ORIGINAL ARTICLE

Traud Winkelmann Æ Dimitri Heintz

Alain Van Dorsselaer Æ Margrethe Serek

Hans-Peter Braun

Proteomic analyses of somatic and zygotic embryosof Cyclamen persicum Mill. reveal new insights into seedand germination physiology

Received: 1 December 2005 / Accepted: 24 January 2006 / Published online: 22 February 2006� Springer-Verlag 2006

Abstract In the horticulturally important ornamentalspecies Cyclamen persicum Mill., somatic embryogenesisis an efficient vegetative propagation method and thedevelopment of artificial seeds is an ultimate aim. Thisstudy aims at a systematic comparison of the proteomesof zygotic embryos, somatic embryos grown in liquidmedium containing 30 or 60 g l�1 sucrose, germinatingembryos of both types and endosperm in order to obtainnovel insights into seed and germination physiology.Using high resolution two-dimensional isoelectricfocussing/sodium dodecylsulfate polyacrylamide gelelectrophoresis (2D IEF/SDS PAGE), 74% of the pro-teins expressed in zygotic embryos were found in similarabundance in somatic embryos grown in 60 g l�1 su-crose. Somatic embryos grown in 30 g l�1 sucroseaccumulated fewer protein species than those grown in60 g l�1. Selected proteins were identified followingmass spectrometry (nano-LC-MS/MS). Four enzymesinvolved in glycolysis (UDP-glucose pyrophosphorylase,fructose bisphosphate aldolase, triosephosphate isom-erase and glyceraldehyde-3-phosphate dehydrogenaseGAPDH) were specifically induced in somatic embryos.11S globulin proteins identified by MS were present inhigh levels in somatic embryos, zygotic embryos and

endosperm, whereas 7S globulins were detected mainlyin endosperm and zygotic embryos. These are the firststorage proteins identified in C. persicum. Xyloglucansare known to be another group of seed storage com-pounds in C. persicum. Interestingly, xyloglucan endo-transglycosylases were found to be highly expressed inendosperm tissue. We discuss the physiological impli-cations of these observations.

Keywords Endosperm Æ Mass spectrometry ÆOrnamental plant Æ Storage proteins Æ Two-dimensionalpolyacrylamide gel electrophoresis Æ Xyloglucanendotransglycosylase

Abbreviations ACN: Acetonitril ÆDTT: Dithiothreitol Æ end: Endosperm Æ GAPDH:glyceraldehyde-3-phosphate dehydrogenase Æ MS: Massspectrometry Æ se: Somatic embryos Æ SOD: superoxidedismutase Æ ze: Zygotic embryos Æ 2,4-D: 2,4-dichlorophenoxyacetic acid Æ 2iP: 6-(c,c-dimethylallylamino) purine

Introduction

Somatic embryogenesis has been shown to be an efficientvegetative propagation pathway in vitro. It has beeninvestigated in detail in some model species like carrot(Daucus carota) and alfalfa (Medicago sativa) aiming todescribe the embryogenesis pathway in terms of mor-phological events and also of gene expression (forreviews see Zimmerman 1993; Mordhorst et al. 1997).Somatic embryos show a striking parallelism to theirzygotic counterparts in their development. Severalcomparisons have been done between both types ofembryos, but mainly at the level of gene expressionand the occurrence of specific storage proteins, e.g. inalfalfa (Pramanik et al. 1992; Krochko et al. 1994), inmaize (Thijssen et al. 1996), or in pelargonium

T. Winkelmann (&) Æ M. SerekInstitute of Floriculture and Woody Plant Science,Faculty of Natural Sciences, University of Hannover,Herrenhaeuser Str. 2, 30419, Hannover, GermanyE-mail: traud.winkelmann@zier.uni-hannover.deTel.: +49-511-76219236Fax: +49-511-7622654

D. Heintz Æ A. Van DorsselaerLaboratoire de Spectrometrie de Masse Bio-Organique,Unite Mixte de Recherche, 25 rue Becquerel, 7509,67087, Strasbourg Cedex 2, France

H.-P. BraunInstitute of Plant Genetics, Faculty of Natural Sciences,University of Hannover, Herrenhaeuser Str. 2, 30419,Hannover, Germany

Planta (2006) 224: 508–519DOI 10.1007/s00425-006-0238-8

(Madakadze et al. 2000). Only few studies have dealtwith comparisons at the proteome level. This approachwas taken in the model plant carrot with the aim ofevaluating the possibilities of taking the easily acces-sible somatic embryos as a system to study the processof embryogenesis (Dodeman et al. 1998). Few proteo-mic studies have been made with commerciallyimportant agricultural or horticultural species, anexception being the investigation using two-dimen-sional gel electrophoresis for separating total proteinextracts in Picea abies (Hakman et al. 1990). In gen-eral, for most horticultural species little gene andprotein sequence information is available.

Cyclamen (Cyclamen persicum MILL.) is one of themost important flowering potted plants for the Europeanand Japanese markets. The annual production in Europeis estimated to be in the range of 140 million plants(Bongartz 1999). In contrast to many other ornamentals,which are propagated more or less exclusively vegetativelythese days, cyclamen is still propagated via seeds. Thisgenerative propagation is encumbered with some diffi-culties, because cyclamen is sensitive to inbreedingdepression resulting in heterogeneity in the hybrids.Moreover, seed production needs manual labour for cas-tration and crossings, resulting in high prices for seeds(up to 0.20 € for one seed, Schwenkel 2001). Thus, a clearneed for vegetative propagation is given. Since cyclamencannot be propagated from cuttings, in vitro culturetechniques have to be applied. Many genotypes can bemultiplied efficiently via somatic embryogenesis (Wicartet al. 1984; Otani and Shimada 1991; Kiviharju et al. 1992;Kreuger et al. 1995; Takamura et al. 1995; Schwenkel andWinkelmann 1998; Winkelmann and Serek 2005).

This propagation method can be applied to multiplyparental lines for F1 hybrids in lower numbers, but alsofor mass propagation of single elite plants. In the lattercase, the development of artificial seeds would be an

alternative for conventional seed propagation. Differ-ences between somatic embryos and their zygoticcounterparts are the water content and the protectiveseed coat. Desiccation treatments can be devel-oped which allow the drying and re-growth of somaticembryos in carrot (e.g. Tetteroo et al. 1995) but also incyclamen (Winkelmann et al. 2004a; Seyring and Hohe2005). Moreover, somatic embryos can be protectedfrom mechanical stress by artificial capsules. But onemarked difference between somatic and zygotic embryosremains, namely the availability of storage compounds,i.e. carbohydrates, lipids and storage proteins. It hasbeen shown in cyclamen that encapsulation in alginatebeads did not inhibit germination, and that externalsupply of nutrients was necessary for their germination(Winkelmann et al. 2004b). So far no reports are avail-able dealing with seed storage proteins in cyclamen.

Here we present a proteomic approach to investigatesomatic and zygotic embryogenesis in C. persicum.Proteins specifically expressed in somatic and zygoticembryos were systematically identified by mass spec-trometry (MS). Furthermore, the comparison of prote-omes of somatic and zygotic embryos with germinatingembryos together with the analysis of endosperm pro-teins revealed relevant information on storage com-pounds in cyclamen seeds. Thus, our analyses gave newinsights into embryogenesis and storage compoundphysiology in C. persicum.

