Mating behavior and reproductive morphology predict macroevolution of sex allocation in hermaphroditic flatworms

Abstract

Background: Sex allocation is the distribution of resources to male or female reproduction. In hermaphrodites, this concerns an individual's resource allocation to, for example, the production of male or female gametes. Macroevolutionary studies across hermaphroditic plants have revealed that the self-pollination rate and the pollination mode are strong predictors of sex allocation. Consequently, we expect similar factors such as the selfing rate and aspects of the reproductive biology, like the mating behaviour and the intensity of postcopulatory sexual selection, to predict sex allocation in hermaphroditic animals. However, comparative work on hermaphroditic animals is limited. Here, we study sex allocation in 120 species of the hermaphroditic free-living flatworm genus Macrostomum. We ask how hypodermic insemination, a convergently evolved mating behaviour where sperm are traumatically injected through the partner's epidermis, affects the evolution of sex allocation. We also test the commonly-made assumption that investment into male and female reproduction should trade-off. Finally, we ask if morphological indicators of the intensity of postcopulatory sexual selection (female genital complexity, male copulatory organ length, and sperm length) can predict sex allocation.

Results: We find that the repeated evolution of hypodermic insemination predicts a more female-biased sex allocation (i.e., a relative shift towards female allocation). Moreover, transcriptome-based estimates of heterozygosity reveal reduced heterozygosity in hypodermically mating species, indicating that this mating behavior is linked to increased selfing or biparental inbreeding. Therefore, hypodermic insemination could represent a selfing syndrome. Furthermore, across the genus, allocation to male and female gametes is negatively related, and larger species have a more female-biased sex allocation. Finally, increased female genital complexity, longer sperm, and a longer male copulatory organ predict a more male-biased sex allocation.

Conclusions: Selfing syndromes have repeatedly originated in plants. Remarkably, this macroevolutionary pattern is replicated in Macrostomum flatworms and linked to repeated shifts in reproductive behavior. We also find a trade-off between male and female reproduction, a fundamental assumption of most theories of sex allocation. Beyond that, no theory predicts a more female-biased allocation in larger species, suggesting avenues for future work. Finally, morphological indicators of more intense postcopulatory sexual selection appear to predict more intense sperm competition.

Keywords: Comparative morphology; Convergent evolution; Evolution; Local sperm competition; Selfing syndrome; Sexual conflict; Sexual selection; Sperm competition; Traumatic insemination.

Fig. 1

Stylized representation of the mating syndromes, details on the general Macrostomum morphology, and the specific in vivo measurements used in the analyses. Most species can be assigned to either the reciprocal mating syndrome (A), with reciprocal mating, a (usually blunt) non-invasive stylet, sperm with lateral bristles and a thickened antrum epithelium, or they show the hypodermic mating syndrome (B), with hypodermic insemination, a needle-like stylet, simple sperm and a thin antrum epithelium. Note that assignments follow the inferred mating syndrome [43], based on morphological data and observations of the location of received sperm. We use colour throughout to represent species assignment to the inferred mating syndromes: hypodermic (yellow, N=40), intermediate (light green, N=2), reciprocal (green, N=69), and unclear (gray, N=9). C In vivo image of M. lignano with visible internal organs and an indicated outline of how the areas of the paired testes (blue) and ovaries (red) were measured (only one testis and ovary are outlined). These area measurements were used to estimate gonad size and gonadal sex allocation (GSA, as testis area/(testis area + ovary area)). Below are drawings of additional reproductive morphology traits, namely a stylet (D), a sperm with sperm bristles (E), and an antrum with received sperm (F), with representations of the measurements taken or the structures scored, which may represent non-gonadal components of male and female allocation and serve as indicators of the intensity of postcopulatory sexual selection (“morphological indicators”). Linear quantitative measurements were taken as indicated in D and E. Such quantitative measures were not possible for the antrum (F), and we thus scored the indicated structures on an ordinal scale and summed these for an overall measure of antrum complexity (high values indicate a more complex antrum). For details see the “Methods” section.

