The nearness of you: the effect of population structure on siring success in a gynodioecious species



Lynda F. Delph, E-mail:


Theoretically, both balancing selection and genetic drift can contribute to the maintenance of gender polymorphism within and/or among populations. However, if strong differences exist among genotypes in the quantity of viable gametes they produce, then it is expected that these differences will play an important role in determining the relative frequency of the genotypes and contribute to whether or not such polymorphism is maintained. In this issue, De Cauwer et al. (2010) describe an investigation of gynodioecious wild sea beet, which in addition to containing females, contain two types of hermaphrodites: restored hermaphrodites carrying a cytoplasm that causes pollen sterility and a nuclear gene that restores pollen fertility, and hermaphrodites without the sterilizing cytoplasm. The results show that restored hermaphrodites, who have relatively low pollen viability, achieve disproportionately high siring success simply because of where they are located in a patchy population (Fig. 1). Notably, these individuals tend to be close to females because of the genetics of sex determination. These results indicate that population structure caused by drift processes can have an unexpectedly large effect on the fitness of these low quality hermaphrodites, thereby contributing in the short term to the maintenance of gynodioecy in this population. While these results indicate that population structure caused by drift processes can have a large effect on the relative fitness of genetic variants, whether these effects promote or discourage the maintenance of polymorphism in the long term is still up for debate.

Figure 1.

 A stretch of beach along which wild sea beet can be seen to be growing among the rocks above the splash zone. This linear arrangement enhances the potential for mating success to depend on proximity to other plants (Photo: J.-F. Arnaud).

Nuclear-cytoplasmic gynodioecy is a breeding system of plants that involves polymorphism. The first part of the name refers to the fact that there are two genomes involved in the sex determination of individuals, the nuclear and mitochondrial genomes. In order for this type of gynodioecy to be maintained, the sex-determining loci in each of these genomes must be polymorphic. The second part refers to there being two morphs in the populations, females and hermaphrodites. Given that any polymorphism is difficult to maintain, it is a wonder that any species exhibits nuclear-cytoplasmic gynodioecy; however, gynodioecy is common and widespread, occurring in approximately 7% of flowering plants (Richards 1986).

Gynodioecy evolves when a female mutant enters a population consisting of only hermaphrodites and spreads. We are now aware that, for the majority of cases, these invading female mutants are female or ‘male-sterile’ because they contain a gene causing male sterility, which is located in their mitochondria (Bailey & Delph 2007). Species can contain more than one type of male-sterility gene, but maintaining these genes requires that they confer enhanced seed fitness (either quality and/or quantity) for females relative to hermaphrodites, to compensate for the loss of fitness via pollen. This is termed ‘compensation’.

When one of these genes first enters a population, females are female because they have the gene and hermaphrodites are hermaphrodites because they do not. However, this scenario fixes positive selection for a second mutation, this time in the nucleus, to counteract the male-sterility gene and restore the ability to make pollen. These genes are termed nuclear restorers. Once a restorer enters a population, then there may be two types of hermaphrodites—those with no male-sterility gene and those with a male-sterility gene and the restorer that counteracts it. Importantly, carrying the nuclear restorer alleles must incur some cost (e.g. lower pollen viability), such that hermaphrodites that carry the alleles have lower fitness than hermaphrodites without restorers in order for both types to be maintained (Delannay et al. 1981; Bailey et al. 2003; Dufay et al. 2007). This is termed ‘cost’.

The particular values of the compensation and cost terms will affect the equilibrium frequency of females and whether nuclear-cytoplasmic gynodioecy is maintained. Of course, it is also perfectly possible that gynodioecy exists not as a universal equilibrium condition but rather is stable only at a metapopulation level. This possibility is appealing considering the remarkable sex-ratio variation often observed among gynodioecious populations of the same species, and the correlation seen within many species between population size or disturbance regimes and female frequencies (Byers et al. 2005; Nilsson & Ågren 2006). For example, fairly young populations often contain high frequencies of females as a consequence of founder effects, and as the population ages and restorer alleles migrate in, the frequency of females declines (Belhassen et al. 1989).

