Chromosome Homology in Tetraploid Southern Highbush x Vaccinium elliottii Hy brids

A yellow-leaf seedling marker, r, was used to determine if there was pref er en tial chromosome pairing in a group of tetraploid southern highbush blueberry hybrids. Plants with four copies of r (no copies of R) fail to develop anthocyanins, and cotyledons, hypo co t yls, leaves, stems, and other vegetative tissues have a bright yellow-green color. In the hybrids that were studied, two genomes were from the diploid wild species, V. elliottii Chapman, and both carried the recessive mark er r. The other two genomes were from southern highbush cultivars and both carried the dominant wildtype allele, R. When RRrr hybrids were intercrossed or crossed to rrrr yellow-leaf plants, the number of yellowleaf rrrr seed lings obtained usually equalled or exceeded the number predicted from nonpreferential chromosome pairing. Since rr gametes can only be produced by RRrr plants when R and r chromosomes pair at Meiosis I, there was no evidence that the chromosomes de rived from V. elliottii were pairing at a higher-than-random rate. species can pair and un der go recombination during meiosis. Vaccinium elliottii is a tall-growing, smallleafed, highly deciduous diploid blue ber ry species that occurs as far south as Gainesville, in the northern Florida peninsula, and is widespread in river bottoms and on rolling hills of the coastal plain and piedmont regions of the southeastern United States. It is of interest in blueberry improvement pro grams because of upright growth habit, very early ripening, tol er ance for mineral soils and up land sites, and for its high-quality, albeit small berry. A re ces sive seedling marker, which removes red pigments from all plant parts, was found in two wild plants, one from northern Florida and one from southwestern Alabama (Lyrene, 1988), and the responsible allele was shown to segregate as a simple recessive in crosses within V. elliottii. The mutant and wildtype alleles at the locus were initially called y and R, but here are called r and R to conform with conventions for naming alleles. Crosses be tween plants homozygous for the yellow-leaf allele and a V. elliottii clone from northeast Florida that had normal foliage, but berries that were pink when ripe produced F 1 pop u la tions that were normal and segregated in de pen dent ly for the two alleles in the F 2 (Lyrene, 1988 and Lyrene, unpublished). Cross es be tween yellow-leaf V. elliottii and a V. darrowi selection that had normal foliage but pink-green mature fruit, produced an F 1 population with normal foliage but which seg re gat ed 1:1 for normal vs. pink-green fruit color (Lyrene, unpublished). This indicated that the V. darrowi clone was heterozygous for a dom i nant allele that reduced anthocyanin in the fruit, and that this allele was not at the same locus as the yellow-leaf allele in V. elliottii. Yellow-leaf V. elliottii plants have not been crossed with any of the northern highbush clones or V. angustifolium (Hall and Aalders, 1963) clones that have white, pink, or green berries. The purpose of this study was to in ves ti gate the homology between the chro mo somes of southern highbush blueberry and those of V. elliottii. This was done by hy brid iz ing tet ra p loid southern highbush cultivars and diploid V. elliottii plants that were ho mozy gous rr, taking advantage of 2n gamete pro duc tion in V. elliottii to produce tetraploid hybrids. The ratio of yellow-leaf to normal seedlings in segregating generations fol low ing the crosses was used to study the pairing and segregation of the chromosomes bearing the marked locus. Materials and Methods Between 1987 and 1989, a large num bers of fl owers of the tetraploid southern high bush blueberry selections FL 6-19 and ‘Mistyʼ, neither of which carried the r allele, were pollinated with pollen from diploid V. elliottii selections ‘Silverhill ̓ and ‘Olenoʼ, both of which were homozygous rr. Evidence that neither FL 