| Abstract: | Female control of nonrandom mating has never been genetically established, despite being linked to inbreeding depression and sexual selection. In order to map the loci that control female-mediated nonrandom mating, we constructed a new advanced intercross recombinant inbred line (RIL) population derived from a cross between Arabidopsis (Arabidopsis thaliana) accessions Vancouver (Van-0) and Columbia (Col-0) and mapped quantitative trait loci (QTLs) responsible for nonrandom mating and seed yield traits. We genotyped a population of 490 RILs. A subset of these lines was used to construct an expanded map of 1,061.4 centimorgans with an average interval of 6.7 ± 5.3 centimorgans between markers. QTLs were then mapped for female- and male-mediated nonrandom mating and seed yield traits. To map the genetic loci responsible for female-mediated nonrandom mating and seed yield, we performed mixed pollinations with genetically marked Col-0 pollen and Van-0 pollen on RIL pistils. To map the loci responsible for male-mediated nonrandom mating and seed yield, we performed mixed pollinations with genetically marked Col-0 and RIL pollen on Van-0 pistils. Composite interval mapping of these data identified four QTLs that control female-mediated nonrandom mating and five QTLs that control female-mediated seed yield. We also identified four QTLs that control male-mediated nonrandom mating and three QTLs that control male-mediated seed yield. Epistasis analysis indicates that several of these loci interact. To our knowledge, the results of these experiments represent the first time female-mediated nonrandom mating has been genetically defined.The process of pollination offers plants the opportunity to selectively breed. For example, in pollinations that include more than one pollen population, pollen often show differential siring ability. This process is called nonrandom mating. Although pollen may fail in pollinations because they are self pollen in an obligate outcrossing plant or pollen from a different species, we focus our studies on differential siring ability of compatible, conspecific mates (Hogenboom, 1973, 1975; Williams et al., 1999; de Nettancourt, 2001; Husband et al., 2002; Wheeler et al., 2009; Meng et al., 2011; Nasrallah, 2011). Nonrandom mating at this level has received intense interest for its potential to avoid inbreeding depression and its potential to be the result of sexual selection (Charnov, 1979; Mulcahy, 1979; Willson, 1979; Queller, 1983; Stephenson and Bertin, 1983; Willson and Burley, 1983; Marshall and Ellstrand, 1986; Charlesworth and Charlesworth, 1987; Mulcahy and Mulcahy, 1987; Cruzan, 1990; Quesada et al., 1993; Snow, 1994; Paschke et al., 2002; Skogsmyr and Lankinen, 2002; Stephenson et al., 2003; Armbruster and Rogers, 2004; Bernasconi et al., 2004; Lankinen and Armbruster, 2007). Despite a long history of theoretical and experimental attention, very little is known about the underlying genetics that govern the process (Carlson et al., 2011).One challenge in understanding the genetics of nonrandom mating lies in its complexity, potentially involving multiple distinct pathways specific to either female or male tissues. Physiologically, postpollination nonrandom mating may be a result of intrinsic differences in pollen competitive abilities (male-mediated nonrandom mating). A number of experimental strategies have been employed to demonstrate male-mediated control of nonrandom mating. For example, experiments in radish (Raphanus sativus) found that some pollen sire more seeds than others in mixed pollinations across a range of maternal plants, demonstrating consistency of male function (Marshall and Ellstrand, 1986, 1988; Mitchell and Marshall, 1998). More direct measures of male function, such as in vitro and in vivo pollen tube growth rates, verify variation in male function and demonstrable impact on nonrandom mating (Snow and Spira, 1991a, 1991b; Pasonen et al., 1999; Skogsmyr and Lankinen, 1999; Stephenson et al., 2001; Lankinen and Skogsmyr, 2002; Lankinen et al., 2009). Finally, recent work in our laboratory has directly mapped the genetic loci responsible for the control of male-mediated nonrandom mating in Arabidopsis (Arabidopsis thaliana; Carlson et al., 2011).Alternatively, or concurrently, nonrandom mating can be the result of differential interaction between the female tissue and competing pollen populations or seeds (female-mediated nonrandom mating). Establishing the female role in nonrandom mating has been more challenging, as most study designs involve the deposition of pollen from multiple donors and thus include the confounding variable of pollen competition. Despite this challenge, a number of experimental strategies have been devised to explore the role of the female in nonrandom mating. For example, a number of studies demonstrate