Convergent rates of protein evolution identify novel targets of sexual selection in primates

Abstract

Sexual selection is the differential reproductive success of individuals, resulting from competition for mates, mate choice, or success in fertilization. In primates, this selective pressure often leads to the development of exaggerated traits which play a role in sexual competition and successful reproduction. In order to gain insight into the mechanisms driving the development of sexually selected traits, we used an unbiased genome-wide approach across 21 primate species to correlate individual rates of protein evolution to relative testes size and sexual dimorphism in body size, 2 anatomical hallmarks of sexual selection in mammals. Among species with presumed high levels of sperm competition, we detected strong conservation of testes-specific proteins responsible for spermatogenesis and ciliary form and function. In contrast, we identified accelerated evolution of female reproductive proteins expressed in the vagina, cervix, and fallopian tubes in these same species. Additionally, we found accelerated protein evolution in lymphoid tissue, indicating that adaptive immune functions may also be influenced by sexual selection. This study demonstrates the distinct complexity of sexual selection in primates revealing contrasting patterns of protein evolution between male and female reproductive tissues.

Introduction

Sexual selection, the process by which traits that increase successful reproduction are favored, has acted on the anatomy and physiology of male primates to produce adaptations including weaponry, coloration, body size dimorphism, large testes, and copulatory plugs (Caro et al., 2021; Darwin, 1859, 1871; Dixson, 2012; Dixson & Anderson, 2002; Kappeler, 1997; Leutenegger & Cheverud, 1985; Lindenfors & Tullberg, 2011). Sexual selection has also resulted in adaptations in the behavior and physiology of female primates, including vocalizations, sexual swellings, and female choice (Maestripieri & Roney, 2005; Small, 1989; Zinner et al., 2004). In contrast to many other orders of mammals, primates exhibit a wide range of mating systems and their associated social groupings from monogamy to single-male polygyny to multimale/multifemale mating (Dixson, 2012; Reichard, 2003; Yamamoto et al., 2009). As such, primates provide a rich opportunity to identify related adaptations resulting from sexual selection, including those driven by intra- and inter-sexual competition, and pre- and post-copulatory sexual selection (Martinez & Garcia, 2020).

Sperm competition, a type of intrasexual postcopulatory selection, occurs in primate species wherein females mate with more than one male within the relatively short period of time surrounding ovulation and fertility (Birkhead & Kappeler, 2004; Dixson, 1995; Gomendio et al., 1998). The most overt male adaptation due to sperm competition is the development of large testes in order to produce large amounts of sperm (Kappeler, 1997; Kenagy & Trombulak, 1986). There is a well-established correlation between the degree of female polyandry and the size of testes relative to body size in primates (Harcourt et al., 1981; Møller, 1988). The largest testes in primates are found in chimpanzees (an African ape) and muriquis (a South American monkey); females of these species have been observed to routinely mate with over ten males in a single day during estrus (Matsumoto-Oda, 1999; Strier, 1999; Tutin, 1980). Other adaptations to sperm competition found among polyandrous primates that are less easily observed include sperm with larger midpiece volumes (Anderson & Dixson, 2002), sperm that swim with greater force (Nascimento et al., 2008), reduced sperm pleiomorphism (Seuánez, 1980), larger seminal vesicles (Dixson, 1998), and a greater degree of coagulation of the ejaculate including the production of a copulatory plug (Dixson & Anderson, 2002). Many of these traits are themselves correlated with each other; as such, relative testes size provides an accessible and reliable proxy for the intensity of sperm competition in primates (Dixson, 2012).

In some primate species males compete overtly with other males through fights, threats, and displays to achieve reproductive access to a number of adult females and/or territory. This is common in species that live in single-male polygynous social groups, where males monopolize access to matings and therefore experience little sperm competition. Rather, this form of precopulatory sexual selection favors males with larger body size or weaponry for use in direct male-male competition (Clutton-Brock, 1985; Harvey et al., 1978). As such, large body size dimorphism in primate species serves as a reliable indicator of considerable intrasexual precopulatory sexual selection (Leigh, 1995; Leutenegger & Cheverud, 1985). However, it is important to note that the converse is not true—that lack of sexual dimorphism is not necessarily indicative of a lack of male-male competition (Lawler, 2009; Plavcan, 2000). Large sexual dimorphism in body size is most commonly found among the Old World monkeys and apes (catarrhines), and usually associated with single-male harem-style polygyny such as that found in gorillas and proboscis monkeys.

