Abstract Background and Aims Capsella is a model genus for studying the transition from outcrossing to selfing, with or without change in ploidy levels. The genomic consequences and changes in reproductive traits (selfing syndrome) associated with these shifts have been studied in depth. However, potential ecological divergence among species of the genus has not been determined. Among ecological traits, competitive ability could be relevant for selfing evolution, as selfing has been shown to be statistically associated with reduced competitiveness in a recent meta-analysis. Methods We assessed the effect of competition on three Capsella species differing in their mating system and ploidy level. We used an experimental design where fitness related traits were measured in focal individuals with and without competitors. Key Results The diploid selfer (C. rubella) was most sensitive to competition, whereas the tetraploid selfer (C. bursa-pastoris) performed the best, with the diploid outcrosser (C. grandiflora) being intermediate. Conclusions These results add to the detailed characterization of Capsella species and highlight the possible roles of ecological context and ploidy in the evolutionary trajectories of selfing species. Capsella, competition, mating system, outcrossing, ploidy, selfing INTRODUCTION Self-fertilization has frequently evolved from outcrossing in plants and also in some groups of animals, fungi and algae (Jarne and Auld, 2006, Billiard et al., 2012, Igic and Busch, 2013). If not counteracted by too severe inbreeding depression, a selfing variant can rapidly invade thanks to its 50 % advantage in gene transmission and the reproductive assurance it procures when mate availability is limiting (Charlesworth, 2006). The shift to selfing is often associated with a suite of reproductive, morphological and physiological changes – such as flower size reduction or reduced investment in male function – that collectively define the so-called selfing syndrome (Sicard and Lenhard, 2011). In addition to the reduction in pollen gene flow induced by selfing, divergence in reproductive traits can promote reproductive isolation, making selfing a potential speciation mechanism (Wright et al., 2013). This has been shown in Solanaceae where shifts from outcrossing to selfing were preferentially associated with speciation (cladogenic shifts) rather than species replacement (anagenic shifts) (Goldberg and Igic, 2012). Beyond reproductive divergence, species must also diverge ecologically to allow species coexistence (Coyne and Orr, 2004). If two species diverge only in their reproductive mode, one species should exclude the other through competition. Habitat shift is also observed between selfers and outcrossers. For example, selfers are more frequent under conditions in which obligatory outcrossers pay the demographic cost of mate limitation, such as disturbed and patchy habitats or newly available environments (Baker, 1967; Munoz et al., 2016). Selfing is thus often associated with weedy habit (Clements et al., 2004) and invasiveness (van Kleunen et al., 2008). Based on Grime’s. competitor, stress-tolerator, ruderal (CSR) ecological strategies, a recent meta-analysis confirmed this trend and found an excess of selfers among ruderal species, corresponding to early colonizers in an ecological succession, whereas more competitive species tend to be outcrossers (Munoz et al., 2016). This could be explained by a colonization/competition trade-off, with the evolution of better colonizing ability being at the cost of lower competitive ability (Burton et al., 2010). Alternatively, selfing can have strong negative genetic effects in the long run, including the accumulation of weakly deleterious mutations (Wright et al., 2008; Glémin and Galtier, 2012), which could reduce competitive ability (Agrawal, 2010; Agrawal and Whitlock, 2012). For example, inbred individuals often suffer more from competition than outbred individuals (Cheptou et al., 2000, 2001; Yun and Agrawal, 2014), but whether this pattern also applies at the species level is unknown. The negative genetic consequences of selfing are also classically proposed to explain the high extinction rate of selfing lineages, according to the dead-end theory of selfing evolution (Stebbins, 1957; Igic and Busch, 2013; Wright et al., 2013). However, selfing species can appear ecologically and demographically successful, yet at the same time show signatures of genomic degradation (e.g. in the plant genus Collinsia: Randle et al., 2009; Hazzouri et al., 2013, or in some fresh-water snails: Burgarella et al., 2015). A better characterization of ecological strategies associated with the evolution of selfing could improve our understanding of this apparent paradox. Under weak competition – when selection is softer (Agrawal, 2010) – selfing species could cope with high genetic loads whereas their persistence would be compromised under more intense competition. As it is often associated with selfing/self-compatibility (Barringer, 2007; Robertson et al., 2011) polyploidy could also play a role in the fate of selfing lineages. Polyploidy is reported to increase competitiveness (Comai, 2005; te Beest et al., 2012) and to reduce genetic load, at least for a transient period after the formation of the polyploid species (Fisher, 1935; Otto and Whitton, 2000). Thus, polyploidy could buffer the deleterious effects of selfing. Whatever the underlying causes, reduced competitive ability, modulated by the ploidy level, could also be a component of the selfing syndrome. However, while changes in flower traits associated with the evolution of selfing have been well characterized among closely related species (e.g. Fishman et al., 2002; Sicard and Lenhard, 2011; Kalisz et al., 2012), we are not aware of any study characterizing competitive ability of selfing versus outcrossing-related species. Capsella (Brassicaceae) is a model genus for studying the transition to selfing (Foxe et al., 2009; Guo et al., 2009; Slotte et al., 2013), polyploidization (Slotte et al., 2008; Douglas et al., 2015) and their association with speciation mechanisms (Rebernig et al., 2015; Sicard et al., 2015). The