cisgenic apple plant. Journal of Biotechnology
154:304-311.
Vanneste, J.L. 2000. What is fire blight? Who is Erwinia amylovora? How to control it?, p. 1-
8, In J. L. Vanneste, ed. Fire blight: the disease and its causative agent, Erwinia
amylovora. CABI Publishing, Wallingford, UK.
Vanneste, J.L. 2011. Biological Control Agents of Fire Blight: Successes and Challenges.
Acta Hortic 896:409-416.
Vanneste, J.L., and S. Eden-Green. 2000. Migration of Erwinia amylovora in host plant
tissues, p. 73-86, In J. L. Vanneste, ed. Fire blight: the disease and its causative agent,
Erwinia amylovora. CABI Publishing, Wallingford, UK.
Vavilov, N.I. 1951. The origin, variation, immunity and breeding of cultivated plants; selected
writings. Chronica Botanica 13:1-364.
Velasco, R. 2010. The apple genome: origin of the species and perspectives International
Congress Interpoma 2010, Bolzano, Italy.
Velasco, R., A. Zharkikh, J. Affourtit, A. Dhingra, A. Cestaro, A. Kalyanaraman, P. Fontana,
S.K. Bhatnagar, M. Troggio, D. Pruss, S. Salvi, M. Pindo, P. Baldi, S. Castelletti, M.
Cavaiuolo, G. Coppola, F. Costa, V. Cova, A. Dal Ri, V. Goremykin, M. Komjanc, S.
Longhi, P. Magnago, G. Malacarne, M. Malnoy, D. Micheletti, M. Moretto, M.
Perazzolli, A. Si-Ammour, S. Vezzulli, E. Zini, G. Eldredge, L.M. Fitzgerald, N.
Gutin, J. Lanchbury, T. Macalma, J.T. Mitchell, J. Reid, B. Wardell, C. Kodira, Z.
Chen, B. Desany, F. Niazi, M. Palmer, T. Koepke, D. Jiwan, S. Schaeffer, V.
Krishnan, C. Wu, V.T. Chu, S.T. King, J. Vick, Q. Tao, A. Mraz, A. Stormo, K.
Stormo, R. Bogden, D. Ederle, A. Stella, A. Vecchietti, M.M. Kater, S. Masiero, P.
Lasserre, Y. Lespinasse, A.C. Allan, V. Bus, D. Chagne, R.N. Crowhurst, A.P.
Gleave, E. Lavezzo, J.A. Fawcett, S. Proost, P. Rouze, L. Sterck, S. Toppo, B.
Lazzari, R.P. Hellens, C.-E. Durel, A. Gutin, R.E. Bumgarner, S.E. Gardiner, M.
Skolnick, M. Egholm, Y. Van de Peer, F. Salamini, and R. Viola. 2010. The genome
of the domesticated apple (Malus × domestica Borkh.). Nature Genetics 42:833-839.
Venisse, J.-S., M. Malnoy, M. Faize, J.-P. Paulin, and M.-N. Brisset. 2002. Modulation of
Defense Responses of Malus spp. During Compatible and Incompatible Interactions
with Erwinia amylovora. Molecular Plant-Microbe Interactions 15:1204-1212.
Vergne, E., X. Grand, E. Ballini, V. Chalvon, P. Saindrenan, D. Tharreau, J.L. Notteghem,
and J.B. Morel. 2010. Preformed expression of defense is a hallmark of partial
resistance to rice blast fungal pathogen Magnaporthe oryzae. BMC Plant Biology
10:206.
Vinatzer, B.A., A. Patocchi, S. Tartarini, L. Gianfranceschi, S. Sansavini, and C. Gessler.
2004. Isolation of two microsatellite markers from BAC clones of the Vf scab
resistance region and molecular characterization of scab-resistant accessions in Malus
germplasm. Plant Breeding 123:321-326.
Visser, T. 1964. Juvenile phase and growth of apple and pear seedlings. Euphytica 13:119-
129.