Male reproductive genes in mammals have often been reported to be evolving rapidly (Clark & Swanson, 2005; Ramm et al., 2007; Wong, 2010) with sperm competition usually given as the ultimate cause. However, this may be an overgeneralization; it appears that subsets of these genes such as those expressed in the seminal vesicles and those that code for extracellular secreted proteins are frequent targets of strong positive selection while others are not necessarily evolving more rapidly than nonreproductive genes (Carnahan-Craig & Jensen-Seaman, 2014; Dean et al., 2009; Good et al., 2013; Ramm et al., 2007; Wong, 2011). Approaches to identifying targets of sexual selection in primates have usually relied on measures of interspecific sequence divergence in the form of the dN/dS ratio, or variations thereof, often focusing on candidate genes chosen based on a priori knowledge of their function (Carnahan & Jensen-Seaman, 2008; Dorus et al., 2004; Finn & Civetta, 2010; Jensen-Seaman & Li, 2003; Wyckoff et al., 2000). Although dN/dS-based tests have been widely used to infer modes of selection, they are limited in their reliance on the assumption that selection works predominantly through repeated sweeps (Hughes, 2007), and that synonymous substitutions are neutral or very nearly so despite evidence to the contrary (Shen et al., 2022).

RERconverge is an alternative approach to identifying positive and purifying selection that does not rely on the assumption that synonymous substitutions are neutral. It has been successfully used to identify genomic adaptations to marine and subterranean living, extended lifespan, hairlessness in mammals, obligate scavenging, diurnality, and hibernation (Chikina et al., 2016; Christmas et al., 2023; Kowalczyk et al., 2020, 2022; Partha et al., 2017; Richardson et al., 2023; Zou et al., 2021). This approach relies on the assumption that selection for the same phenotype occurring independently in different lineages will target the same proteins, and therefore these proteins will evolve in a similar tempo, when corrected for branch- and gene-specific rates (Kowalczyk et al., 2019; Partha et al., 2019). Proteins that are observed to change more slowly than expected across species with a similar phenotype are inferred to be under strong selective constraint related to that phenotype. Conversely, those proteins evolving more rapidly could be doing so either because of positive selection or relaxation of constraint. Using this method to compare relative rates of protein evolution with quantitative phenotypes of relative testes size and sexual dimorphism in body size, we identified numerous targets of pre- and post-copulatory sexual selection in primates in an unbiased genome-wide approach.

Materials and methods

Dataset

Analyses were conducted using the Mammals Multiz Alignment & Conservation (27 primates) amino acid alignments from the UCSC genome browser. This dataset includes 27 primate species and 3 outgroups, containing 19,975 amino acid sequence alignments of known human genes aligned against the translated sequence from the genome assemblies of other species. In order to reduce the size and complexity of the dataset, especially where including multiple closely related species adds little additional information, we removed six primates (mouse lemur (Microcebus murinus), Sclater’s lemur (Eulemur flavifrons), black snub-nosed monkey (Rhinopithecus bieti), crab-eating macaque (Macaca fascicularis), pig-tailed macaque (Macaca nemestrina), and bonobo (Pan paniscus)), five of which were still represented by another member of the same genus, leaving a final dataset including 21 primates along with mouse and dog as outgroups (Mus musculus and Canis lupus familiaris). Primates include the northern greater galago (Otolemur garnettii), black lemur (Eulemur macaco), Coquerel’s sifaka (Propithecus coquereli), Philippine tarsier (Carlito syrichta), Nancy Ma’s night monkey (Aotus nancymae), white-faced sapajou (Cebus capucinus), squirrel monkey (Siamiri boliviensis), marmoset (Callithrix jacchus), golden snubnose monkey (Rhinopithecus roxellana), Angolan colobus monkey (Colobus angolensis), proboscis monkey (Nasalis larvatus), drill (Mandrillus leucophaeus), green monkey (Chlorocebus sabaeus), olive baboon (Papio anubis), sooty mangabey (Cercocebus atys), rhesus macaque (Macaca mulatta), northern white-cheeked gibbon (Nomascus leucogenys), Sumatran orangutan (Pongo pygmaeus), western lowland gorilla (Gorilla gorilla), common chimpanzee (Pan troglodytes), and human (Homo sapiens). The tree topology was generated by the UCSC Mammals Multiz Alignment project based on genome-wide fourfold degenerate sites. Alignments were trimmed to eliminate species with poor sequencing coverage and gaps extending the length of more than 50% of the protein, including those caused by a premature stop codon terminating translation. After the initial pruning, proteins with less than 10 species were also removed from the dataset leaving 18,876 aligned amino acid sequences. For each alignment branch lengths were estimated using the aaml program in the Phylogenetic Analysis by Maximum Likelihood (PAML) package version 4.9 (Yang, 2007), using an empirical model of amino acid substitution rates with rate variability between sites modeled as a gamma distribution approximated with four discrete classes and an additional class for invariable sites (aaml model “Empirical + F”). Amino acid alignments with 80% of the branches having zero value branch lengths were removed from the final dataset, leaving 16,984 proteins for further analysis (Supplementary Table 1). Data trimming, manipulation, and automation was performed with custom python scripts.