genus comprises four closely related species with contrasting mating systems and ploidy. The only outcrossing species in the genus, C. grandiflora, is restricted to western Greece and Albania. The three selfing species include two diploids: C. orientalis diverged from C. grandiflora about 1 Mya and C. rubella much more recently (30000–50000 years ago; Foxe et al., 2009; Guo et al., 2009). The third selfing species, C. bursa-pastoris, is an allo-tetraploid with disomic inheritance that originated about 200000 years ago as a hybrid between C. orientalis and C. grandiflora (Douglas et al., 2015). A second tetraploid species, C. thracica, has also been considered (Hurka et al., 2012), but its distinction from C. bursa-pastoris is not clear as the proposed scenarios for the two tetraploid species incorrectly assumed an autopolyploid origin of C. bursa-pastoris. The selfing species have much larger ranges than C. grandiflora, especially the tetraploid C. bursa-pastoris, with an almost worldwide distribution owing to recent colonization (Cornille et al., 2016). The reproductive changes associated with the selfing syndrome and their genetic bases have been studied in depth, especially in C. rubella (Sicard et al., 2011, 2016; Slotte et al., 2012; Neuffer and Paetsch, 2013) and all four species have been well characterized genetically. Capsella grandiflora has a very large effective population size and selection is very efficient (Slotte et al., 2010; Williamson et al., 2014) whereas the selfing species have much less genetic diversity than C. grandiflora and a strong genomic signature of accumulation of weakly deleterious mutations due to relaxed purifying selection (St Onge et al., 2011; Brandvain et al., 2013; Slotte et al., 2013; Douglas et al., 2015). To better characterize the ecological divergence among these species, we evaluated competitive ability of three of the four species, C. grandiflora, C. rubella and C. bursa-pastoris, whose distributions overlap in Greece and the Balkans. In a competition experiment we evaluated the sensitivity of the three species to competition and what level of competition they imposed on the other species. Our working hypothesis was that C. rubella should be more sensitive to competition than C. grandiflora and C. bursa-pastoris, and that C. grandiflora and C. bursa-pastoris should impose stronger competition than C. rubella. We discuss the potential implications of the results for the evolution of selfing species. MATERIAL AND METHODS Plant sampling and preparation We studied three species of the genus Capsella: C. grandiflora (Fauché & Chaub.) Boiss. (diploid outcrosser), C. rubella Reut. (diploid selfer) and C. bursa-pastoris (L.) Medik (tetraploid selfer). We used 23 C. grandiflora maternal plants from ten populations, 20 C. rubella from 11 populations and 14 C. bursa-pastoris from six populations (listed in Supplementary Data Table S1). The different numbers among species were due to differences in the percentage of germination and survival of seedlings. To minimize ecological differences among species, all plants were sampled in the Greece/Balkans area where the three species can co-occur at the same locality (our personal observations). We also included an ‘outgroup’ competitor to impose a similar inter-species competition treatment to the three focal species. We chose Matricaria chamomilla, an annual Asteraceae species, which was found to co-occur with Capsella species in Greek populations and at least with C. bursa-pastoris in other European populations (our personal observations). More generally, Capsella species and M. chamomilla have overlapping distributions (Europe and temperate Asia) and a weedy habit, and both form rosettes before flowering.. In addition, commercial seeds were available for M. chamomilla, which ensured good germination and homogeneity among plants. Seeds were surface-sterilized and sown into agar plates. Following sterilization, seeds were stratified in agar plates by keeping them for 5 d at 4 °C in complete darkness. Agar plates were then put into a growth chamber with 12-h/12-h artificial light/darkness cycles at 22 °C, with germination occurring approx. 5 d later. Seedlings that presented a radicle of at least 1 cm and well-developed cotyledons were used for the experiment. For an accession to be used as a focal plant, at least 15 of its seeds had to germinate to provide three replicate offspring for each of the five treatments (see below). All remaining seedlings were randomized and used as competitors. Experimental design We compared the performance of the three Capsella species when grown alone and under four competition conditions: intra-specific competition, inter-specific competition with the two other Capsella species, and inter-specific competition with M. chamomilla. For each treatment, the focal individual was sown in the middle of a square pot of approx. 11 × 11 × 11 cm, either alone or surrounded by four competitors sown at the corners of the pot. The number of competitors and the dimensions of the pots were chosen after preliminary tests that demonstrated competition effects. Each accession was replicated three times, with replicates randomly assigned to three separate growth chambers (blocks). All pots were randomized within each growth chamber. We quantified both vegetative and reproductive performances. The diameter of the rosette of focal individuals was measured 3 weeks after seedling transfer into pots (diameter at t1) and the following 2 weeks (diameter at t2 and t3). Bolting started approximately 1 month after seedling transfer (depending on accessions) and flowering ranged from 3 to 4 weeks. The two selfers C. bursa-pastoris and C. rubella set fruits but the self-incompatible C. grandiflora set almost no fruit, because of the absence of insect or experimental pollination in growth chambers. The plants senesced (visually assessed as yellowing of the leaves) approximately 3 weeks after flowering began. At this time, we recorded the total number of flowers produced by focal individuals, corresponding to the total number of flower peduncles, at the end of the experiment. We used flower number, rather than fruit or seed