Exploring the Role of x = 7 Species in Brassica Evolution: Hybridization with B. nigra and B. oleracea C. F. Quiros, 0 Ochoa, and D. S. Douches

From the Department of Vl'~et~ble Crops, Univers,ly of Caillornia, Davis. We are Lndebted to Leslie CottlLel.> and Kawyoshi Hosaka for ",,,eful suggestions and d,s­cussion; to K. Hinata, C. Gomez-Campo, and Paul Wil­liams for supplying the seed for the accessions USl";,

in this stvdy; to Martha (rouch for prOVIding the B napus genomic library; to VLn,-,- D' Antomo, Janet Mar· tinez, Sh"hriar Kia"ian, and M[ll'1L ,\il'C,rath lor tech­nical assistance: and to Jane Johnson lor lvping the manuscript. Address reprinl requests 10 Or. Qu:ros. Department 01 Vegetable Crops, University or C"lifOT­ma, Davis, CA 9S616

JoUTOal of Heredlty t 9&J;79;~" 1 -bH. 0022-1503,138/$2.00

The role of the x' 7 species, B. adpressa and Diplotaxis :;:rucoides in Brassica evolution, was investigated by hybridizing them to the cultivated species B. nigra (.'It" = 8) and B. oleracea (x = 9). In general, the hybrids displayed a low frequency 01 bivalents and unbalanced reductional division during meiosis, resulting in poor f~i­

tility. This cytological behavior was also observed in hybrids between the .~ = 7 species, indicating strong genome divergence of both genomes. Among all the hy­brids, those of D. erucoides x B. nigra had the highest level of fertility, permittinfl the generation of D. erucoides-B. nigra allen addition lines. One of the hybrid plants variegated for yellow-white-red petal color and red pigmentation lor other org"ns. The same enzyme activity zones were found to have multiple bands in all the species including those with x = 7 chromosomes, suggesting possible gene duplications. These mUltiple banded patterns persisted in pollen leachates and Viere transmitted to the hybrids. The hybrids reported in this study provide additional information on intergenomic relationships among Brassic8 and related species 01 allied genera.

The genomic number in Brassiea ranges 7 species in Brassica evolution, we report from x = 7 to x = 19.':'·2' Most or the cy­ the hybridization of two 01 the';(, sl hodes, togenetic evidence' '3,'1 supports Calche­ B odpressa and D erueOi des, to the culti ­side's and Robbelen's hypothesis U1 that vated diploids 8 nigra and B oleracea. the basic genomic number in Brassica is x = 6 hecause of similarities in chromosome

Materials and Method~morphology and autosyndelic pairing. Tllere is no record, however. on the past Plant MaterIals or present existence of x = 6 Brassica The accessions lor the species B adpressa, species. The lowest genomic number ob­ Diplotaxis erucoides. B. nigra, and B ole­served in Brassieo and related genera is x weea used in this study are listed in = 7, which is found in the species B. ad­Table I. pressa Moench. (Hlrschfeldia incano (L.) Lag.-Foss,), B, deflexa Boiss., Diplo/axis HybrIdization erucoides (L.) DC, D. virgmo (Cav) DC, D. The following crosses and reciprocals were aens (Forks,) Boiss., Erueastrum virgatum attempted: D erueo/des x B. adpressa, D. Pres!., £. varium Durieu., ar.c1 Synapsis au­erueoide.s x B. mgra, B. adpresSQ x B. nig­chen (Boi55.) O. E. Schulz.~·L3.1i Thus, the ra, and B. odpressa x S. a/erace(/ Crosses lineage represented by these species can were per/armed in the greenhouse by hand be considered ancestral to the lineage with emasculation and pollination. We rescued higher chromosome numbers, including ovule embryos from hearl-shaped stages the cultivated diploids B_ nigra L., B. ole­to fully developed stages and cultured thnl racea L., and B. campestris L. Only a lew i:l Nitsch and Nii,;ch medil.·.m without hor­reports dealing with the hybridization and mOnes. L~ We transplanted the resulting chromosome pairing relationships of the p~anllets in soil and transferred them to a x - 7 species with other Bras$ica diploid greenhouse. species exist in the literature. To OUT

knowledge, the following hybrids have Hybrid Conftrmaiion been reported: B. nigra (x - 8) x Erucas­We continued the resulting plants from the (rum virgowm," B. eampeSf!'is (x = 10) x crosses as hybrids by isozyme marker', B. adpressa, and Diplotaxis erueo/des x B and/or Southern blotting analysis of ri ­oleracea (x = 9). LA In an effort to gain fur­ bosomal RNA genes. for the isozyme ther understanding on the role 01 the x- markers, horizontal starch gl'1 electropho­

351

Table J. Acce nil and lIpedell used In lbe hybridization lIt8:dJes

Accession ~pecjes Origin

INIA 1235-67 OiplOlo.:.:is erocoides Teruel. Spain

Ad-ll4 Brassica adpresso Marin. (A

CRGC-02 Brassica nigro rapid cycling"