196
Volk, G.M., C.M. Richards, A.A. Reilley, A.D. Henk, and P.A. Reeves. 2008. Genetic
Diversity and Disease Resistance of Wild Malus orientalis from Turkey and Southern
Russia. Journal of the American Society for Horticultural Science 133:383-389.
Volz, R.K., E. Rikkerink, P. Austin, T. Lawrence, and V.G.M. Bus. 2009. “Fast-Breeding” in
Apple: a Strategy to Accelerate Introgression of New Traits into Elite Germplasm.
Acta Hortic 814:163-168.
Voorrips, R.E. 2002. MapChart: Software for the graphical presentation of linkage maps and
QTLs. The Journal of Heredity 93:77-78.
Wang, A., H. Aldwinckle, P. Forsline, D. Main, G. Fazio, S. Brown, and K. Xu. 2011. EST
contig-based SSR linkage maps for Malus × domestica cv ‘Royal Gala’ and an apple
scab resistant accession of M. sieversii, the progenitor species of domestic apple.
Molecular Breeding:DOI: 10.1007/s11032-011-9554-1.
Wang, D., S.S. Korban, and Y. Zhao. 2010. Molecular Signature of Differential Virulence in
Natural Isolates of Erwinia amylovora. Phytopathology 100:192-198.
Wenzl, P., J. Carling, D. Kudrna, D. Jaccoud, E. Huttner, A. Kleinhofs, and A. Kilian. 2004.
Diversity Arrays Technology (DArT) for whole-genome profiling of barley.
Proceedings of the National Academy of Sciences of the United States of America
101:9915-9920.
Wisser, R.J., P.J. Balint-Kurti, and R.J. Nelson. 2006. The genetic architecture of disease
resistance in maize: a synthesis of published studies. Phytopathology 96:120-129.
Wittenberg, A.H.J. 2007. Genetic mapping using the Diversity Arrays Technology (DArT) -
Application and validation using the whole-genome sequences of Arabidopsis
thaliana and the fungal wheat pathogen Mycosphaerella graminicola. PhD thesis,
Wageningen University, Wageningen, The Netherlands.
Xu, M.L., and S.S. Korban. 2000. Saturation mapping of the apple scab resistance gene Vf
using AFLP markers. Theoretical and Applied Genetics 101:844-851.
Xu, M.L., E. Huaracha, and S.S. Korban. 2001. Development of sequence-characterized
amplified regions (SCARs) from amplified fragment length polymorphism (AFLP)
markers tightly linked to the Vf gene in apple. Genome 44:63-70.
Yamamoto, T., T. Kimura, J. Soejima, T. Sanada, Y. Ban, and T. Hayashi. 2004a.
Identification of quince varieties using SSR markers developed from pear and apple.
Breeding Science 54:239 - 244.
Yamamoto, T., T. Kimura, S. Terakami, C. Nishitani, Y. Sawamura, T. Saito, K. Kotobuki,
and T. Hayashi. 2007. Integrated Reference Genetic Linkage Maps of Pear Based on
SSR and AFLP Markers. Breeding Science 57:321-329.
Yamamoto, T., T. Kimura, T. Saito, K. Kotobuki, N. Matsuta, R. Liebhard, C. Gessler, E.
Weg, and T. Hayashi. 2004b. Genetic linkage maps of Japanese and European pears
aligned to the apple consensus map. Acta Hortic 663:51-56.
Yu, J., J.B. Holland, M.D. McMullen, and E.S. Buckler. 2008. Genetic design and statistical
power of nested association mapping in maize. Genetics 178:539-551.
Zhang, Y., and K. Geider. 1997. Differentiation of Erwinia amylovora strains by pulsed-field
gel electrophoresis. Applied Environmental Microbiology 63:4421-4426.
Zhou, J.M., X.Y. Tang, and G.B. Martin. 1997. The Pto kinase conferring resistance to
tomato bacterial speck disease interacts with proteins that bind a cis-element of
pathogenesis-related genes. EMBO Journal 16:3207-3218.
Zhu, Y., and B.H. Barritt. 2008. Md-ACS1 and Md-ACO1 genotyping of apple (Malus ×
domestica Borkh.) breeding parents and suitability for marker-assisted selection. Tree
Genetics & Genomes 4:555-562.