Trait identification

For each primate species, relative testes size was calculated with consideration to the allometric relationship between testes mass and body mass. This was calculated as the logarithm of the ratio of observed testes size compared to the expected testes size generated using the function Y = 0.034X^0.68 where Y is the expected mass of both testes in grams and X is the observed body mass in grams (Kenagy & Trombulak, 1986). Thus, relative testes size, as used here, is equivalent to the residuals from a plot of testes mass to body mass where both variables are log-transformed. Observed body mass and testes mass values for this trait were collected as described in Supplementary Table 2. Values of body mass for males and females of each species were taken from Smith and Jungers (1997), except for those of the drill because of their small sample size; instead, drill body mass was taken from Setchell (2017). Body size dimorphism was calculated as the log-transformed ratio of male body weight over female body weight. Ancestral states were estimated through phylogenetic modeling based on extant species values using the phytools R package (Revell, 2012).

Relative evolutionary rate analysis

To perform our primary analysis, we used the RERconverge pipeline (Kowalczyk et al., 2019) which quantifies the correlation between the rate of change of a trait and the rate of protein evolution allowing for the identification of proteins which are evolving in response to selection related to a particular phenotype along independent lineages. With internal and terminal branch lengths corresponding to the average rate of amino acid changes, we estimated a master tree for the 21 primate species representing the average rate of amino acid substitution across the proteome. Using linear regression, we identified the degree of divergence for each protein from its expected rate of evolution to calculate the relative evolutionary rate (RER). Phenotype vectors and calculated RERs for all branches were correlated using Pearson linear correlations and were adjusted using the Benjamini–Hochberg procedure as well as 1,000 phylogenetically restricted permutations, or permulations, as described in Saputra et al. (2021).

Tissue and pathway enrichment

Expression summaries were collected from the Human Protein Atlas (proteinatlas.org) for each nominally significant protein for both traits (uncorrected p < .05) including tissue expression (Uhlén et al., 2015) and single cell type expression data (Karlsson et al., 2021). Tissue enrichment analyses were performed for 38 tissues where a Wilcoxon Rank-Sum statistic was estimated using RERConverge commands to compare the rho value multiplied by the negative log of the p-value for correlations of genes within the tissue against the same value for all genes. Tissue enrichment analysis was performed to identify tissue expression patterns for conserved (rho < 0) and accelerated (rho > 0) proteins for each trait, with p-values adjusted with the Benjamini–Hochberg procedure and phylogenetic permulations (Saputra et al., 2021). Metascape (https://metascape.org) with default parameters was used for Gene Ontology enrichment analysis of conserved and accelerated proteins with uncorrected p-values and permulation p-values less than .05 (Zhou et al., 2019).

Likelihood ratio tests for selection

To differentiate between positive selection and relaxed constraint for the top 200 proteins identified in the RER correlation for each trait, ranked by p-value, we used codeml from the PAML package (Yang, 2007) to fit the data (alignment of coding regions from the Mammals Multiz Alignment & Conservation nucleotide alignments) to a model constrained to a single ratio of the nonsynonymous substitution rate to the synonymous substitution rate (dN/dS) across all branches of the phylogeny (uniform ratio model, M0). The likelihood of two other models were estimated, a branch model where two sets of branches (foreground and background) were each allowed to have their own dN/dS, and another branch model where the foreground was constrained to neutrality (dN/dS = 1). For each trait, in these models the terminal branches leading to any extant species in the top one-third of the trait value were specified as foreground, and the remaining branches as background. To determine if the foreground and background branches are evolving differently as fit to these models, a likelihood ratio test (LRT) compared the likelihoods of the two-branch model to the uniform model with significance assessed using a chi-square test with one degree of freedom. To test for positive selection across the entire gene, an LRT compared the two-branch model to the neutral foreground model; the was considered as evidence for positive selection if the result was statistically significant and the foreground dN/dS was greater than one.

Results

Dataset and phenotype evolution

Our analysis was performed using the University of California Santa Cruz Genome Browser Mammals Multiz Alignment & Conservation (27 primates) dataset including 19,975 aligned orthologous amino acid sequences. After pruning, our final dataset included 16,984 proteins from 21 primate species and 2 outgroups (mouse and dog). For each primate species, we calculated residual relative testes size and body size dimorphism and used phylogenetic modeling to estimate ancestral states of extinct ancestors (Figure 1). This approach reconstructed the last common ancestor (LCA) of the New World monkeys to have moderately sized testes relative to body size, with a reduction occurring in the monogamous Ma’s night monkey. The LCA of the cercopithecoids is predicted to have relatively large testes, with notable further increases in the baboon and rhesus macaque lineages, and a reduction in relative testes size in the proboscis monkey, which live in single-male polygynous social units. Finally, our phylogenetic modeling predicted the last common ancestor of the African apes, including humans, to have intermediate-sized testes relative to body size, with a strong reduction in the polygynous gorilla and a pronounced increase in the polyandrous chimpanzee. There is relatively little sexual dimorphism in body size among the strepsirrhines, tarsiers, and platyrrhines, and consequently little change during their evolutionary history. Among the catarrhines, our phylogenetic modeling inferred independent increases in sexual dimorphism along the branches leading to the drill, orangutan, gorilla, and proboscis monkey, while identifying a reduction in body size sexual dimorphism during the evolution from the hominoid LCA to the socially monogamous gibbons.

