Impact of pre-copulatory sexual cannibalism on genetic diversity and efficacy of selection

Abstract

Factors that increase reproductive variance among individuals act to reduce effective population size (Ne), which accelerates the loss of genetic diversity and decreases the efficacy of purifying selection. These factors include sexual cannibalism, offspring investment and mating system. Pre-copulatory sexual cannibalism, where the female consumes the male prior to mating, exacerbates this effect. We performed comparative transcriptomics in two spider species, the cannibalistic Trechaleoides biocellata and the non-cannibalistic T. keyserlingi, to generate genomic evidence to support these predictions. First, we estimated heterozygosity and found that genetic diversity is relatively lower in the cannibalistic species. Second, we calculated dN/dS ratios as a measure of purifying selection; a higher dN/dS ratio indicated relaxed purifying selection in the cannibalistic species. These results are consistent with the hypothesis that sexual cannibalism impacts operational sex ratio and demographic processes, which interact with evolutionary forces to shape the genetic structure of populations. However, other factors such as the mating system and life-history traits contribute to shaping Ne. Comparative analyses across multiple contrasting species pairs would be required to disentangle these effects. Our study highlights that extreme behaviours such as pre-copulatory cannibalism may have profound eco-evolutionary effects.

1. Introduction

Female sexual cannibalism is common during courtship and mating, predominantly in predatory arthropod species [1–3]. This behaviour occurs when a female consumes a male immediately before, during or after mating. Sexual cannibalism impacts both individual- and population-level processes. At the individual level, sexual cannibalism impacts the frequency and opportunity for females to encounter mates and to exert mate choice, since the sex ratio of mating individuals becomes female-biased and may ultimately leave some females in the population unmated [4,5]. At the population level, sexual cannibalism negatively impacts population growth [6], and both the individual- and group-level effects reduce the effective population size (Ne) of the population [7]. Overall, any process that biases the sex ratio or increases the variance in individual reproductive success will decrease Ne [8,9]. Pre-copulatory cannibalism has the most dramatic effects [6], as it removes males from the pool of reproducing individuals [10,11]. Post-copulatory cannibalism may confer substantial fecundity benefits to females but may still have negative effects on population growth [6]. The female's mating system (i.e. whether a female mates with a single or multiple males) further interacts with the effective number of males that sire the female’s offspring, where polyandry (multiple mates) may counteract the negative effect on Ne compared to monandry (single mate) [8,12,13].

Declines in Ne come with elevated genetic drift, which results in loss of population genetic diversity and reduced efficacy of selection [14]. These processes are expected to cause a reduction in fitness and may ultimately threaten population persistence, owing to the accumulation of weakly deleterious alleles [15–17].

Everything else being equal, the relationship between sexual cannibalism and Ne generates the prediction that species with high frequencies of sexual cannibalism should experience relatively lower population genetic diversity and reduced efficacy of selection. To our knowledge, this is the first time the prediction has been investigated. Here, we used two spider species of the genus Trechaleoides, which markedly differ in their level of sexual cannibalism, to make an assessment of this prediction. These two species are the unique species of the genus; they have a shared evolutionary history; individuals are morphologically highly similar; and they occur in sympatric populations. The close relationship between the two species reduces the confounding effects of other environmental and life-history factors in affecting population genetic processes. Trechaleoides biocellata females are very aggressive and exert pre-copulatory sexual cannibalism at high frequencies (57%); a female rarely mates with more than a single male and may even die without mating [18]. In contrast, both males and females of T. keyserlingi are polygamous, and females rarely cannibalize males (14%) [18].

We generated transcriptomic (mRNA) data to perform comparative genomic analyses. We examined both Trechaleoides species using Paratrechalea ornata as an outgroup to determine heterozygosity as a measure of population genetic diversity (H_obs) and dN/dS ratios as a measure of purifying selection [19–21]. Our main objectives were to (i) estimate heterozygosity to assess genetic diversity and (ii) estimate dN/dS ratios as a measure of purifying selection, to assess predictions of the effect of pre-copulatory sexual cannibalism on Ne.