Zimmerman, R.H. 1973. Juvenility and flowering of fruit trees. Acta Hortic. 34:139-142.
197
Zini, E., F. Biasioli, F. Gasperi, D. Mott, E. Aprea, T. Märk, A. Patocchi, and C. Gessler.
2005. QTL mapping of volatile compounds in ripe apples detected by proton transfer
reaction-mass spectrometry. Euphytica 145:269-279.
198
199
Acknowledgements
200
This PhD thesis was achieved at the Swiss Federal Institute of Technology Zürich (ETHZ,
Switzerland) in close collaboration with the Federal Research Station Agroscope Changins
Wädenswil ACW (Wädenswil, Switzerland) in the framework of the ZUEFOS project
(Züchtung feuerbrandtoleranter Obstsorten) funded by the Federal Office for Agriculture
(FOAG, Switzerland).
First of all, I express my greatest gratitude to Dr. Andrea Patocchi (Phytopathology,
Agroscope ACW, Wädenswil), for his supervision, advice and help during the last 4 years.
Despite his various projects and duties, he always found time to discuss constructively the
experimental designs, the results and the manuscripts and also to help more practically during
the experiments when needed. I deeply appreciated his ability to find a good balance between
guiding students and letting them work independently. I am also grateful to him for the nice
atmosphere and the dynamism he instilled into his research group. Merci Andrea de m’avoir
offert la possibilité de faire ce doctorat et de m’avoir tant appris; travailler avec toi au cours
des 4 années passées a été un vrai plaisir et, je pense, un privilège. J’espère que bien d’autres
doctorant(e)s, étudiant(e)s, stagiaires ou collaborateurs/-trices auront la même chance que moi
de bénéficier de ton encadrement dans les années à venir.
I would like to thank Prof. Cesare Gessler (Plant Pathology, ETH Zürich) for the opportunity
he gave me to make my PhD thesis at the ETHZ in his research group “Perennial Crops”
within the Plant Pathology group. I benefited a lot from our discussions and I really
appreciated his interest in my work and his willingness to help and solve problems. Un grand
merci Cesare pour ta disponibilité malgré tes multiples responsabilités, ainsi que pour ton
enthousiasme à discuter de mes projets; mon expérience à l’ETHZ aura été très
complémentaire de mon travail à Agroscope ACW.
I gratefully acknowledge Prof. Bruce McDonald (Plant Pathology, ETH Zürich) for accepting
to be a co-examiner at my PhD exam, as well as for the discussions we had on my work and
on research in general.
I gratefully acknowledge Prof. Magda-Viola Hanke for accepting to be a co-examiner at my
PhD exam. Moreover, I am grateful to her and also to her colleagues Drs. Henryk Flachowsky
and Andreas Peil (Julius Kühn Institute, Dresden, Germany), for the very nice collaboration
we had on the early flowering and the ‘Idared’ × ‘Rewena’ projects. I deeply appreciated their
201
willingness to share results as well as their enthusiasm to carry out these projects, no matter
the difficulties.
I am grateful to Dr. Markus Kellerhals, Group Leader “Apple breeding and genetic resources”
(Agroscope ACW, Wädenswil), for having very efficiently managed the ZUEFOS Project he
was in charge of. I regularly benefited from his precious advice and from our discussions,
whether on my work or on plant breeding in general. Merci encore Markus pour l’aide que tu
m’as apportée et pour l’intérêt que tu as manifesté à mon travail au cours de ces années
passées à l’ACW. Merci aussi de m’avoir tant facilité les choses par la qualité de ton français
et ta gentillesse.
I am grateful to Eduard Holliger, Group leader “Phytopathology” (Agroscope ACW,
Wädenswil). He was always present to solve administrative or technical issues concerning
different aspects of my work. I also greatly appreciated his efficient management of the
“Phytopathology” group. Merci beaucoup Edi pour ta disponibilité et ton aide constante.