Materials and methods

Plant material

Proteins were extracted from six different types of tis-sues of C. persicum M ILL.: zygotic embryos, endo-sperm, germinating zygotic embryos, globular somatic

Fig. 1 a–d: Embryogenesis inC. persicum. Seeds with zygoticembryos (arrow) surrounded byendosperm (a), somaticembryos after 3 weeks ofdifferentiation in liquidhormone-free mediumcontaining 60 g l�1 sucrose (b),19 day-old germinatingseedlings (c) and germinatingsomatic embryos (d); barsrepresent 2 mm

509

embryos grown in medium containing 30 g l�1 sucrose,globular somatic embryos grown in medium containing60 g l�1 sucrose and germinating somatic embryos(Fig. 1). Seeds were derived from self-pollinated flowersof I1 plants of the diploid genotype 3738 (one singleplant out of the cultivar ‘Purple Flamed’, obtainedfrom Royal Sluis, Enkhuizen in 1990). About 9–11 weeks after pollination seed capsules were harvested.At this stage, the endosperm had already become solid,but the seeds had not reached the fully desiccated stage.The zygotic embryos were elongated and in the torpedostage. These seeds were split under a stereomicroscopeand endosperm together with the seed coats and zy-gotic embryos were collected separately in 2 ml Ep-pendorf tubes and directly frozen in liquid nitrogen.Since zygotic embryos only represented a small portionof the whole seed, only limited amounts of fresh masswere available for this type of tissue, namely 12–24 mgper extraction. Ripe seeds of the same genotypes weregerminated after surface sterilization (2 min 70% eth-anol, 45 min 1.0% NaOCl) on hormone-free medium(see below) and after about 3 weeks culture at 20�C inthe dark, small tubers emerged from the seeds, whichare the first visible signs of germination in cyclamen.These tubers, the first small roots, and the cotyledonswere taken for protein extraction from germinatingzygotic embryos (Fig. 1c, average weight: 10–15 mg),while the residues of the seed coat and the endospermwere discarded.

Somatic embryos were derived from embryogenicsuspension cultures, which had been initiated from callusand maintained in liquid medium containing 9.05 lM2,4-dichlorophenoxyacetic acid (2,4-D) and 3.94 lM 2iP[6-(c,c-dimethylallylamino) purine] according to Winkel-mann et al. (1998). The embryogenic callus cultures wereinduced on somatic tissue from ovules of an I1 plant ofgenotype 3738 as described by Schwenkel and Winkel-mann (1998). About 14 days after subculture, the sizefraction 500–1000 lm was inoculated in liquid hormone-free medium (half strength Murashige and Skoog (1962)macro and micro nutrients, Fe EDTA full strength,2 g l�1 glucose, 250 mg l�1 peptone, 0.5 mg l�1 nicotinicacid, 0.1 mg l�1 thiamine-HCl, 0.5 mg l�1 pyridoxine-HCl, 100 mg l�1 inositol) containing either 30 or 60 g l�1

sucrose at a density of 1% packed cell volume (deter-mined after centrifugation at 720g for 5 min). Thedeveloping somatic embryos were harvested after 21–24 days of culture at 24�C at 100 r.p.m. in completedarkness. The size fraction of 500–1,000 lm was blotteddry between filter paper and portions of 500 mg wereimmersed in liquid nitrogen and stored at �80�C untiluse for protein extraction. Nearly all embryos were in theglobular stage (Fig. 1b) and weighed approximately0.5 mg each.

After 3 weeks of differentiation in medium with30 g l�1 sucrose, somatic embryos were placed on solidhormone-free medium of the same composition andincubated at 20�C in the dark for germination. After 10–12 weeks, the entire germinating embryos including tubers,

roots and cotyledons (Fig. 1d) were taken for proteinextraction. They had an average weight of 15–20 mg.

Protein extraction

To minimize loss of proteins, 150 mg of plant materialwas pulverized by grinding and directly dissolved in1.2 ml extraction buffer (see below), centrifuged twiceand the supernatant was directly loaded on the gel strips.Therefore, the amount of protein extract loaded to eachgel was based on the fresh weight of the plant material.Total protein extraction was performed following theprotocol of Gallardo et al. (2002): 150 mg of tissue wasground thoroughly in a mortar in liquid nitrogen. Withzygotic embryos, the amount of tissue had to be reducedto 12–24 mg, homogenization was done in Eppendorftubes using micropistils and lysis buffer volume was re-duced to 160 ll. In all other cases, the powder wassuspended in 1.2 ml thiourea urea lysis buffer andincubated at 4�C for 10 min. The buffer consisted of7 M urea, 2 M thiourea, 4% (w/v) Chaps, 1% (v/v)Pharmalyte pH 3–10 (NL=nonlinear) carrier ampholyte(IPG buffer, Amersham Biosciences), 32 mM Tris,53 U ml�1 DNase I and 4.9 Kunitz U ml�1 RNase A.There after phenylmethylsulphonylfluoride was added togive a concentration of 1 mM and the tubes were furtherkept at 4�C for 5 min, before 14 mM dithiothreitol(DTT) was added. This mixture was vortexed for 20 minat 4�C. Then the tubes were centrifuged at 18,000g for20 min at 4�C, the supernatant transferred to a freshcool tube and centrifuged again at the same conditions.The final supernatant was collected and contained theproteins for further analyses.

Proteins were isolated from all tissue types in threeindependent extractions from different charges of tissueand each extraction was analysed by one gel replicate.

Two-dimensional IEF/SDS-PAGE and staining

Separation of proteins by their isoelectric point wasperformed using the IPGphor system (Amersham Bio-sciences, USA) with immobiline dry strip gels (18 cm)with a NL pH gradient (pH 3–10) (Mihr and Braun2003). About 50–70 ll of the protein extracts weremixed with re-hydration buffer (7 M urea, 2 M thiourea,2% (w/v) Chaps, 0.2% (v/v) Pharmalyte pH 3–10 (NL)carrier ampholyte (IPG buffer), 20 mM DTT and a traceof bromophenol blue) to give a total volume of 350 ll.Focusing conditions are described in detail by Werhahnand Braun (2002). After equilibration for 15 min twicewith equilibration solution [50 mm Tris–Cl (pH 8.8),6 M urea, 30% (v/v) glycerin, 2% (w/v) SDS] supple-mented first with 1% (w/v) DTT and second with 2.5%(w/v) iodoacetamide, dry strip gels were placed hori-zontally on a Tricine SDS-PAGE gel. The two-dimen-sion electrophoresis was carried out according toSchagger and von Jagow (1987). Proteins were stained

510

with colloidal Coomassie Blue (Neuhoff et al. 1985,1990).

Evaluations

All gels were scanned and comparatively evaluated byvisual inspection. Only spots, which were reproduciblyfound in the three biological replicates, were included infurther examinations. In pairwise comparisons of theproteomes of different tissues, the relative abundance ofproteins was monitored. Only remarkable differences inthe accumulation of proteins or the new appearance ofproteins, which were repeatedly observed in all threereplicates, were encountered to be specific to a certaintissue type or developmental state. The most interestingspots, which differentially accumulated in somatic em-bryos, zygotic embryos and the endosperm (see Fig. 3)were excised from the gel and prepared for identification.