2. Methods

2.1. Spider collections and RNA extractions

We collected six adult males from each species (T. keyserlingi and T. biocellata), consisting of two males from three populations where the two species co-occur in Uruguay: (i) Parque San Miguel (33°41′51″ S, 53°32′00″ W) Rocha, (ii) Quebrada de los Cuervos (32°55′39″ S 54°27′25″ W), Treinta y tres and (iii) Paso Centurión (32°8′2″ S 53°47′55″ W) Cerro Largo. We additionally collected two males from Paratrechalea ornata, from Quebrada de los Cuervos, as an outgroup. We chose only males to avoid any biases related to possible sex differences in gene expression. Once in the laboratory, we flash-froze them in liquid nitrogen and stored them at −80°C until RNA extraction. We extracted RNA from half of the cephalothorax using RNeasy Kit, QIAGEN, following the manufacturer’s recommendations. Then, we stabilized the extractions in an ethanol precipitation. Library preparation (TruSeq Stranded mRNA LT Sample Prep Kit) and sequencing (PE, 101 bp, Illumina platform) were done by Macrogen Inc.

2.2. Quality control and de novo transcriptome assembly

Prior to de novo transcriptome assembly, reads were quality trimmed with Trimmomatic v. 0.36 [22] using default parameters. De novo species assemblies were performed using Trinity v. 2.12.0 [23] using default parameters. We used BUSCO v. 5.2.2 (Benchmarking Universal Single-Copy Orthologous) [24] to evaluate the proportion and quality (complete, fragmented and duplicated) of the spider orthologue genes present in each assembly (Parasteatoda tepidariorum reference). We then used CD-HIT [25,26] with EST mode with a threshold of 0.95 on each assembly to reduce redundancy. From now on, we will refer to these last CD-HIT output assemblies as just assembly. Individual raw read statistics are shown in electronic supplementary material, table S1. BUSCO completeness results for species assemblies are shown in electronic supplementary material, figure S1.

2.3. Genetic diversity

We aimed to assess heterozygosity for each Trechaleoides species and P. ornata. First, we mapped all Trechaleoides and P. ornata individuals’ reads to their corresponding species assemblies using bowtie2 [27]. Second, we performed the SNP calling and created the vcf files using the Bcftools package [28] with a combination of mpileup (-g) and call (multiallelic-caller (-m)) methods. We filtered sites using the Bcftools package, keeping only those with cover higher than 20× and mapping quality (as output by bowtie) over 20. We calculated H_obs as the number of heterozygous sites within the total number of sites for each individual. To test the difference in H_obs between species, we used R software [29] and performed a generalized linear mixed model with binomial distribution (GLMM (B)) including species as a fixed effect and population as a random effect.

2.4. Selection intensity

We aimed to estimate dN/dS ratios in species with and without pre-copulatory sexual cannibalism. First, we used Transdecoder v. 5.5.0 [30] on each species assembly to predict the longest isoforms and the nucleotide coding sequences (CDS), then we used these nucleotide CDS to identify putative 1-to-1 orthologues shared between species using OrthoFinder v. 2.5.4 [31]. Afterwards, we translated the nucleotide sequences to their corresponding amino acid sequences and aligned them using MUSCLE v. 3.8 [32]. We created nucleotide sequence alignments using the protein alignment as a reference in PAL2NAL v. 14 [33] and performed a test of positive selection at the species level using the Codeml program implemented in the PAML package [34]. We evaluated the evolutionary rate of each terminal branch in the species tree by concatenating all individual gene alignments and estimating the rates of non-synonymous (dN) and synonymous (dS) substitutions and their dN/dS ratio for each branch. With PAML, we bootstrapped codon columns from the concatenated alignment (n = 1000), and using the branch-free ratio model (model = 1), we estimated an independent dN, dS and dN/dS for each terminal branch in the phylogeny for each bootstrapped alignment [35]. We, then, estimated 95% confidence limits of dN, dS and dN/dS for each terminal branch using the distribution of the 1000 bootstrapped estimates using R [29]. For the model, we used the phylogenetic tree inferred by OrthoFinder: ((T. biocellata, T. keyserlingi) P. ornata).

