Experimental sexual selection reveals rapid evolutionary divergence in sex‐specific transcriptomes and their interactions following mating

Abstract Post copulatory interactions between the sexes in internally fertilizing species elicits both sexual conflict and sexual selection. Macroevolutionary and comparative studies have linked these processes to rapid transcriptomic evolution in sex‐specific tissues and substantial transcriptomic post mating responses in females, patterns of which are altered when mating between reproductively isolated species. Here, we tested multiple predictions arising from sexual selection and conflict theory about the evolution of sex‐specific and tissue‐specific gene expression and the post mating response at the microevolutionary level. Following over 150 generations of experimental evolution under either reduced (enforced monogamy) or elevated (polyandry) sexual selection in Drosophila pseudoobscura, we found a substantial effect of sexual selection treatment on transcriptomic divergence in virgin male and female reproductive tissues (testes, male accessory glands, the female reproductive tract and ovaries). Sexual selection treatment also had a dominant effect on the post mating response, particularly in the female reproductive tract – the main arena for sexual conflict – compared to ovaries. This effect was asymmetric with monandry females typically showing more post mating responses than polyandry females, with enriched gene functions varying across treatments. The evolutionary history of the male partner had a larger effect on the post mating response of monandry females, but females from both sexual selection treatments showed unique patterns of gene expression and gene function when mating with males from the alternate treatment. Our microevolutionary results mostly confirm comparative macroevolutionary predictions on the role of sexual selection on transcriptomic divergence and altered gene regulation arising from divergent coevolutionary trajectories between sexual selection treatments.

tissues and substantial transcriptomic post mating responses in females, patterns of which are altered when mating between reproductively isolated species. Here, we tested multiple predictions arising from sexual selection and conflict theory about the evolution of sex-specific and tissue-specific gene expression and the post mating response at the microevolutionary level. Following over 150 generations of experimental evolution under either reduced (enforced monogamy) or elevated (polyandry) sexual selection in Drosophila pseudoobscura, we found a substantial effect of sexual selection treatment on transcriptomic divergence in virgin male and female reproductive tissues (testes, male accessory glands, the female reproductive tract and ovaries). Sexual selection treatment also had a dominant effect on the post mating response, particularly in the female reproductive tract -the main arena for sexual conflictcompared to ovaries. This effect was asymmetric with monandry females typically showing more post mating responses than polyandry females, with enriched gene functions varying across treatments. The evolutionary history of the male partner had a larger effect on the post mating response of monandry females, but females from both sexual selection treatments showed unique patterns of gene expression and gene function when mating with males from the alternate treatment. Our microevolutionary results mostly confirm comparative macroevolutionary predictions on the role of sexual selection on transcriptomic divergence and altered gene regulation arising from divergent coevolutionary trajectories between sexual selection treatments.

| INTRODUC TI ON
Sexual reproduction involves both pre-and post mating interactions between the sexes, and sexual selection influences male and female traits that mediate the fitness outcome of these interactions. While aspects of reproduction can be cooperative, the sexes can diverge over the optima of reproductive traits, such as courtship signals, fertilization and offspring production (Arnqvist & Rowe, 2005). The intensity of sexual selection is linked to the extent to which reproductive fitness optima differ between the sexes and can generate sexual antagonism, in which selection acts in opposing directions on the sexes (Holland & Rice, 1999;Rice, 1996). Comparative genomic studies have found that genes showing rapid divergence and stronger signatures of positive divergent selection are often sex-biased or sex-limited in expression (e.g., Cheng & Kirkpatrick, 2016;Ellegren & Parsch, 2007;Pröschel et al., 2006;Zhang et al., 2007). This is especially true of species showing signs of strong sexual selection, such as increased sexual dimorphism (Harrison et al., 2015;Wright et al., 2019).
