Speciation and changes in male gene expression in Drosophila

Genome-wide analyses have shown that gene regulatory changes and adaptive evolution of gene expression is common among species of Drosophila (Rifkin et al. 2003; Graze et al. 2009; Coolon et al. 2014; Nourmohammad et al. 2017). While changes in expression of many different genes can contribute to speciation, male reproduction genes are most likely major contributors to speciation for at least three reasons: (i) sexual selection and sexual conflict can lead to rapid gene sequence divergence in male genes (Swanson and Vacquier 2002; Clark et al. 2006; Haerty et al. 2007), (ii) early stage breakdowns to gene flow between species predominantly involves disruptions in male reproductive hybrid fitness attributable male reproduction genes (Singh 1990, 1999; Wu and Davis 1993), and (iii) disruption of male gene expression in male sterile hybrids is likely to contribute to the establishment of reproductive isolation (Gomes and Civetta 2015; Alhazmi et al. 2019; Go et al. 2019).

Comparative analyses of gene expression among species of Drosophila has shown that expression of male-biased genes (e.g., testis- and accessory gland-specific genes) diverges faster than female-biased genes and nonsex-biased genes (Meiklejohn et al. 2003; Ranz et al. 2003a; Nuzhdin et al. 2004; Jagadeeshan and Singh 2005; Zhang et al. 2007; Grath et al. 2009; Jiang and Machado 2009; Assis et al. 2012; Nourmohammad et al. 2017; Llopart et al. 2018). One of these studies showed that there is also more variation in expression within species if the gene is male biased in that species (Zhang et al. 2007). Moreover, male-biased genes show high levels of gene expression variation between different populations of D. melanogaster (Hutter et al. 2008; Kohn et al. 2008) and latitudinal clines of D. serrata (Allen et al. 2017, 2018). Divergent gene regulatory mechanisms causing gene expression divergence could be driven by different selective pressure in different species (Mack and Nachman 2017; Signor and Nuzhdin 2018). Interestingly, DNA sequence evolution of genes expressed primarily or uniquely in the male reproductive tract tissue such as testis-specific genes, male seminal fluid, and spermatogenesis genes have been shown to evolve faster and show positive selection in lineages of Drosophila (Haerty et al. 2007).

The rapid evolution of male genes such as those with roles in sperm development could be intuitively linked to hybrid disfunction that often involves male sterility. In fact, studies that surveyed differences in gene expression between species and F₁ fertile versus sterile hybrids, or the use of crossing designs that create hybrids in which the sterility phenotype could be manipulated, have in general found evidence for the rapid evolution of regulation of expression of sperm production genes and an overall enrichment for misregulation of spermatogenesis genes associated with sterility (Maside et al. 1998; Ting et al. 1998; Michalak and Noor 2003; Moehring et al. 2007; Ferguson et al. 2013; Gomes and Civetta 2014; Brill et al. 2016; Civetta 2016; Alhazmi et al. 2019) (Table 1). Despite possible uncertainties of effects unlinked to sterility, OdsH, a gene that functions in sperm production in young males but causes sterility in the hybrids between D. mauritiana and D. simulans (Table 1), has shown misregulation of expression in the sterile condition with accumulation of transcripts in the testes (Sun et al. 2004).