number, to allow comparison between the three species. Note that we were not interested in absolute fitness comparison among species but rather the relative change in fitness due to competition. Data analysis Data were analysed with generalized linear mixed models in R. For all variables, block, species, treatment and all pairwise interactions were included as fixed effects and accessions as a random effect nested within species. Significance of the different factors was tested with deviance analyses using the Anova function of the car package in R (Fox and Weisberg, 2011). Interactions were tested first; significant interactions were then kept to test for main effects (type III ANOVA). Rosette diameters and growth rate were analysed using a mixed linear model fitted by maximum likelihood with the lme function of the nlme package (Pinheiro et al., 2017). Growth rate was analysed as the difference between diameters at time t3 and t1, adding diameter at time t1 as a covariable to control for possible allometric effects. A preliminary analysis showed strong heteroscedasticity so we added a constant variance structure (using the varIndent function). We chose the variance structure giving the best fit according to Akaike information criterion (AIC) values. The flower number distribution was bimodal with a mode at 0 and another around 700 (see Results). It was therefore analysed in two steps. First, we analysed the proportion of flowering plants with a binomial model and a logit link using the glmer function of the R package lme4 (Bates et al., 2015). Second, we excluded plants that did not flower and analysed flower number with a negative binomial model and a log link with the glmmadmb function of the glmmADMB package (Fournier et al., 2012; Skaug et al., 2013). In this analysis, we were specifically interested in the species × treatment interaction term to determine whether the effect of competition differed among species. To obtain a more direct and intuitive estimate of the sensitivity to competition of Capsella species, we defined the following competition index, comparing the performance of an accession under competition and without competition: Ic=WcompetitionWalone where Walone and Wcompetition are the fitness-related traits of the focal accession without and with competition, respectively. The lower Ic, the more sensitive the accession is to competition. Ic was computed separately for each accession and for the four competition treatments in the three growth chambers. We analysed the Ic index with a gamma model and a log link using the glmer function (Bates et al., 2015). Interpretation of effects in this analysis is more straightforward: the species effect directly means that the species differ in sensitivity to competition and the treatment effect that the species impose different competitive effects. Mean effects were estimated with the lsmean function of the lmerTest package (Kuznetsova et al., 2017). When significant effects were detected, differences between levels of the significant factor were tested using the contrast function, with false discovery rate (FDR) correction for multiple testing. We compared the results of this analysis with competition indices obtained as the ratios of mean flower numbers estimated in competition and alone from the previous analysis (instead of mean effects of ratio). We also used the product of the proportion of flowering and flower number estimates as a proxy of total female fitness and, similarly, took the ratio of estimates in competition and alone. RESULTS After transplantation, seedlings had high proportion of survival (99, 91 and 99 % for C. bursa-pastoris, C. grandiflora and C. rubella, respectively) and flowering (82, 82 and 91 %, respectively, not significantly different, see Table 1). The three species produced a large number of flowers, 607 for C. bursa-pastoris, 1296 for C. grandiflora and 666 for C. rubella on average (Fig. 1). For all traits we detected a significant competition effect, with plants alone performing better than with competitors. The accession effect was significant for all traits and block effect for all traits except diameter at t3 and flowering rate (Table 1 and Supplementary Data Table S2). Table 1. Analyses of deviance for reproductive traits d.f. Proportion of flowering plants Flower number LR P-value LR P-value Block 2 0.87 0.646 13.25 0.001 Species 2 2.92 0.233 52.43 4.13e-12 Treatment 4 11.37 0.023 37.83 1.22e-07 Block × species 4 5.47 0.242 19.26 6.97e-04 Block × treatment 8 8.39 0.396 32.13 8.83e-05 Species × treatment 8 6.55 0.586 17.84 0.022 Accession (random) 6 96.17 < 2.2e-16 84.24 < 2.2e-16 d.f. Proportion of flowering plants Flower number LR P-value LR P-value Block 2 0.87 0.646 13.25 0.001 Species 2 2.92 0.233 52.43 4.13e-12 Treatment 4 11.37 0.023 37.83 1.22e-07 Block × species 4 5.47 0.242 19.26 6.97e-04 Block × treatment 8 8.39 0.396 32.13 8.83e-05 Species × treatment 8 6.55 0.586 17.84 0.022 Accession (random) 6 96.17 < 2.2e-16 84.24 < 2.2e-16 LR, likelihood ratio. View Large Table 1. Analyses of deviance for reproductive traits d.f. Proportion of flowering plants Flower number LR P-value LR P-value Block 2 0.87 0.646 13.25 0.001 Species 2 2.92 0.233 52.43 4.13e-12 Treatment 4 11.37 0.023 37.83 1.22e-07 Block × species 4 5.47 0.242 19.26 6.97e-04 Block × treatment 8 8.39 0.396 32.13 8.83e-05 Species × treatment 8 6.55 0.586 17.84 0.022 Accession (random) 6 96.17 < 2.2e-16 84.24 < 2.2e-16 d.f. Proportion of flowering plants Flower number LR P-value LR P-value Block 2 0.87 0.646 13.25 0.001 Species 2 2.92 0.233 52.43 4.13e-12 Treatment 4 11.37 0.023 37.83 1.22e-07 Block × species 4 5.47 0.242 19.26 6.97e-04 Block × treatment 8 8.39 0.396 32.13 8.83e-05 Species × treatment 8 6.55 0.586 17.84 0.022 Accession (random) 6 96.17 < 2.2e-16 84.24 < 2.2e-16 LR, likelihood ratio. View Large Fig. 1. View largeDownload slide Boxplots of the number of flowers produced by focal individuals for all combinations of species and treatments. Each panel corresponds to a focal species and the treatment is presented on the x-axis. Cbp: Capsella bursa-pastoris; Cg: Capsella grandiflora; Cr: Capsella rubella; out: ‘outgroup’ = Matricaria