that maternal identity influences nonrandom mating patterns (Marshall and Ellstrand, 1986, 1988; Snow and Mazer, 1988; Johnston, 1993; Marshall et al., 2000; Carlson et al., 2009, 2013). Studies have also established that manipulation of watering or nutrient regimes of maternal plants changes the patterns and magnitude of nonrandom mating (Marshall and Diggle, 2001; Shaner and Marshall, 2003; Haileselassie et al., 2005; Marshall et al., 2007). These studies and others implicate the identity and condition of the female in the process of nonrandom mating. Despite a long history of research, genetic control of female-mediated nonrandom mating has never been demonstrated, and the identity of the genes involved remains unexplored.In previous work, we developed a system in Arabidopsis to assay nonrandom mating and showed its utility for genetically mapping the loci responsible (Carlson et al., 2009, 2011). Pursuing the genetics of nonrandom mating in a largely selfing plant such as Arabidopsis provides both theoretical and practical advantages. First, outcrossing plants carry higher levels of heterozygosity that produce pollen populations that display different phenotypes because of segregating alleles. This complicates genetic analysis. Also, in outcrossing plants that carry genetic load, reproductive success is context dependent. Pollinations with self pollen or pollen from genetically similar plants often lead to poor reproductive outcomes. For example, in mixed pollinations in generally outcrossing self-compatible plants that include self pollen, self pollen often sire a disproportionally low number of seeds (Bateman, 1956; Weller and Ornduff, 1977; Bowman, 1987; Eckert and Barrett, 1994; Jones, 1994; Hauser and Siegismund, 2000; Teixeira et al., 2009), but other findings have been reported (Sork and Schemske, 1992; Johnston, 1993). Thus, in outcrossing plants, gene variants that influence reproductive success, parental relatedness, and segregating heterozygosity all influence reproductive outcomes. Two of these factors are essentially eliminated by studying plant populations that have historically selfed. As outcrossing populations become increasingly self-fertilizing, they both lose heterozygosity, and their genetic load is purged (Lande and Schemske, 1985; Schemske and Lande, 1985; Charlesworth and Charlesworth, 1987; Lande et al., 1994; Byers and Waller, 1999; Crnokrak and Barrett, 2002). This is the case for Arabidopsis, whose tested populations show relatively low levels of heterozygosity and little evidence for the early-acting inbreeding depression that is indicative of genetic load (Bakker et al., 2006; Bomblies et al., 2010; Platt et al., 2010; Carlson et al., 2013). Thus, this system provides an excellent opportunity to identify and explore the genetic variation in differential reproduction that develops or persists in plant populations unrelated to inbreeding depression.Using this system, we previously identified potential female control of nonrandom mating in mixed pollinations between Vancouver (Van-0) and Columbia (Col-0) accessions of Arabidopsis (Carlson et al., 2009). When Van-0 and genetically marked Col-0 (Col-NPTII) pollen compete on Col-0 pistils, Col-NPTII pollen sire 43% of the progeny, while Van-0 pollen sire 57%. When these pollen compete on Van-0 pistils, Col-NPTII pollen sire 67.5% of the progeny, while Van-0 pollen sire 32.5%. This system offers us, to our knowledge for the first time, the opportunity to genetically define female-mediated nonrandom mating and map the loci responsible.In order to genetically map female control of nonrandom mating, we constructed a new advanced intercross recombinant inbred line (RIL) mapping population derived from a cross between Van-0 and Col-0 accessions of Arabidopsis. RILs are powerful tools that allow high-resolution genetic mapping of loci that direct complex traits. Each RIL contains chromosomes that are defined homozygous patchworks of parental DNA, in this case Van-0 and Col-0. By phenotyping these lines, we can statistically associate nonrandom mating and seed yield phenotypes with chromosomal regions. We chose these two accessions because (1) our previous experiments predict clear female control of nonrandom mating and (2) we have previously mapped male-mediated nonrandom mating controls using a Col-4/Landsberg mapping population (a population that does not display female control of nonrandom mating; Carlson et al., 2011). Thus, this new population provides us the opportunity to map loci that control female nonrandom mating and investigate the degree of conservation of loci that affect male-mediated nonrandom mating. We use this new mapping population to perform quantitative trait locus (QTL) mapping and identify multiple loci that direct both female- and male-mediated control of nonrandom mating and seed yield traits. |
