HeredilY 60 (1988) 183-191 @ The Genetical Society of Great Britain Received 22 Apri I 1987

Genetic recombi ation in a diploid synaptic mutant and a Solanum tuberosum x S. chacoense diploid hybrid

D. S. Douches* and Depanment of Vegetable Crops, University of California, Davis, CA 95616. U.S.A. C. F. Quiros

.'. diploid synaptic mUlant, 116 wilh the genotype syJ/syJ,ps/ps, was Identified as heterolygous for four enzyme-coding loci, lind for the tuber fllsh colou r locus (Y). Through 4x-2x CrOsses, Ilalf-tetrad analysis was applied to determine the ability of t ~,2 M6 clone to transmit its heterozygosity to tetraploid offspring. On the average, an 89·03 per cenl reduction in recombination was fouud within the chromosome segments sampled, resulting in 98 per cent transmission :Jl hererorygosity. Cytological observations of microsporogenesis revealed frequent pachytene associations and a high prollortion of univalenls at diakinesis. These findings show that chromosome pairing in M6 is mostly desynaptic. Diploid and 4x-2x testcrosses were made to study recombination in a 2"-pollen producing diploid hybrid 1704 (pslp.v); :.... tuberosum :.... x:..;. chacoense Bitt. For the loci compared, recombination was reduced 34·56 per ceot on the average. Chis liybtid transmitted on the average 89·2 per cent of its helerozygosity which was significantly higher thM tile value previously olJserved in non-hybrid individuals. Unlike M6, 1704 produced both viable nand 2" pollen. In addition, cytological e~·.~mination of PMCs indicate normal bivalent associations. It is suggested that genomic differentiation between 5. fuberOIum aull S. chacoense could ac('ount for the reduced recombination levels.

INTRODUCTION of the diploid parenCs heterozygosity intact 10 Ihe tetraploid offspring (Okwuagwa and Peloquin,

Meiotic abnormalitie which lead to the formalion 1981 ).of 2n gametes offer unique opportunities 10 Since Iwanaga and Peloquin (1979) reported hybridise diploid Solanum species with the culti­ the fi rst synapti c mutant (5)'1) ".ffecting mega­vated tetraploids for the purposes of increasing sporogenesis in a diploid potato clone, three new genetIc diversity and ultimately yield. The benefits synaptic mutants have been identified: sy2of such hybridisations aimed to capture novel al., (Okwaugwu and(J ohnston et 1981); sy3genetic diversity are direclly related to the meiotic Peloquin, 198 I); aod sy4 (Iwanaga, 1984), all mechanism responsible for diplandroid formalion. aflecting microsporogenesis. In these new synaptic Al present, three meiotic mechanisms have been mutants, general cytological abnormalities are identified thilt lead to maL: 2n 3ametes in the common. Lack of pairing was noted during diplo­diploid potato species: parallel andlor fused lene and diakinesis stages, followed by random spindles at metaphase 11 (First Division Restitu­ distribution of the chromosomes during the first tion or 1DR), premature cytokinesis I, and pre­ division leading to pollen sterility. Okwuagwa and matur<;; .;ytokinesis II (Second Division Restitution Peloqu in (1981) identified the parallel spindle (;1 SDR) (M~,k and Peloquin, 1975; Ramanna, mechanism operating during metaphase II wilhin 1979). The discovery of synaptic ~enes (fwanaga the sy3sy3 genotype. This resulted in the fusion of and Peloquin, 1979) wilhin the diploid potalO unbalanced nuclei, ultimately leading to balanced species has expanded the potential of Ihe FDR and viable dyads. mechan I~ms. The combination of synaptic ]n asynaptic mutants, homologous pairing fails chromosomes with the parallel spindles gene has completely or is incomplete. On the other hand,opened the possibility of transferring 100 per cent the desynaptic condition can alter the maiOlenance • Present address: Crop and Soli SCIences Departmenl. of chromosome pa iring du ring pachylene, ulti­Michigan Slale University. ~.,.", i .""sing, MI4S824, U.S.A. mately reduci ng fertility as a consequence of

184

univalents at diakinesis (Hermsen, 19840). Definitive cytogenetic characterisation of these synaptic mutants has not been possible. Difficuilies lie in the inability to differentiate cytologically between the asynaptic and desynaptic condition. The utilisation of synaptic mutants for the purpose of transferring intact genotypes from parents to offspring rests solely on a theoretical basis, ~Ince

recombination has been known to occur in synaptic phenotypes of other higher plants I,as reviewed by Koduru and Rao, 1981). A study of recombination within chromosome segments of these synaptic mutants should provide experimental data to test the utility of these mutants.

With the production of 2n pollen through the parallel or fused spi ndle mechanisms, gene· centromere mapping is possible through 4x-:: x crosses. Mok and Peloquin (1982) proposed the use of electrophoretic variants for mapping while lwanaga (984) suggested their application :0 analyse the genetic consequences of the desynaptic gene sy4. Recently, Douches and Quiros (1987b) determined gene-centromere relationships for 10 enzyme-coding loci. Positioning of electrophoretic markers on the chromosomes enables the iden­tification and characterisation of meiotic mutants by genetic analysis in the potato.

