Drift in small populations predicts mate availability and the breakdown of self-incompatibility in a clonal polyploid

Abstract

Introduction

Self-incompatible (SI) breeding systems promote outcrossing, maintain genetic diversity (Igic et al., 2008), and alleviate inbreeding depression (Charlesworth & Charlesworth, 1979). Two primary SI mechanisms in plants are recognized, sporophytic (SSI) and gametophytic SI (GSI). Both are widely distributed across angiosperms spanning phylogenetically diverse families and geographic regions (Igic et al., 2008). Despite their benefits, the long-term stability of SI mechanisms within populations is dependent on ecological conditions and population genetic diversity. Specifically, sexual reproduction in SI populations can be depressed compared with populations capable of self-fertilization due to a limitation of compatible mates. Not only must SI flowers receive an adequate number of pollen grains, but the grains deposited must be genetically distinct from the maternal plant at the self-incompatibility locus (hereafter, S-locus) for successful reproduction (de Nettancourt, 1977). Low S-allele diversity within a population causes mate limitation which can reduce reproductive fitness (Byers & Meagher, 1992; DeMauro, 1993; Busch & Schoen, 2008). For instance, seed production is more strongly limited by the receipt of low-quality pollen in SI species than self-compatible species (Cisternas-Fuentes et al., 2023), which is likely driven by the receipt of pollen with shared S-alleles. Strong mate limitation can severely limit population growth and increase the probability of extinction (Busch & Schoen, 2008). These negative effects of mate limitation are especially exacerbated in small populations where sexual reproduction can be lost altogether (Routley et al., 1999; Barrett, 2015). Even with high amounts of pollen transfer, reproduction may fail in small populations harboring very few S-alleles (Young & Pickup, 2010).

Under severe mate limitation, the breakdown of SI systems can rescue population decline. The loss of the SI mechanisms is common, directional (Nasrallah et al., 2002), and irreversible (Stone, 2002; Igic et al., 2008). The most common causes for the breakdown of the SI mechanisms are mutations in the S-locus itself, or modifiers of the S-locus (Stone, 2002). An intermediate step toward the breakdown of SI is the expression of pseudo-self-compatibility (Levin, 1996) through leaky SI. Leaky SI refers to seed production through self-fertilization in SI species that occurs at a lower rate than seed production through outcrossing (Busch & Schoen, 2008). Mutations enabling self-reproduction are often present at low frequencies in SI populations (Busch et al., 2010) and leaky SI often occurs at different rates across populations (Good-Avila & Stephenson, 2002; Nielsen et al., 2003; Mena-Ali & Stephenson, 2007). Because leakiness in self-recognition systems varies within populations and is heritable (Good-Avila & Stephenson, 2002; Mena-Ali & Stephenson, 2007; Baldwin & Schoen, 2017), it should be favored under severe mate limitation because it promotes reproductive assurance.

Theory predicts that the breakdown of SI systems should be more common in polyploids than diploids for two primary reasons. First, selection should favor self-compatibility in mate-limited minority cytotypes following genome duplication events (Sutherland et al., 2018). Second, polyploids may experience reduced inbreeding depression relative to diploids due to multiple gene copies at loci causing inbreeding depression (Layman & Busch, 2018). Experimental evidence shows reduced inbreeding depression in tetraploids relative to diploids (Ozimec & Husband, 2011) and wider variability in self-incompatibility in synthetic neotetraploids (Siopa et al., 2020). In GSI systems in particular, the breakdown of SI may also be driven by the potential for diploid (or greater) pollen to carry more than one S-allele with the failure to reject pollen grains that are heterozygous at the S-locus (i.e. ‘heteroallelic’; Robertson et al., 2011). Heteroallelic breakdown of self-incompatibility systems appears to be relatively restricted in the angiosperm family tree, yet the relative paucity of studies investigating these reproductive systems in polyploids limits a fuller understanding of its importance (Bleeker, 2004; Willi et al., 2005).

Regardless of a population's ploidy level, more S-alleles should be maintained in larger populations (Vekemans et al., 1998; Busch & Schoen, 2008), since genetic drift causes the loss of S-allele diversity (Wright, 1939; Lawrence, 2000). Theoretical connections between S-allele number and mate availability have been established in simple diploid models (Vekemans et al., 1998; Vallejo-Marín & Uyenoyama, 2004; Young & Pickup, 2010). To our knowledge, relationships between population size, S-allele number, and mate availability have not been extended to higher ploidy levels (e.g. tetraploids). On the one hand, the fact that polyploid pollen harbors more S-alleles would predict relatively low mate availability in tetraploids compared with otherwise identical diploids. On the other hand, tetrasomic inheritance generates a greater array of pollen genotypes (Bever & Felber, 1992), which increases the probability that any one pollen grain is compatible with a maternal S-locus genotype. For instance, a heterozygous maternal tetraploid genotype at the S-locus (S₁S₂S₃S₄) can produce offspring using diploid pollen made by genotypes sharing two S-alleles in common (e.g. an S₁S₂S₅S₆ pollen parent, which generates S₅S₆ pollen grains). By contrast, all plants sharing two S-alleles in common are fully cross-incompatible in an otherwise identical diploid system – suggesting relatively high mate availability in tetraploids compared with diploids. Given these important considerations, there is a general need to develop simple population genetic simulations that aid the interpretation of studies on the function of SI systems in natural polyploid populations (Bleeker, 2004; Willi et al., 2005).

Here, we evaluated patterns of seed production from self- and outcross pollinations in 13 populations of a tetraploid herb, Argentina anserina, with genomic estimates of effective population size (Cisternas-Fuentes & Koski, 2023) to quantify mate limitation and the self-compatibility index (SCI). We couple crossing results with detailed histological analyses at the pollen–pistil interface to assess the causes of reproductive failure and the magnitude of leakiness in self-recognition. Previous studies in this system demonstrate that populations in Southwest Colorado are effectively small (1–15 N_E; Cisternas-Fuentes & Koski, 2023), and that recognition and arrest of self-pollen occurs after pollen germinates, often in the upper ⅔ of the style which is characteristic of GSI (Franklin-Tong & Franklin, 2003). Some level of leakiness in the SI mechanism may exist due to the occasional production of seed from self-pollinations (Cisternas-Fuentes et al., 2023) though among-population differences in leakiness have not been assessed. We addressed the following questions:

1. Is mate availability reduced in populations with small effective population size?

   1. At what stage of the pollen–pistil interaction does outcross reproductive failure occur?

2. How do the empirical relationships between S-allele diversity and mate availability relate to that predicted from population genetic simulations in tetraploids?
3. Do smaller populations with lower mate availability exhibit leakier self-recognition?
4. Does leakiness predict variation in the SCI and contribute to variation in seed production at the individual plant and population levels?