Desynapsis and FDR 2N-megaspore formation in diploid potato : potentials and limitations for breeding and for the induction of diplosporic apomixis

The cultivated potato, Solanum tuberosum L., is a highly heterozygous autotetraploid (2n=4x=48) plant species, which after its introduction into Europe in the 16th century has become one of the world's major food crops. The potato has traditionally been grown from tubers. However, as tubers are an excellent substrate for many pathogens and parasites, it is extremely difficult and expensive to produce healthy seed tubers. Most developing countries lack both the knowledge and infrastructure required for the efficient production of healthy seed tubers and the currency for its importation and distribution. As a consequence small farmers in developing countries are generally forced to use diseased tubers from a previous harvest, which may result in dramatic yield losses. In response to the urgent need of cheap but healthy plant material the International Potato Center (CIP) in Peru has, since the Planning Conference on this subject in 1979, propagated the new technology of growing potatoes from true seeds, True potato seeds are relatively easy and cheap to produce and even when harvested from heavily diseased plants generally do not carry any diseases. However one of the major problems in breeding potato varieties that can be maintained and grown from true seeds is the lack of uniformity. True seed progeny of existing varieties or inter-varietal hybrids is mostly highly heterogeneous due to the extreme heterozygosity of the potato.Several methods to synthesize sufficiently uniform true potato seed varieties have been proposed. One of these methods takes advantage of the frequent occurrence of numerically unreduced (2n) gametes in wild and cultivated diploid potato species, which enables the production of hybrid tetraploid progeny from tetraploid-diploid (unilateral sexual polyploidization) or diploid-diploid (bilateral sexual polyploidization) matings. The vigour and uniformity of tetraploid populations produced in this way, largely depends on the mode of 2n-gamete formation in the selected diploid parents. Depending on the genetic consequences of meiotic abnormalities that result into 2n-gamete formation, two distinct modes, viz. first division restitution (FDR) and second division restitution (SDR), can be distinguished. In general FDR is considered superior to SDR because of its abilitity to preserve a relatively large amount of favourable parental heterozygosity, including complex types of epistasis. In this respect the combination of FDR 2n-gamete formation with mutant synaptic genes that substantially reduce gene recombination, is of particular significance as it would provide a means to enhance the ability of FDR 2n-gametes to pick up the genetic constitution of the selected diploid parents with a minimum amount of reassortment. Using synaptic mutants with a virtually complete lack of gene recombination maximum performance and nearly complete uniformity may thus be expected from 2xFDR-2xFDR crosses.Complete or nearly complete uniformity might also be achieved by the induction of apomictic seed formation. Apomixis sensu stricto is the asexual development of maternal embryo's and seeds and thus combines the advantages of both vegetative propagation (uniformity) and generative propagation (disease free plant material). Apomictic embryos may arise either directly from somatic cells outside the embryosac (adventitious embryony) or from unreduced and unfertilized (parthenogenesis) cells, usually egg cells, of the embryosac (gametophytic apomixis). In the latter case unreduced embryosacs are formed that may be of either aposporic or diplosporic origin. In apospory it develops directly from a somatic, mostly nucellar cell of the ovule. In diplospory the unreduced embryosac derives from a generative archesporial cell of the ovule, either directly by omission of meiosis or indirectly by modified meiosis in which neither reduction in chromosome number nor (substantial) gene recombination takes place. Fertilization of the secondary embryosac: nucleus may or may not be required as a stimulus for endosperm formation and subsequent parthenogenetic development of the unreduced egg cells into mature embryos and seeds (pseudogamous and autonomous apomixis respectively).Although apomictic seed formation has never been observed in Solanaceae its importance for growing potatoes from true seeds justified an attempt to breed for apomictic reproduction in potato. The research described in this thesis focussed on the perspectives for inducing gametophytic apomixis, in particular pseudogamous diplosporic apomixis, because:1. The manipulation of gametophytic apomixis (i.e., apospory and diplospory) is to be preferred to adventitious embryony as the presence of embryosacs; enclosing apomictic embryos is an efficient barrier against most viruses and thus greatly facilitates the production of virus free true potato seeds.2. Apospory has never been detected in potato and thus would require an extensive mutation breeding