Seminal fluid protein (SFP)-coding genes have been shown to evolve under positive selection likely linked to post-mating sexual selection (Swanson et al. 2001; Kern et al. 2004; Wong et al. 2008; Civetta and Reimer 2014). This is also a very interesting group of genes as their rapid evolution often leads to a lack of orthology among more distantly related species of Drosophila (Haerty et al. 2007; Ahmed-Braimah et al. 2017). How sequence and expression changes in SFP genes might influence speciation is an open area of inquiry, given their rapid divergence. An early study suggesting their potential role in post-mating reproductive isolation comes from Fuyama (1983), who showed that stimulation of ovulation in D. suzukii females via transferred seminal fluid of males did not occur when males mated with D. pulchrella females. This caused an effective barrier to hybridization and highlighted the potential role of genetic divergence in seminal fluids in the establishment of post-mating reproductive barriers between species. Recently, the role played by specific SFP genes in speciation has been demonstrated by experimentally perturbing their expression. The study found that knock-down of two SFP genes AcpDE36 and CG9997 (Table 1), but not another seminal fluid gene Sex Peptide, affected conspecific sperm precedence (CSP) when sibling species of Drosophila mated (Castillo and Moyle 2014). CSP is a post-mating prezygotic barrier between species caused by the preference of females to utilize sperm from males of the same species as opposed to heterospecific sperm (Price et al. 2000). The potential role of other seminal fluid genes in CSP is unknown. SFP genes with known roles in sperm competition (sexual selection) are interesting candidates to test whether gene perturbation through knock down or knock out also contributes to reproductive isolation via CSP (speciation). Testing connections between sexual selection/conflict and speciation is particularly desired given that speciation driven by sexual selection has been proposed and challenged (Ritchie 2007; Kraaijeveld et al. 2011; Gavrilets 2014; Servedio and Bürger 2014).

Gene expression involves interactions between genetic elements that lead to the production of specific gene products (e.g., transcripts and proteins). Thus, gene expression is expected to be susceptible to incompatibilities arising when two divergent genomes come together in a hybrid. Typically, genetic elements that influence the expression of a gene can be partitioned into cis- and trans-regulatory elements. Cis-regulation is exerted by sequences that are proximal to the gene being regulated and can commonly be thought as promoter or promoter-proximal elements. Trans-regulation is exerted by diffusible products (e.g., proteins) that are not necessarily linked to the gene being regulated. Genome-wide studies of gene expression have identified more divergence between species in cis-regulatory elements than trans-regulatory elements, and their potential contribution to hybrid incompatibilities during the establishment of isolating barriers (Wittkopp et al. 2008; Coolon et al. 2014; Osada et al. 2017; Signor and Nuzhdin 2018). Besides the preponderance of cis over trans divergence, the decomposition of cis versus trans effects has also shown a large proportion of compensatory changes that contribute to maintain similar levels of gene expression in different species (Landry et al. 2005; Coolon et al. 2014). Given that adaptive evolution of gene expression is common in species (Ellegren and Parsch 2007; Parsch and Ellegren 2013), a key question to understanding speciation is as follows: what adaptive changes accumulate within different populations/species and how can these changes induce post-mating reproductive barriers through hybrid incompatibilities?

One possibility is that directional selection in male-expressed genes towards different adaptations in incipient species contributes to reproductive isolation by hybrid divergent genomes becoming incompatible. This is supported by the fact that, in general, the rate of protein-coding gene evolution and gene expression divergence correlates between species of Drosophila (Nuzhdin et al. 2004; Lemos et al. 2005), but more importantly that the correlation is significant between the rate of nonsynonymous substitutions leading to protein content change, a proxy of positive selection, and male gene expression but not by changes in neutrally evolving synonymous changes (Ranz et al. 2003b; Artieri et al. 2007). However, one caveat is that these correlation studies have been conducted with species pairs of the D. melanogaster group that have attained complete or nearly complete isolation. An alternative is that stabilizing selection working to maintain an optimal level of gene expression in different lineages might cause misregulation of gene expression in hybrids and cause dysfunction. Different deleterious mutations can disrupt the level of expression of a gene, but such effect can be compensated by mutations that work to restore levels of expression back to their norm. If different compensatory changes occur in different lineages, the result is a lack of divergence in gene expression between divergent populations or species but misregulation of expression in hybrids. In Drosophila, some genome-wide comparisons of expression that have teased apart how cis- and trans-regulatory elements influence the expression of genes have more often found that their effects work in opposite directions as expected under compensatory mechanisms to keep levels of expression within an adaptive norm (Landry et al. 2005; McManus et al. 2010; Takahasi et al. 2011). This is supportive of the hypothesis that gene expression evolves under a house of cards model of stabilizing selection (Hodgins-Davis et al. 2015). An interesting couple of studies using divergent populations of copepods has proposed that misexpression in hybrids of genes linked to metabolic pathways might work to buffer physiological hybrid dysfunction and possibly ameliorate fitness loss (Barreto et al. 2015, 2018).