I would like to thank Dr. Giovanni Broggini (Plant Pathology, ETH Zürich) for its scientific
guidance on linkage mapping with AFLP markers and for all the inspiring discussions we had
on various aspects of my work. Un grand merci Giovanni pour ton enthousiasme, ta
disponibilité et tes précieux conseils en cartographie.
I would like to acknowledge Dr. Muhammad Awais Khan for his kind and patient guidance
when I took over the work on ‘Florina’ × ‘Nova Easygro’ after him. I learned a lot from our
discussions on QTL analysis and quantitative genetics and the Chapter 2 was significantly
improved thanks to him.
I gratefully acknowledge Dr. Brion Duffy (Phytopathology, Agroscope ACW, Wädenswil)
who was of precious help, especially during my first months at the Agroscope ACW. Many
thanks Brion for the scientific discussions as well as the practical help with the fire blight
screening tests in quarantine greenhouse.
I express my gratitude to Dr. Danilo Christen, Group Leader “Fruit growing” (Agroscope
ACW, Conthey), who gave me the opportunity to do my Master Thesis at the Agroscope
ACW. Merci beaucoup Danilo pour m’avoir accueilli en Suisse et « lancé » vers la recherche
202
en doctorat; le projet sur les poiriers nous a réservé des surprises mais cela en valait largement
la peine; merci aussi pour ta disponibilité et ton dynamisme. I am also grateful to Drs Stefano
Tartarini and Luca Dondini (University of Bologna, Bologna, Italy) for their input in the
project “fire blight resistance QTLs in ‘Harrow Sweet’”, and also for hosting me so kindly
during fruitfull COST 864 STSMs. I also acknowledge Aviad Freiman and Dr. Moshe
Flaishman (Agricultural Research Organization, Bet-Dagan, Israel) for giving me the
opportunity to contribute to their project “Early pear (Pyrus communis L.) flowering by RNAi
silencing of MdTFL1”.
I am grateful to Dr. Juerg Frey, Group Leader “Molecular diagnostics and epidemiology”
(Agroscope ACW, Wädenswil), for his advice on technical and scientific topics related to
MAS, high-throughput genotyping and microsatellites. I also express my gratitude for
equipment loans during the course of my PhD, without which nothing would have been
possible.
Many thanks are due of course to the “Mycology”, “Bacteriology” and “Molecular
Diagnostic” Teams of the Agroscope ACW (Wädenswil) for having introduced me to the
richness of their work. A particular “thank you” to all people working on Erwinia amylovora
for having taught me some knowledge on this pathogen. I am grateful to Dr. Theo Smits for
his constructive advice and interesting discussions on research in general and microbiology
and genomics in particular. I am also grateful to Dr. Joël Pothier for all the discussions we
had on various topics during almost 3 years at the Agroscope ACW. Un grand merci Joël pour
les pauses qui « reboostent », mais aussi pour ton sens de la pédagogie et de la persévérance
dans la recherche que tu sais transmettre. Merci d’avoir ainsi contribué à mon apprentissage
de la démarche scientifique. I express my greatest gratitude to Dr. Cosima Pelludat, Bea Frey,
Maja Hilber-Bodmer and Markus Oggenfuss (Agroscope ACW, Wädenswil) for their
technical assistance, their kindness as well as their excellent lab management that was of great
help throughout the course of my PhD. I am also grateful to Bea Schoch for her help with the
preparation of bacterial inocula and for her very efficient management of the THX lab.
Many thanks are due to all the past and present members of the “B10 room” and Labor 4 at
Agroscope ACW: Kerstin Brankatschk, Melanie Jänsch, Didier Socquet-Juglard, Tim
Kamber, Juliane Weger, Andreas Bühlman, Andy Lehman, Verena Knorst, Jan Candreia,
Nicola Schley, but also Drs. Fabio Rezzonico, Martijn Holterman, Andrea Braun-Kiewnick,
Fiona Walsh, Frédérique Pasquer and Yvonne Möller-Steinbach for their support, enthusiasm
203
and for the very pleasant atmosphere during all this time, at Agroscope ACW as well as
outside.
