Incipient speciation in Oncocyclus irises: Eco-geographic isolation and genetic divergence with no reproductive isolation?
Introduction
The establishment of reproductive isolation is one of the most important steps in speciation because it drives the initial divergence and maintains the divergent forms as distinct species. Therefore, understanding the conditions that promote the evolution of reproductive isolation is a major goal in evolutionary biology (Dobzhansky, 1937; Coyne and Orr, 2004; Lowry et al., 2008; Schluter, 2009; Baack et al., 2015; Butlin and Smadja, 2018). In ecological speciation, reproductive isolation between populations is believed to evolve ultimately as a consequence of divergent selection and resulting environment-specific adaptations, while making no assumptions about the initial spatial structure of populations (Schluter, 2001; Rundle and Nosil, 2005; Nosil, 2012). From this, it follows that when individuals occupy distinct environments and are distinguished by a suite of traits, this might be attributed to ecological speciation only if some form of reproductive isolation has evolved as a result of local selection against migrants. The degree and specific form of reproductive isolation (e.g. pre- or postzygotic) will depend on the strength of selection and intensity of gene flow (Fig. 1A and B). However, if gene flow is absent or extremely low due to spatially limited seed and pollen dispersal, the process of divergence may not necessarily lead to reproductive isolation as the latter is unnecessary for maintaining phenotypically-distinct allopatric populations (Fig. 1D and E). This theoretical possibility, although unlikely for long-term speciation and where environments impose strong natural selection, might be rather common in incipient speciation of isolated populations subjected to subtle environmental differences and relatively weak selection. Some genetic mechanisms of local adaptation agree with this possibility, for example conditional neutrality, when at locus 1, the native allele confers high fitness in a native environment but has no effect on fitness in another environment, whereas at locus 2, the situation is exactly opposite (Savolainen et al., 2013). As a result, due to lack of gene flow, the ‘high-fitness’ alleles do not spread from one environment to the other and there is no trade-off in performance of reciprocally introduced individuals.
To test our hypothesis that eco-geographical divergence may not necessarily lead to pre- or post-zygotic reproductive isolation, we studied four species of Iris section Oncocyclus - I. atrofusca Baker, I. atropurpurea Dinsmore, I. petrana Dinsmore and I. mariae Barbey - that ideally suit investigating a role of reproductive isolation in a process of eco-geographical divergence. These species have non-overlapping distributions, grow in different environmental conditions, and possess different suites of diagnostic traits, which have been assumed to be the result of local selection (Avishai and Zohary, 1980). Although natural hybridization between the four species has never been observed, implying no or very limited gene flow between the species, no sterility barrier exists among them according to Avishai and Zohary (1980). Based on floral morphology, Avishai and Zohary (1980) subdivided section Oncocyclus into seven aggregates and considered each aggregate to represent a single variable species for which one or more subspecies might be recognized. The four species listed above were placed by Avishai and Zohary (1980) in the Haynei aggregate and therefore may be considered as subspecies rather than species. Nonetheless, they are readily distinguished (despite some variation among individuals and populations) by their distinct floral phenotypes (different colour of petals and signal patch) and vegetative traits (Dorman et al., 2009), their different geographic distributions in Israel and adjacent territories, and by occupying habitats that differ in soil type and climatic conditions (Fig. 2).
As part of a study of Oncocyclus phylogeny and speciation, we previously reported the results of a series of experiments and genetic analyses using two desert species from the Haynei aggregate (I. atrofusca and I. mariae) that occupy distinct, in terms of soil type and precipitation, environments in Israel (Volis et al., 2019). These results showed, that in the case of these two species, genetic and phenotypic divergence has occurred in the absence of any strong genetic incompatibility or ecological isolating barrier. The lack of naturally occurring hybridization between I. atrofusca and I. mariae appears almost entirely due to geographical isolation unlinked to local adaptation.
In this study we expand the number of species in our experiments and analyses investigating the role of reproductive isolation in ecological, genetic and phenotypic divergence within Oncocyclus section from two species to four, of which three are exclusively or predominantly desert species (I. atrofusca, I. mariae, I. petrana) and one is a Mediterranean species (I. atropurpurea). Our study includes estimation of molecular genetic (AFLP) and quantitative trait divergence as well as degree of reproductive isolation among the four species. The quantitative traits studied comprised morphological (flower and leaf shape) and two life history traits, germination and flowering timing. In order to help understand the role of habitat specific selection in the current non-overlapping distributions of these four closely related species we also used ecological niche modeling (ENM), the importance of which when used in combination with other evolutionary approaches in speciation studies is now fully recognized (Kozak et al., 2008; Sobel et al., 2010; Schemske, 2010; Sobel, 2014; Alvarado-Serrano and Knowles, 2014). In cases of taxa with non-overlapping distributions, ENM can indicate how predicted differences in species’ niches might translate into ecogeographic isolation (Sobel et al., 2010; Schemske, 2010).