programme. In view of the scanty and in many cases contradictory knowledge about its genetic control (monogenic-polygenic; dominant-recessive) and the lack of efficient screening methods for identifying 'positive' mutations, the outcome of such a trial and error procedure would be highly uncertain.3. It has been suggested that apomixis consists of a number of distinct and genetically controlled elements that lie within the reproductive potentialities of sexual plants. In case of diplosporic apomixis the elements that can roughly be distinguished are a strongly reduced chromosome pairing and/or reduced gene recombination (asynapsis/desynapsis), the formation of unreduced megaspores and embryosacs through FDR, and parthenogenetic development of unreduced egg cells into mature embryos and seeds. As some of these elements - viz. recessive genes for asynapsis/desynapsis, genes for (SDR) 2n-megaspore formation and genes for parthenogenesis along with pseudogamy - are available in potato, their combination in a single genotype may be feasible.4. An attempt to induce diplospory requires detailed studies on the, as yet unkown, effects of mutant synaptic genes in female meiosis and the largely obscure causes and mechanisms of 2n- megaspore formation. Such studies might thus provide vital 'spin off' results to be used in pursuing the as yet utopian production of vigorous and uniform tetraploid true potato seed varieties from 2xFDR-2xFDR crosses.An accurate knowledge of the normal pattern of female meiosis and embryosac formation is essential for deciding whether or not an unreduced embryosac is of diplosporic origin and for recognizing abnormalities in female meiosis that are associated with the expression of mutant synaptic genes and the formation of 2n-eggs. Therefore, normal meiosis and embryosac formation were studied in several diploid potato clones (Chapter 1). In contrast to results reported in the literature, this study indicated that the archesporium of potato cannot be delimited to a single cell. A surplus of archespores sometimes developed into normal sexual embryosacs. So the occurrence of multiple embryosacs within a single ovule need not necessarily be due to apospory as had previously been suggested by some potato researchers. On the basis of the normal sequence of female meiosis it was inferred that 2n-megaspores, if formed by meiotic abnormalities in normal synaptic plants, are likely to be of exclusive SDR origin, whereas meiotic abnormalities resulting in consistent FDR 2n-megaspore formation, and thus the induction of diplospory, would actually require mutant synaptic conditions. Subsequent research was therefore primarily focussed on (1) the identification and characterization of mutant synaptic genes that express in female meiosis, (ii) the identification of synaptic mutants with consistent FDR 2n-megaspore formation and (iii) the elucidation of the cytological mechanisms of FDR 2n-megaspore formation.For this purpose, first quick routine methods allowing for large scale screening and detailed studies of female meiosis had to be developed, because conventional embedding-sectioning techniques are very laborious and hamper quantitative interpretation of meiotic processes. Two techniques were developed. One enabled the screening of female meiosis and embryosac development in intact methyl salicylate cleared ovules and permitted bulk preparation of fixed ovaries within 2 hours (Chapter 2). The other, a 30 minute enzyme-squash procedure, allowed for detailed studies on the effect of mutant synaptic genes and genes for 2n-megaspore formation on chromosome behaviour (Chapter 3).Using these techniques it was established that (at least) three of the six mutant synaptic genes that had been reported for potato were allelic and similarly expressed in both male and female meiosis, whereas a mutant that had been claimed to express in female meiosis only proved to be non- existent. The former three mutants were characterized by normal chromosome behaviour throughout pachytene and a falling apart of bivalent chromosomes at diakinesis and thus displayed a typical desynaptic behaviour. They were therefore reassigned the gene symbol ds -1 (Chapter 4). Asynaptic mutants with a virtually complete absence of chromosome pairing and hence gene recombination are the most attractive candidates for engineering diplosporic apomixis. However, as cogent cases of asynapsis have not yet been reported for potato, desynaptic mutants are the best alternative that is currently available. Therefore the following questions were raised:1. Is it possible to select for consistent 2n-megaspore formation in ds -1 mutants?2. If so, are these 2n-megaspores of exclusive FDR origin as was predicted?3. To what extent does crossing-over and hence gene recombination occur?In chapter 6 the first two questions are addressed. The level of 2n-megaspore formation was determined in 126 ds -1 mutants using seed set from 2x.4x and 2x.2xFDR testcrosses as a criterion. Although the majority formed on the average less than 5 seeds/fruit, 14% of all ds-1 mutants produced 2n-megaspores in