I also gratefully acknowledge the past and present members of the Plant Pathology group at
the ETHZ for the nice and friendly atmosphere, especially during the first year of my PhD
thesis. I especially deeply appreciated the very good group feeling within the Perennial crops
group thanks to Dr. Paolo Galli, Dr. Caterina Matasci, Fabienne di Gennaro, Thalia
Vanblaere, Michele Gusberti and all the others. Merci à Marcello Zala dont la vitesse des
commandes et la bonne humeur sont devenue proverbiales. Many thanks are also due to
Ulrike Rosenberger, Drs. Ueli Merz and Patrick Brunner for their kindness and help to solve
administrative and technical problems whenever needed.
I would like to thank in particular Dr. Gabriella Parravicini Rusca, Johannes Fahrentrapp
(ETH Zürich) as well as Isabelle Baumgartner and Gabriella Silvestri (Agroscope ACW), my
colleagues on the research topic « QTL-apple-fire blight ». Many thanks to all them for their
kindness, good temper and efficient collaboration, Gabriella P. and Johannes on the molecular
aspects, Isabelle and Gabriella S. in the planning and carrying out of fire blight phenotypic
tests. I acknowledge also Carolin Schwer (Agroscope ACW) for her work on genetic mapping
in apple. Thanks to Michael Gasser and Franz Krebs (Agroscope ACW) for their support
during the fire blight tests in quarantine glasshouse. Special thanks go to Verena Knorst for
her precious contribution in the SSR markers genotyping in the Chapter 3 of this thesis.
I express my gratitude to Urs Moser, Daniel Sager, André Imboden, Sabine Klarer for their
precious help in greenhouse maintenance in Eschikon. Special thanks are due to Maja Frei for
having taken care of my trees during several months. I am also grateful to Juergen Krauss,
Reto Leumann, René Total, Lucie Franck and the personal in charge of the maintenance of the
greenhouses at the Agroscope ACW in Wädenswil for their efficient and kind collaboration.
Solving technical problems was much easier with the help of all these people. Ich möchte
Rolf Blapp besonders bedanken für seine Gastfreundschaft, sein Einsatz und Kenntnisse mit
Pflanzen im Gewächshaus, und insbesondere für seine groβartige grüne Daumen.
I gratefully acknowledge the skillfull support from the Informatik Team at the Agroscope
ACW for softwares installation, informatic tips and technical support, in particular from
Walter Riesen, Martin Kast and Toni Zürcher. I am also grateful to Dr. Benno Graf,
204
Department Head “Crop Protection and Extension Fruit and Vegetable crops”, Anneliese
Britschgi and Anita Rahm-Gasser (Agroscope ACW) for their efficiency to deal with
administrative issues. Markus Bünter, Group Leader “Phytosanitary inspectorate” at
Agroscope ACW, is gratefully acknowledged for his important advice and support for
graftwoods imports and biosafety issues.
Last but not least, I would like to thank my parents, my brother and my family for their
continuous and untiring support during all my studies, without which I would not have
achieved this PhD.
205
Curriculum Vitae
Pierre-Marie Le Roux
Citizen of France
Place of birth: Sablé-sur-Sarthe (72, France)
Date of birth: 11.10.1982
Education
May 2008-October 2011: PhD student at the Swiss Federal Institute of Technology,
Zürich (ETHZ), in collaboration with the Federal Research Station Agroscope Changins-
Wädenswil ACW (Wädenswil, Switzerland).
Topic: Molecular breeding for fire blight resistance in apple (Malus spp.).
Supervisors: Dr. Andrea Patocchi (ACW), Prof. Cesare Gessler (ETHZ).
2007: Diploma thesis at the Federal Research Station Agroscope Changins
Wädenswil ACW (Wädenswil and Conthey, Switzerland).
Topic: Identification of microsatellite markers tightly linked to quantitative trait loci for fire
blight resistance in pear (Pyrus communis L.).
Supervisors: Dr. Andrea Patocchi (ACW), Dr. Danilo Christen (ACW).
2007: Graduate engineer in Agronomy, major Plant Science and Production, Ecole
Nationale Supérieure d’Agronomie, Agrocampus Rennes (France).