6-19 nor ‘Misty ̓carried the r allele comes from the fact that no anthocyanin-free plants have been found among 200,000 related seedlings that have been fruited in the Florida Received for publication1 Feb. 2002. Accepted for publication 19 May 2002. Flor i da Ag ri cul tur al Exper i ment Station Jour nal Series No. R-08621. The evolution of many species with in genus Vaccinium has involved poly p loidy. Dip loid, tetraploid, and hexaploid spe cies are found in Vaccinium sections Cyanococcus (blue ber ry), Oxycoccus (cran ber ry), and Vaccinium (bilber ry), and section Myrtillus has diploid and tetraploid species. Two kinds of tetraploid species are possi ble in plants—autotetraploids and al lotet ra p loids, also called amphidiploids (Allard, 1960; Stebbins, 1950). The distinction is of im por tance to plant breeders because the type of polyploidy determines whether or not cer tain types of hybrids will be fertile or sterile and whether genetic recombination during mei o sis will occur freely among all four sets of chromosomes or will be confi ned to ex chang es between homologous, as opposed to homoeologous chromosomes. The dif fer ence between autotetraploidy and al lotet ra p loidy is based on two contrasting patterns of chro mo some pairing during meiosis. Both au totet ra p loid and allotetraploid species, as contrasted with some tetraploid plants pro duced in the laboratory, have regular bivalent chro mo some pairing. However, in au totet ra p loids, the de ter mi na tion of which two of the four ho mol o gous chromosomes will pair dur ing meiosis in a particular meiocyte is ran dom, whereas in allotetraploids, chromo some pairing is strictly predetermined, occuring only between ho mol o gous and never be tween homoeologous chro mo somes. If an au totet ra p loid species rep re sent ed by the genome formula AAAA is crossed with a second autotetraploid species, BBBB, from which it is genetically so di ver gent that chro mo somes of the two species cannot pair, the AABB hybrid could still be highly fertile because the hybrid contains two sets of A chromosomes and two sets of B. If, however, two al lopoly p loid species similarly related were crossed, e.g., AABB x CCDD, chromosomes in the ABCD F 1 hybrid would not pair, and the hybrid would be highly ster ile. Haploids derived from au totet ra p loids can be somewhat fertile; haploids de rived from allotetraploids are highly sterile. Recessive alleles in autotetraploids normally give 35:1 segregation in the F 2 generation after a cross in which one parent of the F 1 was ho mozy gous for the dominant allele and the other was homozygous for the recessive allele. In con trast, recessive marker genes in al lotet ra p loid species normally segregate 3:1. Camp (1945), who studied the tax on o my and evolution of the genus Vaccinium before the differences between au to poly p loids and allopolyploids were fully understood, pos tu lat ed that allopolyploidy was widespread in several of the sections of Vaccinium, in clud ing Cyanococcus. More recent studies of spe cies and interspecifi c hybrids in Cyanococcus have shown that several of the species thought by Camp to be allopolyploids are actually autopolyploids. Both isozyme markers (Krebs and Hancock, 1989) and RAPD markers (Qu and Hancock, 1995; 1998) have been used to show that V. corymbosum and a tetraploid hybrid between V. corymbosum and V. darrowi Camp have tetrasomic inheritance, as would be expected in autotetaploids. In ter spe cifi c hybrids have been made between various dip loid species in section Cyanococcus. High fertility in the diploid F 1 generations has in di cat ed that the chro mo somes of the crossed species were able to pair well. When V. corymbosum is crossed with V. angustifolium Ait., V. darrowi, V. myrsinites L., or V. elliottii, and tetraploid F 1 hybrids are backcrossed to V. corymbosum, the back cross-1 seedlings are usually quite fertile. This suggests that chro mo somes of these 26-7210, p263-265 263 