CRGe-03 Brassico olemceo rapid cycl ing"

resis waS used to separate the enzymes obtained from young leal crude extracts. Details of this technique have been de­scribed elsewhere.23,26 The following en­zymes were assayed: 6-phosphogluconate dehydrogenase (6PGD), phosphogluco­isomerase (PGI), leucine aminopeptidase (LAP). phosphoglucomutase (PGM), triose phosphate isomerase (l'I'!), glutamate ox­alacetate transaminase (GOT), and malate dehydrogenase (MDH). Phenotypes and cellular localization for most of these en­zymes have been reported by Arus and Orton 2 and Quiros et al. 26 lor B, oleracea and 8 compeslns. Pollen leachate/leal electrophoretic pattern comparisons were performed in the parental species to de­termine duplicated loci following the cri ­teria described by Gottlieb. 1O

For the Southern analySIS, the technique used by Quiros et al. was employed,2<1 Total genomic DNA was iSQ;ated from leaves of individual plants i~nd digested with the ell ­donuclease EcoRI. We probed the mem­branes resulting from Southern blotting wilh rONA from 8. napus (clone 9-4), (M. McGrath, unpublished data) extracted {rom a lambda (Charon 35) genomic library kit supnlied by Dr. Marlhil Crouch from In­diana University, We sized DNA fr;,,~ments

usinrt Hind IlIlu;obda DNA digests. Radish DNA was used as a control because the rRNA genes of this Brassica cJose relative are well characterized.s

Cytological Studies We fixed flower buds in propionic acid: absolute ethanol (I :3) containing ferric diloride as a mordant,26 Arter 24 hours we rinsed the buds and stored then: in 70% ethanol. Ant hers were dissected and squashed in a drop or 1% acetocarmine. We based chromosome counts on at least 20 cells, unles~; otherwise specified. Pollen viability was estimatl'u oy pollen slain­<l,bility of at least 100 grains in 1% aceto­

'; armine

Results

The lallowing is an account of the char­acteTlstics 01 the hybrids obtained after

352 The Journal of He1edily 198879(5)

..B

.. ,. .­

"

o

F Figure I. Pollen mOlller cells for the spedes used In ltllS stully and rhen hybrids (A) Melaphas~ I cell of B udprI'.'.lO 2n ~ 2x ~ 14, w]\h three pairs assoclated in a hexavalenl (6) lli~kmeSlS (".'1 of Dip/maxIS emeO/des, 2n

2.< 14, (C) Metaphase I (ell from D em(O/des x B. adpresso hybrId displaying 1<1 I~ (D) Unbalanced tetrads ami nllC"':Iuclei in the same hybrid. (F:) Metaphase I cell of hybrirl B nigra x B odpres.'i(J with 2n r 15 chromosomes associated III I "' (arrow). 511 and 21. (F) MClaphase I cells (top and, "oler) Irom B oler(J('co x R. adpresso hybnds dIsplaying 16 Is.

crossing the various species involved in to confirm chromosome numbers. These the study, Before executing the hybridiza­ agreed with the expected numbers {or each tions, we per/armed chromosome counts of the species (Figure I, A and B). Some in several plants of the parental species of the plants of B. adpresso, however, had

Table 2. Chroroowme 8sso<;latloos In diakinesis and melaphQ8(, II tor four hybrids

Balanced

8ivaleol' on dlakmesls (%)" dIstribution io metaphase

Hybrid" N 0 2 3 4 5 6 8 X 11(%)'

D x adp 23 17 4 9 17 22 22 9 0 3.2 7 (7-7) D x ng 19 6 0 6 6 32 27 17 6 4,0 80 (7-8) ng x atlp 13 15 i5 8 IS 8 30 8 0 3.0 12 (7-8) 01 x adP 25 8 16 24 24 16 8 4 0 0 2.6 n (7-9)

" D D, erucoides; adp ~ 8 udpresso: ng B nigra; 01 ~ B olerQceo , R, 'LnamHlg chromosomes were unovalent. , OlSlnbu\,on 01 expecled genomic numbers for each SpecleS in eaell pole Expected tI\stribulLon of ~enomic

numbers ,n opposite poles

reciprocal translocations involVing three chromosome pairs, judging from the pres­ence of hexavalents (Figure IA) ancl re­duced pollen fertility. For the hybridiza­tlons, we selected B. adpressa plants (ree of chromosomal aberrations,

The absence of isozymes in pollen leachates from all the species for the most anodal electrophoretic zones, TPI-l, 6PGD­1. MDH-l, PGI-I GOT-I, and PGM-I, sug­gested that these are located in plastids. 'o

Conversely, the isozymes of lOnes 6PGD­2, TPI-2, MDH-2, PGI-2, PGM-2, and GOT-3 expressed activity in pollen leachates and leaves, disclosing their location in the cy­tosoL