In internally fertilizing species, the main arena for post ejaculatory molecular interactions is the female reproductive tract (FRT), which includes sites of sperm transfer, storage and subsequent fertilization of eggs transiting from the ovaries. Post mating female responses are extensive, influencing female behaviour, morphology and physiology. These responses are mediated by interactions between components of the male ejaculate, including sperm produced in the testes and seminal fluid proteins (Sfps) produced by accessory glands, and female reproductive proteins in the FRT and ovaries (Oku et al., 2019;Sirot et al., 2015;Wolfner, 2009). Post mating changes in females include altering investment in oogenesis (Wolfner, 2009), remating propensity (Chapman et al., 2003), and sperm storage and usage (Avila et al., 2010(Avila et al., , 2015. Other aspects of female physiology are also altered, for example hunger (Carvalho et al., 2006), aggression (Bath et al., 2017) and physiological homeostasis (Cognigni et al., 2011;Ribeiro & Dickson, 2010). Genes associated with immunity and stress change gene expression upon mating in females, which may impact susceptibility or resistance to pathogens and/or parasites (Oku et al., 2019;Zhong et al., 2013).
An increasing number of studies have characterized transcriptomic post mating changes in females and the associated gene functions. Many studies typically examine either whole bodies or abdomens of females (Delbare et al., 2017;Fowler et al., 2019;Hollis et al., 2014Hollis et al., , 2016Innocenti et al., 2014;Innocenti & Morrow, 2009;Lawniczak & Begun, 2004;McGraw et al., 2008;Veltsos et al., 2017). Alternatively, some studies examine only one component of the FRT, either the "lower" reproductive tract, defined by the female sperm storage organs (Mack et al., 2006;Prokupek et al., 2008) or the "upper" reproductive tract, defined by the oviducts (Kapelnikov et al., 2008); but see (McDonough-Goldstein et al., 2021). Receipt of the male ejaculate affects sperm storage dynamics, oogenesis and oviposition (Sirot et al., 2015) making them all subject to sexual selection (and sexual conflict). Likewise, for males, the main focus on the role of sexual selection and sexual conflict has been on Sfp evolution given that they are among the most rapidly evolving proteins known (Ahmed-Braimah et al., 2017;Ellegren & Parsch, 2007).
However, sperm and the cellular architecture of the testes also can be subject to rapid morphological evolution and sexual selection (Lüpold et al., 2009).
The strong post mating sexual selection and sexual conflict associated with the reproductive interactions between the sexes also cause rapid evolutionary changes between lineages, which could influence reproductive isolation (Ahmed-Braimah et al., 2020;Manier et al., 2013;Markow, 1997). In particular, post mating prezygotic (PMPZ) reproductive isolation in which gametes do not interact properly prior to fertilization (Ahmed-Braimah et al., 2017;Garlovsky et al., 2020) or affect egg production (Matute & Coyne, 2010) is hypothesized to result from divergent coevolutionary trajectories of sexual selection and sexual conflict in isolated populations. If different populations experience different population coevolution over time, then there will be gene expression or proteomic mismatches in sexual interactions between independently evolved lineages when the male ejaculate interacts with noncoevolved female reproductive tissues (Ahmed-Braimah et al., 2020;Diaz et al., 2021;McCullough et al., 2020). However, such studies have focused on comparisons of the mating response of either heterospecific crosses or conspecific crosses where the sexual selection history of the populations are unknown.
The role of sexual selection in altering the coevolutionary dynamics between the sexes can be addressed experimentally.
Experimental sexual selection manipulates the opportunity and strength of sexual selection by subjecting isolated populations to either polyandrous conditions, which promotes strong sexual selection, or enforced monandrous conditions, which reduces it. This approach has been used to examine the evolution of gene expression in response to sexual selection, and link it to macroevolutionary patterns of sex-biased and sex-limited gene expression evolution (Hollis et al., 2014;Immonen et al., 2014;Veltsos et al., 2017). These previous studies supported the role of sexual selection in divergent sex-biased gene expression, but whether male-or female-biased genes responded the most varies between species, sexes, tissues and sexual experience, for unknown reasons (Hollis et al., 2014(Hollis et al., , 2019Immonen et al., 2014;Parker et al., 2019;Veltsos et al., 2017).
