The Use of Inter-Sectional Hybrids in Blueberry Breeding

Abstract

The tetraploid species in Vaccinium section Cyanococcus and tetraploid V. uliginosum in section Vaccinium are autotetraploid. The same is probably true of the tetraploid species in other sections of the genus. Chromosome pairing at metaphase I in these species is normally regular and bivalent, but each gamete contains two homologous chromosomes for each of the 12 basic chromosome types (x = 12 in Vaccinium). Thus, F₁ hybrids between tetraploid plants from different Vaccinium sections can have regular bivalent chromosome pairing during meiosis and high fertility, even though diploid hybrids involving the same sections are highly sterile. Such fertile tetraploid hybrids are called amphidiploids. Amphidiploidy can give rise to new species; several important crop species are domesticated amphidiploids. The conditions for amphidiploid formation are narrow. The two species that hybridize must be divergent enough to insure faithful homologous bivalent chromosome pairing in the hybrid, where each bivalent consists of chromosomes derived from the same parent species. However, the parent species must be closely enough related to permit formation of vigorous hybrids. The first indication that amphidiploidy could be useful in blueberry breeding was a report by Rousi in 1963 of vigorous, fertile hybrids between tetraploid V. uliginosum (section Vaccinium) and tetraploid V. corymbosum (section Cyanococcus). In Florida, crosses of colchicine-induced tetraploid V. arboreum (section Batodendron) with tetraploid highbush cultivars and with tetraploid Florida V. myrsinites (section Cyanococcus) indicate that Vaccinium sections Batodendron and Cyanococcus also have the right degree of divergence to produce vigorous, fertile tetraploid hybrids. The feasibility of producing other intersectional tetraploid combinations in Vaccinium, and the vigor and fertility of the hybrids, can best be determined by trial and error.

Keywords:

• Vaccinium arboreum
• Vaccinium corymbosum
• Vaccinium myrsinites
• amphidiploids
• autopolyploids
• speciation

INTRODUCTION

In his ground-breaking monograph on the blueberries of North America, Camp (1945) stated his belief that many of the polyploid species in section Cyanococcus were allopolyploid (amphidiploid), and he speculated on the identity of the parent species involved in the formation of many of the polyploid species. At the time Camp wrote, the cytogenetic implications of autopolyploidy versus allopolyploidy were not widely appreciated, and Camp believed that polyploids derived from interspecific hybrids were necessarily allopolyploids. Since then, it has become clear that speciation can occur without significant differentiation of chromosome structure, and, indeed, this is what has happened in Vaccinium section Cyanococcus. Thus, diploid hybrids between any two species in section Cyanococcus are normally completely fertile, and backcrossing the diploid hybrids to either parent results in backcross-1 seedlings of high fertility. Segregation ratios for both molecular and morphological markers have shown that tetraploid V. corymbosum L. and tetraploid hybrids between V. corymbosum and V. darrowii Camp are both autotetraploids (Qu et al., 1998). A modern understanding of the difference between autopolyploid and allopolyploid species is based on chromosome pairing behavior (Clausen et al., 1945). If, in a tetraploid hybrid, chromosome pairing is always or nearly always limited to pairs involving chromosomes that came to the hybrid through the same gamete, then crossing over will not occur between chromosomes of the parent species, and there will be little or no recombination of species' differences in the F ₂ and subsequent generations. Thus, amphidiploid species have been called “true breeding hybrids.”

‘Nessberry’, ‘Veitchberry’, and ‘Loganberry’ are manmade amphidiploids involving raspberry-blackberry hybridization in Rubus (Clausen et al., 1945). Many other natural and manmade amphidiploids have been useful in agriculture.

In 1961, in Finland, Rousi (1966) crossed tetraploid V. uliginosum L. (section Vaccinium) with tetraploid V. corymbosum, hoping to combine the cold hardiness of V. uliginosum with the better fruit qualities of V. corymbosum. The cross was not difficult to make, the hybrids were fertile, and meiosis in the hybrids was surprisingly regular. The frequency of univalents and multivalents at Metaphase I of meiosis was roughly as low as in the parent species. Rousi postulated that autosyndesis was occurring in the hybrid, V. uliginosum, chromosomes pairing with one another and Cyanococcus chromosomes doing the same (Rousi, 1966, 1967).