Dip/otaxis erucoides x B. adpresso A total of 17 seedlings were obtained Irom th is cross when using Diplotaxi.s as th e pis­tillate pa rent. The "allen viability ob­served in these plants was very low, rang­ing from 0 to 8%. All had lhe expected chromosome number of 2n "" 14. For the determination of chromosomal associa­tions i, meiosis, we pooled the counts of five plants because of the frequent degen­eralion of pollen mother cells. The num­ber of bivalents observed in diakinesis ranged from zero f) six (rigure Ie), but most of the cells had three to five, with an average 01 3.2 bivalents per cell (Table 2). We observed no cells with seven bivalents. The most common configuration in meta­phase II was six and eight chromosomes at each pole. One out of 15 cells was found to have the balanced seven to seven con­figuration (Table 2). Among other meiotic abnormalities, we observed oremalure equational division in some 01 the cells. Unbalanced tetrads were commonly seen in telophase ll. often displaying micro­nuclei (Fic;ure 10). We atso observed dyads all( 'riad s at this stage,

The hybrids were phenotypically atike. They had white nowers like the Diplolaxis parent hut with yellow throats, most likety coming from B. adpressa, which has yellow nowers. The plants were intermediate for several other characteristics. such as pu­bescence of nower buds (Figure 2B), prominence 01 stigma and position of pod constriction (Figure 20), and angle of pod attachment 10 the rachis (Figure 2C), They combined the larger leal size of 8, adpressa and the larger nower size of D. erucoides. The red dot in the anlhfr tips 01 B ad­pressawas manifested in the hybrids. Both the leaf asymmetry of B. adpressa. char­acterized by odd-pinnately parted leaves, and the incised leaves of Dip/a/axis, ex­pressed in the hybrids, resulted in asym-

FIgure 2. Morphology of various organs in Ihe species slud led and their hybrids, (A) I.<'aves of D, em,mdes (top), 8 m/pressa (bot!om), and hybrid (center), (B) flower buds 01 D, eruw,d~s (Ielt), 8 odpressa (right), and hybrid (middle) (C) RachIS of B. adpressa (lell). D era(oide.' (right), and hybrid (mIddle) CD) Slyles of 8 adpressa (lell). D, eruco,d~s (nghl), and hybrid (mIddle), showmg stIgma SOle and posillon of style conslrlcllon (asterISks) (E) Flowers of D. erucoides (lelt). 8. nigra (right), and hyhrld (""ddle), (F) Leavl'" of B nigra (left), B, adp.-essa (nghl), and hybrid (middle), (G) Styles of B. adpressa (lop), 8, oleracea (boltom), and hybnd (,enler), showing constrictIons (asl"risks) (H) L"aves of B adpri!ssa (Jeft), 8. o!eracea (right), and hybrid (mIddle) (I) flower buds 01 B olera,-ea (Ietl), B. adpressa (righl). and hybrid (mIddle)

metrically incised leaves (Figure 2A), hybrid origin (Figure 3). III general, the The electrophoretic analysis revealed zymograms for E. adpresso were more

isozymes characteristic of both parental complex than those of D erucoides(Figure species in these plants. confirming their 3, Band 0). In B. adpresso. multiballded

Quiros et al • BraSSIca Species Hybnds 353

A

TPi-l

Tpi-2

HI ad I

Figure 3. Zymograms (or lou, enzymes in D. emcordes (D), H odpresso (ad), and hybrids (H) Two lines per plant unle,s otherwise specIfied Anode is above. (A) Phosphoglucomutase. PGM-2 Note segregatton [0' the slow iso;ryrn'" ,,( B. adpresso. (13) Malate dehydrogenase. MDH' Note the hJ~t1er complexity o( the f! adpressQ [lhenotype <)Ycr Ihal o( D, erlJ(oid,,~ Th~ additIonal bands of B adpre>so were present in the hybnds (C) Tuos'" phosphate ISomerase, TI'I. Parental specIes dJfiere<:i only for the TPI- t plastid lso;rymes No segregation was observed for the three-banded phenotype of the TPI·2 cytosolic Iso;ryrnes. suggesting duplicated loci (D) C· phosphogluconale dehydrogenas~. 6PGD-2 Cylosohc lSo;rymes only show". The muillband~d phenotype of B adpressa is transmitted to the hybrJd progeny.

phenotypes lor the GOT and MDH and for except for the plastid isorymes of 6PGD-1 the cytosolic isozymes GPGD-Z and TPI-2 (Figure 4A). The hybrids segregated for were n')served in both pollen leachates one PGM-2 (Figure 3A) and one GOT-3 and leaves. Furthermore, these multiple isozyme from the B odpressa parent. bands did not segregate. They expressed As expected from the cytological obser­in all the re~IJlting hybrids, indicating that vations, the hybrids were steri Ie. Sel fings, these enzymes lItay be coded by dupli­ sibs, and crossing to both parental species cated loci. D. erUCOIdes showed mufti­ at either mature nower or bud stages failed banded phenotypes for the ~ame enzymes to yield any seed or culturable ovules. Sinee

both pare ntal species are selH -leO ITt pati­bie. the hybrids were expected to be self· incompatible This assumption could not be tested due to their sterility. Alter chro­mosome doubling with colchicine, the hy­brids became fertile, producing seed from seltings