In this paper, M6, a diploid synaptic mutant with the ability to generate FOR 2n Dollen through the parallel spindles mechani m (Il'J/.IT.d,ps/ps), was analysed through 4x-2x CrOsses and sub­sequent half-tetrad rationale, in conjunction with cytological observations of microspOlogenesis. The resulting recombinational frequencies were compared with recombinational values obtained in non-mutant plants to quantify the reduction in recombination. The same approach was applied to determine the level ofrecombination in an inter­specific hybrid of S. tuherosum L. and S. chacoense Bitl. as a means to detect genom ic differentiation between these two species. This work provides ex perimentaJ data on the extent of transmission of heterozygosity for synaptic mutants and inter­specific hybrids to their progenies,

MATERIAL AND METHODS

Pfant material

The diploid synaptic, parallel spindles, double mutant, M6 (syJ/sy3,ps/ps) (2n = 2x = 24), and the interspecific hybrid, n04 (ps, ps) (2n: 2x­24), were kindly surnlied by Dr S. J, Peloquin

D S, DOUCHES AND C F QUIROS

(University of WisconSIn, Madison). M6 was selec­ted fran! the progeny of the cross W529S.7 (( l x W5337.3 (J), both of which were species hybrids between S. luberosum dihaploids (2n x 2x, ~4)

and S. phureJa Juz. et Buk. (2n =2x =24). n04 is a diploid species hybrid bdween a S. (uberowm dihaploid and S. chacoense. These diploids were crossed to tetraploid clones. The tetraploid parents used in the 4.~ -~x and 2x-4x crosses were Lemhi Russet, NO D277 -2, 8C8370-4, T245.7. and PASJ006-7. These clones welL' killdh supplied by Dr R. E. Voss (University of California, Davis). Oi ploid cros:;cs were made in the direct ion 84S lOx n04 and T704 x 85S056. 845 10 and 85 SD56 are diploid clones of S. phurein origin.

Efee trap/wres is

Employing horizontal starch gel electrophoresis, eight enzyme systems were sludied which revealed nine enzyme-coding loci: Dia-l (dIaphorase), GOI­1 (glutamat~ oxaloacetate transminase), ldh-I (isocitrate dehydrogenase), Mdh·J \ malate dehy. drogenase), 6-Pgdh-3 (6-phosphogluconate dehy­drogenase), Pgm-2 (phosphoglucomutase), Prx-2, Prx-3 (peroxidase), and Sdh-l (shikimic acid dehydrogenase). Inheritance data and allozyme pattern, for these loci are described by Quiros and Me Hal e ( 1985) and Douches and Qui ros (1987 (l). Excluding Din-I and Prx-2. gene-centromere map distances were previously estimated (Douches and Qui ros, \987 b), Speci fic electrophoretic pro­cedures are described by Qui lOS (1981). Enzyme stains used are according to Vallejos (J 933 l.

To test for recombination, 4x-2x(FDR) tcst. crosses were ma<;le for the enzyme loci listed above and Y determining tuber flesh colour (Howard, 1970). Since a single 4x parent was not identified to be nulliplex for all the loci tested. a series of crosses were made using Lemhi Russet, NDD277­2, 8C8370-4, Y245.7, and PAS3006-7, The genotypes of these clones are shown in tables I and 2. Gene-centromere map distances for these loci including Di(l-l and Prx-2 were previously estimated by Douches and Quiros (1987 b) Ilsing diploid stocks from the S. luber05um group. These values were used as standards 10 compare recombi­national frequencies in M6 and n04, Gene centromere map distances are simply another way of expressing the recombination frequency for a chromosome segment between a locus and its cen­tromere.

Under normal synaptic conditions, a 4x­2x(FDR) testcross would generale progeny that

-----

185 RECOMBINATION IN DIPLOID POTATOES

can be divided into three classes (assuming dial­lelic loci): simplex, nuJJiplex, and duplex genotypes, with the second and third classes indicating a recombinational event. I n case of dominance, only the nuJliplex progeny would be diagnostic; however, with these co-dominant elec­trophoretic markers, the duplex genotype can be distinguished from the simplex on the basi~ of asymmetric banding intensities (Martinez.Zapater and Olivier, 1985). Under asynaptic conditions assuming zero recombination, these 4x-2x test­crosses should yield exclusively simplex genotypes_ If crossing-over does occur, nulliplex and duplex progeny should be recovered. The greater the distance of a gene from its c~ntromere,

the better the opportunity for detecting crossing­over in M6 and n04.