Additionally, these studies were limited in that post mating responses and/or sex-specific tissue responses were rarely examined.
Consequently, understanding how sexual selection impacts sexlimited and reproductive tissue-specific gene expression, the consequences of this divergence on post mating responses, and whether such divergence results in altered regulation of gene expression when mating between sexual selection treatments is limited.
In this study, we used replicate populations of D. pseudoobscura after 150 generations of experimental evolution in which either monandry is enforced (referred to as M), which reduces sexual selection and conflict, or the opportunity of polyandry is elevated (referred to as E), which may increase the strength of selection and conflict, to test the hypotheses of the role of sexual selection on tissue-specific and sex-specific gene expression. Our previous study found substantial phenotypic responses to the manipulation of the sexual selection environment (Crudgington et al., , 2009(Crudgington et al., , 2010Debelle et al., 2014Debelle et al., , 2016Snook et al., 2005), including those that could influence post mating responses such as investment in male accessory glands (Crudgington et al., 2009) and ovariole number and subsequent offspring production in females (Crudgington et al., 2010;Immonen et al., 2014). Sex-biased gene expression evolution has also occurred but our previous studies were based on either whole bodies of only one sex (Immonen et al., 2014) or heads and abdomens of each sex in the virgin and courted condition, but not following mating (Veltsos et al., 2017). These phenotypic responses may result from evolution of tissue-and sexspecific gene expression.
Here, a quantitative transcriptomic approach was made to investigate the impact of sexual selection on gene expression divergence, sampling male testes and accessory glands separately, and separating the FRT into the lower reproductive tract (including the uterus and sperm storage organs) and the ovaries. We first determined whether sexual selection treatment impacts virgin gene expression in all four tissues, testing the prediction that polyandry selects for upregulation. It has previously been suggested that males subjected to intense post copulatory sexual conflict should upregulate seminal fluid proteins for manipulation of female reproductive investment (Hollis et al., 2016). Likewise, polyandrous females should be poised for mating in anticipation of receipt of a manipulative male ejaculate that interacts within the FRT and thus should exhibit anticipatory upregulation of reproductive genes (Heifetz & Wolfner, 2004;Hollis et al., 2016;McGraw et al., 2004). We then determined the relative impact of sexual selection, mating per se, and their interaction for each female reproductive tissue. We tested the prediction that post mating interactions will diverge between sexual selection treatments, with poised polyandry females showing less upregulation upon mating relative to monandry females (see previous predictions). Such priming could be reflected in the types of genes that alter expression, including genes acting later during oogenesis (Immonen et al., 2014;Veltsos et al., 2017) and immune and stress response genes, all of which may show reduced changes in expression following mating in females already poised for mating.
Immune and stress response genes have been commonly identified in female post mating transcriptomic responses, are assumed to indicate costs of mating, receipt of a foreign ejaculate and sexual conflict (Innocenti & Morrow, 2009;Zhong et al., 2013). Mating involves interactions between the sexes so we asked whether the post mating expression response of females arises from an interaction between the sexes or is primarily driven by one sex. Given that sexual selection is stronger on males (Winkler et al., 2021), we tested the prediction that polyandrous males will induce a larger female post mating transcriptomic response, especially with monandry females.
Related, we tested the prediction that divergence of coevolutionary trajectories between monandry and polyandry populations will generate unique or more pronounced responses in crosses between these populations, which could form the basis of post mating prezygotic reproductive isolation.