Mass spectrometry analysis and data interpretation

In situ digestion of protein spots was performed usingthe Mass PREP Station (Micromass, Manchester, UK)as described by Sarnighausen et al. (2004). Briefly, se-lected gel plugs were excised from preparative gels andwashed three times in a mixture containing 25 mMNH4HCO3: ACN (1:1, v/v). The cysteine residues werereduced by 50 ll of 10 mM DTT at 57�C and alkylatedby 50 ll of 55 mM iodoacetamide at room temperature.After gel dehydration with ACN, proteins were digestedovernight at room temperature in 15 ll of a solutioncontaining 12.5 ng ll�1 of a modified porcine trypsin(Promega) prepared in 25 mM NH4HCO3. Finally, adouble extraction was performed, first with 60% (v/v)ACN in 5% (v/v) formic acid, and subsequently with100% (v/v) ACN. Nano-LC-MS/MS analysis of theresulting tryptic peptides was performed using a capil-lary LC system (Micromass) coupled to a hybrid quad-rupole orthogonal acceleration time-of-flight tandemmass spectrometer (Q-TOF II, Micromass) according toHeintz et al. (2004). Chromatographic separations wereconducted on a Pepmap_C18, 75 lm i.d. 315 cm length,reverse-phase (RP) capillary column (LC Packings,Sunnyvale, CA, USA) with a flow rate of 200 nl min�1,accomplished by a pre-column split. An external cali-bration was performed using a 2 pmol l�1 GFP [(Glu1)-Fibrinopeptide B] solution. Mass data acquisition waspiloted by MassLynx 4 software (Micromass) usingautomatic switching between MS and MS/MS modes.Classical protein database searches were performed on alocal Mascot (Matrix Science, London, UK) server. Tobe accepted for the identification, an error of less than100 ppm on the parent ion mass was tolerated and thesequences of the peptides were manually checked. Onemissed cleavage per peptide was allowed and somemodifications were taken into account: carbamidome-thylation for Cys and oxidation for Met. In addition, the

searches were performed without constraining proteinMr and pI, and without any taxonomic specifications.These searches did not always lead to a positive identi-fication since the C. persicum genome has not yet beensequenced. In such cases, the use of a de novosequencing approach was necessary for a successfulidentification. For this purpose, the interpretation of theMS/MS spectra was performed with the PepSeq toolfrom the MassLynx 4 (Micromass) software, as well asthe PEAKS studio software (Bioinformatics Solutions,Waterloo, Canada). The resulting peptide sequenceswere submitted to the BLAST program provided at theEMBL site (http://www.dove.embl-heidelberg.de/Blast2/msblast.html) in order to identify them byhomology with proteins present in the databases. Weused the MS-BLAST specifically modified PAM30MSscoring matrix, no filter was set and the nrdb95 databasewas used for the searches as described by Castro et al.(2005). The statistical evaluation of the results and thevalidation of the matches was performed according toShevchenko et al. (2001). Sequence similarity searcheshave been employed to identify proteins via their knownhomologues in other species. All the MS BLAST con-ditions used in our work are those described as the‘default’ by Shevchenko et al. (2001). Thus, in Table 1only matching peptides with corresponding homologousproteins are indicated, matches are listed with a per-centage of at least 15% of amino acids in referenceproteins covered by matching peptides from nano-LC-MS/MS analysis.

Results

The protocol of Gallardo et al. (2002) allowed for effi-ciently extracting proteins from very small amounts ofplant material, as it was the case for zygotic embryos.Zygotic embryos in the torpedo stage were preparedfrom seeds 9 to 11 weeks after pollination, when theendosperm already had solidified. They had an averagelength of 2–4 mm (Fig. 1a) and an average weight of0.2–0.35 mg. For zygotic embryos, portions of 12–24 mg were homogenized in 2 ml Eppendorf tubesresulting in protein extractions sufficient for at least twogels. Protein extraction was equally efficient from so-matic embryos (Fig. 1b), but lower protein yields on thefresh mass basis were obtained from germinating zygoticand somatic embryos (Figs. 1c, d). Protein extracts ofthe endosperm had an extremely high viscosity and hadto be diluted about twofold with extraction buffer beforebeing used for electrophoresis.

Pairwise comparisons of different tissues

The evaluation of the two-dimensional-gels of zygoticand somatic embryos resolved over 200 spots, whichwere reproducibly found in three independent replica-tions (Fig. 2). Considerable similarity between embryos

511

Table

1ListofC.persicum

endosperm

(end),somatic(se)

andzygoticem

bryos(ze)

proteinsidentified

bynano-LC-M

S/M

Safter

IEF/SDS–PAGE

No.

Spot

no.a

Isolated

from

Abundance

Protein

(nameand

identifier)b

Mrc

pId

Matchingpeptides

e%

f

198

sehighforalltissues

a-tubulingi|17402471

49721

4.86

TVGGGDDAFN-TFFSETGAGK-A

VFVDL-EPTVID

EVR-

GTYRQLFHPEQLISGK-EIV

DLCLD

R—

VGGGTGSGLG-LLERLSVD

YGK-A

IYDIC

RRSLDIE

RPTYTNLNR-FDGALNVDVNEFQTNLV

PYPRIH

FMLSSYAPVISAEK-D

VNA

AVATIK

-IQ

FVDWCPTG

FKCGIN

YQPPTVVPGGDLAK-A

VCMISN

STSVAEVFSR-FDLM-

YAK-A

FVHW

YVG

EGMEEGEFSE

AREDLAALEK

51%

299

sehighforalltissues

b-tubulingi|5668671

49359

4.68

FWEVV-I

NVYYNEASGGR-W

AKGHY-C

HSLG-V

VEPY-L

TTPSFGDLNH

LISATMSGVT

CCLRFPGQLN

SDLRKLAVNL-IP-

FPRLHFFM

VGFAPLTSR-Q

QMWD-Y

LTASAIY

R-EVD

EQ-

MLNVQNKN

SSYFVEWIP

NNVK-V

SEQFTA

MFR

32.4%

349

sese>

segerm.;

ze>

zegerm.

Fe-SOD

gi|3599469

20960

6.14

ALEQLDDA-FNGGGHIN

HSIF

WK-E

GGGE-FGSFEALLQK-M

NAE-

GAALQGSG

WVWLGLDKEL

K-LVVETTANQDPLVTK-

LVPLLGID

VWEHAYYL-IW

KVIN

WKYASE

55%

4171

seze>

zegerm.

ferritin

1gi|5758041

23696

5.20

AMFAYFD

R-P

PSEF—

GDALY

AMELALSLEK-I

AEYVSQLR

17.5%

5105

sese>

zeUDP-glucose

pyrophosphor-

ylase

gi|32527831

51778

5.68

QISESEK-Y

LSGE–QQVEWSK-LNGGLG

TTMGCTGPK-Y

GCSVP

LLLMNSFNTH

D-Y

NSNIE

IHTF-L

VA

DDFVPLPSK-D

GWYPP

GHGDVFPSL-L

DALLSQGKEYVFVAN

SDNLGAVVDL

K-N

EY-

CMEVTPK-G

GTLISYEGK-F

KIF

NTNNLWVN-LK-R

LVEA

DALK-

MEIIPN

PK-V

LQLETAAGAAIRFFDHAIG

INVPR-A

NP

ANPSIE

LGPEFKK-IPSIIEL

DSLKVAGDVW

FGVNVTLK-E

I-PEGVVLE

NK

30%

610

sese>

ze;se>

segerm.;ze>

zegerm.

fructose-bisphosphate

aldol-

ase

gi|22620

38447

5.96

YAD

ELIA

NASYIA

TPGK-L

AAD

ESTGTIG

K-SIN

VE-PFV

DAMK-

GTVE

LAGTNGETTT

QGLDGLAQR-A

ILENANGLAR-V

PI-PEI-

LVDG-P

EVIA

EYTVR-P

WTLSFSYGR-A

QEVFLAR-S

EATLGKYQGG

AG

35%

ze38134

6.49

YAD

ELIA

NAAYIG

TPGKGIL

AAD

ESTGTIG

KR-INVENV-E

LLF

TTPGA-L

FEETLY

Q-V

EVLK-EGNVLPGIK

-ELAGTNGETTT

QGLDGL-LAIL

ENANGLAR-V

PIV

EPEIL

VDG

PH-V

APEVIA

EYT-

VRA-N

LNA

MN-K

PW

TL-G

R-A

QA

VFLAR-G

ADASESLHVK

44%

7159

sese>

zetriose

phosphate

isomerase

gi|38112662

27023

5.73

TFFVGGN

WK-IP

WVIL

GHSERR-ILGESNEFVGDK-V

IACVGE

TLEQRESGST

MDVVAAQTK-A

YEPVWA

IGTGK-V

ATPA

QA-

QEVHAELR-IIY

GGSVSGANCKE

LAGQPDVDGF

LVGGASLKPE

FID

IIK

49%

8119

sese>

ze;se>

segerm.