All crosses were made :n the greenhouse. Offtypes observed in 4x-2x offspring were subjec­ted 10 chromosome counts in ~ither root-tip or anolh\'1 squ,·shes 10 rule oul triploids. No triploids were observed, presumably due to a strong triploid block (M arks, 1966; Hanneman and Peloquin. 1968; Hermsen, 1984&). Germinated seedlings

Table 1 Half-retrad amdysts of 1... I\.."Clrophorehc markers 1(1 4x-2x crosse!> Wtlh M6, :}

Record

N'l. (rOll Paren[

R60S32 Lemhi Ru:s~~··.

xM6

~"SDI" NDD277·2 xM6

8hSD.12 .... ~'mhl

Russe:( xM6

4 T,,(al

5 l>6SDJ" NDl)277-2 xM~

6 86S060 ,10:nO·4

xM6 7 TOlal g ~6SD16 NOO277·2

"M~

9 ~!iS [)f;O BC3.nU·4 xM6

10 T01~J

II 86S016 NOD: . , 0 .~

xM6

Gcnorypc; 5imrl" Duple,

Pr.\-3'J I 3'.l' X" .1 1.1" 20R (178)" 5 (1951

VY.l\ "( y~, :157

.......'y" y,. ~2

&.Pgdh..I:J'.J';' x ;'3' 409(375)

401

6-PgJI,.J'J'J'3' x J'3' 60 0

Idh·1 J I:1'1-' X I' r' 461 (.\28) 412

7 (72)

Idli.,'I',·'I·- x I'I' 58

Sdlr.,'I' I" I' x I' J'< 470 (.1851 411 (:1951

2 (4.1.51

wen.~ transplamed into Irays (50 plants per tray). In 4 10 6 weeks, root or leaf samples from healthy and vigoroosly growing seedlings were sampled for electrophoretic analysis. To determine tuber flesh colour (Y locus), the seedlings were allowed to tuberise in the Irays. Small mature tubers were harvested ;0 lhree to four months. Tubers were (hen visually compared with parental tubers for flesh colou r determination.

To cytologically study diplandroid formation, flower buds were collected and immediately fixed in a freshly prepared ethanol/propionic acid solu­lion (3: 1) with ferric chloride added as a mordant (Swaminathan el al., 1954). After 24 hours, the buds were removed from the fixative, rinsed. and stored in 70 per cenl ethanol al 4°C. Anthers were dissected and squashed in I per cent acetocarmine. All meiolic stages of Ihe pollen mother cells were examined. The occurrence of parallel and/ or fused spindles duri ng Metaphase [I. followed by dyad formation identified the FDR mode of 2n pollen formation (Mok and Peloquin, 1975; Ramanna, 1979). Pollen staining was in accordance with Quinn el aJ. (1974).

syn<.lptl(,_/P~T3.lkl ,;plodle$ mutant

Observed E,pected Percent re-com· n;COm· reductlofl

btnallon binauon in

frcq Ut'oc) frequency~ r{'oom­

Null'pJcx Iper cent) Iper c"nt) X b,nalJoo

4 (I ~.51 4'15 18'00 26 (f........ 76·90

6(5) 289 16.84 27 l ...... .t 8289 0

0

4 (72) 2.)) 3050 '824'- ­ 92.36 0

0

0143.5) 0.21 1840 1009'" ':lS.HI> 1 fl7) 049 8,30 I:> 7·"'" 9410

., From Douches and Quiros. 19Mh '" Expected values in p(lrcc'Hhe~es

186 0 S DOUCHES AND C F QUIROS

Table 2 Half-tetrad analysis of eleclrophorettC markers ,n 4x :' x leslcrosses with D04, an Jnterspeci fic hybrid of S. luberosum x S c1wcoel1se

Cros' Parent Locus Nulliplex progeny

Progeny S1ze

Observed recom bma lion frequency (per cent)

Expecteda

recom blOat ion frequency tper cent) x'

Red uClion in reco mbi nallOn (per cent)

8650t9 865046 865054

1'245.7 PAS3006-7 BC8370-4

GOT-I GOT-I GOT-I

'/0 3/0" 1/1

341 224

59

1 17 1 34 339

~n 624 1.12 865D19 1'245.7 PGM-2 .; (2)' 223 (221) 3.59 2.0 NS 000 865D46 PAS3006-7 PRX-2 28 (21 ) 224(203) 25.0 I ~ 0 NS 000 865019 1'245.7 IDH·I 3 220 5.38 865046 PAS30067 IOH-1 10 224 893 86SD54 BC8370-4 lOH 4 57 14.04

20 (41) 501 (460) 7.98 184 22 05~M 56.60 86S [) 19 Y245.7 SOH-l J 220 2·69 S6S))46 PAS3006-7 SDH-l 6 224 1·25 86$D54 BC8370-4 SOH-I I 59 3·39

86$D19 865D54

1'245.7 BC370·4

OIA-1 DIA-I

10 (23) 8 :3

503 (480) 119 59

:3 98

10 17

8.3 7.70""" 52.05

865046 86SI)54

PA5J006-7 BC8370-4

6·PGOH-3 6·PGOH·3

II 15 5

178 221

5 I

12.36 13.50 1960

20(42) 272 (230) 1465 305 12.18··~ 5200 865D46 PAS~006-7 MDH-I 20 (38) 224(186) 1786 33.5 9.81"'''' 4670

'From Douches. D. S Jnd C. F. Quiru" 19X7" .. Nulliple"jduplex ~~nOlyp~s

, Expecled value in fJarenthc~e~

,1 Expected value ror Prx-l-('en\romer~ linkJge block ... ProbabdilY >0001

RESULTS

Extent of recombination in the synaptic mutant M6

Ekctrophoretic ~Illalysis of M6 revealed that it was heterozygous for four isozyme loci: Idh- J 112

, Sdh· J J /-', Prx11J,l and 6-Pgdh-3 IJ 2, In addition, M6 was heterozygous for yellow tuber flesh as deter­mined by the genotype of its parents (Masson, 1985), Five chromOSOme segments of varying lengths were: therefore 3vailable to study and test for the occurrence of recombination in the synaptic mutant M6 (table I).