D. erucoides x B. nigra We obtained a total or '2 plants from this cross alter ovule culturing. The reciprocal cross also yielded cufturable ovules, hltt these were not grown to full plants. Pollen viability was low, ranging from 2.5 to 10%. All the plants had the expected chromo­some number of 2n = 15 Studies on chro­mosomal associations revealed a maxi­mum 01 seven bivalents, although the most frequent associations were four and five bivalents, with an average of 4.0 per celf (Table 2). In melaphase II, the most com­mon configuration was seven and eight chromosomes at each pole. Similar to the previous hybrid, some cells of the D. er­ucoitles x B nigra hybrid displayed lag­gards in anaphase L Premature equational division of chromatids was also frequently observed. Dyads, triads, and unbalanced tetrads were also evident in telophase II.

Morphologically, the hybrid plants had a similar appearance, with Ihe white now­ers of Dip/ataxis but with yellow th roats, comparable to those observed lor the pre­vious hybrid. The petals had a claw of in­termediate size when compared to those of B. nigra, which had long claws. and to Dip/%xis, which had ses;;ile petais (Fig­ure 2E). The leal morphology of the parents was not distinct enough to be di­agnostic for the hybrids. They mostly dif­fered in size, with the B. nigra parent hav­ing smaller leaves. The hybrids had the lighter green leaf color of the B. nigra par­ent and were intermediate lor leal size and for site of pod constriction. An exceptional hybrid plant was observed, displaying white-yellow-red variegation in the petals and red variegation in other organs such as anthers, style, leaves, and stems, form­ing stripes or sectors (rigurf .J). Most of Ihe Oowers i:l this plant were white, and the variegated ones were dispersed throughout the plant. The three-color var­iegation could be observed in the same petal or independently in the petals of the same or different Oowers. The yellow pig­ment is normally observed in the petals 01 the B. mgra parent, while the red is observed in the sterns and the veins of the senescing petals and leaves of Dip/o/axis. NOlle of these pigments variegated in any 01 the plants 01 the parental species.

354 The Journal of Hered,ty 1988 79(5)

lectrophoretic analysis confirmed the hybrid origin of the plants resulting from this intergeneric cross, which displayed the isozymes of both parental species (Fig­ure 4). The multiple banded isozyme pat­terns found in D. eru(oides and B. adpre.ssa were also present in B. nigra, including those coding for the 6PGD-I plastid iso­zyrnes. They were expressed in the hybrid without evidence of segregation (Figure 4A).

Because of the low fertility of the hy­brids, no seed was obtained from selfings. However, a plant with 30% pollen stain­ability and 2n = 20 chromosomes was ob­tained by embryo culture after backcross­ing one of the hybrids to D. erucoides. This plant had the same phenotype as the hy­brid for the isozyme loci tested, except that the activity or the Dl'plolOXIS isozymes was stronger ill the former because of the higher dose of Dip/o/axis alleles (Figure 4A-B). The most common chromosomal association observed at diakinesis in this plant was 811 + 41. Although it resembled the Dip/o/axis parent morphologically, it had some abnormal characteristics such as thick .and deformed leaves and the flat­ler pods typical of B. nigra. The presence of 8. nigra isozymes for the enzymes 6PGD-l and 6PGD-2 (Figure 4A), MOl :-2 (Figure 4B), TPI-2, PGM-2 (Figure 4C), LAP-2, and GOT-3 indicated that the chromo­somes carrying these marken> were pres­ent in this plant. Sixteen plants were ob­tained after selling by bud pollination. or these, four were diploid (2n = 14) and 10 were hyperploids with 2n ~ '\ 16, and 20 chromosomes (Table 3). The chromo­some numbers of the other two were not determined, The extra chromosome in the 2n = 15 plants was observed either as a univalent or associated ,JL a trivalent in metaphase l. The frequency of these as­sociations differed for each plant tested, indicating differential chromosome ho­mology of specific B nigra chromosomes with those of D. erucoides (Table 3). lim­ited observations of cells in anaphase I indicated that the disjunction in these plants was normally seven to eight at each pole, Some of the plants had specific B. mgra isozymes ill addition to those Irom D. erucoides, suggesting that the exira chromosomes ori~inated from B. nigra (Figure 4). These U erucoldes-B. nigra ad­dition lines disclos~d synteny tor the loci coding lor GOT-3 anu LAP-2 iSQzymes and for 6PDG-I and 6PDG-2 (Table 3). The fer­tility of the trisomies was quite high, rang­ing from 60 to 85'Y" when:as in the diploids it was close to 100%. Most of the plants in