heatshock

protein

70

gi|34908140

70907

5.10

GEGPA-PMIV

VQ-TLSSTAQTTIE

IDSLYEGID

FYS-T

ITRARFEEL

NMDLFR-SSVHDVV

LVGGSTRI-VQQLLQDFFN

GKELCK-SIN

PDEAVAYGAAV

QAAIL

SGEGN

EK-V

QDLLLLDVTPLSLGLET

AG-

GVMTVLIPR-E

QVFSTYSDNQ

PGVLIQ

VYEGER-FELSGIP

-NALE-

NYAYNMR-W

LDSNQ

LAEADEFEDK-M

KELESIC

NPIIAK

32%

9157a

sese>

segerm.

dnaK-typemolecularchaper-

onehsp70gi|1076746

71068

5.13

VE

IIANDQGNRT

TPSYVGFTDSER-N

QVAMNPIN

TV

FDAK-V

VQ

YK-Q

FAA

EEISSMVLIK

-EIA

EAYLG

TTIK

NAVVTVPAYFNDSQR-

DAGVIA

GLNVMRIINEPTAAAIA

YGLD

KK-A

TAG

DTHLGGEDFD

NR-L

FRKCM

20.9%

10

44

sese>

zeglutathione-S-transferase

gi|34914768

20850

7.53

MSSG

YNIL

TR-C

VLASSGF-M

SCP-FFYSFNVLGG

LDNEGK-V

FTY-

DAVGSYGR-SPSPLLLPAKDAVTPLSE

SEAID

LVK

39%

11

14

sese>

ze;se>

segerm.

GAPDH

gi|66012

36662

8.30

IGIN

GFGR—

AGI-

ALNDNFVK—

AASFNIIPSSTGAAK—

LVSWYDNEWGYSTR—

FGI-

VEGLMTTVHSIT

ATQK—

GIL

GYTEDDVVSTDFVGDSR

DFVVESTGVFTDK-TLLFGEKPVTVFG-NCLAPLAK-

LFGDKPVTVFGVR—

YDSVHGVWK—

FNIIPSST—

IGIN

-GFGR—

LTGMAFR-VNFVVEST—

LVSNAS—

AGLALND

39.7%

12

18

sese>

segerm.

GAPDH

gi|66012

36662

8.30

VLGYTEDDVVSTDFVGDSR-G

LVEGLMTTVHSLTATQK-

DLVSNASCTTNCLAPLAK-TLLFGEKPVTVFGLR-

FGLVEGLFTTVHSLTAT-V

PTVDVSVVDLTVR-A

ASFNLLPSSTG-

YDNEWGYSTR-LVSWYDNE-PMFVMGVNE-LDLVSNASCGT-

NPEDLPW-LGLNGFGR-LTGMSFR-V

LPALNGK-V

VDLLVH-L

A-

LNDNFVK-V

LLSAPSK

44.2%

512

Table

1(C

ontd.)

No.

Spot

no.a

Isolated

from

Abundance

Protein

(nameand

identifier)b

Mrc

pId

Matchingpeptides

e%

f

13

23

zeze>

seze>

end;ze>

zegerm.

GAPDH

P25861

36685

8.3

DAPYFVMGV-PFLTTDFGLVEGLMTTVHSLTATQK—

VLGY-

TEDDVVSTDFVGDSR-LVSWYDNEW

GYSTR—

VSNASCTTNCLA-

PLAK—

VPTVDVSVVDLTVR—

AASFDLLPSSTGAAK—

LVESTGV-

FTDK—

TLLFGEQPVTVFGLR—

APMFVMGVNE—

PFLTT-

DYMT—

LALNDNFVK—

DAPMFVM

GV—

LGLNGFGR—

VLLS-

APSK—

EFLVEST—

VLPALNGK—

LTGLSFR—

VVDLLV

47%

14

30

zeze>

seze>

zegerm.

7SglobulinQ9AUD0

67069

7.5

LVLEGEAYLELACPHMS—

AGTTVYLVNPDK-

NER—

FSPELLEAAFN—

FGPGGENPES—

LALVLEGEAYL-

ELA—

WPFGGEGK-V

ALLEAEP—

TFLVPNH—

NPESFF—

GSM-

TAPH—

EGMLVEASEEQ-H

EEGGL—

ESFFKDPR

15.2%

15

157

zeze>

zegerm.;end>

ze11Sglobulin-likeprotein

Q8W1C2

59127

6.46

HTSNYQNQLDNNPR—

GLLLPQYTNAPTLFYVVR—

NDKQFQCAG-

VA—

GVLFSGCPETF—

EGDLLALPA-D

AELLAELFGVDLD-

TAR—

CQIE

15.7%

end

11Sglobulin-likeprotein

Q8W1C2

59127

6.46

AELLADLFGVDLETAR-N

DQQFQC-EGDLLALPA-LPKYTNAPTL-

FYVVR-G

CPETF-LFAGCPE-N

YQNQLD-C

PETFE-FLLAGNPKDE-

NKLDENPR-V

LVAMEHTSNYQN-C

SLRL

15.6%

16

47

zese>

segerm.;ze>

zegerm.

11SglobulinQ6QJL

152296

6.14

GQLVVVPQNFVDVK—

DNPSHTLFFNPR—

LEENMCTAT

16.3%

17

48

seandze

se>

segerm.;ze>

zegerm.

11Sseed

storageprotein

Q84ND2

53641

7.68

SAIH

AMPLDVLS-NANFQTLSGR-NAHTIIY-EWIA

F-GKIQ

IVDA-

YIE

PGN-LEATNR-LEENMCTA

15.2%

SAIH

AMPLDVIS-L

SADHGVLYTNA-G

VLYTNAIL

NP-G

LEWIA

F-

NAHTIIY-FQTLSGR—

NNANQL-LVVVP-LEENM

CTA

15.8%

18

50

sese

germ.>

seglobulinprecursorQ6Q385

53663

8.62

GQLVVVPQNF—

LQLSDAHGVLYTNA—

NSHSVAY—

LNG-

LEASQ—

EQEEFSLN—

INNPSRPDFFNP

17.0%

19

190

zese>

segerm.;ze>

zegerm.;