Raw reads were submitted to the National Center of Biotechnology Information (NCBI) in the Sequence Read Archive with the number PRJNA977065 and BioSample accessions SAMN35448646 to SAMN35448659. Tables with intermediate results (estimates of dN, dS, dN/dS and H_obs and R codes for these estimations) are accessible as electronic supplementary material in this manuscript and in Dryad [36].

3. Results

3.1. Genetic diversity

The H_obs was significantly lower in T. biocellata (range: 0.00037–0.00058; mean = 0.0005) than in T. keyserlingi (range: 0.00070–0.00087; mean = 0.0008) and intermediate in P. ornata (range: 0.00069–0.00074; mean = 0.00072) (table 1; figure 1).

Table 1.

Generalized Linear Mixed Model analysing the number of heterozygous sites within the total number of sites in relation to species (fixed effect) and population used as a random effect. Significant p-values are shown in bold.

	fixed effects		random effects
number of heterozygous sites within the total sites	species		populations
	estimate	p ‐value	variance	s.d.
Trechaleoides keyserlingi	0.459	<0.0001	0.012	0.109
Paratrechalea ornata	0.249	<0.0001

Figure 1.

Trechaleoides and Paratrechalea ornata species heterozygosity. The x-axis shows the two species; the black boxplot corresponds to the species T. biocellata and the white boxplot corresponds to the species T. keyserlingi. The y-axis shows the number of heterozygous sites within the total number of sites for each individual.

3.2. Selection intensity

From the 80 774 coding transcripts submitted to OrthoFinder to identify orthologous groups, it identified 70 183 genes (86.9% of the total) and 20 355 orthogroups (containing both orthologues and paralogues). There were 13 873 (17.1%) orthogroups shared among the three species, and 8.037 of these were 1-to-1 orthologues all of which were analysed for selection intensity. The free ratios model indicated that T. biocellata had higher values of dN/dS compared with T. keyserlingi (0.1921 versus 0.1833). We found a significantly lower synonymous substitution rate in T. biocellata compared with T. keyserlingi (0.0177 versus 0.0178). Full dN, dS and dN/dS values for each branch are reported in table 2.

Table 2.

Estimates of dN, dS and dN/dS from the branch-free ratios model. It shows the mean for each one and the low and high 95% confidence interval (CI) in brackets for T. biocellata, T. keyserlingi and P. ornata. Significance differences are met when the 95% CI do not overlap.

species	dN	dS	dN/dS
	mean (95% CI)	mean (95% CI)	mean (95% CI)
T. biocellata	0.0034 (0.00341–0.00341)	0.0177 (0.01777–0.01778)	0.1921 (0.1920–0.1922)
T. keyserlingi	0.0032 (0.00326–0.00327)	0.0178 (0.01783–0.01784)	0.1833 (0.1832–0.1834)
P. ornata	0.0101 (0.01016–0.01016)	0.0696 (0.06968–0.06970)	0.1458 (0.1458–0.1458)

4. Discussion

Our findings reveal that the sexually cannibalistic T. biocellata has significantly lower heterozygosity (about two-thirds) and evidence of relaxed purifying selection compared with the non-cannibalistic sister species T. keyserlingi. This result is in accordance with the prediction that pre-copulatory sexual cannibalism (as in T. biocellata) can result in a lower Ne. With only one species comparison, this evidence is preliminary, and other factors may also play a role. For example, we cannot disentangle the mating system from the presence of sexual cannibalism, and there may be other confounding factors such as offspring investment or life-history factors that influence Ne [37]. Nevertheless, these results highlight the potential significant effects that pre-copulatory sexual cannibalism may have on demography and evolutionary processes.