The 6- rgdh-J locus was calcubted to be 30,1 m.u. from its centromere in crOS5es between G rO.J r Tuberosum accessions (Douches and QUiros, 1987b). Th us, this locus provided the best opportun ity to detect Cfoss-over in M6 because of

its distal chromosome arm position. A recombina­tion value of 2· 33 per cent was obtained for M6 after pooling segregation data from two tetraploid progenies, 865016 and 865060 (X:': 3,071, P ='

0,07) (table 1). Segregation for the Jdh-l locus, located at

18·4 m.u. from the centromere, was also amenable to half-tetrad analysis in the M6 progenies. 1n the family 86DS60, two recombinant duplex individuals were observed from 60 plants. However, in the family, 865016, the analysis was not straightforw(lrd. The tetraploid parent, NDD277-2, as determined by progenv testing, was a si mplex heterozygote ( Idh-I ) 12 J2/' J rathe r lha n homozygous for the locus as required by half­tetrad analysis. As a result, only half of the progeny was amenable to the half-tetrad rationale. Rec­ombination was not detected in this cross as nul· liplex genotypes were not found.

RECOMBINATION IN DIPLOID POTATOES :87

Segregation of the Sdh·J locus, located 8·3 m.U. from the centromere, was studied in the family 860S16. Since the tetraploid parent was a duplex

1 5 5heterozygote (Sdh_l 1/ / / ) and M6 W~l\

heterozygous with the alleles Sdh-l l 15, only one

of the recombinant progeny classes was observable in the cross. A 0·49 per cent recombination rate was detected for this gene-centromere segment ill a progeny of 412 plants.

The cross between Lemhi Russet and M6 (86SD32) estimated a 4-15 per cent recombination rate in the chromosome region between the Prx-J locus and its centromere. The expected rate fur this locus was 18,0 per cent.

Recombination between the Y locus :".!1d its centromere was evaluated in the same family, 8650532, along with 86S0 16. Based upon the frequency of white-fleshed tubers (table I), pooled families (X 2 = 3,83, P = 0,06) revealed a 2·89 per cent rate for this proximal arm region compared with the normal location of 16·8 m.L1. from the centromere.

Cytological analysis of microsporogenesis in M6(sy3/sy3,ps/ps)

General asynchrony was found in the meiotic pro­cess. At times, ~ingle anther squashes revealed meiotic stages from prophase I to dyad formation. Ouring pachytene, meiocyles were identiLed with varying degrees of chromosome pi:i1ring. Complete synapsi~ was identified frequently, suggesting that some level of recombination may be occurring. (~g. IA). Oiaki ne~is was generally characterised by cells showing mainly univalents, hence reveal­ing the desynaptic nature of thiS clone. Bivalent formation was very low with a range of zero to three bivalents per cell (fig. I B).

The chromsomes of M6 dUling metaphase J lacked normal orientation on the equatorial plate, hence anaphasf I was not clearly distinguishable from metaphase I. Spindles were elongated with a curved orientation (fig. Ie). Lagging chromo­somes were regularly observed duri ng the firSl division, leading to abnormal chromosome distri-

A B /

... f. ,­" , .

c

J 'I ;:..•• ". I ,.,1"

J''II...~.;.,

,I

./,;

J

~"

" . .:t o

,....

FiRure 1 Various stages of meJO!I$ in pollen mother cells of the desynaptlc, d\plold clone M6. (A) Pachytene show,ng mostly bi~alcnts. (B) Diakine is wllh cell showinG mainly univalenls (C1 Metaphase/anaphase 1 characterised by unbalanced chromosome segregation and cu,,'~d, elongated spindles (U) Telophase 11 ",th fu\ed spindle or;entation (f5). Note tetrad fQrmat;on.

188

bution. Bivalent frequenci>.'s within these meiocytes ranged from zero to [our; however, most cells lacked clear bivalent associations. Chromo­some imbalance was noted at the eod of the first division. At times bivalent separation was observed during late anaphase (fig. Ie). Precocious chromatid separation of univalents was not evident. Metaphase 11 was characterised by a strong expression of the parallel/fused spindles mechanism. Anaphase II and telophase II wete typical of psps. Expression of tripolar spindles and normal spindle orientation with balanced tetrads was evident at low frequencies (fig. lD). An esti­mate of pollen fertility, based upan the frequency of plump, stainable pollen, was 36·0 per cenl.