Figure 1.

Primate cladogram with branches colored to correspond with the change in logged A) relative residual testes size B) body size dimorphism. Branch lengths are not representative of evolutionary distance. See Materials and Methods for scientific names of all species.

Relative evolutionary rate analysis

The relative evolutionary rate (RER) for each branch and each of the 16,984 proteins was correlated to each trait using Pearson linear correlation (Supplementary Table 3). The direction of the correlation is indicated by a positive or negative correlation statistic, or rho value. For the testes size analyses, we interpret a positive rho value as accelerated evolution in species with relatively large testes. This acceleration could be indicative of positive selection or relaxed constraint of a given protein in species with relatively large testes; alternatively, it could indicate strong conservation in primates with relatively small testes—with these two possibilities not mutually exclusive. Inversely, a negative rho value suggests stronger conservation in primates with relatively large testes than that seen in species with relatively small testes. For the sake of brevity, proteins more accelerated or more conserved in primates with relatively large testes than those with relatively small testes will be described as “large testes accelerated” and “large testes conserved,” respectively. The same naming convention will be used for proteins accelerated and conserved in primates with a large degree of sexual dimorphism in body size. In species with large relative testes sizes, we found 1,000 proteins with rates that deviated significantly from expectation (uncorrected p < .05). Of those, 444 were more conserved (rho < 0) and 556 were more accelerated (rho > 0). Similarly, in species with larger males than females, 761 proteins are evolving significantly faster or slower (uncorrected p < .05) with 350 being more conserved and 411 more accelerated than expected (Supplementary Table 3).

Tissue expression

Of the thirteen proteins with Benjamini–Hochberg adjusted permulation p-values less than .001 in the relative testes size analysis, six are most highly expressed in reproductive tissues including testis, epididymis, breast, cervix, and vagina. (Table 1). Single cell type specificity shows a similar trend with seven of these proteins highly expressed in reproductive cell types. Of the proteins with defined functions, CDC123, CCNA1, UBE2S, and OASL have all been shown to be involved with cell cycle regulation and cell division, and DEFB126 aids in sperm transport and fertilization. In the analyses based on sexual dimorphism in body size, we detected four proteins with Benjamini–Hochberg adjusted permulation p-values less than .001: GPR148, LYPD8, MAATS1, and ZIM2 (Table 2). LYPD8 is enriched in the intestines, MAATS1 is enriched in the fallopian tube, and ZIM2 is enriched in brain, testes, and ovary. ZIM2 and MAATS1 are both highly expressed in ciliated and reproductive tissues and cell types.

Table 1.

Top accelerated and conserved proteins in primate species with relatively large testes.

Protein	Rho	Tissue specificity	Tissue expression cluster	Single cell type specificity	Single cell type expression cluster	Molecular function	Biological process
Large relative testes conserved
CDIP1	−0.977	Low specificity	Nonspecific	Early spermatids	Bipolar cells: visual perception		Apoptosis
CDC123	−0.770	Low specificity	Cell cycle regulation	Low specificity	RNA binding		Cell cycle, cell division
CCNA1	−0.588	Testis, parathyroid	Ciliated cells: cilium organization	Spermatocytes, early and late spermatids, endometrial ciliated cells	Spermatocytes and spermatids: spermatogenesis	Cyclin	Cell cycle, cell division, mitosis
Large relative testes accelerated
APOOL	0.745	Low specificity	Mitochondria	Cardiomyocytes	Smooth muscle cells: muscle contraction
UBE2S	0.741	Bone marrow, testis	Bone marrow: mRNA splicing & cell cycle	Dendritic cells, breast myoepithelial cells	Nonspecific: mitochondria, early and late spermatids	Transferase	Cell cycle, cell division, Ubl conjugation pathway
DEFB126	0.690	Epididymis	Epididymis: antimicrobial	Not detected	Squamous epithelial cells: keratinization	Antibiotic, antimicrobial, defensin	Fertilization
GPRIN3	0.554	Brain	Cerebellum: gene expression regulation	T-cells, NK-cells, macrophages	Macrophages, prostate T-cells, adipose T-cells, skeletal muscle endothelial cells
ARHGEF15	0.553	Adipose, breast	Adipose: ECM organization	Adipocytes, endothelial cells	Endothelial cells: angiogenesis	GTPase activation, guanine-nucleotide releasing factor	Transport
ARSL	0.573	Kidney, liver, pancreas	Vesicular transport	Proximal tubular cells, hepatocytes, undifferentiated cells, proximal enterocytes, Paneth cells, enteroendocrine cells, distal enterocytes	Intestinal epithelial cells: absorption	Hydrolase
C1orf74	0.561	Skin	Skin: epidermis development	Squamous epithelial cells, glandular and luminar cells, suprabasal keratinocytes	Squamous epithelial cells, prostate basal glandular cells
SPRR3	0.543	Cervix, esophagus, vagina	Esophagus: epithelial cell function	Suprabasal keratinocytes	Keratinocytes: cornification	Keratinization
OASL	0.516	Bone marrow, stomach	Bone marrow: mRNA splicing and cell cycle	Gastric mucus-secreting cells, distal enterocytes, Paneth cells, T-cells, intestinal goblet cells, Hofbauer cells, early spermatids
KLHDC9	0.493	Testis	Testis: unknown function	Late spermatids, early spermatids, spermatocytes	Spermatids: spermatogenesis