In 1981, with the long-term objective of incorporating desirable features from V. arboreum Marshall (diploid section Batodendron) into cultivated highbush blueberry, Lyrene (1991) crossed diploid V. darrowii (section Cyanococcus) with diploid V. arboreum. Sixteen hybrid plants were selected and grown for 10 years in a field containing many tetraploid highbush blueberry selections. After the plants became established, they flowered heavily each year, but produced few berries. Open-pollinated seeds were collected from the diploid hybrids, and the few seedlings obtained proved to be tetraploid or near-tetraploid outcrosses with tetraploid highbush (Lyrene, 1991). Another 100 V. darrowii x V. arboreum hybrids were planted in the same field about 1985, and after 25 years, most of the plants are still alive and vigorous, but produce little or no fruit. Chavez and Lyrene (2009) found that pollen from these diploid plants, when viewed at 250× was mostly aborted. Meiosis, studied from immature anthers, had some bivalents at Metaphase I, but also many irregularities of chromosome pairing and separation, and most microspores were inviable.

It was suspected that sterility, as seen in these diploid F₁ hybrids between sections Cyanococcus and Batodendron, might not occur if the cross was made at the tetraploid level. Several attempts to produce hybrids by pollinating tetraploid southern highbush cultivars with pollen from various diploid V. arboreum selections gave no hybrids (Lyrene, unpublished). It had been hoped, apparently without foundation, that V. arboreum would produce enough 2n gametes to give a few tetraploid hybrids. Failure of these crosses led to the work reported here, in which the crosses were repeated using colchicine-induced tetraploid V. arboreum.

MATERIALS AND METHODS

In 2004, approximately 10,000 seeds of V. arboreum were imbibed in 0.2% aqueous colchicine for 4 days, and then sown on peat under mist. In 2006, three fertile tetraploid plants were selected based on large-diameter pollen tetrads (Lyrene, unpublished). In 2006 and 2007, 3,000 flowers of tetraploid highbush cultivars and selections were emasculated and hand-pollinated in a greenhouse with pollen from the three tetraploid V. arboreum selections. The resulting seedlings were grown in a greenhouse for 4 months and in a field nursery for 10 months. Hybrids were confirmed by vegetative morphology in late autumn and winter. Fifty of the most vigorous hybrids that produced flowers were dug, potted, and placed in a greenhouse. These were backcrossed to tetraploid highbush selections. The F₁ hybrids were used as pollen parents in some backcrosses and as seed parents in others.

In February 2008, rooted cuttings of three plants of V. myrsinites Lam. were pollinated with pollen from one of the tetraploid V. arboreum plants. V. myrsinites (section Cyanococcus) is a tetraploid lowbush, evergreen, Florida native blueberry. The V. myrsinites plants used were propagated from a forest near Jacksonville, in northeast Florida. Resulting seeds were planted in November 2008, and seedlings were transplanted to a field nursery in May 2009. The hybrids flowered in the field in the spring of 2010 and fruit resulting from open pollination were examined in July 2010.

RESULTS AND DISCUSSION

The crosses between tetraploid highbush cultivars and colchicine-induced tetraploid V. arboreum gave highly variable results. Some crosses gave no seedlings. Some gave a few seedlings that, based on morphology in the field nursery, were not inter-sectional hybrids. Other crosses gave a few seedlings that were hybrid but lacked vigor. Still other crosses gave a mixture of vigorous hybrids, weak hybrids, and non-hybrids. Among the plants that were clearly inter-sectional hybrids and were medium to high in vigor, some produced few or no flower buds after 9 months in a field nursery, even though highbush seedlings in the same nursery had produced numerous flower buds. Pollen from hybrids that did flower was examined under the microscope at 250× The hybrid plants differed greatly in the amount of pollen shed and in the percentage of microspores that were plump and apparently viable. Few F₁ hybrids had full pollen fertility, but most shed at least some pollen and had between 10 and 70% of the microspores that appeared to be fully developed.

Approximately 50 of the F₁ hybrids that were highest in pollen shed and pollen fertility (as judged by microscopic examination of pollen) were hand pollinated or used as pollen sources in backcrosses to tetraploid highbush cultivars. The results of the crosses were highly variable. About half of the highbush x V. arboreum F₁ hybrids used in crosses had medium to high fertility, with a high percentage of fruit set in backcrosses to highbush and numerous seeds per berry. Several thousand backcross seedlings have been produced. Some are weak, but after 6 months in the field, most are as vigorous as highbush x highbush seedlings in the same nursery.