end>

zevicilin-likeprotein

Q6QJL

115656

5.48

ELIC

FEVNADDNE—

GTVFVVPA

15%

20

201

end

end>

zexyloglucan-endo-trans-gly-

cosylase

Q84JX

331751

5.66

LWMDPSADFHTYSIL

WNP-L

LPGNSAGTVTTYYLSSK

MIY

NYCT-

DANR-W

SADDWATR-P

YTLHTNV-L

YSTLWDAD-V

DGTPLR-

YLYGK-FQNWE

19.4%

21

202

end

end>

zexyloglucan-endo-trans-gly-

cosylase

Q84JB

932693

4.99

PPEC-TDIL

WG-IVFYVDGT-STLW

NADD-LSLDQDSGSGF-

YVDGTPLR-PYTLHTNV-LWFDPSADF-LWSADDWATR-

LVPGNSAGTVTTYYLSS-Q

NNHMIY

NYCTDANR-FDPSADFH-

TYSLLWNP

29%

22

203

end

end>

zexyloglucan-endo-trans-gly-

cosylase

Q84JB

932693

4.99

FDPSADFHTYSLLWNP—

QNNHMLYNYCT-

DANR—

LLPGNSAGTVTTYYLSS—

LWSADDWATR—

LWFDP-

SADF—

PYTLHTNV—

YVDGTPLR—

LSLDQDSGSGF—

YL-

FGQLDM—

STLWNAAD—

LLW

YVDGT—

TDLLW

G—

PPEC

29%

23

204

end

end>

zexyloglucan-endo-trans-gly-

cosylase

Q6RHY0

32155

5.23

FDPSADMHTYSLLWNP—

LLPGDSAGTVTTYYLSS—

WSADD-

WATR—

LWFDPSADF—

MLYNYCTDANR—

PY-

TLHTNVYAAG—

LYSTLWSADD—

VDGTPLR—

LLWYVDGT—

YL-

FGK—

FDLLWG

13.6%

aSpots

are

named

accordingly

toFigs.2and3

bAccessionnumber

inNCBIorSWISS-PROT

databasesandprotein

assignmentafter

BLAST

homologysearches

cTheoreticalmolecularmass

(Da)

dTheoreticalisoelectricalpoint

ePeptides

identified

via

theMascotsearchengineandconfirm

edbydenovosequencingare

indicatedin

regularcharacters,whereaspeptides

only

identified

bydenovosequencingandBLASTsearchare

initalic

f Percentageofaminoacidsin

reference

proteinscovered

bymatchingpeptides

from

nano-LC-M

S/M

Sanalysis

513

of both types was found, as expressed in the percentageof spots in similar relative intensity of about 74%. Inzygotic embryos approximately 11% of the proteinsdifferentially accumulated, while in somatic embryos15% were found in clearly increased abundance orexclusively.

When proteomes of somatic embryos grown inmedium with 30 or 60 g l�1 sucrose were compared, thevast majority of spots were of similar abundance, onlyfew proteins (about 7%) were found in higher abun-dance in somatic embryos, which had differentiated inthe higher sucrose concentration. Only 1% of the pro-teins specifically accumulated in somatic embryos grownin medium with 30 g l�1 sucrose.

During germination of somatic embryos approxi-mately 22% of proteins showed decreasing abundanceand 2% completely disappeared. Interestingly, similarproportions of down-regulated proteins were found inthe comparison of zygotic embryos and germinated zy-gotic embryos. On the other hand, about 13% of thespots were observed in higher intensity and 2% exclu-sively in germinating somatic embryos. For zygoticembryos the number of proteins accumulating in ger-minating embryos was even lower (4% up-regulated, 1%new).

The comparison of proteomes of zygotic embryosand endosperm has revealed about 76% of the spotswere of comparable relative intensity. About 7% of theproteins were found in higher abundance in zygoticembryos and the endosperm, respectively.

Identification of proteins by nano-LC MS/MS

According to the objectives of this study, 83 spots wereprepared from gels containing the proteins of zygoticembryos, somatic embryos grown in medium with60 g l�1 sucrose and endosperm and analysed by liquidchromatography-tandem MS. These 83 spots werechosen due to their accumulation profiles: (a) spotsshowing significant differences in abundance betweensomatic and zygotic embryos by visual evaluation, (b)spots of high intensity in the endosperm and (c) spotsof high abundance in embryos and decreasing duringgermination. Two bioinformatic approaches were fol-lowed to identify these proteins: (1) MASCOT searcheswere performed to directly identify matching proteinsin databases and (2) MS/MS data were used to gen-erate peptide de novo sequence data. As a result, 26 ofthese proteins were identified with at least 15% cov-erage. The number of high-quality de novo peptidesequence data was higher (41 spots), but due to lack ofmolecular information (genomic and proteomic) avail-able for cyclamen only 26 gave significant alignmentsto proteins from other plants. All identified proteinsare listed in Table 1, some of which will be describedbelow.

Two proteins were present in high relative abundancein all tissues analysed, and they aligned with high cov-erage percentages to a-tubulin (spot no. 99) and b-tubulin (spot no. 98, Fig. 2).

For some proteins, a decreased abundance was ob-served during germination for both somatic and zygoticembryos. Among these, superoxide dismutase (SOD,spot no. 49), and ferritin (spot no. 171) were detected(Figs. 2, 3).

Identified proteins of differential abundance in somaticand zygotic embryos

Besides the striking similarities in the proteomes of so-matic and zygotic embryos, especially if the former haddifferentiated in medium with 60 g l�1 sucrose, somespots were found to vary in intensity when both types ofembryos were compared. Some of the proteins specifi-cally accumulating in somatic embryos were identified(spot nos. 105, 10, 159, 119, 44, 14, 157a, Table 1).Interestingly, three were enzymes involved in carbohy-drate metabolism, i.e. UDP-glucose pyrophosphorylase(105, Fig. 2), fructose bisphosphate aldolase (10, Fig. 2)and triosephosphate isomerase (159, Fig. 3). Three spotsgave significant matches to GAPDH (spots nos. 14, 18and 23), of which no. 14 was specific for somatic em-bryos (Fig. 2), 18 was higher in somatic embryos than ingerminated ones and 23 was up-regulated in zygoticembryos (Fig. 3, Table 1). The molecular mass of threeof them was comparable (about 37 kDa), but they dif-fered in charge.

Furthermore, two heat shock proteins 70 of differentsize (spots 119, Fig. 2 and 157a, Fig. 3) and a glutathi-one-S-transferase (spot 44, Fig. 3) were found to bepresent in higher relative levels in somatic than in zy-gotic embryos.

Identification of storage proteins and enzymes involvedin storage compound metabolism

Storage proteins are accumulating in seeds and aredecreasing in abundance during germination. Becauseone of our intentions was the identification of storageproteins, proteins following this accumulation profilewere prepared from the gels and submitted to MS/MS.In the endosperm gels (Fig. 2e) five groups of proteinswere especially prominent in spot intensity (indicated byletters in Fig. 2e). In each group different isoforms werefound, which have similar molecular masses, but whichdiffer with respect to their pI. In four of these fivegroups, homologies to storage proteins or precursorswere observed: group A with proteins of about 35 kDaand basic pI (7–8) expressed similarity to 7S globulins(spot no. 30, Fig. 3, Table 1). Group A storage proteinswere highly expressed in zygotic embryos and endo-sperm, but less abundant in somatic embryos grown in

514

medium with 60 g l�1 sucrose and very weak or even notdetected in somatic embryos derived from medium with30 g l�1 sucrose and germinating zygotic and somaticembryos.

Group B proteins were monitored in the acidic part(pI 5–6) of the gels and had a size of about 27 kDa(Fig. 2). One of them (spot 157, Fig. 3) was identified as11S globulin-like protein (Table 1; isolated from endo-sperm and zygotic embryos). These 11S globulin-likeproteins were most abundant in the endosperm and so-matic embryos.

Storage proteins of group C had a molecular size ofabout 20 kDa and pI of 7–8 (Fig. 2). Spots isolated fromzygotic embryos and somatic embryos (nos. 47, 48, 50)were identified as 11S globulins and globulin precursor(no. 50, Fig. 3). They showed a similar expression profileas proteins of group B.

The last groupDwas present in high concentrations inthe endosperm, in zygotic embryos and somatic embryosgrown in medium with 60 g l�1 sucrose. These proteinswith pI of 4.5–5.5 and a size of 10–15 kDa were found tobe down-regulated during germination. For one protein(190, Table 1, Fig. 3) an alignment to a vicilin-like pro-tein has been detected, suggesting that the proteins ofgroup D are belonging to the group of 7S globulins.