Table 2.

Top accelerated and conserved proteins in primates with body size sexual dimorphism.

Protein	Rho	Tissue specificity	Tissue expression cluster	Single cell type specificity	Single cell type expression cluster	Molecular function	Biological process
Dimorphic body size conserved
ZIM2	−0.591	Brain, ovary, testis	Ciliated cells: cilium & cell projection	Late spermatids, early spermatids, spermatogonia, inhibitory neurons, theca cells, spermatocytes	Spermatogonia & spermatocytes: spermatogenesis	DNA-binding	Transcription, transcription regulation
GPR148	−0.539	Not detected	Not detected	Not detected	Not detected	Membrane	G-protein coupled receptor, receptor, transducer
MAATS1	−0.513	Fallopian tube	Ciliated cells: cilium & cell projection	Cardiomyocytes	Nonspecific: transcription regulation
Dimorphic body size accelerated
LYPD8	0.541	Intestine	Intestine	Paneth cells, distal enterocytes	Intestinal epithelial cells

Using a dataset of enriched proteins for each human tissue from the Human Protein Atlas (proteinatlas.org) we conducted a tissue enrichment analysis comparing nominally significant proteins (uncorrected p < .05) for each trait to the full background dataset, with the enrichment statistics corrected with both 1,000 phylogenetic permulations and the Benjamini–Hochberg procedure (Supplementary Table 4). Here we detected significant tissue enrichment of testicular proteins in the large testes conserved group with 102 of the 444 proteins being testes-enriched. Interestingly, we also detected enrichment of testicular proteins in the large body conserved group, though the association is not as strong. Female reproductive tissues were enriched in large testes accelerated (vagina and cervix) and large body conserved proteins (fallopian tube). Large testes accelerated proteins expressed in lymphoid tissue, endocrine tissue (pituitary gland), and digestive tissues (liver, intestine, and stomach) are also significantly enriched (Figure 2).

Figure 2.

Tissue enrichment analysis was conducted for all proteins with uncorrected p-values below .05 and tissue expression was compared to a background dataset including all 16,984 proteins investigated. Heatmap colors were generated using tissue enrichment scores (ES) where the ES for each tissue denotes the degree by which tissue-enhanced proteins are overrepresented in the list of accelerated or conserved proteins in primates with large relative testes or large body size dimorphism. The top 10 tissues for each trait are listed with their corresponding p-values and Benjamini–Hochberg adjusted p-values. Female and male reproductive tissues are represented above each table. Figure produced using gganatogram (Maag, 2018).

Gene ontology enrichment analysis

Given that there is little overlap between primates with both large relative testes and large body size dimorphism, we sought to further investigate the large number of testes-specific proteins that are significantly conserved in both species with large testes and with large size dimorphism. To do so, we performed a gene ontology enrichment analysis comparing both sets of conserved testicular proteins to one another as well as a background including all 16,984 proteins (Figure 3 and Supplementary Table 5). We detected functional overlap in proteins involved in meiotic cell cycle, male gamete generation, and male infertility. Interestingly, we found significant functional enrichment for proteins involved in DNA methylation during gamete generation and sperm to egg plasma membrane fusion in large testes conserved proteins. In large body dimorphism conserved proteins, we found that female gamete generation and nuclear chromosome segregation are also enriched.

Figure 3.

Pathway enrichment heatmap displaying the negative logged p-values of the top pathways enriched in testes-specific proteins which are conserved in primates with large relative testes and large body size dimorphism. The summary of these results is represented in Supplementary Table 5.