| Experimental evolution lines
The origin, establishment, and maintenance of the selection lines are described in detail elsewhere . Briefly, 50 wild-caught females of D. pseudoobscura from a population in Tucson, Arizona, USA were brought into the laboratory and reared for three generations, then four replicate lines of two different sexual selection treatments were established. We modified the opportunity for sexual selection by manipulating the adult sex ratio in food vials (2.5 × 80 mm) by either confining one female with a single male (enforced monogamy treatment; M, monadry) or one female with six males (elevated polyandry treatment; E, polyandry). This species is naturally polyandrous with wild-caught females frequently being inseminated by at least two males at any given time (Anderson, 1974). We successfully equalized effective population sizes between the treatments . At each generation, offspring were collected and pooled together within each replicate line for each treatment, and a sample from this pool was used to start the next nonoverlapping generation in the appropriate sex ratios. Thus, this proportionally reflected the differential offspring production across families within a replicate and treatment. Generation time was 28 days and all populations were kept at 22°C on a 12L;12D cycle, with standard food media and added live yeast. Note that "monandry" versus "polyandry" as used here refers to the evolutionary history under which the individuals have evolved, not their current reproductive status.

| Sample preparation
To generate experimental males and females, parents were collected from each replicate at generation 157-158. We standardized for maternal and larval environments as previously described (Crudgington et al., 2010). Briefly, parents were mated en masse in food bottles, transferred to containers with oviposition plates, allowed to oviposit for 24 h, and then 48 h later, 100 first instar larvae were seeded in standard food vials (Crudgington et al., 2010). Virgin males and females were collected under light CO 2 anaesthesia on the day of eclosion and kept separate in vials of 10 individuals for 5 days to ensure reproductive maturity (Snook, 2001). On Day 5, within a 2 h window after lights turned on, one virgin female was placed in a food vial with one virgin male that was from either the same experimental replicate ("coevolved"; MM, EE where the first letter is the female) or the other treatment ("noncoevolved"; ME, EM). Our previous studies analysed gene expression in either whole body females or heads and abdomens of males and females. A potential criticism of this approach is that observed responses potentially confound changes in gene expression with allometric changes in relevant tissues (Montgomery & Mank, 2016). Most importantly, we know that male and female reproductive tissues are key to evolutionary responses to sexual selection and involved in sexual interactions. Therefore, here we carried out analyses of dissected male testes, accessory glands, and female ovaries and reproductive tracts (see Supporting Information S1 which illustrates the experimental design).
We dissected age-and circadian rhythm-matched virgin males and females from the same collections. Each treatment was represented by 100 individuals, the tissues of which were equally split into 4 separate tubes, for easy pooling. Each pool contained the dissected tissues of the four biological replicates of the E or M treatments. For the mating treatments, males were put first in individual vials with fly food and allowed to settle. Females were then added, and were dissected 6 h after the first couple mated, in the order of mating, within a 2 h block.
Dissections were performed under ether anaesthesia in RNAlater (Ambion) on ice blocks. We separately collected the ovaries and the remainder of the FRT, including the sperm storage organs (seminal receptacle and spermathecae). We refer to these different female tissue sets as ovaries and the FRT. The male accessory glands and testes were also dissected separately (ejaculatory bulbs were not included).
All tissues were left at 4°C in RNAlater (Ambion) for one day and then transferred to -80°C until RNA extraction. The pools were processed for RNA extraction using Trizol (Ambion) following the manufacturer's instructions. RNA extractions were cleaned up in Qiagen RNeasy kit columns according to the manufacturer's protocol, including the 15 min DNase treatment. The quality of RNA extractions was checked with Nanodrop and Bioanalyser.

| Sequencing and mapping
Illumina libraries were prepared using the ScriptSeq kit (Illumina Inc) following the manufacturer's protocol. rRNA was depleted using Epicenters' Ribozero kit. Paired-end second stranded libraries were sequenced at 100 base pair (bp) read length using an Illumina HiSeq 2000. Reads were mapped to the D. pseudoobscura genome v3.1, and indexed using bowtie2 (Langmead et al., 2009). Paired-end reads were aligned using option "-g 1 -library-type fr-secondstrand" with TopHat2.0.8b (which calls bowtie2.1.0; Kim et al., 2013) and instructs TopHat2 to report the best alignment to the reference. Exon features were counted using HTSeq-count (Anders et al., 2015) and the reads of all exons of each gene were combined to provide overall measures of gene expression (Veltsos, 2021).