2004: Bachelor of Science, major Cellular Biology and Physiology, Université de
Bretagne Occidentale, Brest (France).
2000-2003: Competitive French “Classes Préparatoires” Biologie, Chimie, Physique et
Sciences de la Terre, Lycée Chateaubriand, Rennes (France).
2000: Baccalauréat S (equivalent A-level) with scientific focus.
Publications
2011:
Flachowsky H, Le Roux P-M, Peil A, Patocchi A, Richter K, Hanke M-V. Application of a
high-speed breeding technology to apple (Malus × domestica) based on transgenic early
flowering plants and marker-assisted selection. Published online in the journal New
Phytologist: 8 July 2011.
206
Freiman A, Shlizerman L, Golubowicz S, Yabloviz Z, Korchinsky R, Cohen Y, Samach A,
Chevreau E, Le Roux P-M, Patocchi A, Flaishman M A. Early pear (Pyrus communis L.)
flowering by RNAi silencing of MdTFL1: a platform for fast and compact breeding. Published
online in the journal Planta: 28 December 2011.
Le Roux P-M F, Flachowsky H, Hanke M-V, Gessler C and Patocchi A. Use of a transgenic
early flowering approach in apple (Malus × domestica Borkh.) to introgress the major
quantitative trait locus for fire blight resistance of the apple genotype ‘Evereste’. Published
online in the journal Molecular Breeding: 22 November 2011.
Le Roux P-M F, Christen D, Duffy B, Tartarini S, Dondini L, Yamamoto T, Nishitani C,
Terakami S, Lespinasse Y, Kellerhals M, Patocchi A. Redefinition of the map position and
validation of a major quantitative trait locus for fire blight resistance of the pear cultivar
‘Harrow Sweet’ (Pyrus communis L.). Submitted to the journal Plant Breeding.
2010:
Le Roux P-M F, Khan, M A, Broggini G A L, Duffy B, Gessler C, Patocchi A. Mapping of
quantitative trait loci for fire blight resistance in the apple cultivars ‘Florina’ and ‘Nova
Easygro’. Published in the journal Genome, 53: 710-722, 2010.
Oral communications
2011:
Oral presentation: Application of a fast breeding approach in apple (Malus × domestica
Borkh.). Thirteenth Eucarpia Symposium on Fruit Breeding and Genetics, September 11-15,
Warsaw, Poland.
Poster presentation: QTL analysis of fire blight resistance in an apple F1 progeny ‘Idared’ ×
‘Rewena’. Thirteenth Eucarpia Symposium on Fruit Breeding and Genetics, September 11-15,
Warsaw, Poland.
Oral presentation: Application of a fast breeding approach in apple. COST Action 864
PomeFruitHealth Final Meeting, February 1-3, Diepenbeek, Belgium.
2010:
Poster presentation: Introgression of a major QTL for fire blight resistance using early
flowering transgenic apple trees. 12th
International Fire Blight Workshop, August 16-20,
Warsaw, Poland.
Posters presentation: Amélioration classique du pommier: les défis and Plantes cisgéniques.
Portes ouvertes d’Agroscope Changins Wädenswil ACW, June 18-20, Changins (VD),
Switzerland.
Oral presentation: Experience on the early flowering approach at Agroscope ACW. COST
Action 864 PomeFruitHealth Meeting, Working Group 4, February 11-12, Bennekom, The
Netherlands.
207
2009:
Poster presentation: QTL mapping of fire blight resistance in the apple cultivars ‘Florina’ and
‘Nova Easygro’ Symposium of the Zurich-Basel Plant Science Center “Plant-Microbe
Interactions”, November 13, Basel, Switzerland.
Short-Term Scientific Missions
2008:
Short-Term Scientific Mission COST-STSM-864-04062: Verification of the positions and
effects of three QTLs for fire blight resistance in the cultivar ‘Harrow Sweet’ (Pyrus
communis L.) Host: Dr. Stefano Tartarini, Dipartimento Colture Arboree, Università degli
Studi di Bologna, October 27 to November 1, Italy.