In addition, we performed a gene ontology enrichment analysis where conserved (rho < 0) and accelerated (rho > 0) proteins (uncorrected p < .05) were separated for each trait and compared against a background dataset of all 16,984 proteins. Here we found that there is functional enrichment for male gamete generation, cilium organization and assembly, as well as the CTFR-GOPC-STX6 complex in large testes conserved proteins. On the other hand, proteins which are more rapidly evolving in species with relatively large testes appear to perform functions related to viral and bacterial response and immune system regulation (Figure 4 and Supplementary Table 6). In primates with large sexual dimorphism in body size we detected significant enrichment of proteins involved in the immune response to foreign invaders among accelerated proteins and female meiotic nuclear division in large dimorphism conserved proteins (Figure 4 and Supplementary Table 7).

Figure 4.

Pathway enrichment heatmaps displaying the negative logged p-values of the top 25 pathways which are enriched in the proteins significantly accelerated and conserved in the RER analyses (uncorrected p < .05) in primates with large A) relative testes and B) body size dimorphism. The summary of these results is represented in Supplementary Tables 6 and 7.

Likelihood-based tests of selection

Of the 200 genes with the most significant correlation between their RER and relative testes size, 48 of them showed significant heterogeneity in dN/dS between the foreground and background branches, where the foreground branches were those terminal branches leading to the seven primates with the largest relative testes. Similarly, of the 200 genes with the most significant correlation between their RER and body size dimorphism, 32 had significant heterogeneity in dN/dS. However, in no case for either phenotype was the dN/dS of the foreground branches significantly greater than one, as assessed by comparing a model allowing for a single value of dN/dS for the foreground to a model constraining the foreground dN/dS to 1 (Supplementary Tables 8 and 9).

Discussion

Differing primate mating behaviors and their associated forms of pre- and post-copulatory sexual selection have led to distinct morphological and physiological adaptations (Birkhead & Kappeler, 2004; Darwin, 1859; Dixson & Anderson, 2001). The molecular basis for these adaptations is largely unknown. With our unbiased proteome-wide scan we have identified proteins which may have evolved in response to pre- and postcopulatory sexual selection resulting in large sexual size dimorphism and relative testes sizes, respectively. In contrast with previous work describing the rapid evolution of male reproductive genes in response to postcopulatory sexual selection (Clark & Swanson, 2005; Clark et al., 2006; Torgerson et al., 2002), our results suggest conservation of proteins involved in male gamete generation and cilia formation in species with relatively large testes. On the other hand, we detected accelerated evolution of proteins associated with female reproductive tissues in these same primates (Figure 2). Additionally, we detected a correlation between the rate of evolution of proteins involved in the regulation of the adaptive immune system and phenotypes resulting from both pre- and postcopulatory sexual selection.

Ancestral state reconstruction

Our phylogenetic modeling predicted that the ancestral state of the haplorrhine LCA to have had intermediate relative testes size and low levels of sexual dimorphism in body size (Figure 1). These results are largely congruent with previous attempts at primate-wide reconstructions of ancestral mating systems and group structure, which concluded that the LCA of the New World and Old World monkeys lived in multimale/multifemale social groups and exhibited polygynandry (Opie et al., 2012; Shultz et al., 2011). Extant primates with these behaviors tend to have relatively low body size sexual dimorphism, but relatively large testes due to the sperm competition coincident with polygynandrous mating. These and other studies reconstructed the origin of harem-style polygyny to have arisen later in evolution independently several times among the cercopithecoids, among whom it is now common (Grueter et al., 2012; Opie et al., 2012). The anatomical correlates of mating behavior investigated in the present paper reinforce this conclusion, with the common ancestor of catarrhines showing minimal sexual dimorphism in body size, followed by independent increases in dimorphism in the drill, proboscis monkey, and gorilla (Figure 1B). Among the hominoids, gorillas and chimpanzees provide extreme examples of body size dimorphism and large relative testes, respectively. Our analyses reconstruct their common ancestor to have had an intermediate phenotype in both regards, with these adaptations in body size and testes size occurring along their respective terminal branches leading to the extant gorilla and chimpanzee species. We predict the LCA of humans and chimpanzees to have had rather moderate sexual dimorphism in body size, not surprisingly considering that this is also true of both extant species. No fossil has yet been found that represents the human-chimp LCA, but several early hominins (e.g., Australopithecus sensu lato) appear to have been extremely sexually dimorphic, on par with gorillas, while the evolution of later hominins (i.e., Homo) shows a trend of reduction leading to modern humans (Plavcan & van Schaik, 1997; Villmoare et al., 2019). The conclusion that australopithecines exhibited a large degree of body size sexual dimorphism remains quite contentious, and a full review of this is well beyond the scope of this discussion (Reno et al., 2003; Sasaki et al., 2021). Regardless, if true, it would suggest that this trait has been highly labile throughout hominid evolution. Finally, our phylogenetic modeling predicted that the human-chimpanzee LCA had similar relative testes size to humans with chimpanzees possessing highly derived large testes. Although many authors have suggested that the human sociosexual system evolved from a chimp-like polygynandrous system (e.g., Gavrilets, 2012; Lovejoy, 2009), it is more parsimonious to conclude that the large testes of chimpanzees evolved following the divergence from humans, rather than before followed by a reversion in the human lineage.