| Statistical analysis
We analysed the count data using edgeR v3.18.1 (Robinson et al., 2010) running in R v.3.4.0 (R Development Core Team, 2007) with scripts from (Veltsos, 2021). The tissues were always analysed separately but, for statistical analysis of specific contrasts, all libraries for each tissue were used, including libraries not contributing to that contrast. This was to allow incorporating as much information as possible on gene expression variance, counteracting the fact that data points available for each gene in a specific contrast are limited. Libraries were normalised with the TMM procedure in edgeR. Because the analysis was performed for each tissue separately, not all annotated genes had counts in each analysis. Additionally, normalisation can result in negative counts for some genes. Therefore, for analysis we considered only genes with average 0 normalized counts per million across all libraries used in each analysis were retained. The number of genes retained for analysis were therefore 11,751 genes for testes, 11,754 for accessory glands, 10,272 for female reproductive tracts and 8624 genes for ovaries out of 16,467 annotated genes for D. pseudoobscura. Dispersion was measured with default parameters using a negative binomial model. We considered genes to be differentially expressed (DE) if they were below the 5% false discovery rate (FDR) threshold (Benjamini & Hochberg, 1995). We did not employ a log 2 FC threshold because allometry is unlikely to influence results obtained from specific tissues (Montgomery & Mank, 2016). Differences in gene expression magnitude of genes within a contrast were texted using the wilcox.test function in R.
We performed gene ontology (GO) enrichment analysis using topGO v2.22.0 with the weight01 algorithm to account for GO topology (Alexa & Rahnenfuhrer, 2010). The GO universe was defined from the genes that showed appreciable expression in each tissue.
Results with p < .05 on Fisher's exact tests, corrected for topology, were retained (Supporting Information S3).
For analysis of Sfps, we contrasted the distribution of the change in expression (log 2 FC) of all genes, and DE genes in the contrast between sexual selection treatments for virgin accessory gland transcriptomes using density plots and boxplots, respectively. We compared the distributions of three Sfp-related gene subsets, identified from D. pseudoobscura proteomics (Karr et al., 2019). The largest subset was 3,281 proteins produced in the accessory gland ("proteome"). Of these, 528 had protein secretory signals ("secretome") and 163 were also orthologous to D. melanogaster seminal fluid proteins (putative Sfps or "exoproteome"). The majority of these genes could be crossreferenced to our data (Figure 2 legend), but only proteome genes were detected among the DE accessory gland genes.

| Sexual selection causes gene expression divergence in virgin male reproductive tissues
Previous work has hypothesized that monandry selects for downregulation and polyandry selects for upregulation of genes expressed in male reproductive tissues in the abdomen of D. melanogaster (Hollis F I G U R E 1 Differential gene expression (absolute log 2 FC changes, y-axis) in virgin (a) male or (b) female tissues comparing responses of upregulated genes in either polyandry (E: light grey) or monandry (M: dark grey) selection treatments. Dots indicate differentially expressed (DE) genes and their number is noted above each box Examination of the 80 accessory gland genes that were differentially expressed between treatments, regardless of set, showed we found only a nonsignificant trend towards higher expression in polyandry males (Figure 2b).
In conclusion, virgin male reproductive tissue expression was affected by experimental manipulation of the strength of sexual selection but not in the predicted direction. Both the total number and proportion of DE genes was higher in testes than accessory glands. There was also a weak trend for monandry, not polyandry, males to upregulate testes genes and accessory gland genes with secretory signals.

| Sexual selection causes gene expression divergence in virgin female reproductive tissues
Similar to males, we tested whether monandry selects for downregulation and polyandry selects for upregulation of genes (Hollis et al., 2016)

| Sexual selection causes divergence in the female post mating response
With regard to the female post mating response, we expected polyandry females to already express post mating response genes as virgins, whereas monandry females would increase the expression of these genes after mating (the poised hypothesis; Heifetz & Wolfner, 2004;McGraw et al., 2004). This response would be seen as a significant selection treatment by mating status interaction in our model (Table 1).