H B no

H B n

Mdh­

o n

pgm-2

c FIgure 4. Zymograms of D erucOIde> (0) x 8. nigra (n) hybrids and add'llon plants oeriwd afler hackcross,,,s the hybrid to Diplolax,s. Two hne, per plant unless otherwise sp"dfied Parental species (D, n), hylJmls (H), and backcross (13) derjvallves shown at the teft secti,)O of A and B Arrows show 8, nI!lra lsorymes in addu,ons al the right sectIon of tllese figures. Anode IS above, CA) 6-I"losphogluC(Jnale dehydrogenase, Slngle·banded phenotype oj Dlplor(Jx,s parent suggestlllg Stw!l<' locus Add;hons Jor cyIo,;r,ILC 6PCD·2 'So,\'Jnes d"playing a faster band of lJ. nigra One ILne per plant (B) Malale dehydrogenase Add ition carrying ~ylosollC MDH-2 Isozyme from B fHgm One hne per plant (e) Phosphogl":'<l111ulase isozymes (PGM-2) From parental species, hybrids, and add,tlons Two Jines per plant. Slower PCM-2 ,ylosohc band ,s ~pecJlic for IJ nigra.

this progeny, induding the diploids, set of six bivalents or five bivalents plus one seed spontaneously, indicating self-com· trivalent was observed in diakinesis, with patibility, which is absent in both parental an average of three bivalents per cell (Ta­species. ble 2, Figure I E). Laggards in anaphase [

and uneven chromosome distribution were B. tligra x B. adpressa also the rule for these plants. The leaves We obtained four plants by embryo culture 01 both hybrk' ;)lants were similar in sym­after using B, nigra as the pistillate parent, metry, shape, and light green color to B but reciprocal crosses were unsuccessfuL nigra (Figure 2F). The nowers were yellow Only two of them survived. These had 0% and had the prominent stigma of 8. nigra, pollen stainability and the expected num­ while lhe style had a constriction inter· ber 01 2n 15 chromosomes. A maximum mediate between the two parental species.

Ou,ros et al • 8raSSIC<l Species Hyb"Cls 355

Figure 5. Red-yellow variegation In nowers of an inlergeller,c llybfld 01 Drplolax" ecucoides x 8rassiw nigra. Note varlega(ion in (he petals and olher organs, ,nclud\ng style. anthers. filaments. and sepals

The pods resulting from sell-poll inations of the B. a/eroceo parent (Figure 2H). Their firmed the hybrid ongtn of these plants

ill chromosome-doubled branches had the flower buds were elongated as in the B (Figures 6 and 7). They had the isozymes width and Oatness characteristic of the B odpresso parent (Figure 21). and rONA fragments of both parental wgraparent. The enzymes LAP, TPI, MOH, Bot h isozyme electropho resi sand species induding the 0.9 and 0.2 kilobase 6POG, and GOT were diagnostic for con­ Southern analysis of rRNA genes con- (Kb) fragments cOJltributed by the male firming the hybrid origin of these plants,

B. oleracea x B. adpressa We obtained Ihree plants from this cross Table 3. Chromosome flU III ber and a&\loclallons In me'aphase tin D, erucoMe8-nlgro Olonosomlc

additIon Hnesusing S_ oleracea as the pistillate parent. The plants were ve ry vigo rous and d iffe red Melaphase I

from each other in pubescence and inten­ (%y Fertlllry sity of foliage color. All three had the white Plant 2n UnJV Trlv, B, "'8m markers (%)'

flowers characteristic of the B_ a/emcea parent, but with pale yellow throats. Two

212-2 212-3

15 15 27 73

LAP-2, con 6PGD-l, 6PGD-2

83 64

of these hybrids had the bluish green fo­liage color of the 8 a/eroceo parent. The

212-5 212-6 212-8

15 15 15

80 20

LAP-2. GOT-2

S6 82 77

foliage of the other was lighter in color, 212·10 15 69 29 MDH-2 8-4

intenned iate between its parents. All three plants were intermediate in flower size,

Z12-1 I 212-12

16 20

6PGD-!. 6PGD-2, 1~\P-2

TPI-l. COT-2, LAP-2, PGI-2. 6PGD-\. 6PGD-2, PGM-2

pod constriction (Figure 2G), pod attach­ment to the rachis, and leal morphology.

212-13 212-15 212-J6

15

16

25 75 TPI-! TPI-!