Finally, it is also possible that gene duplication events leading to the relaxation of selective pressures in newly evolved genes could contribute to divergence in gene expression and possibly gene function (e.g., neo-functionalization). Many transcription factors are known to arise by gene duplication (Voordeckers et al. 2015), suggesting gene regulatory divergence can be widely driven by duplication events. According to a comprehensive genome-wide scale study by Wyman et al. (2012), duplicate pairs contain more male-biased genes than female-biased genes in Drosophila. There are, in D. melanogaster, specific examples that illustrate ancestral genes with broad or specific patterns of expression but newly evolved duplicates being male biased (Yanicostas et al. 1995; Bai et al. 2007; Belote and Zhong 2009; Li et al. 2009; Sirot et al. 2014). OdsH is particularly interesting as it is a duplicate of unc-4 that appears to have evolved towards a new, but dispensable, function within species, but able to exert a major role in species differentiation and speciation (Ting et al. 2004). The gene expression is severely affected in sterile male hybrids (Sun et al. 2004).

It is common for reproductive isolation via hybrid dysfunction to occur mainly in the heterogametic sex, and it has been often linked to a major X-chromosome (or Z-) effect (Haldane 1922; Presgraves 2008). The role of sex chromosomes in hybrid dysfunction has been extensively covered elsewhere (Presgraves 2008; Coyne 2018). One interesting characteristic of the X chromosome is its fast rate of evolution, which likely facilitates its effect on hybrid male sterility (Llopart 2012; Llopart et al. 2018). This divergence is strongest for genes with male-biased expression (Baines et al. 2008; Meisel et al. 2009; Llopart et al. 2018), and recessive X alleles appear to have a significant contribution to misexpression in the hybrids and possibly sterility (Llopart et al. 2018). However, male-biased genes of flies are significantly underrepresented on the X chromosome (Parisi et al. 2003; Assis et al. 2012) and seminal fluid genes, as one major group of male-biased genes, are localized only in autosomes (Chapman 2001; Findlay et al. 2008). Nevertheless, the X chromosome contributes to autosomal male-biased gene regulation in this system (Meiklejohn et al. 2011; Kemkemer et al. 2014). This is likely because the faster X-gene expression divergence is not limited to faster cis-, but also involves rapid divergence of trans-diffusible regulatory elements, suggesting the capacity of the X chromosome to regulate both X-linked and autosomal gene expression (Coolon et al. 2015; Llopart et al. 2018). Hybrid male sterility clearly involves interactions between X-linked and autosomal factors, and specific targets have been identified (e.g., Nup genes; Tang and Presgraves 2009). We also have clear examples of male-expressed X-linked speciation genes, such as Ovd (Phadnis and Orr 2009; Phadnis 2011) and OdsH (Ting et al. 1998), that can cause hybrid dysfunction in Drosophila by modulating testes transcript production (Sun et al. 2004; Michalak and Ma 2008) or affecting the expression of specific autosomal testes gene targets (e.g., GA20504 and GA10921) (Alhazmi et al. 2019; Go et al. 2019) (Table 1). It is worth to note that unlike Drosophila, male-biased genes are enriched on the mammalian X chromosome (Wang et al. 2001). A comparison between humans and chimpanzees with respect to tissue-specific differences in expression levels and protein-coding sequences found X-linked genes expressed in testis showing significant expression changes as well as positive selection thus supporting a role for X chromosome male-biased genes in primates evolution (Khaitovich et al. 2005).