If species are genetically and phenotypically distinct, and adapted to different environments, interspecific hybrids exhibiting maladaptive intermediate phenotypes are expected to have reduced fitness in both native and alien environments due to several possible genetic mechanisms (Schluter, 2001; Rundle and Whitlock, 2001; Hufford and Mazer, 2003). Our hypothesis, however, suggests that given a lack of inter-specific gene flow and despite genetic and phenotypic divergence consistent with local selection (Fig. 1E) and in the absence of genetic drift (Fig. 1F), inter-specific hybrids and their parents may not differ in performance. This is because adaptation under subtle differences in local selection at the initial stage of divergence may not require trade-offs in performance across environments. Thus, we expected that (i) the probability of detecting reproductive isolation among the four species will correspond to the degree of the species' environmental similarity; and (ii) for some species pairs with the most similar environments it could be zero.
To understand whether the current distribution of the four Oncocyclus species is a result of habitat specific selection, and degree of reproductive isolation among them, we 1) analyzed molecular (AFLP) and morphological inter- and intraspecific variation; 2) performed habitat suitability analysis for each species via ecological niche modelling; 3) compared the fitness of progeny produced from inter-specific and intra-specific crosses.
Section snippets
Study species
The four species studied, I. atrofusca, I. atropurpurea, I. petrana and I. mariae, are rhizomatous geophytes with deciduous, herbaceous aerial shoots. They are endemic or sub-endemic to Israel, and protected by law as "vulnerable" species. Individual plants of each species usually have several flowering stems, each bearing a solitary large flower. Flowers are devoid of nectar but provide an overnight shelter for male solitary bees (Sapir et al., 2005). All four species are diploid (2n = 20) and
Molecular genetic divergence
Most of the AFLP variation, as revealed by AMOVA, was distributed within populations (76%), with the other two components (among species and among populations within species) comprising 14% and 10%, respectively.
Though the optimal number of genetic groups (K) in the analyzed dataset was determined to be two by the Evanno et al. (2005) method (Suppl. Figure 1), it is of interest to examine barplots for K ranging from 2 to 4. In each case, there was a clear separation of I. mariae from the other
Discussion
Although reproductive isolation is traditionally considered a crucial element of speciation, in some scenarios of conflicting effects of selection and drift it may not evolve (Nosil, 2008; Schluter and Conte, 2009; Räsänen et al., 2012). In cases of geographic isolation or spatially limited dispersal, when gene flow is absent or negligible, phenotypically distinct and non-overlapping clusters can arise due to non-adaptive radiation or drift (Rundell and Price, 2009; Sobel et al., 2010).
Conclusions
In this study, we tested a hypothesis that genetically divergent incipient species having non-overlapping geographical distributions and occupying ecologically distinct habitats may not necessarily evolve reproductive isolation from each other. Our results tend to support this hypothesis with respect to four species of Iris in Israel, but with certain qualifications. The coastal Mediterranean species, I. atropurpurea, was locally adapted to its habitat and showed extrinsic postzygotic isolation
CRediT authorship contribution statement
Sergei Volis: Conceptualization, Data curation, Investigation, Methodology, Writing - original draft. Yong-Hong Zhang: Data curation, Investigation, Visualization. Michael Dorman: Data curation, Investigation. Richard J. Abbott: Writing - original draft.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This project was supported by a grant from the Israel Ministry of Sciences, and a grant from Israel Nature and Parks Authority. We would like to thank Israel National Park and Reserve Authority for permission to sample and use the plants, and two anonymous reviewers for helpful comments on an early version of the manuscript.
References (49)
- et al.
Plant ecotypes: genetic differentiation in the age of ecological restoration
Trends in Ecology & Evolution
(2003) - et al.
Integrating GIS-based environmental data into evolutionary biology
Trends in Ecvology and Evolution
(2008) - et al.
Maximum entropy modeling of species geographic distributions
Ecological Modelling
(2006) - et al.
Adaptive radiation, nonadaptive radiation, ecological speciation and nonecological speciation
Trends in Ecology & Evolution
(2009) Ecology and the origin of species
Trends in Ecology & Evolution
(2001)- et al.
Ecological niche models in phylogeographic studies: applications, advances and precautions
Molecular Ecology Resources
(2014) - et al.
Genetic affinities among the Oncocyclus irises
Botanical Gazette
(1980) - et al.
The origins of reproductive isolation in plants
New Phytologist
(2015) - et al.
Coupling, reinforcement, and speciation
American Naturalist
(2018) - et al.
The Irises still exist: the conservation status of species Iris section Oncocyclus in Israel, a century after their description
Annali di Botanica
(2000)