3/27/03, 10:35:45 AM HORTSCIENCE, VOL. 38(2), APRIL 2003 264 BREEDING, CULTIVARS, ROOTSTOCKS, & GERMPLASM RESOURCES blueberry breeding program, nor among more than a million seedlings that have been seen in germination pots. No anthocyanin-free seedlings have been reported by other blueberry breeders, nor has a V. darrowi or V. corymbosum plant that lacked vegetative an tho cy a nin been reported from wild pop u la tions. Such plants undoubtedly exist, but the allele frequency must be very low in the south ern highbush gene pool and in native Section Cyanococcus species. Seeds from the crosses were sown in a greenhouse in November of the year in which the crosses were made, and the seedlings were transplanted to a fi eld nursery in May of the following year. Based on in ter me di ate hybrid phenotype, isozyme analysis, the ability to produce at least some yellow-leaf segregants in crosses with tetra-allelic rrrr plants, and production of a high percentage of po ten tial ly viable pollen as indicated by staining with 2% acetocarmine dye, nine F 1 hybrids that appeared to be tetraploid RRrr plants were identifi ed from the 1987 crosses. In Dec. 1988, the eight largest plants were potted from the fi eld and were chilled at 5 °C in a dark chamber for ≈6 weeks. They were then placed in a warm greenhouse and hand-crossed in various com bi na tions after emasculation. The resulting seeds were sown in the greenhouse in Dec. 1989, and several hundred yellow-leaf seed lings were selected and transplanted to trays of peat. After several months, the10 most vig or ous were potted and maintained in the green house for crossing the following fall and win ter. In Dec. 1990, 11 hybrids ob tained by pol li nat ing yellow-leaf V. elliottii clones ‘Silverhill ̓ and ‘Oleno ̓with pollen from south ern highbush cultivar ‘Misty ̓were dug from the fi eld, potted, and placed in a chamber at 5 °C. Also included were the in ter spe cifi c hy brids from the previous year and fi ve of the10 yellow-leaf plants obtained by intercrossing the original F 1 hybrids. After ≈6 weeks, the plants were placed in a bee-proof green house to fl ower. In Mar. 1991, additional crosses were made between pairs of F 1 hy brids, and 14 of the F 1 hybrids were pollinated with pollen from the tetraploid yellowleaf progeny of earlier F 1 x F 1 crosses. Seed from the crosses was planted in mid-July 1991 after having been soaked in an aqueous solution of 10.4 mM gibberellin A 3 and 1.0 mM N-benzyladenine for 24 h to enhance high-temperature germination. The germination pots were watered daily and were observed through Feb. 1992 to permit ger mi na tion of seeds that did not respond to the chemical treatments. The red-leaf seed lings were counted and discarded after they emerged. The yellow-leaf seedlings were trans plant ed to trays of peat and were grown in the green house for four additional months to make sure no red pigments appeared. The number of red-leaf and yellow-leaf seedlings were counted from 13 AAaa x AAaa crosses and from 18 AAaa x aaaa cross es. The ratios of red-leaf : yellow-leaf seedlings were tested by Chi-square to determine if the ratios conformed to those expected from ran dom pairing between southern highbush and V. elliottii-derived chromosomes. Re sults and Discussion As expected from tetraploid x dip loid cross es in Vaccinium, which has a strong trip loid block (Galletta, 1975; Sharpe and Sherman, 1971), the crosses between the tet ra p loid southern highbush cultivars and the dip loid V. elliottii clones that were homozygous for rr gave only a small fraction of the number of seedlings normally obtained from homoploid crosses. However, high pollen staining, high fertility, and the phenotypes of the seedlings obtained when the F 1 hybrids were intercrossed indicated that all but one of the hybrids were tetraploid. Most of the hybrids had vigor equal to or greater than the parent species. If the RRrr F 1 hybrids were be hav ing as strict allotetraploids, with