The contribution of the Y chromosome in speciation is less clear; nevertheless, there is suggestive evidence. For example, in contrast to other species of Drosophila, the X chromosome has only a small effect on hybrid male sterility between D. virilis and D. americana; however, the Y chromosome exerts a strong effect on hybrid male sterility (Sweigart 2010). An interesting characteristic of the Y chromosome is that it seems more likely to contribute to gene regulation via epigenetic mechanisms (Lemos et al. 2010), suggesting a possible environment-dependent role in speciation, such as heat-induced male sterility related with Y chromosome variation (Rohmer et al. 2004; David et al. 2005). While its role in male gene regulation has remained unclear, variation in the Y chromosome in D. melanogaster has been found to correlate with male-biased gene expression of autosomal and X-linked genes and contribute to divergence between D. melanogaster and D. simulans (Lemos et al. 2008).

Phenotypic plasticity, that is the ability of a genotype to produce different phenotypes depending on environmental conditions, is a central topic in evolution. Gene expression is sensitive to variation in physical environments (e.g., crowding, temperature) as well as genetic background (i.e., different genotypes respond in different ways to the same environment) (Pavey et al. 2010; Pfennig et al. 2010). It is therefore possible that differences in gene expression triggered by environmental changes might result in hybrid incompatibilities. Two lines of evidence offer some support for this hypothesis. First, models of speciation that incorporate genotype-by-environment interactions (GEI) have been shown to speed up the accumulation of BDM incompatibilities (Bordenstein and Drapeau 2001). Second, environmental changes have been shown to contribute to reproductive isolation in Drosophila by altering hybrid fitness. For example, the expressivity of hybrid inviability between D. melanogaster and its sibling species is influenced by temperature, and the temperature effect can also be modulated by the genotype of the parental strains (genome background) (Hutter and Ashburner 1987; Coyne et al. 1998; Miller and Matute 2017).

Plasticity in male ejaculate traits has been shown in Drosophila. For example, in D. melanogaster, the manipulation of larval density results in changes in size of the gland that produces seminal fluids as well as variable production and expenditure of SFPs (Bretman et al. 2016; Wigby et al. 2016). In studies where adult D. melanogaster male density was manipulated, complex transcriptomic responses and changes in expression of SFP genes were also identified (Fedorka et al. 2011; Mohorianu et al. 2017). We also know that social environment manipulations lead to plasticity of Drosophila sperm transfer (Garbaczewska et al. 2013). Previous studies have focused on genotypic response to environmental manipulation, but we lack knowledge on the role of genotypic background and environmental interactions in Drosophila. However, male gene expression in nematodes is sensitive to both environment and genotypic background and displays significant GEI affecting SFP expression (Patlar et al. 2019; Patlar and Ramm 2020). Considering that male-expressed genes, such as those producing SFP, often express condition dependent, one gap in our knowledge is the relationship between their variation in expression and plasticity in phenotypic output, particularly as it relates to reproductive isolation.

Ideally, whether the male-expressed genes contributing to reproductive isolation vary depending on environmental conditions and genetic background should be determined. In this context, male genes preventing hybridization between divergent populations via CSP can be important mediators of speciation (Civetta and Ranz 2019). For example, two genes known to function in sperm competition and CSP in Drosophila, i.e., Acp36DE and CG9997 (Table 1) (Castillo and Moyle 2014), might be plastic in their expression, and thus, their contribution to reproductive isolation among populations/species of Drosophila might be dependent on environmental conditions, genotypic differences, and genotype × environment interactions (GEI). Moving forward, tools such as the Drosophila Genetic Reference Panel of inbred lines offer an opportunity to simultaneously test the differential expression of candidates genes across genotypes and over different environments, as well as to phenotypically assay the effect of such changes on reproductive isolation phenotype (e.g., CSP). This approach should provide an opportunity to test GEI of male gene expression and associations with speciation phenotypes. In the long term, an avenue of inquiry pertains to testing the possibility that specific epigenetic changes might contribute to post-mating reproductive isolation. For example, some studies in plants have provided evidence to support the hypothesis that trans-generationally heritable epialleles can contribute to hybrid incompatibility (Durand et al. 2012; Lafon-Placette and Köhler 2015; Blevins et al. 2017).