Extent of recombination in the Interspecific hybrid Tl04

Electrophoretic analysis of T704, a S. tuheros~m x S. chacoense species hybrid, revealed eight heterozygous enzy lne-codi ng loci: DiQ-l, GO(-],

ldh-l, Mdh-I, 6-Pgdh·3, Pgm-2, Prx·2 and Sdh-1. Unlike M6, T704 yielded progenies in both 4><-2x and 2x-2x crosses because of viable 2n and n pollen. Recombination frequencies were studied in progenies of bOlh ploidy ;evels.

The 4x-2x rationale applied to the synaptic mutant was similarly extended to the 4x-2x crosses ::1Vol vi ng TI04 as the pollen parent. A series of three 4x-2x crosses and two 2x-2x crosses were made to estimate the recombination ra: ,'S for the eight di !ferent loci and two linkage blocks (table 2).

The standard recombination rate for the chromosomal segment between Mdh·! and its cen­tromere was estimated to be 33·5 per cent. In a 4x-2x cross involvinQ n04, a rate of 17·86 per cent was estimated flOm a progeny of 224 plants ir family 860$46 (table 2).

For 6-Pgdh-3, located at 30'1 m.ll. from its cen­tromere, a recombination value of 14·65 per cent was obtained for n04 after pooling the segregating

D. S. DOUCHES AND C F. QUtROS

data from progenies 86S046 and 86S054 (X;: 0·48, P ----' 0'53).

A combined esti mate of 7· 98 per cent was found lor TI04 after pooling the three families for segre­gation of the ldh-I locus (x 2 2· 514, P '-" O· 30)._'C

This ratc was significantly reduced from the expec­ted rate 01 18 -4 per cent. The Sdh-l-centromere recombination value in the same three families (X 2 = 0,987, P = 0'63) was reduced to 3·90 per cent from the standard rate of 8·3 per cent.

No reduction in recombination was found for the chromosome region between Pgm-2 and its centromere. Based upon a fit to chi-square expecta­tions, the present estimate of 3·58 per cent was not significantly different from the expected value of 2·00 per cenl.

The highest recombination rate (25 per cent) for any chromosome segments studied in T704 was found withi n the Prx-2 -centromere segment in family 865D546 of 224 plants.

Segregation of the Dia-J and Gol·1 loci were also studied in these 4x-2x testcrosses: however, no standard estimates of the recombination frequencies exist to compare these values. The pooled families 86SD 19 and 86S 054 (X 2

- 0·182, p O' 70) estimated a 12· 36 per cent rate for the Via-} -centromere segment in T704. Segregation of the Got·] locus was studied in 86S019, 865046, and 86SD54. A pooled estimate of all three families (X 2 "2'S2, P=O'12) places Got-! proximal to its centromere with a \·12 per cent ratc.

The reciprocal 2x-4x cross involving T704 was unsuccessful because 01 little or no expression of 2n egg formation, however, n04 was used as the pisti lIate parent in a diploid cross (86S 047). A 17 per cent recombination ratc was observed for the Idh-l/Sdh-J linkage block, which was reduced from the standard value of 36·8 per cent (table 3). An earlier cross between 84S 10 and TI04 (86S0 \0) estimated a 10·4 per cent recombination frequency for this same linkage block, while a 32·4 per cent rate was obtained between 6-Pgdh-3 and DiQ-l (Douches and Quiros, 1987Q).

Table 3 Diplold \~$tcross data wIth n04, an interspecific hybnd of S. rubersllm x S. chacoerlse

Red uetion In

Cross Linkage Rf per celli RF expected· Progeny Size

reeom bi nation (per cent)

86$D31 86SD47 86SDJ I

84$IOxT704 T704 x SSSD56 84SIOxT704

fdIJ-ljSd!l-l ld!l-I/Sdh-I 6-Pgdh-J/ Olll- J

10.4 17,0 32.4

46.8 36.8

78 80 78

71.7 53.8

, Fro m L:>ouches an d Quiros (19860)

189 RECOMBINATION IN DIPLOID POTATOES

Cytological observations of T704

Examination of PMCs confirmed a strong expression of the parallel/fused spindles mechan­ism leading to 2n pollen formation. Analysis of early stages found normal chromosome pairing during pachytene and bivalent formation in diakinesis. Normal chromosome disjunction fol­lowed. Pollen fertility was high (>90 per cent).

DISCUSSION

Genetic and cytological analyses of M6

In 4x-2x testcrosses with M6, five opportunities were available to sample this clone's ability to transmit its heterozy~()sity. Despite general pachytene pairing, all five chromosome segments of M6 showed severe reductions in crossing-over throughout the genome_ The average reduction in crossing-over from expected was 89·03 per cent with a range of 76·96. -98· 86 per cent. Under nor­mal synaptic conditions, these five regions would have transferred an average of 81·59 per cent of their heterozygosity. ConsiJering the reduction in ;ecombination under the abnonnal synaptic condi­tions conferred by the syJ gene, 97 ·99 per cent of M6's heterozygllsity was transmitted to its progeny.