For the FRT, we found both the main effects of sexual selection treatment and mating, and their interaction, to be significant (Table 1). To illustrate the interaction, the main effect of mating within each female treatment was plotted on separate axes, and the genes that always respond to mating and those that show significant interaction effects between mating and sexual selection treatment are indicated separately (Figure 3a). Most (143/146) of the significant genes for the interaction were upregulated upon mating in monandry females (and downregulated upon mating in polyandry females; Table 1; Figure 3a

| Female, not male, sexual selection treatment drives the female post mating gene expression response
The post mating gene expression response in female reproductive tissues represents an interaction between the sexes. To examine these interactions in more detail, we partitioned gene expression in TA B L E 1 Outcome of contrasts testing the effect of sexual selection, mating status, and their interaction on the female post mating response in the female reproductive tract (FRT) and ovaries  Table 2). Since the strength of sexual selection is stronger in males than females (Winkler et al., 2021), we predicted that both the male effect and the interaction would have the strongest impact on the female post mating response ( Table 2).
Contrary to predictions, in both female tissues, we found no significant interaction effect and little to no male effect ( Table 2).
There was a small male effect in the FRT with seven DE genes, en- tab "Main female effect OV").
In summary, female sexual selection history plays a critical role in determining the extent and function of differences in the post mating response, with the effect of male and interaction between the sexes minimal. Again, mated polyandry females had relatively suppressed gene expression changes relative to mated monandry females in the FRT. However, in the ovaries, mated polyandry females had relatively more highly expressed genes although these were related to egg production, perhaps suggesting that polyandry females can "gear up" for oogenesis more quickly than monandry females. One caveat to this analysis is that it combines females from the two sexual selection treatments to test for male and female effects. Given that we showed above that sexual selection treatment influences the female post mating response, consequences of interactions between males and females may be obscured.

| Sexual selection asymmetrically alters post mating gene expression
To further decompose the effect of sexual selection treatment origin on the female post mating response, we examined gene expression in coevolved and noncoevolved matings separately for females of different sexual selection history. This also allowed us to test the prediction that divergent coevolutionary trajectories between sexual selection treatments would generate unique or more pronounced responses when mating with a noncoevolved male. We predict this effect will be asymmetric as monogamous females have TA B L E 2 Outcome of contrasts testing the effect of female sexual selection treatment, male sexual selection treatment, and their interaction on the female post mating response in the female reproductive tract (FRT) and ovaries   Table 3) whereas in polyandry females there was no significant differential post mating response based on sexual selection treatment of the male (Table 3). We tested the prediction that polyandry selects for upregulation in virgin male and female reproductive tissues (Heifetz & Wolfner, 2004;Hollis et al., 2016;McGraw et al., 2004). Given the role of accessory glands in manipulating female reproductive investment, and that the FRT is the main site of molecular interactions between the sexes, we expected that these tissues would show more divergent gene regulation than testes or ovaries. Support for these predictions varied. In males, sexual selection resulted in divergent gene regulation in both testes and accessory glands. Contrary to predictions, a larger proportion of genes change expression in the testes and relaxation of sexual selection resulted in more upregulated genes in the testes and higher expression of accessory gland genes with a signal of a secretory function. However, the function of genes that changed expression in the testes and accessory glands were distinct for each sexual selection treatment, suggesting differences in the potential consequences on male reproductive fitness and effects on female mates. In females, sexual selection resulted in divergent gene regulation in both the FRT and ovaries, but a larger proportion of ovary genes were differentially expressed compared to the FRT. These results are contrary to predictions. However, the magnitude of differential gene expression is larger in the FRT and for polyandry females, which supports the prediction.
We previously showed that virgin polyandry D. pseudoobscura females have more ovarioles than monandry females (Immonen et al., 2014).
