A genetic dissection of domestication-related traits in pea

Crop domestication refers to the process in which wild plants have become adapted for agricultural purposes. In different crops, common selective pressures unique to cultivation environments or agricultural practises have led to the accumulation of similar sets of traits, known as the domestication syndrome. Understanding the underlying genetic basis for domestication syndrome traits has been of interest for both plant breeders and evolutionary geneticists, as it can provide a window into the evolutionary history of a trait, lineage or crop species and assist in crop improvement efforts. This study investigated the genetic control of domestication and diversification in pea (Pisum sativum L.); an important global legume crop and prominent genetic model system. As in other crop legumes, major phenotypic changes during this transition have included loss of seed dormancy and pod shattering, both considered to be critical domestication traits, as well as other changes such as earlier flowering, reduced branching and development of a more robust growth habit, which have been considered diversification or crop improvement traits. This study examined the genetic control of several of these traits in a biparental cross between the domesticated P. sativum var sativum line NGB5839 (a near-isogenic derivative of cultivar Torsdag) and the wild P. sativum var humile line JI1794, a representative of the northern humile‚ÄövÑvp lineage which is considered a major contributor to the modern day domesticated pea. It consisted of the generation of a high-density linkage map, QTL mapping, and the evaluation of specific QTL regions for dormancy-related and flowering time traits. In Chapter 3 a high density, high confidence genetic linkage map was developed for a population of 137 F\\(_8\\)+ Recombinant Inbred Lines (RILs) derived from the JI1794 x NGB5839 cross. This map incorporated 4599 DArT markers spanning a total of 1617cM. This exhibited good coverage, with an average of roughly three markers per cM with no markers >10cM apart. Map assembly was validated and assessed by a detailed examination of synteny with the well-characterized Medicago truncatula genome, and with four other legume genomes: Lens culinaris, Trifolium pratensis, Cicer arietinum, and Phaseolus vulgaris. This is the first study to compare the syntenic relationship of pea with recently sequenced T. pratensis and L. culinaris. L. culinaris is currently the most closely related species to pea for which a genome sequence is publicly accessible, and the high level of synteny makes this a useful new reference point for pea genetics. The resulting map was used in chapter 4 as the basis for QTL analysis of a range of domestication-related and other traits in the RIL population when grown under extended natural long-day conditions. Most traits were found to show relatively simple genetic control and, with only a few exceptions, conditioned by one or two major loci (PEV ‚Äöv¢‚Ä¢15%) and several minor loci (PEV ‚Äöv¢¬ß 15%). QTL clustering was observed in four genomic regions near LF, Hr, LE and a region on LGVI, possibly in part reflecting pleiotropic effects of flowering time genes on other traits such as development rate and growth habit. This chapter provides new insight into the basis for pea domestication and subsequent diversification episodes and identifies major loci for dormancy and flowering time. Chapters 5 and 6 explore in more detail the genetic control of several traits potentially related to seed dormancy. QTL analysis for permeability identified two previously unreported loci; a major locus on LGII (Perm2) and a minor locus on LGVII (Perm7). Perm2 mapped close to major testa thickness QTL TT2 and the well-known Mendel A locus. In Chapter 5, the effect of Perm2 and TT2 were validated in segregating progenies derived from the JI1794 x NGB5839 F\\(_2\\)population used for the generation of RILs, which were confirmed to have strong linkage to the A loci. The widespread existence of free-germinating pigmented domesticated lines suggests that the major change in permeability during domestication may not be caused by the A gene itself. Mutants for the A-interacting protein A2 showed reduced testa thickness but no difference in permeability. This indicating that reduced testa thickness, may depend on biosynthetic pathways regulated by the A/A2 complex however permeability remains inconclusive. Loss of seed coat roughness and its co-ordinated emergence with the domesticated pea has previously been understood to implicate it in its control of seed dormancy. A major locus for roughness corresponding to the GTY locus was detected on LGVI, while a second novel QTL was identified on LGVII. Chapter 6 investigated the genetic control of seed coat roughness in more detail and explored its relationship with testa thickness and permeability. Using advanced generation segregating populations, GTY was confirmed to co-segregate with increased testa thickness and was fine mapped to a region inferred to contain around 50 genes based on a pea transcriptome linkage map. A second testa thickness QTL in this region which had not been observed in the original QTL analysis was identified and shown to be closely linked to but genetically independent of GTY. Permeability experiments showed seed coat roughness and testa thickness on LGVI had no effect on hardseededness. This has been the first time the effect of testa roughness and thickness on permeability has been tested using advanced segregating populations and has provided the most detailed mapping to date of the GTY locus. Chapter 7 explored in more detail the genetic control of newly identified flowering time QTLs on LGII and LGV (DTF2 and DTF5). Like many other species, domesticated peas in general flower earlier with reduced photoperiod sensitivity. Among five long-day flowering time QTLs identified in Chapter 4, two (DTF3a and DTF6) had previously been identified in short-day conditions in F\\(_2\\) progeny of the same cross, while the third (DTF3b) mapped to the region of Mendel's Le. Using advanced generation segregating populations and expression analysis, floral inhibitory gene LATE FLOWERING (LF) was considered a likely candidate for DTF2, while DTF5 mapped to a region known to include the FTa1-FTa2-FTc cluster of florigen genes. Plants carrying the NGB5839 alleles at DTF2 and DTF5 showed elevated expression of both LF and FTa1 genes respectively and is consistent with the direction of the respective QTL effects, therefore suggesting that DTF2 and DTF5 might represent gain-of function variants in NGB5839. Involvement of the LGV FT gene cluster in natural variation for flowering time has not previously been documented. Overall the results in this thesis make a significant contribution to our understanding to domestication and diversification of pea, and significantly extend previous studies on the genetic control of traits related to seed dormancy and flowering. In addition, the generation of a high-quality map will provide an ongoing resource for future analyses of a wider range of other domestication-related traits.