Data sets appropriate for testing whether equilibrium or metapopulation processes rule the day have come slowly as they rely on in-depth knowledge of genotypic, phenotypic, demographic and ecological variables. Most studies deal with a limited set of variables. For example, some gynodioecious species have been shown to have fine-scale patch structure, which has been interpreted as being caused by limited dispersal of seeds from parent plants (Gehring & Delph 1999). And the siring success of males of a dioecious lily was shown to vary depending on proximity to females and not on the number of flowers they made, a somewhat counterintuitive result given that plants that make more flowers should be producing more pollen (Smouse et al. 1999). Given these results, one could, in theory, ask the following question. If (i) a gynodioecious population is patchy (because of founder effects/limited seed dispersal), and (ii) restorer alleles carry a cost in terms of pollen characteristics, what will be the major factor affecting the siring success of the various hermaphrodites? Will it be their relative pollen quality and/or quantity or will patchiness override such equilibrium inducing forces? Answering such a question requires that one knows whether a hermaphrodite is or is not restored. However, knowledge of the underlying genetics controlling whether an individual is a female or a hermaphrodite, although well characterized for many crop species, is rarely known for naturally occurring gynodioecious species (Delph et al. 2007).

One exception is a nuclear-cytoplasmic gynodioecious species that exists both as a crop and weed, beet. Using genetic markers developed for crop varieties, as well as some from wild populations, researchers have begun compiling data sets that can differentiate between mitochondrial alleles and types of hermaphrodites in natural populations. This issue of Molecular Ecology presents an investigation that utilizes these data and that has some surprising results. In their study, De Cauwer et al. (2010) demonstrate that the patchy distribution of mitochondrial male-sterility genes of wild sea beet (Beta vulgaris ssp. maritima) places restored hermaphrodites in neighbourhoods with high female frequencies relative to non-restored hermaphrodites. Despite the relatively low quality of pollen from restored hermaphrodites, which may be caused by either incomplete restoration (Dufay et al. 2008) or a constitutive cost of restoration, these hermaphrodites have higher siring success than non-restored hermaphrodites. They owe this unexpected, enhanced success to their proximity to females in a population that exhibits population structure and patchiness. Perhaps most surprising is that this effect remains although beets are wind pollinated and are frequently subject to long-distance pollen dispersal.

The effect of such fine-scale population structure on the stability of nuclear-cytoplasmic gynodioecy is difficult to predict. On the one hand, the increased siring success of restored hermaphrodites increases the average fitness of restorer alleles. Theory predicts that large fitness differences between restorer and maintainer alleles are necessary to maintain both alleles in populations; therefore, we may expect that spatial structure will not be associated with stable nuclear-cytoplasmic gynodioecy. However, the authors argue that ‘strong spatial genetic structure is likely to amplify the expected effects of frequency dependent selection by clustering together similar genotypes’. We interpret this to mean that spatial structuring will enhance the maintenance of nuclear-cytoplasmic gynodioecy when compensation and cost levels allow because selection can more efficiently select against restorer alleles and maintain nuclear-cytoplasmic gynodioecy when hermaphrodites with different nuclear restorer genotypes directly interact and compete within population patches.

A new model by Dufay & Pannell (2010) that specifically addresses the effects of seed and pollen flow on the maintenance of nuclear-cytoplasmic gynodioecy finds that genetic drift does not support stable nuclear-cytoplasmic gynodioecy, even when selection would favour it. Conversely, seed migration and selection can maintain polymorphism within populations, whereas pollen flow and selection can maintain polymorphism among (but not within) demes of a metapopulation. This prediction fits nicely with previous studies that have found that founder effects and drift alone cannot explain patterns of sex-ratio variation in nature (Caruso & Case 2007; Dufay et al. 2009). We think selection also has a role in determining the long-term dynamics of nuclear-cytoplasmic gynodioecious species, although its effects will need to be carefully teased from data sets that include effects of mitochondrial male-sterility alleles, restorer alleles, fine-scale population structure, and short- and long-distance seed and pollen flow. De Cauwer et al. (2010) is an intriguing first attempt at combining most of these factors, and sets the stage for future investigations of this type.

L.F.D. is an evolutionary ecologist interested in breeding-system evolution and sexual dimorphism in plants. M.F.B. is a teacher and population biologist interested in the interplay between selection and diversity, particularly in plant breeding systems. Lynda and Maia have a long history of working together on both theoretical and empirical studies of gynodioecy.