Limits...
Both morph ‐ and species ‐ dependent asymmetries affect reproductive barriers between heterostylous species

View Article: PubMed Central - PubMed

ABSTRACT

The interaction between floral traits and reproductive isolation is crucial to explaining the extraordinary diversity of angiosperms. Heterostyly, a complex floral polymorphism that optimizes outcrossing, evolved repeatedly and has been shown to accelerate diversification in primroses, yet its potential influence on isolating mechanisms remains unexplored. Furthermore, the relative contribution of pre‐ versus postmating barriers to reproductive isolation is still debated. No experimental study has yet evaluated the possible effects of heterostyly on pre‐ and postmating reproductive mechanisms. We quantify multiple reproductive barriers between the heterostylous Primula elatior (oxlip) and P. vulgaris (primrose), which readily hybridize when co‐occurring, and test whether traits of heterostyly contribute to reproductive barriers in unique ways. We find that premating isolation is key for both species, while postmating isolation is considerable only for P. vulgaris; ecogeographic isolation is crucial for both species, while phenological, seed developmental, and hybrid sterility barriers are also important in P. vulgaris, implicating sympatrically higher gene flow into P. elatior. We document for the first time that, in addition to the aforementioned species‐dependent asymmetries, morph‐dependent asymmetries affect reproductive barriers between heterostylous species. Indeed, the interspecific decrease of reciprocity between high sexual organs of complementary floral morphs limits interspecific pollen transfer from anthers of short‐styled flowers to stigmas of long‐styled flowers, while higher reciprocity between low sexual organs favors introgression over isolation from anthers of long‐styled flowers to stigmas of short‐styled flowers. Finally, intramorph incompatibility persists across species boundaries, but is weakened in long‐styled flowers of P. elatior, opening a possible backdoor to gene flow through intramorph pollen transfer between species. Therefore, patterns of gene flow across species boundaries are likely affected by floral morph composition of adjacent populations. To summarize, our study highlights the general importance of premating isolation and newly illustrates that both morph‐ and species‐dependent asymmetries shape boundaries between heterostylous species.

No MeSH data available.


Pollen transfer and sexual organ distance: Mean values and standard errors (estimated from generalized linear mixed‐effects models) of (A) distances (absolute values) from anthers of pollen donors to stigmas of pollen recipients and (B) corresponding number of pollen grains transferred by Anthophora plumipes bees for both intra‐ and interspecific comparisons with Primula elatior (left panels) and Primula vulgaris (right panels) as pollen donors. Significance levels: P ≤ 0.001 (***) or not significantly different P > 0.05 (ns). Sequential Bonferroni correction was used to account for multiple tests. Mean values of number of pollen grains deposited on intra‐ and interspecific stigmas were used to calculate mechanical barriers (RImech; see Table 3).
© Copyright Policy - creativeCommonsBy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC5016645&req=5

ece32293-fig-0004: Pollen transfer and sexual organ distance: Mean values and standard errors (estimated from generalized linear mixed‐effects models) of (A) distances (absolute values) from anthers of pollen donors to stigmas of pollen recipients and (B) corresponding number of pollen grains transferred by Anthophora plumipes bees for both intra‐ and interspecific comparisons with Primula elatior (left panels) and Primula vulgaris (right panels) as pollen donors. Significance levels: P ≤ 0.001 (***) or not significantly different P > 0.05 (ns). Sequential Bonferroni correction was used to account for multiple tests. Mean values of number of pollen grains deposited on intra‐ and interspecific stigmas were used to calculate mechanical barriers (RImech; see Table 3).

Mentions: The intra‐ and interspecific reciprocity values of experimental plants were similar to those of plants from natural populations (Table S1). Anther–stigma distances between pollen donor and recipient flowers were larger between than within species, as expected; the differences were significant for P. vulgaris, but not significant for P. elatior (Fig. 4A; Table S1; for GLMM results see Table S2a). We counted a total of 133 612 pollen grains exported from anthers to reciprocal stigmas across all pollen transfer experiments (P. elatior: 64 526; P. vulgaris: 69 086). High anthers exported significantly fewer pollen grains to inter‐ than intraspecific reciprocal stigmas, as expected, while low anthers exported significantly more pollen grains to inter‐ than intraspecific reciprocal stigmas (Fig. 4B; for GLMM results, see Table S2b). In both species, RImech was thus positive for pollen recipients with high stigmas (L‐morph), but negative for recipients with low stigmas (S‐morph; Table 3). Mechanical isolation was therefore asymmetric between morphs, as predicted (see Introduction), but not between species (Table 4).