One group of proteins was especially prominentexclusively in the endosperm gels (group E, Fig. 3 spots201–204). They could be identified as xyloglucan endo-transglycosylases. Their estimated molecular size wasabout 30 kDa and their pI varied between about 4.0 and4.8. While these proteins were absent in somatic andzygotic embryos, they were observed in germinatingzygotic embryos although in lower intensity than in theendosperm.

Fig. 2 a–f Analysis of proteinsextracted from different tissuesof C. persicum by two-dimensional SDS-PAGE. azygotic embryos, b somaticembryos differentiated inmedium with 60 g l�1 sucrose, cgerminating zygotic embryosendosperm, d somatic embryosdifferentiated in medium with30 g l�1 sucrose, e endosperm,the most prominent proteingroups are marked by lettersand explained in detail inresults, f germinating somaticembryos. Spots with numbersindicate proteins that wereidentified by MS (Table 1).Molecular mass range (verticaldimension): 100 kDa (top)–5 kDa (bottom), non-linearseparation

515

Discussion

In these first detailed proteomic analyses in C. persicumseveral insights have been obtained concerning the pro-tein composition of somatic and zygotic embryos on theone hand and concerning seed and germination physi-ology on the other. Although in this initial study onlyabundant and soluble proteins were analysed and theevaluation was performed qualitatively rather thanquantitatively, valuable novel information on the molec-ular and functional proteomic aspects of embryos andseeds have been obtained for this poorly characterizedspecies. Imin et al. (2004) established a proteomic refer-ence map for globular staged somatic embryos of Med-icago truncatula and resolved more than 2,000 proteins.The pronounced difference in resolution may be due tothe fact that these authors used a different extractionprotocol and applied bigger gels as well as silver staining.

The results will be discussed in three sections belowaccording to the objectives of this study, which were (a)to compare the protein profiles of somatic and zygoticembryos and describe the differentially accumulatingproteins, (b) to identify seed storage proteins and (c) tolook for proteins playing a role in the synthesis ofstorage compounds. To simplify matters, the term pro-teome will subsequently be used, notwithstanding thatonly a sub-proteome was analysed in the present study.

Proteomic comparison of somatic and zygotic embryos

Somatic embryos mimic their zygotic counterparts inmorphological as well as developmental aspects. In this

study it has been shown that cyclamen somatic embryosresemble zygotic embryos in their protein compositionas well. Relatively few proteins were present in differentrelative abundance, especially if somatic embryos, whichhad developed in medium with the higher sucrose con-centration, were compared to zygotic embryos. How-ever, differences in the developmental stage of somaticand zygotic embryos might have caused some of thedifferences in protein abundance observed in this study.Zygotic embryos had reached the torpedo stage, whilesomatic embryos were mainly in the preceding globularstage. Further studies should implicate proteomic anal-yses of somatic and/or zygotic embryos in their differentstages of development.

Four of the proteins, which were found in higheramount in somatic embryos, were identified as enzymesinvolved in carbohydrate metabolism (UDP-glucosepyrophosphorylase, fructose bisphosphate aldolase, tri-osephosphate isomerase and GAPDH). It seems likelythat their higher abundance is evoked by the exogenoussucrose supply by tissue culture media and that theseenzymes are involved in glycolysis. Likewise, theexpression of UDP-glucose pyrophosphorylase gene andthe activity and content of the corresponding enzymewere up-regulated in Arabidopsis thaliana leaves whenfed with sucrose (Ciereszko et al. 2001, 2005).

Other proteins occurring in high concentrations incyclamen somatic embryos were heat shock 70 proteinsof different molecular sizes (spots 119, 157a). Many heatshock proteins are molecular chaperones, which areformed in response to stress and also are developmen-tally regulated. Small heat shock proteins of 15–30 kDahave been reported to accumulate during maturation ofsomatic embryos of cork oak, and by a distinct regula-

Fig. 3 a-f Analysis of selectedproteins of C. persicumextracted from different tissues.a, d endosperm; b, e zygoticembryos; c, f somatic embryosdifferentiated in medium with60 g l�1 sucrose (enlargementsof Fig. 2). Spots with numbersindicate proteins that wereidentified by MS (Table 1)

516

tory control in response to different stress treatments(Puigderrajols et al. 2002).

Besides storage proteins, which will be discussedbelow, some other proteins have been identified, whichare present in embryos (zygotic and somatic) at highlevels. Among these, SOD was very prominent in so-matic and zygotic embryos. This enzyme is involved inthe defence against oxidative stress. During germinationhigh amounts of energy are needed resulting in veryhigh mitochondrial activity and superoxide production.Therefore, this protein is necessary in germinating seeds,but was also reported to be present in dry seeds (Zhuand Scandalios 1993). Moreover, it was suggested thatSOD has a protective role during the acquisition ofdesiccation tolerance. In maize, MnSOD transcripts in-creased after ABA or high osmotic treatments of em-bryos (Zhu and Scandalios 1993). Proteins, which areinvolved in stress response like heat shock proteins orSOD are of interest also in another aspect: their accu-mulation has been reported to be involved in triggeringsomatic embryogenesis. The auxin treatment (mostcommonly by 2,4-D) is thought to induce an oxidativeburst, which results in stress response and induction ofsomatic embryogenesis (Thibaud-Nissen et al. 2003).Another protein, glutathione-S-transferase (spot 44),accumulating in somatic embryos, is involved also in theresponse to auxin or oxidative stress. Similar resultswere observed for embryogenic cell cultures of M.truncatula by Imin et al. (2004), where 18% of theidentified proteins play a role in defence and stress re-sponse.

In soybean, the transcripts of genes for oxidativestress response precede the appearance of globular so-matic embryos (Thibaud-Nissen et al. 2003), whereas inour studies these stress-related proteins still were foundin later stages of somatic embryos. Further investiga-tions should include very early stages of somatic em-bryos or even embryogenic cultures in whichdifferentiation has not been initiated.

Seed storage proteins in somatic and zygotic embryosand in the endosperm

Our study revealed, that in cyclamen 11S globulins werefound in somatic embryos, endosperm tissue and inlower amounts in zygotic embryos, while 7S globulinsshowed higher abundance in extracts from zygotic em-bryos and endosperm. In the literature, the most fre-quently analysed proteins in somatic embryos incomparison to zygotic embryos are seed storage pro-teins. Similar storage proteins in somatic and zygoticembryos were reported early for rapeseed (Crouch 1982)and alfalfa (Krochko et al. 1994). Madakadze et al.(2000) identified an 11S globulin and two low molecularweight proteins as storage proteins in zygotic and so-matic embryos of Pelargonium x hortorum. They de-scribed mature somatic embryos to contain about 80%of the amount of storage proteins found in zygotic em-

bryos. The amount of 7S globulins being the majorstorage proteins in oil palm (Elais guineensis) were foundto be 80 times lower in somatic embryos than in zygoticembryos (Morcillo et al. 1998). The authors claim a lackof maturation in somatic embryos to be the reason forthis observation. Likewise, carrot somatic embryos didnot accumulate the major storage protein (daucin)probably due to in vitro culture conditions (Dodemanet al. 1998). Klimaszewska et al. (2004) were able toshow that maturation in media with 6% sucrose ascompared to 3% resulted in higher levels of the mostimportant storage proteins, 11S globulins and 7S vicillinlike proteins, in Pinus strobus. In agreement with thesefindings, higher abundance of storage proteins for so-matic embryos grown in higher sucrose concentrationswere identified in cyclamen in this study. It can be as-sumed that the treatment with 60 g l�1 sucrose pro-moted maturation of cyclamen somatic embryos.Likewise, this treatment also improved the desiccationtolerance of the embryos (Winkelmann et al. 2004b).