frequencies that resulted in consistent seed set within the 5-25 seeds/fruit range and allowed for routine production of nearly exclusive tetraploid progeny from 2x.2xFDR crosses. Subsequent cytological analyses revealed that 2n-megaspore formation in ds-1 mutants resulted from a direct equational division of univalent chromosomes in the first meiotic division (pseudo-homotypic division), an FDR mechanism that had previously been reported to occur in some of the diplosporic apomictic plant species. Additional data on SDR 2n-megaspore formation in full-sib normal synaptic plants indicated that both SDR and FDR 2n-megaspore formation are likely to be caused by common genes for precocious chromosome division. Depending on the relative timing of cell cycle and chromosome division this precocious chromosome division may impose post-reductional (SDR) or pre-reductional (FDR) 'restitution' of the somatic chromosome number under normal and mutant synaptic conditions respectively.Simply inherited marker traits are required for the analysis of genetic recombination in ds -1 mutants and had to be identified first. Some of the genetic markers used are characterized in chapter 8. On the basis of extensive analysis of chiasma formation and gene-centromere mapping of a number of simply inherited marker genes in normal synaptic plants and desynaptic mutants it could be concluded that the ds -1 gene substantially reduced the overall frequency of crossing-over and thus gene recombination in both male and female meiosis (Chapters 5 and 7). In addition, the genetic analyses revealed that FDR 2n-megaspores and FDR 2n-pollen from ds-1 mutants preserve approximately 94.1 % of the overall parental heterozygosity as opposed to 79.5 % that is preserved by FDR 2n-pollen from normal synaptic plants. The ds -1 gene was further demonstrated to particularly enhance the ability of FDR 2n-megaspores, and 2n-pollen to pick up the genetic constitution of the parental clone, including complex types of favourable epistasis, with a minimum amount of reassortment.Summarizing, it may be stated that the identification of diploid desynaptic mutants with consistent FDR 2n-megaspore formation extends the opportunities for direct transfer of enhanced diploid germplasm to tetraploids by means of sexual polyploidization and, since FDR 2n-megaspores and 2n-pollen from ds -1 mutants are relatively efficient in preserving the genetic constitution of selected diploid parents particularly demonstrates the feasibility of 2x( ds -1;FDR)-2x( ds -1;FDR) crosses for the production of relatively uniform and vigorous true potato seed varieties.Because of the potential for limited genetic recombination the use of the ds -1 gene in the development of diplosporic apomixis seems less obvious. However, as genes for asynapsis have not been identified in potato so far, it may be considered the best alternative that is currently available. Moreover some genetic diversity in apomictic progeny of diplosporic plant species as a consequence of autosegregation is quite common. The finding that some desynaptic clones formed FDR 2n-eggs through pseudo-homotypic division (≈diplospory) strongly supports the hypothesis that gametophytic apomixis consists of a number of distinct and genetically controlled elements which may be combined to attain approximately identical reproduction in largely sexual plant species. The application of this approach to produce completely uniform true potato seed varieties obviously requires breeding for increased levels of FDR 2n-egg formation in synaptic mutants completely suppressing genetic recombination, and in addition either introduction of genes for pseudogamous seed development in such clones or the development of an efficient system for pseudogamous seed production.Finally, it should be recognized that mutant synaptic genes may impose certain limitations. Because they are generally expressed in both male and female meiosis and thus are either largely sterile or produce only functional FDR 2n-gametes; resulting in polyploidization upon crossing, they have to be manipulated in heterozygous condition. Breeding schemes that consist of (i) introducing mutant synaptic genes and genes for FDR 2n-gamete formation in advanced diploids through backcrossing and (ii) subsequent selection of improved mutant synaptic segregants with FDR 2n-gamete formation following intercrossing of advanced heterozygotes, would be appropriate but laborious. As to FDR 2n-megaspore formation it should moreover be realized that mutant synaptic conditions are actually required. Since heterozygous diploid clones are normal synaptic and thus at best form SDR 2n-megaspores the question remains how to predict whether or not such clones carry genes that bring about substantial FDR 2n-megaspore formation in derived synaptic mutants. If the hypothesis holds true, that SDR and FDR 2n-megaspore formation are caused by common genes for division precocity (Chapter 6), the occurrence of SDR 2n-megaspore formation through postreductional precocious chromosome division in normal synaptic heterozygotes, might be a helpful criterion.