Using female whole body microarrays, we found upregulated genes in polyandry females to be associated with oogenesis (and likely to be expressed in the ovary) whereas upregulated genes in monandry were associated with genes in somatic tissues and metabolism (Immonen et al., 2014). In the current analysis, virgin polyandry females had upregulated genes with functions in immune responses and later stages of egg production whereas virgin monandry females upregulated genes associated with BMP signalling pathway which is involved in early patterning the Drosophila eggshell (Niepielko et al., 2012). These results support the hypothesis that polyandry females are poised for receipt of a potentially manipulative ejaculate (Heifetz & Wolfner, 2004;McGraw et al., 2004).
We also examined the female post mating response testing the effect of sexual selection, mating and their interaction on gene expression divergence. In particular, we tested the prediction that polyandry females would show a smaller post mating response relative to monandry females given polyandry females are poised for mating. We additionally asked whether post mating responses are specific to sexual selection treatment with respect to biological processes. As predicted, we found an interaction between sexual selection treatment gene expression and mating status but this was tissue specific, occurring only in the FRT, and asymmetric across sexual selection treatments, with most DE genes upregulated in mated monandry females. Moreover, we saw a contrasting pattern in that genes which are upregulated upon mating in monandry females had lower expression in mated polyandry females, supporting the hypothesis that polyandry selects for females to be poised for mating, perhaps to combat a manipulative ejaculate.
While gene expression in ovaries showed no interaction effect, monandry females also upregulated more genes after mating than polyandry females, with gene functions supporting more advanced reproductive development in polyandry females (also supported in Immonen et al., 2014;Veltsos et al., 2017). Overall, these results support the hypothesis that polyandry females are poised for mating. In the FRT, DE genes associated with the interaction effect and for mating status were enriched for immune function, a commonly observed effect of mating in Drosophila (Hollis et al., 2019;Innocenti & Morrow, 2009;Sirot et al., 2015). It has previously been suggested that upregulation of these genes arises from sexually antagonistic interactions between the sexes, such that males are immunogenic to females (Innocenti & Morrow, 2009;Zhong et al., 2013). However, there remains insufficient data to determine whether these effects are detrimental or beneficial to females overall (Bagchi et al., 2021;Oku et al., 2019).
Given that sexual selection changes the pattern of gene expression in mated females in both female reproductive tissues, that the female post mating response arises as an interaction between the sexes, and that sexual selection is stronger on males (Winkler et al., 2021),  -Braimah et al., 2020;Bono et al., 2011;McCullough et al., 2020). However, it is difficult to infer the historical role of different evolutionary processes from patterns of contemporary divergence between species and therefore whether mismatches generated PMPZ or evolved after divergence cannot be determined. Experimental evolution can address this problem as it can distinguish between current and historical processes, but lacks the full complexity of natural conditions and typically does not result in reproductive isolation over the relatively short time frame studied.
Using experimental sexual selection, we found patterns that support these predictions in the FRT, but not in ovaries. Gene regulation in the FRT varied between monandry and polyandry females and depended on the type of male involved. We inferred altered regulation by the identification of uniquely DE genes when mated with a non- In conclusion, our results tested several predictions arising from sexual selection and sexual conflict theory, and highlight substantial gene expression divergence both in the long-term following 150 generations of altered sexual selection intensity and short-term plastic responses when mating. We found sex-and tissue-specific effects of sexual selection on gene expression and gene function, alterations in gene expression and gene function specific to origin of the female and male partners, and predicted asymmetric altered gene regulation and function arising from divergent coevolutionary trajectories between sexual selection treatments. Changes in gene expression identified here and in sex-biased gene expression in response to sexual selection (Veltsos et al., 2017) have recently been shown to be associated with genomic divergence in these lines (Wiberg et al., 2021). Overall, our results complement studies in natural populations in which sexual selection has been implicated in gene expression and genomic divergence.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The RNAseq data underlying this article are available in the ArrayExpress repository under accession number E-MTAB-10047