77 77

The leaves of the putative hybrids had evenly pinnated leaves with a prominent "Association of the addilionaf chromosome as unovalent (Univ_) or ,,;valen( (Triv,) terminal leaflet, both traits characteristic "% poilen stainabIlity,

Figure 6. Zyrnograms tor 8 Qleroceo (01). 8 adpre.<so (ad). and hybrids. Two lines per plant Anode .sabove. (A) Malate dehydrogenase (MOH-2). Confirming in­terspecific hybrid, H, and backcro,;s 10 8 adpre:;.su derivative. B. (B) 6PGD-I. plashd isozymes, and 6PDG­2, cytosolic isozymes Both mlgration zo"~s suggest duplicated loci. Interspecific heterodimer lor 6PGD-l (asterisk) and possibly the slow 6PGD-l B olerccev isozyme (circle) are lost in backcross derivative. (C) Phosphoglucoisomerase. Cytosolic PGI-2 isozymes confirming interspecific hybrid.

parent, B. adpressa Diagnoslic isozymes thai confirmed the hybrids were MDH-2 (Figure 6A.), 6PGD-I, 6PGD-2 (Figure 6B), and PGI-2 (Figure 6C).

Chromosome counts in the putative hy­brids revealed the expected number o( ~n;= 16 (Figure I F). The plants were highly sterile, with less than 1% pollen stainabil­ity. The most frequent chromosomal as­sociations at diakinesis and metaphase I were two and three bivalents, with a max­imum 01 six and an average of 2.6 bivalents per cell (Table 2). During metaphase JI seven to nine was the most common con­figuration. The moSI unbdianc:ed configu­ration observed during this stage was one to fifteen. No cells with an eight to eight distribution were observed at this stage. Laggards and premature equational divi­sion was often observed in anaphase I. Un­balanced tetrads dyac.'.~., and triads were also observed in inese plants.

'rhe sterility 01 the hybrids precluded obtaining selfed seed. Doubling 01 the

KB H

0.9

0.2

Fill'-lTe 7. rONA restrlClIon fragments, 0 9 and 0.2 Kb from 8. odpresso (all) are evirjenl on the B. oleracea x B. odpressa hybrids D,plmax,s (D) has the same rONA pattern as B. olerocea (01)

chromosomes by colchicine did not re­store the (ertilty of the hybrids. However, a culturable ovule was also obtained after backcrossing the diploid hybrid to B. ad­pressa. This plant had 2n i8 chromo­somes and was completely sterile. It did not produce any seed alter selfing or back· crossing it to any 01 the parental species. Morphologically it looked like B. udpressa except lor darker foliage, anthocyanin pig­mentation in stems and leat veins, and greater vigor. The zymograms at this de­rivative are shown in figure 6, A and B.

Discussion

The high sterility and meiotic abnormali­ties observed in the D. erucoides x B. ad­pressa hybrids suggest that the genomes 01 two x· 7 species, De and Ad,11.18 have diverged considerably tram each other Tile high (requency 01 univalents, unbal· anced disjunctions in anaphase I, and the complete sterility in the hybrid provide an effective breeding barrier, precluding gene flow between these two species.

Our limiled cytological survey of the hy­brids involving the cultivated species, B. nigra and B. olerocea, and the x= 7 species, D. erucofdes and B. adpressa, disclosed, on average, four or less bivalents. The pres­ence o( all unlvalents in 6 10 17% or the cells, depending on the hybrid, indicates thaI the genome affinity of these species might be equivalent to that observed (or B. nigra and B. oJeracea.n Among these hybrids, D. erucoides x B. nigra displayed rhe fewest meiotic abnormalities and enough fertility to produce progeny, re­

suIting in the generation of the alien ad­dition lines.

Taking into account the hypothesis stat­ing that the elementary monogenomic species are secondary polyploids derived from an extinct species 01 x 6 chromo­somes,?··, the number o( bivalents ob­served in the hybrids would be expected to include autosyndetic as well as allosyn­detic pairs. Although priority often is given to the occurrence 01 autopairs over allo­pairs in discussions of chromosome ho­moeology among the Brassica r.:e­oomes,3-5.1S the opposite might be the case. Since the genomes of the cultivated mono­genomic species are assumed to be of re­cent origin, their chromosomes are ex­pected to have higher intergenomic homoeology than intragenomic homoeol­ogy, The events leading to the develop­ment of higher genomic numbers by the generation of additional chromosomes Irom the x = 6 archetype, such as rear­rangements and aneuploidy,' might have taken place very early in Brassica evolu­tion. Thus, the conservation of chromo­some homoeology in the anCestral x ~ 7 genome or the autosyndetic capacity of its chromosomes is expected to be less than the allosyndetic ability o( the chromo­somes constituting the more recently de· rived genomes 01 8, 9, and to chromo­somes. Furthermore, in the absence of a mechanism for pairing suppression of homoeologous chromosomes in Brass/ca,o multivalents would be expected to occur at a higher frequency than that occasion­ally observed (or the elementary mono­genic diploids. I Although several attempts have been made to quantify the degree 01 auto- and allosyndetic pairing in Brassica, either by digenomic triploids,5 tetra­ploids,18 or haploi ds tests, J the results are not always conclusive because of bOlh the possibility at non-homologous pairing and putative genetic (actors limiting the degree 01 pairing.3.• Therefore, assuming that the pairing observed in our hybrids is pre­dominantly al\osyndetic, rhe cytological and fertility data indicate that the D. eru­coidesgenome (De) has higher affinity than that of B. adpressa (Ad) for the B. nigra genome (B).