Reproduction

Much of the previous work on male reproductive proteins has focused on rapid evolution primarily attributed to postcopulatory sexual selection by way of sperm competition (Clark & Swanson, 2005; Ramm, 2014; Wong, 2010). However, in our analyses we identified a large number of proteins that are strongly conserved in response to sperm competition, with nearly 25% of the proteins significantly conserved in species with relatively large testes (uncorrected p < .05) showing enriched or enhanced expression in testes, despite the fact that testicular proteins only make up ~10% of total protein (proteinatlas.org). Unsurprisingly, our pathway enrichment results (Figures 3 and 4) suggest that these proteins are largely performing functions related to the function and regulation of male gamete generation. Thus, we have identified a number of proteins which are more strongly conserved in species with high predicted levels of sperm competition, suggesting that strong purifying selection has consistently removed any variants at these genes, while allowing for more relaxed constraint in species with low sperm competition. In primates with large sexual dimorphism in body size, we notice a similar trend of conservation of testes-enriched proteins which are associated with gamete generation. This result can be attributed to either positive selection or relaxed constraint in sexually monomorphic species and/or purifying selection in sexually dimorphic species. Additionally, we find that seminal vesicle proteins are relatively conserved in single-male polygynous primates which is compatible with previous research suggesting that seminal vesicle proteins are rapidly evolving in response to sperm competition (Carnahan-Craig & Jensen-Seaman, 2014; Karn et al., 2008).

Female reproductive tissues are often left out of the conversation around sperm competition and mating system variation, but evidence suggests that interactions between the female reproductive tract and male ejaculate may serve as a method of female choice (Firman, 2020; Holt & Fazeli, 2016; Ishikawa et al., 2016). Further, sperm–egg interactions have been implicated as the cause of rapid evolution in some reproductive proteins (Vicens et al., 2014). We observed accelerated rates of amino acid substitutions in vaginal and cervical proteins in large testes primates, suggesting either accelerated positive selection or relaxed constraint in these species (Figure 2).

Immunity

Immune genes, similar to reproductive genes, tend to evolve rapidly compared to genome-wide rates of evolution in animals (Castillo-Davis et al., 2004; Waterston et al., 2002). Specifically in primates, positive correlations between white blood count values and mating promiscuity suggest that immune response genes may be rapidly evolving in response to exposure to sexually transmitted diseases in promiscuous species (Nunn et al., 2000). Immune-related proteins that interact closely with pathogens have also been shown to evolve faster in promiscuous primates (Wlasiuk & Nachman, 2010). Similarly, through our enrichment analyses we found that bacterial and viral response proteins show accelerated rates of evolution in primates that experience precopulatory sexual selection as well as in those species experiencing strong postcopulatory sexual selection (Figure 4).

Both the innate and adaptive immune systems play intimate and important roles in spermatogenesis, fertilization, pregnancy, and birth (Weng et al., 2023). During spermatogenesis, the haploid sperm need to be protected against autoimmune attack by the male immune system in the testes; failure to do so may lead to infertility (Hillgarth et al., 1997; Shibahara et al., 2022). Similarly, antisperm antibodies in the vagina and cervix can also result in infertility, and therefore the female immune system must be modulated to simultaneously avoid rejecting sperm while preventing infection by sexually transmitted pathogens (Shibahara et al., 2021; Wigby et al., 2019). Implantation of a blastocyst and subsequent growth of the placenta occurs in a proinflammatory environment, where female immune cells appear to be necessary at the site of implantation for successful establishment and maintenance of pregnancy (Mor et al., 2011). These roles of the immune system in male and female reproductive biology may differ significantly in their action or intensity in species with higher or lower sperm competition. For example, in chimpanzees polyandrous females likely encounter orders of magnitude more sperm cells per mating period than closely related monandrous gorillas, as well as an increased risk of sexually transmitted pathogens. It has also been proposed that the female immune response could serve as a form of cryptic female choice in some animals (Eberhard, 1996); the evolution of cryptic female choice should be more prominent in polyandrous species.

Testosterone suppresses immune function. This suppression may be a consequence of the increase of testosterone necessary for the development of male characteristics including larger body and muscle mass, as well as sexually dimorphic secondary sex characteristics as described by the immunocompetence handicap hypothesis (Folstad & Karter, 1992). Conversely, the immune suppression by testosterone may be a mechanism to avoid autoimmune attack of sperm by the male, as noted above (Hillgarth et al., 1997). Finally, this suppression may simply reflect an evolutionary trade-off, or sexually antagonistic selection (Schroderus et al., 2010). Whether immune system repression is an adaptive response, or an unfortunate repercussion of high testosterone is yet to be determined. Regardless, testosterone levels, their seasonality, and interaction with the immune system could conceivably vary among primate species in a manner consistent with types and strength of sexual selection. Indeed, Anderson and Jones (Anderson & Jones, 2022) recently identified a correlation between the numbers of genome-wide androgen response elements and traits indicative of pre- and post-copulatory sexual selection across a similar set of primate species.