chamomilla. Non-flowering plants excluded. Fig. 1. View largeDownload slide Boxplots of the number of flowers produced by focal individuals for all combinations of species and treatments. Each panel corresponds to a focal species and the treatment is presented on the x-axis. Cbp: Capsella bursa-pastoris; Cg: Capsella grandiflora; Cr: Capsella rubella; out: ‘outgroup’ = Matricaria chamomilla. Non-flowering plants excluded. For vegetative traits we detected no difference in sensitivity to competition treatment among species. The three species did not differ significantly, either for rosette diameter (Supplementary Data Table S2 and Fig. S1) or for growth rate (Table S2 and Fig. S2), and all species suffered equally from competition for vegetative traits (no significant species × treatment interaction, Table S2). Introducing a variance structure significantly increased model fits. However, the results were similar when heteroscedasticity was not taken into account (Table S3). In contrast, all factors and all pairwise interactions were significant for flower number (Table 1 and Fig. 1). In particular, the three species were differently sensitive to competition treatment (significant species × treatment interaction). To facilitate the interpretation of these interactions we analysed the competition index for flower number. Species differed in their sensitivity to competition (species effect) as well as in their capacity to impose competition on others (treatment effect); however, species and treatment only had marginally significant interacting effects (Table 2, P = 0.058). The three Capsella species induced similar levels of competition (no significant difference) but the ‘outgroup’ M. chamomilla induced significantly less competition (P-value of contrasts = 0.005, 0.051 and 0.009 with C. bursa-pastoris, C. grandiflora and C. rubella respectively, after FDR correction). Table 2. Analysis of deviance of the competition index (Ic) for flower number d.f. LR P-value Block 2 12.08 0.002 Species 2 11.41 0.003 Competitor 3 17.52 5.51e-4 Block × species 4 8.30 0.081 Block × competitor 6 18.18 0.006 Species × competitor 6 12.21 0.058 Accession (random) 6 36.43 2.27e-06 d.f. LR P-value Block 2 12.08 0.002 Species 2 11.41 0.003 Competitor 3 17.52 5.51e-4 Block × species 4 8.30 0.081 Block × competitor 6 18.18 0.006 Species × competitor 6 12.21 0.058 Accession (random) 6 36.43 2.27e-06 Ic (535 observations) was analysed with a gamma model with gamma link. The ‘competitor’ factor is equivalent to the ‘treatment’ factor in the first analysis without the ‘alone’ level. LR, likelihood ratio. View Large Table 2. Analysis of deviance of the competition index (Ic) for flower number d.f. LR P-value Block 2 12.08 0.002 Species 2 11.41 0.003 Competitor 3 17.52 5.51e-4 Block × species 4 8.30 0.081 Block × competitor 6 18.18 0.006 Species × competitor 6 12.21 0.058 Accession (random) 6 36.43 2.27e-06 d.f. LR P-value Block 2 12.08 0.002 Species 2 11.41 0.003 Competitor 3 17.52 5.51e-4 Block × species 4 8.30 0.081 Block × competitor 6 18.18 0.006 Species × competitor 6 12.21 0.058 Accession (random) 6 36.43 2.27e-06 Ic (535 observations) was analysed with a gamma model with gamma link. The ‘competitor’ factor is equivalent to the ‘treatment’ factor in the first analysis without the ‘alone’ level. LR, likelihood ratio. View Large The diploid selfer C. rubella was the most sensitive to competition with competition index ranging from 0.56 to 0.69 depending on the competitor (least-square estimates based on the model with all interactions, see Figure 2). The tetraploid selfer C. bursa-pastoris, was significantly less sensitive to competition than C. rubella with competition index ranging from 0.70 to 1.06 (P = 0.0013, contrast between C. rubella and C. bursa-pastoris after FDR correction). The diploid outcrosser, C. grandiflora, was intermediate with competition index ranging from 0.59 to 0.87 but was not significantly different from any of the two other species (P = 0.113 and 0.125, compared with and C. bursa-pastoris and C. rubella, respectively, after FRD correction). However, taking into account the marginally significant species × competitor interaction, the difference among the three species depended on whether the competitor is C. rubella or another species. Capsella grandiflora was the most sensitive to competition by C. rubella although the difference was not significant (Fig. 2 and Supplementary Data Tables S4 and S5). For the other competitors, we found the same ranking as for the global analysis but both C. bursa-pastoris and C. grandiflora were significantly less sensitive to competition than C. rubella (P = 0.0008 and 0.0233, respectively, after FDR correction), but not different from each other (P = 0.318, after FDR correction, see Tables S4 and S5). Fig. 2. View largeDownload slide Least-squares mean (± CI) of competition indices (Ic) for all combination of species and competitors. Least-squares were computed on the full model with interactions. Dashed line corresponds to Ic = 1 (no competition effect). White segments correspond to Ic computed as the ratio of the mean flower number in competition and alone and white dots correspond to Ic computed on the product of flowering rate and flower number. In contrast to Fig. 1, each panel corresponds to a competitor species and the focal species are plotted on the x-axis. Rest of the legend as in Fig. 1. Fig. 2. View largeDownload slide Least-squares mean (± CI) of competition indices (Ic) for all combination of species and competitors. Least-squares were computed on the full model with interactions. Dashed line corresponds to Ic = 1 (no competition effect). White segments correspond to Ic computed as the ratio of the mean flower number in competition and alone and white dots correspond to Ic computed on the product of flowering rate and flower number. In contrast to Fig. 1, each panel corresponds to a competitor species and the focal species are plotted on the x-axis. Rest of the legend as in Fig. 1. Finally, we compared these results with competition indices measured as the ratio of mean numbers of flowers for plants alone or in competition estimated in the initial model (Supplementary Data Table S6). Competition indices were slightly lower but very highly correlated (rPearson = 0.989) indicating the robustness of our results (white segments in Fig. 2). Taking flowering rate into account in the computation of competition indices also left conclusions unchanged (white dots in Fig. 2). DISCUSSION Variation in competitive ability among Capsella species Selfing recently evolved from outcrossing in the genus Capsella, with or without change in ploidy level. This was associated with drastic changes in morphological and functional reproductive traits, making Capsella a good model to study selfing syndrome evolution (Sicard et al., 2011, 2016; Slotte et al., 2012; Neuffer and Paetsch, 2013). Here, we characterized further the divergence among three of these species, focusing on competitive ability, which has been found to be negatively associated with selfing in a meta-analysis (Munoz et al., 2016). Competition affected the three species to the same extent for vegetative traits but its effect differed significantly among species for flower number. On average the diploid selfer C. rubella suffered proportionally more from competition than the tetraploid selfer C. bursa-pastoris. The diploid outcrosser C. grandiflora was intermediate but not significantly different from C. bursa-pastoris. Depending on the competitor, C. grandiflora was either significantly less sensitive to competition than C. rubella or not significantly different. The interpretation of these results could be obscured by trade-offs between growth and reproduction, as delay in reproduction usually leads to bigger plants with higher fecundity (e.g. Roff, 2000). The level of competition experienced by individuals could thus depend on their growth rate. However, we also found no difference among species for the three measures of diameter or for growth rate (Supplementary Data Tables S2 and S4). Moreover, this issue is less pronounced in annual species with short life cycles (few months here) (de Jong and Klinkhamer, 2005). No significant difference was observed between the three Capsella species acting as competitors. This could be due to the low power of the experimental design to detect such an effect, but it suggests that the nature of the competitor has a weaker effect than competition itself. Only the ‘outgroup’ M. chamomilla had lower competitive effect, suggesting that the most distant species has different resource requirements, and hence weaker competitive interaction. These results partly match the prediction that competitiveness should decrease with selfing (Munoz et al., 2016). However, additional studies in other species would be necessary to support the idea that competitive ability could also be a component of the selfing syndrome. Importantly, our results point to a stronger effect of polyploidy than mating system, which should be controlled for in future studies. The possible causes of the association between competitiveness and outcrossing remain unclear (see discussion in Munoz et al., 2016). A first possibility is the well-known colonization/competition trade-off. In the short term at least, selfing is often associated with range expansion (van Kleunen and Johnson, 2007; Grossenbacher et al., 2015) and such a trade-off is typically expected to evolve during range expansion (Burton et al., 2010). However, it is important to note that better colonizing ability can be an automatic effect of selfing without physiological costs for competitiveness, simply because it allows colonization of a new habitat with only one individual (no Allee effect). Alternatively, these results can be interpreted in the light of mutation load theory. Theoretical models predict that the ecological context, such as the intensity and nature of competition or environmental heterogeneity, can affect inbreeding depression and mutation load (Agrawal, 2010; Roze, 2012). Experimentally, competition was shown to increase inbreeding depression (Cheptou et al., 2000, 2001; Yun and Agrawal, 2014). Our results suggest that the same could be true for load, including the drift load due to weakly deleterious mutations differentially fixed between species, as observed between C. grandiflora and C. rubella (Brandvain et al., 2013; Slotte et al., 2013). Thus, similar experiments need to be repeated in other species and conditions, and comparisons of the sensitivity to competition among populations of the same species with potentially different load would also be useful to test more directly the load hypothesis. The difference between competitive effects experienced by the two selfers was stronger, more consistent and easier to interpret. The two selfing species present a similar genomic signature of relaxed purifying selection (Douglas et al., 2015), so the observed difference could be partly explained by the masking of deleterious mutations due to genetic redundancy in the tetraploid species. Under disomic inheritance, pure redundancy is not predicted to be stable, except under specific conditions, and one copy is expected to be lost (Fisher, 1935; Nowak et al., 1997). Yet, the time to loss of the redundant state can be long (Fisher, 1935; Kimura and King, 1979) and may not have been reached in C. bursa-pastoris for which the level of gene loss is low (Douglas et al., 2015). Regarding the selfing/outcrossing comparison, the two species may also have evolved contrasting competitive abilities for other reasons. However, they have similar ecology and C. rubella evolved more recently than C. bursa-pastoris, from the ancestral (outcrossing) state. Moreover, C. bursa-pastoris appears to be a better colonizer than C. rubella. Thus, the colonization/competition trade-off hypothesis appears less likely. Potential implications for the evolution of selfing species Whatever the underlying causes, our results may help to understand the evolution of the different Capsella species and suggest implications for the evolution of selfing species in general. In line with the dead-end hypothesis initially formulated by Stebbins (1957), selfing lineages are expected to arise frequently but to go extinct at a higher rate than outcrossing lineages. Such a ‘live