The appearance: 01 three-banded phe­notypes for the cytosolic enzymes in the pollen leachates from all the species in­volved in this study reveals duplicated loci,lOII with the middle band correspond­ing to the interlocus heterodimer. Evi­dence o( duplications lor the plastid iso­zymes is disclosed by the true breeding behavior of three-banded phenotypes.

QUirOS at al • BrassJCiI. Species Hybnds 357

The presence of the same duplicated loci in the two cultivated as well as in the x =

7 species demonstrates that the genome of the latter has already suffered dupli­cations presumably due to repatterning events leading to genome multiplicity. Several models have been proposed to ex­plain the origin of duplicated loci: tandem duplications arising by unequal crossing over, transposon-mediated processes, and overlapping reciprocal trans locations fol­lowed by hybridization and selfing_ lo.2o

Based on the high frequency of translo­cations observed in Brassicd 2 and on our finding of translocations in B. adpressa, it is possible that this type of aberration was one of the mechanisms originating dupli­cated loci in Brassica. Although the self­incompatibility of the diploid Brassica species diminishes the yield of structur­ally homozygous individuals for the trans­located chromosomes, they may still arise by sib crosses.

The origin of the unusual white-red-yel­low petal variegation in one of the D, er­ucoides ,x. 8, nigra hybrids is unknown. Our attempts to obtain progeny from this plant failed. We can only speculate thaI this variegation may be due to somatic chromosomal abnormalities representing genomic modifications caused by their ba­sic incompatibilities, This can result in chromosome losses, the generation 01 chromosome fragments, and the presence of bridges in somatic ilnaphase as report­ed in Nien/iana interspecific hybrids, L6

These events produce somatic cells homo­zygous for the recessive gene determining yellow petal pigmentation that generate sectors or streaks of the same color. The presence of anthocyanin pigmentation in the petals and other organs could be ex­plained in the same manner, although it is more difficult 10 conceive because red pip-mentation is a more compl~x trait. It is manifested as a coduminant trail in part of the stems and oodes and the stigma, However, this pigmentation is never man­ifested in the petals, style, or sepals of the parental species, with the exception of se­ne~dng petals which sometimes display a uniform light pink coloration far from the intensity observed in the hybrid. It could be that a gene controlling the expression of this trait was lost in some o( the cells due to somatic chromosomal abnormali­ties, resulting in the odd expression of the red pigment. Another possibility is that

the activation of silent movable elements alters the regulation of gene expression as a response of the genomes to the stress caused by the wide hybridization,l. This Situation resembles the action of the Dot­ted eDt) element in maize, which causes the mutation of the nonlunclional a allele into new ones that allow the synthesis of anthocyaoin in plant and kerneL'"

The hybrids obtained io our study serve to complete the diagram 01 Mizushima;~

depicting intergenomic retationships among Brossica and related species of al­lied geoera. The problem of identifying the putative ancestral species of the mono­genomic diploid species remains un­solved, By comparing synteny relation­ships of duplicated loci in the x 7 species with those diploids of higher genomic numb<'rs, it may be feasible to determine which species are ancestral.24 ,25 This will make it possible to test the hypothesis that states that evolution in 8rassica occurred in ascending order, that is, (rom lower to higher genomic numberS,2l.27 In order to trace unique events leading to gene du­plication in x = 7 species, it will be nec­essary to constmct alien addition lines to dissect and compare the chromosomes constituting their genomes,

Refercoces

I Armsrrong. K, C.. and W, A Ketler Chromosome pamng In haplOids of Brass;co Qleraceo, Can J, Gpr,el Cylol 24 735- 739. 1982,

2 Mus. P. and T, J, Orton IlLheritance and hnkage relalLonships of Isozyme 10Cl m BrosslCo ol"rac€a J ll<;red 74-405-412,1982,

3. AUia, T , and G, Rf)bbelen, Cytogenetic relatlOnshlp Within cultivated BroSS/co analyzed ill amphlhaploids from the three diploid ancestors, Can, J, Genet Cytol 28:323-329. 1986,

4, AUia, T" and G, Robbelen MeiotiC palling and hap­loids and amphidiploids of spontaneous versus syn­thetic origm on rape. 8rOSSICtJ ntJpus L Can. J. Genet. Cylol. 28:330-334. 1986

5 AUia, T.. C. Busso, awl G. Robbelen, Digenomic trlP­10ids for an assessment of chromosome relationships In the cultivated diploid Brass/co species, Genome 2') 326-330, 1987,

6 Busso. C. T AUla, and G, RObbelen, TrigenomlC combmations lor the analysis of 1L"~I01ic control m the cu[(,valed Brossico species, Genome 29:331-333. 1987

7 Calcheslde. D, G The chromosomal relalionshlps m the Swede and turnip groups of BrGssicG. Ann, Bot. 48:60 1-633, 1934

8 Delseny, M.. R. Cooke, and P Penon Sequellc.e hel­e,,:>geneity in radISh nuclear ribosomal RNA genes. Planl Sci. Lett. 30:107-109.1983.