The protein profile of zygotic embryos could beconsidered as the ideal profile, and culture conditions forsomatic embryos should be adapted in a way to achievethis ideal protein profile.

For cyclamen, no reports on storage proteins havebeen published so far. Therefore, the identification of 7Sglobulins and 11S globulins in the present study as wellas their localisation in endosperm tissue and in embryosreveals novel information for this species. Globulins arethe major seed storage proteins in dicotyledonous plants(Shewry et al. 1995). The polymorphisms observed in thedifferent groups of storage proteins are due to multigenefamilies as well posttranslational modifications (Shewryet al. 1995). The 11S globulins consist of six subunits,each of which is composed of an acidic (a) chain (Mr ofabout 40,000) and a basic (b) chain (Mr of about 20,000)(Shewry et al. 1995). Groups of proteins expressingcharge heterogeneity, which corresponded to 11S glob-ulins (groups B and C, Fig. 2e), were identified incyclamen. They were observed in high relative abun-dance in somatic embryos, zygotic embryos and endo-sperm (Fig. 2). In contrast, the spots corresponding to7S globulins (groups A and D, Fig. 2e), which were alsodescribed for the first time in cyclamen, were of highintensity in endosperm and zygotic embryo protein ex-tracts. In somatic embryos grown in the higher sucroseconcentration, the one group (group D) of the 7Sglobulins accumulated to relatively high levels, but thesecond group (group A) was only weakly expressed insomatic embryos. However, it has to be taken into ac-count that different developmental stages of somatic andzygotic embryos could have an effect on storage proteinquantities. Future studies should pay attention to thetemporal accumulation of storage proteins during thedevelopment of embryos.

The observation of storage proteins in P. strobus re-vealed a different situation: while 7S globulins werefound in comparable concentrations in somatic and zy-gotic embryos, 11S globulin levels were higher in zygotic

517

embryos (Klimaszewska et al. 2004). These differencesare probably due to the plant species.

Xyloglucans as storage compounds in cyclamen

Besides proteins, the major storage compounds incyclamen seeds are carbohydrates. A special type of car-bohydrate polymers was identified in cyclamen in 1889 byReiss> (cited in Hayashi 1989) and called amyloids(current nomination: xyloglucans, Braccini et al. 1995).These hemicelluloses play an important role in cell wallstructure, and in several plant species they were identifiedas cell wall bound storage compounds in either the cot-yledons or in the endosperm as in case of the members ofthe Primulaceae to which cyclamen belongs (Koiiman1960). It has been assumed, that reserve xyloglucans notonly support the developing embryo with energy, but alsoprotect it from desiccation (Hayashi 1989). The last stepof xyloglucan biosynthesis is catalysed by the enzymexyloglucan endotransglycosylase (XTH, Rose et al. 2002).This enzyme can also display hydrolysing activity. In thisstudy high abundance of XTH was observed in cyclamenendosperm, which most likely was due to the building upof storage xyloglucans. The isoelectric point of the XTHsin cyclamen endosperm was low when compared to lit-erature. One well-studied species regarding storage xylo-glucans is nasturtium (Tropaeolum majus), but the reportsconcentrate on their mobilisation during germinationrather than on their assembly (reviewed by Buckeridgeet al. 2000). The observation of XTHs in cyclamenendosperm for the first time, therefore provide newinteresting information regarding the synthesis of xylo-glucans as storage carbohydrates.

In conclusion, this first proteomic analysis of cycla-men embryos, although examining only abundant andsoluble proteins, has revealed several new insights. (1)The similarity of proteomes of somatic and zygoticembryos points to the interesting parallelism not only inmorphology and development but also with regards tomolecular physiology. However, differentially accumu-lating proteins between both types could be furtheranalysed in more detail aiming at the understanding ofphysiological differences, which discriminate somaticembryos from their zygotic counterparts. Culture con-ditions should be adjusted in a way that somatic em-bryos resemble zygotic ones as much as possible. (2) Theidentification of seed storage proteins will improve ourknowledge on reserve proteins in this species and othermembers of this poorly investigated plant family. Fur-thermore, it enables us to adapt the in vitro cultureconditions in a way that somatic embryos accumulatetheir own reserves and may be directly sown into soil infuture. We are planning to identify appropriate treat-ments during differentiation and maturation of cycla-men somatic embryos (e.g. ABA, osmotics) bymonitoring their effects on storage protein accumula-tion. (3) Finally, the novel finding of the high abundanceof XTHs in cyclamen endosperm should initiate new

projects studying their role in assembly and mobilisationof xyloglucans as storage carbohydrates in the endo-sperm. The identification of the genes encoding enzymesin xyloglucan biosynthesis in cyclamen and their tem-poral and tissue specific expression profiles are aimsfor further studies. Moreover, the incorporation ofxyloglucans into an artificial endosperm might allowdeveloping artificial seeds, which can germinate auton-omously in soil.

Acknowledgements The authors would like to thank DagmarLewejohann and Annette Steding for their excellent technicalassistance, and Professor David Collinge (The Royal Veterinaryand Agricultural University in Copenhagen, Denmark) for lin-guistic editing of the manuscript.

References

Bongartz W (1999) Cyclamen. Thalacker-Medien, Braunschweig,pp 42-50

Braccini I, Du-Penhoat CH, Michon V, Goldberg R, Clochard M,Jarvis MC, Huang ZH, Gage DA (1995) Structural analysis ofcyclamen seed xyloglucan oligosaccharides using cellulase diges-tion and spectroscopic methods. Carbohydr Res 276:167–181

Buckeridge MS, Santos HP Tine MAS (2000) Mobilisation ofstorage cell wall polysaccharides in seeds. Plant Physiol Bio-chem 38:141–156

Castro AJ, Carapito C, Zorn N, Magne C, Leize E, Van DorsselaerA, Clement C (2005) Proteomic analysis of grapevine (Vitisvinifera L.) tissues subjected to herbicide stress. Exp Bot56:2783–95

Ciereszko I, Johansson H, Kleczkowski LA (2001) Sucrose andlight regulation of a cold-inducible UDP-glucose pyrophos-phorylase gene via a hexokinase-independent and abscisic acid-insensitive pathway in Arabidopsis. Biochem J 254:67–72

Ciereszko I, Johansson H, Kleczkowski LA (2005) Interactive ef-fects of phosphate deficiency, sucrose and light/dark conditionson gene expression of UDP-glucose pyrophosphorylase inArabidopsis. J Plant Physiol 162:343–353

Crouch ML (1982) Non-zygotic embryos of Brassica napus L.contain embryo-specific storage proteins. Planta 156:520–524

Dodeman VL, Guilloux ML, Ducreux G, de Vienne D (1998)Somatic and zygotic embryos of Daucus carota L. display dif-ferent protein patterns until conversion to plants. Plant CellPhysiol 39:1104–1110

Gallardo K, Job C, Groot SPC, Puype M, Demol H, Van-dekerckhove J, Job D (2002) Proteomics of arabidopsis seedgermination: a comparative study of wild-type and gibberellin-deficient seeds. Plant Physiol 129:823–837

Hakman I, Stabel P, Engstrom P, Eriksson T (1990) Storage proteinaccumulation during zygotic and somatic embryo developmentin Picea abies (Norway spruce). Physiol Plant 80:441–445

Hayashi T (1989) Xyloglucans in the primary cell wall. Annu RevPlant Physiol Plant Mol Biol 40:139–168

Heintz D, Wurtz V, High AA, Van Dorsselaer A, Reski R, Sar-nighausen E (2004) An efficient protocol for the identification ofprotein phosphorylation in a seedless plant, sensitive enough todetect members of signalling cascades. Electrophoresis 25:1149–1159

Imin N, De Jong J, Mathesius U, van Noorden G, Saeed NA,Wang XD, Rose RJ, Rolfe BG (2004) Proteome reference mapsof Medicago truncatula embryogenic cell cultures generatedfrom single protoplasts. Proteomics 4:1883–1896

Kiviharju E, Tuominen U, Tormala T (1992) The effect of explantmaterial on somatic embryogenesis of Cyclamen persicum Mill.Plant Cell Tissue Organ Cult 28:187–194

518

Klimaszewska K, Morency F, Jones-Overton C, Cooke J (2004)Accumulation pattern and identification of seed sorage proteinsin zygotic embryos ofPinus strobus and in somatic embryos fromdifferent maturation treatments. Physiol Plant 121:682–690

Kooiman P (1960) On the occurrence of amyloids in plant seeds.Acta Bot Neerl 9:208–219

Kreuger M, Postma E, Brouwer Y, Van Holst GJ (1995) Somaticembryogenesis of Cyclamen persicum in liquid medium. PhysiolPlant 94:605–612

Krochko JM, Bantroch DJ, Greenwood JS, Bewley JD (1994) Seedstorage proteins in developing somatic embryos of alfalfa: de-fects in accumulation compared to zygotic embryos. J Exp Bot45:699–708

Madakadze RM, Krochko JE, Senaratna T (2000) Identificationand characterization of storage proteins in zygotic and somaticembryos of geranium (Pelargonium x hortorum). J Am SocHortic Sci 125:525–529

Mihr C, Braun HP (2003) Proteomics in plant biology. In: ConnPM (ed) Handbook of proteomic methods, Humana Press,Totowa, pp 409–416

Morcillo F, Alberlene-Bertossi F, Hamon S, Duval Y (1998)Accumulation of storage protein and 7S globulins during zy-gotic and somatic embryo development in Elais guineensis.Plant Physiol Biochem 36:509–514

Mordhorst AP, Toonen MAJ, de Vries SC (1997) Plant embryo-genesis. Crit Rev Plant Sci 16:535–576

Murashige T, Skoog F (1962) A revised medium for rapid growthand bioassays with tobacco tissue cultures. Physiol Plant15:473–497

Neuhoff V, Stamm R, Eibl H (1985) Clear background and highlysensitive protein staining with Coomassie Blue dyes in poly-acrylamide gels: a systematic analysis. Electrophoresis 6:427–448

Neuhoff V, Stamm R, Pardowitz I, Arold N, Ehrhardt W, Taube D(1990) Essential problems in quantification of proteins followingcolloidal staining with Coomassie Brilliant Blue dyes in poly-acrylamide gels, and their solution. Electrophoresis 11:101–117

Otani M, Shimada T (1991) Somatic embryogenesis and plantregeneration from Cyclamen persicum Mill. leaf cultures. PlantTissue Cult Lett 8:121–123

Pramanik SK, Krochko JE, Bewley JD (1992) Distribution ofcytosolic mRNAs between polysomal and ribonucleoproteincomplex fractions in alfalfa embryos—stage-specific transla-tional repression of storage protein synthesis during early so-matic embryo development. Plant Physiol 99:1590–1596

Puigderrajols P, Jofre A, Mir G, Pla M, Verdaguer D, Huguet G,Molinas M (2002) Developmentally and stress-induced smallheat shock proteins in cork oak somatic embryos. J Exp Bot53:1445–1452

Rose JKC, Braam J, Fry SC, Nishitani K (2002) The XTH familyof enzymes involved in xyloglucan endotransglucosylation andendohydrolysis: current perspectives and a new uniformnomenclature. Plant Cell Physiol 43:1421–1435

Sarnighausen E, Wurtz V, Heintz D, Van Dorsselaer A, Reski R(2004) Mapping of the Physcomitrella patens proteome. Phy-tochemistry 65:1589–607

Schagger H, von Jagow G (1987) Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteinsin the range from 1 to 100 kDa. Anal Biochem 166:368–379

Schwenkel HG (2001) Introduction: Botany, Economic Impor-tance, Cultivars, Micropropagation of C. persicum. In:Schwenkel HG (ed) Reproduction of Cyclamen persicum Mill.through somatic embryogenesis using suspension culture sys-tems. European Communities, pp 3–7

Schwenkel HG, Winkelmann T (1998) Plant regeneration via so-matic embryogenesis from ovules of Cyclamen persicum Mill.Plant Tissue Cult Biotechnol 4:28–34

Seyring M, Hohe A (2005) Induction of desiccation-tolerance insomatic embryos of Cyclamen persicum Mill. J Hortic Sci Bio-technol 80:65–69

Shevchenko A, Sunyaev S, Loboda A, Shevchenko A, Bork P, EnsW, Standing KG (2001) Charting the proteomes of organismswith unsequenced genomes by MALDI-quadrupole time-of-flight mass spectrometry and BLAST homology searching.Anal Chem 73:1917–26

Shewry PR, Napier JA, Tatham AS (1995) Seed storage proteins:structure and biosynthesis. Plant Cell 7:945–956

Takamura T, Miyajima I, Matsuo E (1995) Somatic embryogenesisof Cyclamen persicum Mill. ‘Anneke’ from aseptic seedlings.Plant Cell Rep 15:22–25

Tetteroo FAA, Hoekstra FA, Karssen CM (1995) Induction ofcomplete desiccation tolerance in carrot (Daucus carota L.)embryoids. J Plant Physiol 145:349–356

Thibaud-Nissen F, Shealy RT, Khanna A, Vodkin LO (2003):clustering of microarray data reveals transcript patterns asso-ciated with somatic embryogenesis in soybean. Plant Physiol132:118–136

Thijssen MH, Spoelstra P, Emons AC (1996) Immunodetectionand immunolocalization of globulin storage proteins duringzygotic and somatic embryo development in Zea mays. PhysiolPlant 98:539–549

Werhahn W, Braun HP (2002) Biochemical dissection of themitochondrial proteome from Arabidopsis thaliana by three-dimensional gel electrophoresis. Electrophoresis 23:640–646

Wicart G, Mouras A, Lutz A (1984) Histological study of organ-ogenesis and embryogenesis in Cyclamen persicum tissue cul-tures: evidence for a single organogenetic pattern. Protoplasma119:159–167

Winkelmann T, Serek M (2005) Genotypic differences in callusformation and regeneration of somatic embryos in Cyclamenpersicum Mill. Euphytica 144:109–116

Winkelmann T, Hohe A, Schwenkel HG (1998) Establishingembryogenic suspension cultures in Cyclamen persicum ‘PurpleFlamed’. Adv Hortic Sci 12:25–30

Winkelmann T, Meyer L, Serek M (2004a) Desiccation of somaticembryos of Cyclamen persicum Mill. J Hortic Sci Biotechnol79:479–483

Winkelmann T, Meyer L, Serek M (2004b) Germination ofencapsulated somatic embryos of Cyclamen persicum. Hort-Science 39:1093–1097

Zhu D, Scandalios JG (1993) Maize mitochondrial manganesesuperoxide dismutases are encoded by a differentially expressedmultigene family. Proc Natl Acad Sci 90:9310–9314

Zimmerman JL (1993) Somatic embryogenesis: a model for earlydevelopment in higher plants. Plant Cell 5:1411–1423

519