fast but die young’ pattern has been well characterized in Solanaceae (Goldberg et al., 2010) and Primulaceae (de Vos et al., 2014). The negative genetic effect of selfing (accumulation of deleterious mutations and reduced adaptive potential) could be one of the main drivers of selfing lineage extinction (e.g. Glémin and Galtier, 2012; Wright et al., 2013). Yet, despite strong evidence for genomic degradation (low genetic diversity, accumulation of deleterious mutations), the selfers C. rubella and especially C. bursa-pastoris appear more ecologically successful than the outcrosser C. grandiflora that exhibits very high genetic diversity, low genetic load and efficient selection (Slotte et al., 2010, 2013; Brandvain et al., 2013; Douglas et al., 2015). Similar patterns were also observed in the genus Collinsia where selfing species showed evidence of genomic degradation (Hazzouri et al., 2013) despite larger species ranges (Randle et al., 2009). This could be explained by the time taken by negative effects of selfing to be effectively expressed, leading to increased extinction risk over time (Ho and Agrawal, 2017). In addition, the ruderal habit of Capsella and other selfing species (Munoz et al., 2016) and their low competitive habitats allow mutation load to build up without demographic consequences if competition remains low (Agrawal, 2010; Agrawal and Whitlock, 2012). The ecology of the three Capsella species partly matches these conditions. Although they can be found in sympatry, C. rubella and C. bursa-pastoris usually occupy smaller and more disturbed patches than C. grandiflora, which can be found in large and dense field populations (our personal observations). To explain the increasing extinction risk over time, a recent meta-analysis also suggested that selfing species could experience diminishing niche breadth despite geographical expansion (Park et al., 2017). This could be explained by the negative genetic effect of selfing mentioned above, but reduced competitive ability could also contribute to preventing their establishment in new, especially more competitive, habitats. One reason would be that the accumulated genetic load is expected to have stronger demographic impact under more competitive conditions (Agrawal, 2010; Agrawal and Whitlock, 2012). In the long run, it could also contribute to explain how selfing lineages could ‘senesce’ in diversification rates (Ho and Agrawal, 2017). If selfing species derived from outcrossing ancestors are rapidly trapped into restricted non-competitive habitats, subsequent newly formed selfing species (from already-selfing ancestors) would also inherit restricted ecological niches without benefiting from higher reproductive assurance compared to ancestors. Although still speculative, we suggest that scenarios that better integrate ecological strategies (including competitive ability) could help to resolve the paradox of selfing species that appear ecologically and demographically successful in the short term, but are an evolutionary dead-end in the longer term (Stebbins, 1957; Igic and Busch, 2013; Wright et al., 2013). Under such a scenario, polyploidy could buffer the negative effect of selfing and delay the extinction risk. This could contribute to explain the ecological success of polyploid selfing species, benefiting from the reproductive assurance of selfing without paying its full genetic cost, at least temporarily. The association between selfing and polyploidy (Barringer, 2007; Robertson et al., 2011) could be due both to the facilitation of the shift to polyploidy by selfing (Rodriguez, 1996; Rausch and Morgan, 2005) and to the reduction of extinction risk in selfers by polyploidy. SUPPLEMENTARY DATA Supplementary data are available online at www.aob.oxfordjournals.org and consist of the following. File S1: Full dataset. Table S1: List of accessions and their locality of origin. Table S2: Analyses of deviance for vegetative traits. Table S3: Analyses of deviance for vegetative traits without variance structure. Table S4: Analyses of deviance for flower number without C. rubella as a competitor. Table S5: Analyses of deviance for flower number with only C. rubella as a competitor. Table S6: Mean effect estimates of the proportion of flowering plants and flower number for all combinations of species and treatments. Figure S1: Boxplots of the three diameters (in mm) for all combinations of species and treatments. Figure S2: Boxplots of growth rates for all combinations of species and treatments. ACKNOWLEDGEMENTS We thank Amandine Cornille and Dmytro Kryvokhyzha for technical advice and help during the experiment, and Marion Orsucci, Jurriaan de Vos and an anonymous reviewer for helpful comments on the manuscript. This study was supported by the Swedish Research Council and the Philip Sörensen Foundation. For this project, S.G. was supported jointly by the French CNRS and the Marie Curie IEF Grant ‘SELFADAPT’ 623486. The authors declare no conflict of interest. LITERATURE CITED Agrawal AF. 2010. Ecological determinants of mutation load and inbreeding depression in subdivided populations. The American Naturalist 176: 111– 122. Google Scholar CrossRef Search ADS Agrawal AF, Whitlock MC. 2012. Mutation load: the fitness of individuals in populations where deleterious alleles are abundant. Annual Review of Ecology and Systematics 43: 115– 135. Google Scholar CrossRef Search ADS Baker HG. 1967. Support for Baker’s law as a rule. Evolution 21: 853– 856. Google Scholar CrossRef Search ADS Barringer BC. 2007. Polyploidy and self-fertilization in flowering plants. American Journal of Botany 94: 1527– 1533. Google Scholar CrossRef Search ADS Bates D, Maechler M, Bolker B, Walker S. 2015. Fitting linear mixed-effects models using lme4. Journal of Statistical Software 67: 1– 48. Google Scholar CrossRef Search ADS Billiard S, Lopez-Villavicencio M, Hood ME, Giraud T. 2012. Sex, outcrossing and mating types: unsolved questions in fungi and beyond. Journal of Evolutionary Biology 25: 1020– 1038. Google Scholar CrossRef Search ADS Brandvain Y, Slotte T, Hazzouri KM, Wright SI, Coop G. 2013. Genomic identification of founding haplotypes reveal the history of