Here, we investigated a species with pre-copulatory sexual cannibalism, which reduces the number of reproducing males and may leave some females unmated. These effects bias operational sex ratio and increase variance in individual reproductive success, processes that act to reduce Ne [8,9]. Mathematical models predict that high frequencies of pre-copulatory sexual cannibalism can hamper population growth [6]. We note that whether cannibalism occurs pre- or post-copulatory may have implications for its effect on Ne. Pre-copulatory cannibalism should have more negative impact on Ne than post-copulatory cannibalism because the male does not reproduce. When cannibalism occurs after mating, the male leaves offspring contributing to Ne and its impact would depend on whether post-copulatory cannibalism can function as a paternal investment [38], as species with high investment in offspring tend to have lower Ne [37]. Differences in mating systems also impact Ne [8]. Observations of the two species suggest that the cannibalistic T. biocellata has a monogamous mating system, while T. keyserlingi females are polyandrous [18]. Therefore, it is possible that monogamy in T. biocellata may act to further reduce Ne and thereby reinforce the negative effects of pre-copulatory sexual cannibalism on Ne [8]. Polygamy, on the other hand, could also contribute to explain changes in Ne among species. This emphasizes the complexity of deciphering the multiple interacting factors that shape Ne. We performed analyses on two species that are closely related, and therefore share a common evolutionary history, which reduces the impact of confounding effects. As we sampled spiders from three distinct locations where the two species co-occur, we did our best to rule out environmental factors that cause differences between species. Interestingly, at the population level, we observed that both species have the lowest heterozygosity in San Miguel. We do not know which factors are affecting individuals in this locality, but these can additionally erode genetic variability, maybe owing to increased random genetic drift, inbreeding and reductions in gene flow [39].

Populations that suffer a loss of genetic diversity can experience reduced fitness, which may have sex-specific effects. This could potentially also impact the expression of sexual traits, for example, the ability of males to produce and display sexual traits [40] (reviewed in [41]). If the evolution of sexual cannibalism has dramatic consequences for Ne, such processes have the potential to change sexual selection processes in small populations. We found evidence of a higher dN/dS ratio in T. biocellata than in T. keyserlingi, which likely reflects a relaxation of the intensity of purifying selection in the former species. Comparable results were found in other species that differ in their mating system, for example, in social versus sub-social spiders [42] and in selfing plant species compared to outcrossing ones [43].

Purifying selection is the dominant force responsible for the evolution of protein-coding sequences such as those studied here. If sexual cannibalism causes a reduction in Ne, the accompanying decrease in the efficacy of selection against weakly deleterious mutations causes an increase in their accumulation [16,42,44]. The accumulation of deleterious mutations can reduce population fitness [17,45].

In conclusion, we provide genomic evidence to assess the impact of sexual cannibalism on Ne. We found support for a reduction in population genetic diversity and relaxed purifying selection consistent with a reduction in Ne in a sexually cannibalistic spider species compared to a non-cannibalistic species. This highlights the important effects that pre-copulatory sexual cannibalism may have on demographic processes by altering the operational sex ratio and potentially increasing variance in reproductive success among individuals, which interact with evolutionary forces to shape the genetic structure of populations. Comparative analyses across multiple contrasting species pairs would be required to establish the effect of sexual cannibalism on population processes and to disentangle effects of other factors such as the mating system.

Ethics

This work did not require ethical approval from a human subject or animal welfare committee.

Data accessibility

Raw reads were submitted to the National Center of Biotechnology Information (NCBI) to the Sequence Read Archive with the number PRJNA977065 and BioSample accessions SAMN35448646 to SAMN35448659. Tables with results (estimates of dN, dS, dN/dS and H_obs) and R codes for these estimations are accessible as electronic supplementary material of this manuscript [46] and in Dryad [36].

Declaration of AI use

We have not used AI-assisted technologies in creating this article.

Authors’ contributions

M.M.V.: conceptualization, data curation, formal analysis, funding acquisition, investigation, writing—original draft, Writing—review and editing; J.B.: conceptualization, formal analysis, methodology, writing—original draft, writing—review and editing; T.B.: supervision, writing—original draft, writing—review and editing; M.J.A.: conceptualization, funding acquisition, investigation, project administration, resources, writing—original draft, writing—review and editing; I.H.T.: conceptualization, funding acquisition, investigation, project administration, resources, supervision, visualization, writing—original draft, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

MMV was supported by Comisión Sectorial de Investigación Científica (CSIC), Agencia Nacional de Investigación e Innovación (ANII), Comisión Académica de Posgrado (CAP), and Programa de Desarrollo de las Ciencias Básicas (PEDECIBA), Montevideo, Uruguay. MJA and IHT were supported by Sistema Nacional de Investigadores (SNI) ANII, and PEDECIBA Montevideo, Uruguay. MJA was supported by Caldeyro-Barcia National Science Award (PEDECIBA).