Crosses between V. myrsinites and V. arboreum were conducted on a much smaller scale than the highbush x V. arboreum crosses, but produced hybrid seedlings with less effort. The several hundred hybrids grown in a field nursery were vigorous, flowered heavily (in contrast to the highbush x V. arboreum seedlings in the same nursery, which made few flower buds after 9 months in the field), and most plants set full crops of berries when open pollinated. Berries of the hybrids ripened about 1 month later than those of V. myrsinites and about 3 months earlier than V. arboreum. The berries on most of the hybrids contained as many or more well-developed seeds as open-pollinated V. myrsinites berries, and the seeds were much larger. Many morphological features of the hybrids were intermediate between V. myrsinites and V. arboreum.

The contrast between the almost total sterility of diploid V. darrowii x V. arboreum hybrids and the high fertility of the tetraploid V. myrsinites x tetraploid V. arboreum hybrids was extreme. Some highbush x tetraploid V. arboreum hybrids were also highly fertile, but others were either male sterile, female sterile, or both. The low fertility of some of the highbush x V. arboreum hybrids could have resulted from aneuploid gametes from the colchicine-produced V. arboreum parents. Colchicine-induced tetraploids in other genera have produced aneuploid gametes at high frequency (Clausen et al., 1945). In colchicine-produced tetraploid rye (Secale cereale L.), for example, only about 50% of the open-pollinated progeny are tetraploid. Most of the rest are aneuploid, predominately lacking one chromosome or having one extra.

Because Vaccinium is a large genus, with many sections, and with species native on various continents and on many islands, opportunities for producing and testing amphidiploid combinations are great. Section Myrtillus has several species of horticultural interest, including the bilberry (V. myrtillus L.), V. deliciosum Piper, V. membranaceum Douglas ex Torr., and others. Some of these may make fertile hybrids with V. corymbosum at the tetraploid level. Vaccinium stamineum L., a diploid from the southeastern U.S. (section Polycodium), is a highly variable species with large, juicy, fruit and small seeds. We once had a few V. darrowii x V. stamineum hybrids of low to medium vigor that flowered in the field, but this work was not carried forward.

Many questions remain regarding the use of Vaccinium amphidiploids in breeding. What will happen when fertile F₁ tetraploid intersectional hybrids are backcrossed to tetraploids of the parent species? If the fertility of the F₁ hybrid is due to having two genomes from each parent section, reduced fertility might be expected in backcross plants having three genomes from one section and one from the other. When Hiirsalmi (1977) backcrossed F₁ V. uliginosum x V. corymbosum tetraploid hybrids to V. corymbosum, the seedlings were extremely variable, even more variable than in the F₁ population. Rousi did not comment on any reduction in the fruitfulness of the BC₁ population. He noted that berry size, color, and flavor were highly variable in the BC₁ generation (Hiirsalmi, 1977). Four tetraploids obtained in Florida by open pollination of V. darrowii x V. arboreum diploid hybrids in the presence of tetraploid V. corymbosum produced berries abundantly when open pollinated, but were only partially male fertile (Brooks and Lyrene, 1998). These four plants had three Cyanococcus and one Batodendron genome. Further backcrossing gave seedlings of varying fertility, but open-pollinated fruit set in the field was high in many of the seedlings. It appears that fertility will be an important selection criterion in backcross generations following tetraploid intersectional crosses, but not an insurmountable problem.

Another question regarding amphidiploids in Vaccinium is why, despite widespread polyploid speciation in several sections, no natural amphidiploid species have been reported. Further research may indicate that polyploids with amphidiploid gene segregation ratios exist in Vaccinium, but so far, none have been reported.

CONCLUSIONS

Inter-sectional diploid hybrids in Vaccinium are normally sterile. However, tetraploid species in Vaccinium are autotetraploid, and inter-sectional tetraploid hybrids can be fertile, presumably due to autosyndetic chromosome pairing. Fertile, tetraploid intersectional hybrids could be used by blueberry breeders in several ways. Probably most important is the unleashing of vast amounts of genetic and phenotypic variability in segregating generations following wide hybridization. Wide hybridization is highly mutagenic, in that the interaction of large numbers of genes that normally do not occur together creates phenotypes not seen in seedlings from intraspecific crosses. Genetic variability is the raw material from which breeders create cultivars, and wide hybridization is a goldmine of variation. It is unlikely that any first-generation amphidiploid combinations in Vaccinium would produce berries of commercial quality, but a broad-based foundation population of amphidiploids could be the starting point for recurrent selection that could eventually lead to commercial cultivars. Alternatively, fertile amphidiploid hybrids could be backcrossed to one of the tetraploid parent species with the goal of adding new traits or enhancing variability.