the rr chro mo somes from V. elliottii always pairing with each oth er, the F 1 plants would be expected to produce no rr gametes, and no yellow-leaf seedings would appear in the F 2 generation or in the progeny obtained by crossing RRrr and rrrr plants. Thus, any tendency toward pref er en tial pair ing of the V. elliottii-derived chro mo somes in the tetraploid hybrids would re sult in a de fi cien cy in the number of yellow-leaf seedlings obtained in the next generation, compared to the number expected from com plete ly non-preferential pairing. When there were de vi a tions from the expected values, they were mostly in the opposite direction (Tables 1 and 2), that is, more than the expected number of yellow-leaf plants were observed. Segregation ratios were determined for 31 crosses, and 11,831 seedlings were scored for leaf color. In the F 2 intercross pop u la tions, where the 35:1 ratio expected from random chromosome pairing would have pro duced 72 yellow-leaf seedlings, 115 were ob served (Ta ble 1). Chi-square indicated that the proba bil i ty of the discrepancy being due to sam pling error was exceedingly small. In the cross es where RRrr F 1 hybrids were pol li nat ed with pollen from yellow-leaf rrrr plants, 1539.5 yellow-leaf seedlings were expected, and 1656 were observed (Table 2). Again, the Chi-square test indicated the probability of this difference Table 2. Yellow-leaf phenotypic seg re ga tion ratios, with Chi-square analysis, for crosses be tween F 1 in ter spe cifi c hybrids heterozygous (presumably RRrr) for the yellow-leaf allele, used as female parent, and homozygous (rrrr) yellow-leaf hybrids, de rived from F 1 x F 2 cross es, used as pollen parent. Yellow-leaf : Red-leaf Cross Expected Observed χ P FL 88-174 x FL 91-212 19.7 : 98.3 18 : 100 0.18 >0.75 FL 89-6 x FL 91-203 461.3 : 2306.7 478 : 2290 0.73 >0.50 FL 89-12 x FL 91-213 91.3 : 456.7 78 : 470 2.33 >0.25 FL 91-70 x FL 91-213 102.2 : 510.8 121 : 429 3.95 <0.05 FL 91-104 x FL 91-211 50.5 : 252.5 61 : 242 2.62 >0.10 FL 91-104 x FL 91-213 47.5 : 237.5 37 : 248 2.78 >0.10 FL 91-105 x FL 91-203 163.7 : 818.3 191 : 791 5.46 <0.05 FL 91-105 x FL 91-214 238.5 : 1192.5 286 : 1144 11.35 <0.01 FL 91-106 x FL 91-212 77.3 : 386.7 30 : 434 34.73 <0.01 FL 91-107 x FL 91-203 101.3 : 506.7 123 : 465 5.58 <0.05 FL 91-107 x FL 91-213 23.3 : 116.7 26 : 114 0.38 >0.75 FL 91-108 x FL 91-209 6.2 : 30.8 9 : 28 1.52 <0.25 FL 91-109 x FL 91-205 37.5 : 187.5 42 : 183 0.65 >0.25 FL 91-110 x FL 91-213 17.0 : 85.0 23 : 79 2.54 >0.10 FL 91-111 x FL 91-201 59.2 : 295.8 61 : 294 0.07 >0.75 FL 91-111 x FL 91-215 9.8 : 49.2 9 : 50 0.08 >0.75 FL 91-112 x FL 91-213 28.3 : 141.7 38 : 132 3.99 <0.05 FL 91-113 x FL 91-209 18.8 : 94.2 25 : 88 2.45 >0.10 All combined 1539.5 : 7697.5 1656 : 7581 10.58 <0.005 Theoretical expected ratio, based on tet ra som ic chromosome behavior, is 1 yellow-leaf : 5 red-leaf. Table 1. Yellow-leaf phenotypic seg re ga tion ratios, with Chi-square analysis, for crosses among southern highbush x V. elliottii hybrids het erozy gous (presumably duplex tetraploids) for the yellow-leaf allele. Total seedlings = 2,594. Yellow-leaf : Red-leaf Cross Expected Observed χ P FL 88-174 x FL 89-4 0.9 : 32.1 2 : 31 1.38 <0.25 FL 88-174 x FL 89-6 5.4 : 189.6 6 : 189 0.07 >9.75 FL 88-174 x FL 91-105 5.2 : 180.8 9 : 177 2.86 >0.05 FL 89-2 x FL 88-174 5.5 : 191.5 10 : 187 3.79 >0.05 FL 89-4 x FL 89-12 2.9 : 100.1 3 : 100 0.00 >0.99 FL 89-5 x FL 88-174 5.0 : 176.0 11 : 170 7.41 <0.01 FL 89-5 x FL 91-105 7.2 : 252.8 13 : 247 4.81 <0.05 FL 89-6 x FL 88-174 6.2 : 216.8 14 : 209 10.10 <0.01 FL 89-6 x FL 91-70 12.0 : 421.0 22 : 413 8.58 <0.01 FL 89-12 x FL 88-174 3.8 : 134.2 4 : 134 0.01 >0.99 FL 89-12 x FL 89-4 2.6 : 126.4 3 : 127 0.06 >0.75 FL 89-94 x FL 89-6 6.8 : 237.2 2 : 242 3.49 <0.10 FL 89-94 x FL 91-109 7.5 : 261.5 16 : 253 9.91 <0.01 All combined 72.1 : 2521.9 115 : 2479 26.26 <0.005 Theoretical expected ratio, based on tet ra som ic chromosome behavior, is 1 yellow-leaf : 35