The genotypes combining the sy3 gene and the ps genes were first described by Okwuagwu and Peloquin (198 j J, however, pachytene stages were not mentioned. In many respects, our cytological oh~ervations in M6 agree with those of Okwuagwu and Peloquin. The s)'1, sy2, and sy3 phenotypes ha ve similar expression, but sy 1 was ex pressed only ;,., megasporogene~is. Ramanna (1983) observed from 1-1 to 8·3 II per cell during meta­phase r in desynaptic plants derived from the parents (I x J) of M6. This range of restricted pair­ing could suggest a variable expression for the desynaptic genes as influenced by other genotypic factors and interactions.

The cytological similarities between sy4 and sy3 were striking. 1wanaga 's (1984) cytological description of the synaptic abnormalities matched closely with the present description of the syJ in M6. He noted a consistent low frequency of synap­sis during pachytene and diakinesis. Thus, rwanaga predicted:! 97 per cent transmission of heterozygosity for this mutant, based upon a frequency of 3 chiasmata per meiocyte.

Hermsen and Ramanna (1983) suggested that desyn:lptic genes would be expressed in both micro- and megasporogenesis. No fruit set was obtainable in 2x4x crosses with M6. Lack of seed

set may be attributed to either lack of expression of the sy3 gene in the female side or lack of a restitution mechanism. Cytological examination of megasporogenesis would be needed to discrimi­nate between these two possibilities.

Although the cytogenetic data indicated that M6 is desynaptic; i.e. generalised pairing during pachytene and high frequency of univalents at diakinesis, the 4x-2x genetic data revealed a reduced recombination typical of an asynaptic mutant. This indicated that the syJ mutation inter· feres with chiasma formation leading to a pre­mature separation of the bivalents in diakinesis. The theoretical expectation of 100 per cent transfer of the parenl's heterozygosity to the offspring was close to being realised in M6. Thus, this restricted recombination caused the transfer ofa significantly greater (81·54 per cent vs. 97·99 per cent) amount of heterozygosity than the ps gene alone. The potential value of the !>yJ/sy3,ps/ps genotype for 4x-2x breeding schemes to develop highly uni­fonn, heterozygous progenies for true seed produc­tion (TPS), is demonstrated. However, the utility of the syJ gene in the development of an apomictic system would be questionable because of its poten­tial for recombination.

Genetic analysis of T704

In comparison to M6, crossing-over occurs much more frequently in the interspecific hybrid n04. Of the loci sampled, an average recombinational frequency of 2·01 per cent was found in M6, while n04 had 10·82 per cent, a five-fold increase, whereas recombination in normal FDR clones had a 18·55 per cent recombination frequency. In 4x­2X testcrosses, the recombination rate for six chromosome regions were compared to expected val ues for their respective regions. Reduction in crossing-over in n04 ranged from 0 to 56·60 per cent, with an average of 34·56 per cent. Consider­ing the eight gene-centromere segments sampled in these 4x-2x testcrosses, the average heterozy­gosity transferred was 89 ·18 per cent. Based upon X 2 comparisons of observed and expected recombi­nation rates for individ ual loci, a significantly greater amount of heterozygosity is transferred to the tetraploid offspring (89.18 per cent vs. 81 ,45 per cent respectively) using n04.

For the four chromosome arm regions where reduction in recombination was significant, the decrease regressed linearly. For each recombina­tion value observed, the reduction averaged 51·84 per cent regardless of the expected recombina­tional frequency [or the respective segments.

190

The chromosome arm region involving Pgm-2 did not deviate from the expected value. Based upon chi-square tests, a 3·59 per cent recombi na· tion frequency for the segregation of the Pgm-2 locus was not significantly di fferent from the expected valu of 2·0 per cent. It can be a~gued

that a significant reduction in crossing-over would be difficult to detect statistically in small chromo­some segments with population sizes of approxi­mately 200.

A pooled recombination frequency of 1·12 per cent was observed for the segregation of the GOI·/ locus in 4x-2x testcrosses. No additional 4x-2x crosses have been available using different diploid parents for a comparison. Applying the reasoning set forth for the Pgm-2 segregation along with 2x-4x (unpublished) and 4x-4x data (Douches and Quiros, 1987b), it is likely that the COI-J recombination frequency obtai ned expresses a true proximal po.sition for this locus.

The highest recombination rate (25 per cent) for any chromosome segments studied in ]'704 was found within the Prx-2-centromere segment. At this time no direct estimates for this segment exist, but it is known that Prx-2 <lnd Prx·J are tightly linked with 0·6 map unib ~eparating the two loci (D-:l\lches and Quiros, 19870; Quiros and McHale, 191))). Hence, the Prx·]-centromere linkage (18·0 m.u.) was substituted as the reference. Know· ing thi~ relationship, the observed Prx-2­centromere recombination rate was deduced to be not significantly different than expected. With this linkage block of 18·0 m.ll., a reduction in recombi­nation would have been expected. rn comparison, fdl~ 1 segregation has a si mi lal' rate of 1S-4 per Cenl under normal instances. Within 1704, a rate 'or this locus was reduced to 7· 98 per cent. The difIerential rates for the two loci situated at equivalent distances from their centro meres indi­cate different degrees or chromosome homology between the two genomes, This conclusion is sup­po,ted by higher recombination frequencies found in ceMain chromosome segments of higher ex hibit­ing de- or asynaptic conditions (Rick, 1971; Koduru an(: Rao, 1981).