Ciliated cells and ion transport

Our results also suggest a link between cilia formation and postcopulatory sexual selection. Cilia are centriole-based projections used for cellular movement as well as chemosensation and proprioception (Satir & Christensen, 2007). The spermatozoan flagellum is formed with the same core structural components as motile cilia and, subsequently, male fertility is often impacted in cases of ciliopathies such as Primary Ciliary Dyskinesia (Fliegauf et al., 2007). In primate species with relatively large testes, proteins involved in the development and function of cilia are significantly conserved. We expect that ciliary proteins are under relaxed constraint in species with small relative testes due to reduced pressure for high quality sperm.

While proteins related to the formation and function of cilia are largely conserved, we detected acceleration in an essential voltage-dependent calcium channel known as the cation channel of sperm (CatSper) channel. CatSper is responsible for the hyperactivation and chemotaxis of sperm flagellum modulated by intracellular Ca²⁺ concentrations (Carlson et al., 2003, 2009; Granados-Gonzalez et al., 2005; Ren et al., 2001). In mice, the CatSper complex is made up of 4 core components (CATSPER1-4) and at least 6 identified auxiliary subunits (CATSPERb, CATSPERd, CATSPERe, CATSPERg, CATSPERz, and EFCAB9). It has been shown that knockouts of any of the 4 core components or CATSPERd, lead to loss of CatSper function and, ultimately, infertility (Avenarius & Hildebrand et al., 2009; Chung & Navarro et al., 2011; Hildebrand & Avenarius et al., 2010; Jin et al., 2007; Luo & Chen et al., 2019; Qi & Moran et al., 2007). Knockouts of CATSPERz and EFCAB9 as well as deletions in CATSPERe have been shown to cause asthenospermia in both murine and human models (Brown et al., 2018; Chung et al., 2017). Of the six characterized auxiliary subunits, three were identified in the present study as evolving significantly more rapidly in species with relatively large testes (CATSPERb, p < .001, CATSPERe, p = .003, CATSPERg, p = .028), and a fourth nearly so (CATSPERd, p = .07), while there was no correlation in the rate of evolution of the core components (CATSPER 1-4) with either trait.

The patterns of evolution observed in the CatSper complex through our analysis are consistent with previous work showing that the CatSper complex has undergone strong positive selection throughout the primate lineage (Cooper & Phadnis, 2017). Our work extends that observation to reveal that the rates of amino acid substitution differ in accordance with mating system and modes of sexual selection. Much of the previous work on the CatSper complex has been focused on the core components, but this compelling result highlights the necessity for further investigation into the auxiliary components. Even further than the direct implications, here we have observed an interesting example of an essential pathway where auxiliary and core components are evolving differentially in response to a strong selective pressure. Perhaps across structures and pathways more generally, auxiliary components of essential complexes are more likely to be targets for positive selection to drive adaptive change while purifying selection maintains the function of core components.

Conclusions

Through our use of an unbiased genome-wide approach, we have identified targets of pre- and post-copulatory sexual selection, many of which are directly involved in male and female reproduction, ciliary function, and the immune response. Though perhaps of even more interest are the vast number of proteins which are strongly correlated with testes size and sexual body size dimorphism and yet lack known reproductive functions. These proteins serve as promising targets for future characterization to elucidate the genetic mechanisms driving adaptation via sexual selection. In addition, although the present study focused solely on protein evolution, many or even most of the targets of sexual selection may be noncoding regulatory elements, and as such should be the focus of future work (e.g., Anderson & Jones, 2022; Roy, 2022; Tosto et al., 2023). One ongoing challenge when identifying proteins that evolve rapidly is distinguishing between accelerated evolution due to recurrent selective sweeps (positive selection) and that due to relaxed constraint. Codon-based branch model and branch-site model likelihood approaches can be used for this purpose (Álvarez-Carretero et al., 2023), but are limited by the requirement to group branches into a small number of discrete categories. Our approach used in the present paper allows for continuous variation in any phenotype, and can successfully identify proteins that are evolving differently in different species that experience a range of intensities of sexual selection, with the limitation that the patterns produced under positive selection cannot be easily distinguished from those under relaxed constraint.

Data availability

Data tables can be found in the Supplementary Materials. Additional large data files have been deposited in Dryad: 10.5061/dryad.fn2z34v23.

Author contributions

B.L.P. and M.I.J.S. conceived the project. B.L.P. collected the data, carried out all analyses, and created figures. B.L.P. and M.I.J.S. wrote the manuscript.

Conflict of interest: The authors declare no conflict of interest.