9 f,omcz·Campo, C, and K. Hinala In 8rassico crops and WIld allies (S, Tsunoda. K. Hinata. and C. Gomez-

Campo, eds.) Japan Scientific SocIetIes Press, Tokyo. 1980. pp. 3-31

)0 GOlllieb. L D, Isozyme numbet and phylogeny In Proteins anet nucleir ac,,)s 10 plant syslemalics (U, Jensen amJ [). E Fairbrothers, eds) Spnnger-Verlag. Berll n. 1983, pp 209-221.

II Gollileb. I .. D.. and L C Greve, BIochemical prop­erlles of duphraled ,sozym~of phosphoglucose 15010'

erase 10 the plant ClarlllO xontiana Blochem, Genel. 19'155-172,1981.

12, Gustalsson. M" B Benlzer, R. Von Borhmer. and S, Snogarup MeIosis ,n Grcl'k IJros:<ica 01 the a/crocco group Bot [\nt 73-84, 1976,

13, Harberd. 0 J A contribution to the cyto-taxonomy 01 Brasska (Cruclferae) and ilS allies Bol. LIOn Soc. 65,1-23, 1972­

14 Harberd. D J • and I: D, MacArlhur MeIotic anal­ySIs 01 some specIes and genus hybm)s 10 T1L~ Bras­siceae, In Brossrca crops and Wild allies (S Tsunoda. K. Hinata, and C Gomel·Campo. "ds ). Japa.n SCIentific Societies Press. Tokyo. 1980. pp, 65~7.

15, Keller. W, A.. and K, C. Armstrong, Production 01 haploids via anther culture In 8ra$sico o<eracea var, ltahea Euphytica 32,151-159, 1983,

16 McClintOCk, B, The sigmficance 01 responses of the genome 10 challenge. SCIence 226,792-801, 19134.

17Mizushlma. U PhylogenetIc studies ot some WII<J B",.",co species, Tohuku J Agr Reo, 19:83-99.1968

18 Mizushlma. U, Genom~ analysis in Brass!Co and allIed genera In 8rasSlw crops and WIld alhes (S Tsunoda, K. Hinala. and C Gomez-Campo, eds) Japan SCIentific SOCIeties Press, Tokyo. 1980. DP 89-106,

19. Nllsch, J P, and C Nitsch Ifaploid plants Irom pollen gral ns. Science 163:85. 1969

20 Prchersky, E, and L 0 Got1l1eb, EVIdence for du­plication of the structural genes coding plaslid and cytosollc isozymes of triose phosphate isomerase 'n "'I,loid sp('c les of ClarkuJ Genetics 105:421-436, 1983

21 Prakash. S" and K HlOata Taxonomy, cytogenetics and orig," of crop Brassle"s. a reView, O!"'ra Botanica 55 I-57. 1980

22 PrakaSh. S" and S, Tsunoda Cytogen..tlcs 01 Bras­siUJ In Cytogenelics 01 crop plants (M. S Swamina­than. P K, Bupta. and V. Sinha, eds.) MacMillan India Ltd, 1983, pp 482-513

23 QUiros. C. F , and N, McHale, Genetic <; ~alysis ot isozyme vananls 1O dIplOId and telraploid potatoes, GenetlCs 111:13(-145, 1985

24 Quiros, C F" S F Kianlan, 0, Ochoa, and 0, Douch' es Genome evolution III lJrasslca Use of molecular markers and CytogenetIC slocks Crucilerae Newsletter 10:21-23. 1985

25. Quiros. C. F . 0, Ochoa. S. K,anian, and 0 Douches Evolutionary Irends in Brassico: Ciathenng eVIdence from chromosome addItion lines, Cruciferae Newslet­ler II :22-23. 1986

26, QI Li ros, C F ,0 Ochoa, S Kiaman, and D. Douch~

Analysis of the Brassiw o/<:rocea genome by the gen· ..ratlOn 01 8, compestris-olerocea chromosome additIon lines: Characterimtion by isolymes and rRNA g"Jl~S

Theor. AppJ. Genet. 74,758-766, 1987,

27, Robbelen. G Beitrage zur Analyse des Bross/c&­Genomes, Chromosoma 11.205-228. 1960.

28. SwamlOathan, M, S,. M L Magoon. and K. L Mehra, A ,·"nple propionic-carmme PMC sme", melhod tor plants w,th small chromosom('$, Jndian J, Genet, PI, Breed_ 14:87-88. 1954.

29. Williams. P, H.. and C. B. Hili, Rapid-rychng pop­ulations 01 8rossr(o SCIence 232: 1385-1389. 1986,

358 The Journal or Heredity 1988:79(5)