Fig. 2

Evolution of gonadal sex allocation (GSA) across 120 species of Macrostomum. Colors along the branches show results of an ancestral state reconstruction for GSA. Some clades are marked with Roman numerals (see the “Results” section), and species names are abbreviated (three letters for the genus, and three letters or a three-digit number, respectively, for named species and new species; see Additional file 1: Table S2 for full names). The three panels show, from left to right, residual testis size, residual ovary size, and GSA (i.e., testis area/(testis area + ovary area)), with dots representing species means and whiskers representing standard errors (no SEs are given for the former two since the residuals were calculated across all species using species means). Dot color indicates the inferred mating syndrome that the species are assigned to hypodermic (yellow), intermediate (light green), reciprocal (green), and unclear (gray). The last column (N) gives the number of specimens with a GSA estimate. Note that no SEs are drawn for species with one sample only (14 of 120) and in many cases the SEs are small and thus do not extend beyond the symbols

Fig. 3

Distribution of gonadal sex allocation (GSA) across all species and links to heterozygosity. A The black solid line represents the GSA density across all species, and the colored curves show the densities for the species assigned to the hypodermic (yellow) and reciprocal (green) mating syndrome. Points below show raw data, jittered on the y-axis for visibility. B Distribution of GSA by inferred mating syndrome. C Macroevolutionary landscapes inferred using BBMV. The macroevolutionary landscape represents the normalized evolutionary potential, which determines how fast species evolve towards a trait value (see the “Methods” section). Peaks in the landscape correspond to trait values that species are attracted towards. Given are the three models evaluated with their respective AIC values. The full BBMV model was strongly supported. D Relationship between heterozygosity and GSA. E Distribution of heterozygosity by inferred mating syndrome. Color shows the inferred mating syndrome: hypodermic (yellow), intermediate (light green), reciprocal (green), and unclear (gray). D includes the fit and corresponding statistics including all species, as well as separate fits for only the hypodermic and reciprocal mating syndrome, and B, E include results of comparing the hypodermic and reciprocal mating syndrome (i.e., excluding species assigned as intermediate and unclear). Boxplots show the second and third quartile with whiskers extending up to 1.5 times the interquartile range. R² values represent R²_pred of the full PGLS including the phylogeny

Fig. 4

PGLS analysis of body size and gonad size. A, B Regression of testis size and ovary size on body size (left) and distribution of gonad size by inferred mating syndrome (right). The slope was significantly different from unity for the testis (i.e., negative allometry indicated by an *) but not for the ovary (see Additional file 1: Table S7 for details). C Regression of body size on gonadal sex allocation. D Distribution of body size by inferred mating syndrome. E Regression of residual testis size on residual ovary size. Points are scaled based on sample size to illustrate the weights used for the PGLS model. Color shows the inferred mating syndrome: hypodermic (yellow), intermediate (light green), reciprocal (green), and unclear (gray). Scatterplots include the fit and corresponding statistics for analysis with all species. Boxplots include results of comparing the hypodermic and reciprocal mating syndrome (excluding species assigned as intermediate and unclear) and show the second and third quartile with whiskers extending up to 1.5 times the interquartile range. R² values represent R²_pred of the full PGLS including the phylogeny

Fig. 5

PGLS analysis of four sexual traits against gonadal sex allocation. Each panel (A–D) includes the fit and corresponding statistics for analysis with all species (solid line, top), only species assigned to the hypodermic mating syndrome (dotted line, middle) or only species assigned to the reciprocal mating syndrome (dashed line, bottom). Color shows the inferred mating syndrome: hypodermic (yellow), intermediate (light green), reciprocal (green), and unclear (gray). Values are plotted on a logarithmic scale in panels B–D but the axis is labelled with back transformed values. R² values represent R²_pred of the full PGLS including the phylogeny. Detailed results are in Additional file 1: Table S6