Both morph ‐ and species ‐ dependent asymmetries affect reproductive barriers between heterostylous species
Pollen transfer and sexual organ distance: Mean values and standard errors (estimated from generalized linear mixed‐effects models) of (A) distances (absolute values) from anthers of pollen donors to stigmas of pollen recipients and (B) corresponding number of pollen grains transferred by Anthophora plumipes bees for both intra‐ and interspecific comparisons with Primula elatior (left panels) and Primula vulgaris (right panels) as pollen donors. Significance levels: P ≤ 0.001 (***) or not significantly different P > 0.05 (ns). Sequential Bonferroni correction was used to account for multiple tests. Mean values of number of pollen grains deposited on intra‐ and interspecific stigmas were used to calculate mechanical barriers (RImech; see Table 3).
© Copyright Policy - creativeCommonsBy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC5016645&req=5

ece32293-fig-0004: Pollen transfer and sexual organ distance: Mean values and standard errors (estimated from generalized linear mixed‐effects models) of (A) distances (absolute values) from anthers of pollen donors to stigmas of pollen recipients and (B) corresponding number of pollen grains transferred by Anthophora plumipes bees for both intra‐ and interspecific comparisons with Primula elatior (left panels) and Primula vulgaris (right panels) as pollen donors. Significance levels: P ≤ 0.001 (***) or not significantly different P > 0.05 (ns). Sequential Bonferroni correction was used to account for multiple tests. Mean values of number of pollen grains deposited on intra‐ and interspecific stigmas were used to calculate mechanical barriers (RImech; see Table 3).
Mentions: The intra‐ and interspecific reciprocity values of experimental plants were similar to those of plants from natural populations (Table S1). Anther–stigma distances between pollen donor and recipient flowers were larger between than within species, as expected; the differences were significant for P. vulgaris, but not significant for P. elatior (Fig. 4A; Table S1; for GLMM results see Table S2a). We counted a total of 133 612 pollen grains exported from anthers to reciprocal stigmas across all pollen transfer experiments (P. elatior: 64 526; P. vulgaris: 69 086). High anthers exported significantly fewer pollen grains to inter‐ than intraspecific reciprocal stigmas, as expected, while low anthers exported significantly more pollen grains to inter‐ than intraspecific reciprocal stigmas (Fig. 4B; for GLMM results, see Table S2b). In both species, RImech was thus positive for pollen recipients with high stigmas (L‐morph), but negative for recipients with low stigmas (S‐morph; Table 3). Mechanical isolation was therefore asymmetric between morphs, as predicted (see Introduction), but not between species (Table 4).

View Article: PubMed Central - PubMed

ABSTRACT

The interaction between floral traits and reproductive isolation is crucial to explaining the extraordinary diversity of angiosperms. Heterostyly, a complex floral polymorphism that optimizes outcrossing, evolved repeatedly and has been shown to accelerate diversification in primroses, yet its potential influence on isolating mechanisms remains unexplored. Furthermore, the relative contribution of pre‐ versus postmating barriers to reproductive isolation is still debated. No experimental study has yet evaluated the possible effects of heterostyly on pre‐ and postmating reproductive mechanisms. We quantify multiple reproductive barriers between the heterostylous Primula elatior (oxlip) and P. vulgaris (primrose), which readily hybridize when co‐occurring, and test whether traits of heterostyly contribute to reproductive barriers in unique ways. We find that premating isolation is key for both species, while postmating isolation is considerable only for P. vulgaris; ecogeographic isolation is crucial for both species, while phenological, seed developmental, and hybrid sterility barriers are also important in P. vulgaris, implicating sympatrically higher gene flow into P. elatior. We document for the first time that, in addition to the aforementioned species‐dependent asymmetries, morph‐dependent asymmetries affect reproductive barriers between heterostylous species. Indeed, the interspecific decrease of reciprocity between high sexual organs of complementary floral morphs limits interspecific pollen transfer from anthers of short‐styled flowers to stigmas of long‐styled flowers, while higher reciprocity between low sexual organs favors introgression over isolation from anthers of long‐styled flowers to stigmas of short‐styled flowers. Finally, intramorph incompatibility persists across species boundaries, but is weakened in long‐styled flowers of P. elatior, opening a possible backdoor to gene flow through intramorph pollen transfer between species. Therefore, patterns of gene flow across species boundaries are likely affected by floral morph composition of adjacent populations. To summarize, our study highlights the general importance of premating isolation and newly illustrates that both morph‐ and species‐dependent asymmetries shape boundaries between heterostylous species.

No MeSH data available.