the selfing species Capsella rubella. PLoS Genetics 9: e1003754. Burgarella C, Gayral P, Ballenghien Met al. 2015. Molecular evolution of freshwater snails with contrasting mating systems. Molecular Biology & Evolution 32: 2403– 2416. Google Scholar CrossRef Search ADS Burton OJ, Phillips BL, Travis JM. 2010. Trade-offs and the evolution of life-histories during range expansion. Ecology Letter 13: 1210– 1220. Google Scholar CrossRef Search ADS Charlesworth D. 2006. Evolution of plant breeding systems. Current Biology 16:R 726– 735. Google Scholar CrossRef Search ADS Cheptou P-O, Imbert E, Lepart J, Escarré J. 2000. Effects of competition on lifetime estimates of inbreeding depression in the outcrossing plant Crepis sancta (Asteraceae). Journal of Evolutionary Biology 13: 522– 531. Google Scholar CrossRef Search ADS Cheptou PO, Lepart J, Escarre J. 2001. Inbreeding depression under intraspecific competition in a highly outcrossing population of Crepis sancta (Asteraceae): evidence for frequency-dependent variation. American Journal of Botany 88: 1424– 1429. Google Scholar CrossRef Search ADS Clements D, DiTommaso A, Jordan Net al. 2004. Adaptability of plants invading North American cropland. Agriculture, Ecosystems & Environment 104: 379– 398. Google Scholar CrossRef Search ADS Comai L. 2005. The advantages and disadvantages of being polyploid. Nature Reviews Genetics 6: 836– 46. Google Scholar CrossRef Search ADS Cornille A, Salcedo A, Kryvokhyzha Det al. 2016. Genomic signature of successful colonization of Eurasia by the allopolyploid shepherd’s purse (Capsella bursa-pastoris). Molecular Ecology 25: 616– 629. Google Scholar CrossRef Search ADS Coyne JA, Orr H. 2004. Speciation . Sunderland, MA: Sinauer Associates. de Jong T, Klinkhamer P. 2005. Evolutionary ecology of plant reproductive strategies . Cambridge: Cambridge University Press. de Vos JM, Hughes CE, Schneeweiss GM, Moore BR, Conti E. 2014. Heterostyly accelerates diversification via reduced extinction in primroses. Proceedings of Biological Science 281: 20140075. Google Scholar CrossRef Search ADS Douglas GM, Gos G, Steige KAet al. 2015. Hybrid origins and the earliest stages of diploidization in the highly successful recent polyploid Capsella bursa-pastoris. Proceedings of the National Academy of Sciences USA 112: 2806– 2811. Google Scholar CrossRef Search ADS Fisher RA. 1935. The sheltering of lethals. American Naturalist 69: 446– 455. Google Scholar CrossRef Search ADS Fishman L, Kelly AJ, Willis JH. 2002. Minor quantitative trait loci underlie floral traits associated with mating system divergence in Mimulus. Evolution 56: 2138– 2155. Google Scholar CrossRef Search ADS Fournier DA, Skaug HJ, Ancheta Jet al. 2012. AD Model Builder: using automatic differentiation for statistical inference of highly parameterized complex nonlinear models. Optimization Methods and Software 27: 233– 249. Google Scholar CrossRef Search ADS Fox J, Weisberg S. 2011. An R companion to applied regression . Thousand Oaks, CA: Sage. Foxe JP, Slotte T, Stahl EA, Neuffer B, Hurka H, Wright SI. 2009. Recent speciation associated with the evolution of selfing in Capsella. Proceedings of the National Academy of Sciences of the United States of America 106: 5241– 5245. Google Scholar CrossRef Search ADS Glémin S, Galtier N. 2012. Genome evolution in outcrossing versus selfing versus asexual species. Methods in Molecular Biology 855: 311– 335. Google Scholar CrossRef Search ADS Goldberg EE, Igic B. 2012. Tempo and mode in plant breeding system evolution. Evolution 66: 3701– 3709. Google Scholar CrossRef Search ADS Goldberg EE, Kohn JR, Lande R, Robertson KA, Smith SA, Igic B. 2010. Species selection maintains self-incompatibility. Science 330: 493– 495. Google Scholar CrossRef Search ADS Grossenbacher D, Briscoe Runquist R, Goldberg EE, Brandvain Y. 2015. Geographic range size is predicted by plant mating system. Ecology Letters 18: 706– 713. Google Scholar CrossRef Search ADS Guo YL, Bechsgaard JS, Slotte Tet al. 2009. Recent speciation of Capsella rubella from Capsella grandiflora, associated with loss of self-incompatibility and an extreme bottleneck. Proceedings of the National Academy of Sciences of the United States of America 106: 5246– 5251. Google Scholar CrossRef Search ADS Hazzouri KM, Escobar JS, Ness RWet al. 2013. Comparative population genomics in Collinsia sister species reveals evidence for reduced effective population size, relaxed selection, and evolution of biased gene conversion with an ongoing mating system shift. Evolution 67: 1263– 1278. Ho EKH, Agrawal AF. 2017. Aging asexual lineages and the evolutionary maintenance of sex. Evolution 71: 1865– 1875. Google Scholar CrossRef Search ADS Hurka H, Friesen N, German DA, Franzke A, Neuffer B. 2012. ‘ Missing link’ species Capsella orientalis and Capsella thracica elucidate evolution of model plant genus Capsella (Brassicaceae). Molecular Ecology 21: 1223– 1238. Google Scholar CrossRef Search ADS Igic B, Busch JW. 2013. Is self-fertilization an evolutionary dead end? New Phytologist 198: 386– 397. Jarne P, Auld JR. 2006. Animals mix it up too: the distribution of self-fertilization among hermaphroditic animals. Evolution 60: 1816– 1824. Google Scholar CrossRef Search ADS Kalisz S, Randle A, Chaiffetz D, Faigeles M, Butera A, Beight C. 2012. Dichogamy correlates with outcrossing rate and defines the selfing syndrome in the mixed-mating genus Collinsia. Annals of Botany 109: 571– 582. Google Scholar CrossRef Search ADS Kimura M, King JL. 1979. Fixation of a deleterious allele at one of two ‘duplicate’ loci by mutation pressure and random drift. Proceedings of the National Academy of Sciences of the United States of America 76: 2858– 2861. Google Scholar CrossRef Search ADS Kuznetsova A, Brockhoff PB, Christensen RHB. 2017. lmerTest: tests in linear mixed effects models . Journal of Statistical Software82: 1–26. Munoz F, Violle C, Cheptou P-O. 2016. CSR ecological strategies and plant mating systems: outcrossing increases with competitiveness but stress-tolerance is related to mixed mating. Oikos 125: 1296– 1303. Google Scholar CrossRef Search ADS Neuffer B, Paetsch M. 2013. Flower morphology and pollen