Detection of reproductive interference between closely related Salvia species with small-scale separated distributions by multifaceted pollination and molecular analyses

Abstract

Reproductive interference, an interspecific interaction in reproductive process that exerts an adverse effect, has gained attention as a contributing factor in promoting exclusive distributions between closely related species. However, detailed studies on the possibility of reproductive interference between native plants are still lacking, presumably because strong reproductive interference can rapidly realize exclusive distributions, leaving the two species apparently independent. Salvia japonica and S. lutescens are found in separate localities at a small scale, although their distributions overlap at a large scale. We investigated the possibility of reproductive interference between them through field surveys, hand-pollination experiments, evaluation of hybrid fertility, cpDNA and nrDNA genotyping, and genome-wide DNA analysis. The field survey results did not reveal apparent negative interaction in competition for pollinator services. Mixed pollination with conspecific pollen and counterpart pollen reduced seed set in S. japonica, and hybrid progeny produced by mixed pollination were less than 20% as fertile compared to the pure species. The DNA genotyping results suggested the possibility of hybridization where their distributions overlap, and the genome-wide DNA analysis results showed clear genetic differentiation between the two species as well as the existence of hybrids. These results suggest that bi-directional reproductive interference between S. japonica and S. lutescens may have led to their present separated distributions at a small scale.

Keywords: Salvia japonica; Salvia lutescens; Distributional relationship; Hybridization; Reproductive interference.

Fig.1

Inflorescence of S. japonica (left), S. lutescens (middle) visited by a Halictidae bee, and flowers of a putative hybrid between the two species (right)

Fig. 2

Maps (left) showing the approximate distribution of S. japonica (blue oval) and S. lutescens (orange oval) in Japan and the study site locations, and maps (right) showing the detailed distribution of each species at HS and OY. SA, SO2, Mt. Kasagata, and Miyazaki (circles) are study sites with only S. japonica. SO1 (square) is a study site with only S. lutescens. HS, OY, Sudogawa, Kintoki, Mt. Mikuni, and Mt. Higane (triangles) are study sites with both species, according to our preliminary survey on herbarium specimens

Fig. 3

Seed set following natural pollination and conspecific hand-pollination of a S. japonica and b S. lutescens at HS, OY, and SA (S. japonica only) or SO1 (S. lutescens only). N and C indicate natural pollination and conspecific hand-pollination, respectively. Conspecific hand-pollination of the SA and SO1 populations was conducted at the Nagoya University Museum Botanical Garden. A significant effect (P < 0.05) of conspecific hand-pollination on seed set was determined by GLMM analyses followed by a Wald test. Error bars show the 95% confidence interval. n.s. = not significant

Fig. 4

Seed set following conspecific (gray) and mixed (black) hand-pollination of a S. japonica and b S. lutescens in 2019 and 2021. A significant effect (P < 0.05) of the pollination treatment on seed set was determined by GLMM analyses followed by a Wald test. Error bars indicate the 95% confidence interval. n.s. = not significant

Fig. 5

Proportion of ellipsoid-type (i.e., viable) pollen grains in pure individuals of each species and in hybrid progeny. Significant differences (P < 0.05) between the hybrids and the pure species were determined by GLMM analyses followed by a Wald test. Error bars indicate the 95% confidence interval

Fig. 6

Seed set in hybrid progeny following pollination with S. japonica pollen or S. lutescens pollen and in pure individuals of each species following pollination with conspecific pollen. Significant differences in seed set (P < 0.05) between the hybrids and the pure individuals when pollinated with S. japonica pollen or S. lutescens pollen were determined by GLMM analyses followed by a Wald test. Error bars indicate the 95% confidence interval

Fig. 7

Distributions of chloroplast DNA (cpDNA) haplotypes in a S. japonica and b S. lutescens and putative hybrids. The pie chart for each site shows the proportions and numbers of individuals with the type A (blue) and type B (orange) haplotypes. See Table 3 for the nucleotide substitutions between the types

Fig. 8

Distributions of internal transcribed spacer (ITS) (nrDNA) genotypes in a S. japonica and b S. lutescens and putative hybrids. The pie chart for each site shows the proportions and numbers of individuals with type C (blue), type D (orange), type C/D (green), and type C/D' (purple) genotypes. See Table 3 for the nucleotide substitutions in the types

Fig. 9

Results of the STRUCTURE analysis of the MIG-seq results for S. japonica, S. lutescens, and putative hybrids between the two species (arrows) at HS (top) and OY (bottom). The arrangement of the samples roughly corresponds to their distributional relationships (samples near the center are from a locality close to a locality of the counterpart species), except for the two putative hybrid samples at HS, which were from a locality between two S. lutescens localities (see Fig. 2). The cpDNA haplotype and the nrDNA genotype of each sample are shown in the small boxes at the bottom (see Figs. 7 and 8 for the types indicated by each color; blank boxes indicate no data)