]n the diploid (JOSS where 1704 was used as the pi~tillate parent, a 53·80 per cent reduction in expected recombination was seen for the Jdh­l/Sdh-I linkage hlock. The reduction in crossing­over regressed in a similar fashion as the chromo­some segments studitd in microsporogenesis. Pre­viously half-tetrad analysis and diploid testcross data using standard diploid stocks positioned the Sdh·J and Idh-I on opposite chromosome arms (Douches J,nd Quiros, 1987b). The recombination

D S DOUCHES AND C F. QUIROS

rate for th is lin kage block in 1704 2lgrees with the 4x-2x dat<l for the Jdh-l and 5dh-J gene· centromere segments when TI04 was also used as the pollen parent.

Since Dia-/ :'lnd 6-Pgdh-3 are found on the same chromosome il can be assumed th:lt the recombination frequency within Ihe Via·]­centromere segmenl was reduced at a similar rate as 6-Pgdh-J. Hence, we can estimate a normal recombination frequency of 25·73 per cent for the Din-/-centromere segment. If the predicted re­combination value for the Dia-l-cent romere segment and the expected value for the 6-Pgdh-3­centromere are summed, a 56'23 m.u. map distance is then predicted. Under normal synaptic condi­tions, this linkage block had not been detected. It is possible that in utilisi ng TI04, we were ab Ie to take advantage of the reduced recombination levels to identify this new linkage block and posi­tion the loc; in relation to their centromere. The usefulness oj this interspecific hybrid in assigning new markers to linkage grou ps could conti nue to be realised until more markers sat urate the genome.

Douches and Quiros (1987 b I reported gene­centromere map distances for ten enzyme-coding loci and the Y locus. Their investigation and the current study suggest that these markers are distri­buted over II chromosome arms of the potato genome. With the current 4x-2x linkage data for Dia-I and 6-Pgdh-3, we assign Dia·J and 6-Pgdh.J to opposite chromosome anns. The genomic distri­bution of these 12 loci is now expanded to 12 chromosome arms.

Genomic differentiation in Tl04

Unlike M6, the reduction in recombination in T704 must arise from a different mechanism. Besidt:s the formation of parallel spindles, other meiotic or chromosome abnormalities were not noted in 1704. Red uction in crossing over was not expected in this clone since both parental species (5. luberosum and S. chacoense) are supposed to have the same genome (Hawkes, 1978).

Our findings, however, raise lhe question of genomic differentiation between S. /uberosum and S. chacoense. This is supported by distorted segre­gations for several genes in S. luberosum x S. chocoense progenies (Quiros and McHale, 1985; Douches and Quiros, 1987a). Hawkes (1978) assigns the genomic formula Al to these two species, which belong to the South American sub­group. Most of the diploid species of lhis subgroup hybridise readily and produce fertile F 1 hybrids.

RECOMBINATION IN DIPLOID POTATOES

From pairing studies, Hawkes concluded that genomic differentiation in the conventional sense has not progressed very far among the South American diploid species. Ramanna and Hermsen (1979) suggest differentiation has not resu iled in pairing failure at meiosis. However, when Swaminathan (J 953) doubled the chromosomes of S. rybinii x S. tuberosum and S. chacoellse x S. tuberosum species hybrids, quadrivalent frequen­cIes were reduced in the S. ehe x wo hybrids, sug­gesting low levels of genomic differentiation. To account for observations like these, Dvorak (J 983) suggested that the genotypes of the di ploio species act in a way to suppress heterogenetic painng. The segregation data for n04 reported in this paper may supply evidence (0 support this proposal of genomic differentiation. Further work mUSI be done with other accessions of S. chacoense and other diploid species to ascertain the significance of this question.

REFERENCES

DOUCHES. D. S. A D QU!RI)S. C. 1'. 1987a Addillonal'SllZyme loci in tuber-bearing SulallulI1s: Inherllance .nd linb~c

rclat;on,h;p, J. Hemin)' (m press\ OOUC HeS.lJ ~ ,'N[)QL'IRl)~.' I- IQR7b. L,;.:o14x-2" crosses

to determIne gene-cenlrOJlJ('re Iliap di~l:ln("s or "01.yme loci in Solanum species. Can. J Genel. C)'IQI (Ill press)

[)VORAK. J 1983. Evidence ror :~enelJc suppression or heterogenetic chromosome paIring in polyploid species of Sol~num seCI PeIOiG. Can.} Genel C'yfOl.. 25. 530-539.

HANNEMAN. R. E. AND PELOQUIN. S r 1968 Plord~ lev~l~ or progeny from diploid· tetraploid <'fosses In pOlato. Am. POlalO J.• 45. 255-1'>1

HA WK ES. J. G. 1978. Dlosystenwucs or the Potato. Harris. P. M. (ed } In Ti,e POlata Crop "h,' \clen/i/,( basis of improve­meill. Chapman and II all. 1. ,.,ndon. pp. J5-67

HeRMSEN. J (, TH 19840 Mechanisms and genetic Implica­tions of 2{1-g~mcte fOrmation lo .....a Slal~ 1. Res.. 58,421­434.

HhRMSI:N. J. (,. TH 19841> The pOlell(Jul of meiotic polyploldll ­ation in breedIng allogamous cwps Iowa Slale}e Res., 58. 435-448

HOWARD. H W 197(). Genetics of the potdlO. Spnnger-Verlag. New York

IW'\N,<\C,A. M 1984 Di,covery of a ,\'naptlc mutant ;n potato h~ploids and ils usefulness fn; pOlato breeding. T1,eor, Appl. Genn. 68. 87-93.

I""A GA. M. PeLOQ I . $ J 1979. SynaptIc mutant affecling only megasporogeneSIS I/l pot<Jtoes J. Hered.. 70,385-389.

191

JOHNSTON. S. A.. HANNEMAN, R. E. IK. ANI) RU~IDt,. R. w. 1981 l<kntlfivtion of an asynapllc variant rn Sola"um mmmer<oni, 4m PO/OIr, J.. 58. 506

JON(,EDIJ!<,. E 1983. SeleCt;0n for first dlv;s;on re~tlt\J!lon

2n -egg formation in diploid potatoes. Po/alo Res,. 26. 399 KODURlJ.P R K ANDRAO.M.K.198I.C)'togeneticsorsynap­

lIC mulanlS in hi~her plan IS TIiear Appl. Genet.. 59. 19~

214. MARKS. G. E. 1966. The enigma of triplOId potatOes. EU(Jhym'a,

15, 285-290 MARTI NEZ.APATER, .I. M AND OLIVI ER, J L 1984 Gerwtic

analysis of rsoz)'me loci In tctraplold potatoes (Solllnllm IUbemsum t) Gene/irs. 108.669-679.

M.'\SSO, _M 1985. MapPl/lg. Combmlllg abilities. herllabtlilJc., and hcterosis with 4)< -2x crosses in potato PH. D Thesis, Universit), of Wisconsin. Madison.

MOl<.. D W S. AND PELOQUIN S. I 1975 Three mC'c:hanisms of 2n pollen formation in diploid potatoes. Can J Genel ('vIQI., 17.217-225

MOK. I. G. AND PELOQUIN. ~ J 1982, Sexual polyplo\dll"llon and protem diversit), in !"olatoes. Am. Po/alo 1.. 59.480

OKWL!AGWU. (.0 ANI) PU OQUIN. S. J 1981. A method or lransferring the intact partonlal genOlype to the oHspnng VIa meiotic mUlants. Am, Poralo 1.. 58, 512-5)3,

QUINN. A. A.. 11,10". [) W S AND PELOQUIN. s.J.1974, Distrrbu­lion and Significance of dlplandroids among the diploid Solanums. Am, POIQ/O J" 51. 16-21.

QUIROS. (' F. 1981. Slarch gel electrophoreSIS technIques uscd with alfalra and other Medl('aso species. ('an. J. P/(Jn(. SCI .. 6J,745-749.

QUIROS. ('. F ANI) MC IIAt'. N 1985. GenetIC analysIS of Isozym e vanan IS in d ipi ooid and Ielr;'1Jloid pot aloes, GenellCS, II 1. 131-145,

RAMANNA. M S 1979. A re-examinatIon of tllc mechanisms of 2" gamete formatIon In pot"to and IlS Implications ror breedin~: Euphylica. 28. 537-561.

RAMA'."'A. M s. 1983. First diVIsion restitution gametes Ihrough renile desynapllc mulanl, of pOlalo Ellph,'r;ClJ" 32. 337 -350

RAM"NNA, M. S. AND HERMSf"' . .1 G n,. 1979. Genome Relalionships In tuber-bearing Solanums Hawkes. J. G.. Lesler. R_ N. and Skeldlng, A. D. (cds.) In nle bIOlogy and (oxonomyoJ Ihe Solanceae AcademiC Press. Nt:w York. pr. 647-657.

RIC K. l. M, 197), Some cytogenetic feature, or the genome In

dr plOIO pi ant speci es Siadier Symp".<IQ. Vol, , and 2. I 53­174.

SWAMINi\THAN, M $.1953 Studies on the imer-relatlonshlps belween I;lXonom\csencs in the section TuberarwnI, Genus SQlalwm. 1 CommersonlQna and Tul>erosa Am PalalO 1. 30,271-281.

gWAMINAn-r,"-N. M S.. MAGOON, M L .\"D M£HRA. K. L 1954, A si m pie prop;on;c-carmmc PM l smea r met hod for plants witl1 small chromosomes, Indian 1. Genci Plum Breed.. 14,87-88

VALl 1.105_ ('.' 1983. Enzyme act,,·uy stallling. {eds.) Tank.sky. S. D. and Orton. T. J. 10 Isozymes in plaltl generic) and breedllig Vol A. L1sevler. AmSlerdam. pp. 469-516.