germination in the genus Capsella (Brassicaceae). Flora 208: 626– 640. Google Scholar CrossRef Search ADS Nowak MA, Boerlijst MC, Cooke J, Smith JM. 1997. Evolution of genetic redundancy. Nature 388: 167– 171. Google Scholar CrossRef Search ADS Otto SP, Whitton J. 2000. Polyploid incidence and evolution. Annual Review of Genetics 34: 401– 437. Google Scholar CrossRef Search ADS Park DS, Ellison AM, Davis CC. 2017. Selfing species exhibit diminished niche breadth over time. bioArxiv 157974; doi:https://doi.org/10.1101/157974. Pinheiro J, Bates D, DebRoy S, Sarkar D, R Core Team. 2017. nlme: Linear and Nonlinear Mixed Effects Models . R package version 3.1-131, https://CRAN.R-project.org/package=nlme. Randle AM, Slyder JB, Kalisz S. 2009. Can differences in autonomous selfing ability explain differences in range size among sister-taxa pairs of Collinsia (Plantaginaceae)? An extension of Baker’s Law. New Phytologist 183: 618– 629. Google Scholar CrossRef Search ADS Rausch JH, Morgan MT. 2005. The effect of self-fertilization, inbreeding depression, and population size on autopolyploid establishment. Evolution 59: 1867– 1875. Rebernig CA, Lafon-Placette C, Hatorangan MR, Slotte T, Kohler C. 2015. Non-reciprocal interspecies hybridization barriers in the Capsella genus are established in the endosperm. PLoS Genet 11: e1005295. Google Scholar CrossRef Search ADS Robertson K, Goldberg EE, Igic B. 2011. Comparative evidence for the correlated evolution of polyploidy and self-compatibility in Solanaceae. Evolution 65: 139– 155. Google Scholar CrossRef Search ADS Rodriguez DJ. 1996. A model for the establishment of polyploidy in plants. The American Naturalist 147: 33– 46. Google Scholar CrossRef Search ADS Roff DA. 2000. Trade-offs between growth and reproduction: an analysis of the quantitative genetic evidence. Journal of Evolutionary Biology 13: 434– 445. Google Scholar CrossRef Search ADS Roze D. 2012. Spatial heterogeneity in the strength of selection against deleterious alleles and the mutation load. Heredity 109: 137– 45. Google Scholar CrossRef Search ADS Sicard A, Kappel C, Josephs EBet al. 2015. Divergent sorting of a balanced ancestral polymorphism underlies the establishment of gene-flow barriers in Capsella. Nature Communications 6: 7960. Google Scholar CrossRef Search ADS Sicard A, Kappel C, Lee YWet al. 2016. Standing genetic variation in a tissue-specific enhancer underlies selfing-syndrome evolution in Capsella. Proceedings of the National Academy of Sciences USA 113: 13911– 13916. Google Scholar CrossRef Search ADS Sicard A, Lenhard M. 2011. The selfing syndrome: a model for studying the genetic and evolutionary basis of morphological adaptation in plants. Annals of Botany 107: 1433– 1443. Google Scholar CrossRef Search ADS Sicard A, Stacey N, Hermann Ket al. 2011. Genetics, evolution, and adaptive significance of the selfing syndrome in the genus Capsella. Plant Cell 23: 3156– 3171. Google Scholar CrossRef Search ADS Skaug HJ, Fournier DA, Nielsen A, Magnusson A, Bolker B. 2013. Generalized Linear Mixed Models using AD Model Builder. R package version 0.8.3.3 . Slotte T, Foxe JP, Hazzouri KM, Wright SI. 2010. Genome-wide evidence for efficient positive and purifying selection in Capsella grandiflora, a plant species with a large effective population size. Molecular Biology & Evolution 27: 1813– 1821. Google Scholar CrossRef Search ADS Slotte T, Hazzouri KM, Agren JAet al. 2013. The Capsella rubella genome and the genomic consequences of rapid mating system evolution. Nature Genetics 45: 831– 835. Google Scholar CrossRef Search ADS Slotte T, Hazzouri KM, Stern D, Andolfatto P, Wright SI. 2012. Genetic architecture and adaptive significance of the selfing syndrome in Capsella. Evolution 66: 1360– 1374. Google Scholar CrossRef Search ADS Slotte T, Huang H, Lascoux M, Ceplitis A. 2008. Polyploid speciation did not confer instant reproductive isolation in Capsella (Brassicaceae). Molecular Biology & Evolution 25: 1472– 1481. Google Scholar CrossRef Search ADS St Onge KR, Kallman T, Slotte T, Lascoux M, Palme AE. 2011. Contrasting demographic history and population structure in Capsella rubella and Capsella grandiflora, two closely related species with different mating systems. Molecular Ecology 20: 3306– 3320. Google Scholar CrossRef Search ADS Stebbins GL. 1957. Self fertilization and population variability in higher plants. The American Naturalist 91: 337– 354. Google Scholar CrossRef Search ADS te Beest M, Le Roux JJ, Richardson DMet al. 2012. The more the better? The role of polyploidy in facilitating plant invasions. Annals of Botany 109: 19– 45. Google Scholar CrossRef Search ADS van Kleunen M, Johnson SD. 2007. Effects of self-compatibility on the distribution range of invasive European plants in North America. Conservation Biology 21: 1537– 1544. van Kleunen M, Manning JC, Pasqualetto V, Johnson SD. 2008. Phylogenetically independent associations between autonomous self-fertilization and plant invasiveness. The American Naturalist 171: 195– 201. Google Scholar CrossRef Search ADS Williamson RJ, Josephs EB, Platts AE, Hazzouri KM, Haudry A, Blanchette M, Wright SI. 2014. Evidence for widespread positive and negative selection in coding and conserved noncoding regions of Capsella grandiflora. PLoS Genetics10: e1004622. Wright S, Ness RW, Foxe JP, Barrett SC. 2008. Genomic consequences of outcrossing and selfing in plants. International Journal Of Plant Science 169: 105– 118. Google Scholar CrossRef Search ADS Wright SI, Kalisz S, Slotte T. 2013. Evolutionary consequences of self-fertilization in plants. Proceeding of the Royal Society of London B 280: 20130133. Google Scholar CrossRef Search ADS Yun L, Agrawal AF. 2014. Variation in the strength of inbreeding depression across environments: effects of stress and density dependence. Evolution 68: 3599– 3606. Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of the Annals of Botany Company. All rights reserved. For permissions, please e-mail: firstname.lastname@example.org. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)
Annals of Botany – Oxford University Press
Published: Feb 17, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera