No Evidence of Sexually Antagonistic Coevolution in Drosophila Reproductive Tract Transcriptomes

Abstract

Drosophila seminal fluid proteins (SFPs) are often cited as an example of interlocus sexual conflict, wherein the proteins increase male fitness while decreasing female fitness, spurring recurring female counter-adaptations and rapid molecular evolution. This model predicts that male-expressed genetic variation in the accessory gland, which produces seminal fluid, should generate counter-evolving genetic pathways in females, resulting in sexual coevolution. Using a trio of D. melanogaster populations exhibiting substantial SFP expression divergence due to recent selection, we test for coevolution in the female post-mating transcriptome in the lower reproductive tract and head. Contrasting predictions of sexual antagonism, female post-mating gene expression is indifferent to male population of origin. Instead, our results better support the alternative hypotheses that environmental variation is the source of selection on male SFP gene expression and that population differentiation in the female post-mating transcriptome is generated by female-expressed genotypic differentiation.

Introduction

A central goal of evolutionary genetics is to identify the selective forces responsible for patterns of adaptive molecular evolution. Drosophila seminal fluid proteins (SFPs, Wigby et al. 2020) are often presented as a leading case study for postcopulatory, intersexual conflicts driving rapid gene expression and protein sequence evolution. This model posits that after their transfer to the female reproductive tract (FRT), SFPs manipulate female post-mating physiology to advance male fitness, such as by delaying remating, inducing over-investment in egg-laying, or abbreviating lifespan (Chapman et al. 1995; Rice 1996; Holland and Rice 1998; Lung et al. 2002; Hollis et al. 2019), reviewed in Hopkins and Perry (2022). Females would counter-evolve resistance mechanisms, generating sustained selection on the interacting, male-transferred, and female-expressed genes and proteins. This dynamic is one proposed cause for the rapid SFP evolution observed across many taxa (Wyckoff et al. 2000; Swanson and Vacquier 2002; Swanson et al. 2003; Begun and Lindfors 2005).

Previous work connects SFP dosage to female post-mating physiology and gene expression, but how these genetic interactions may shape sexual coevolution is disputed (Hopkins and Perry 2022). In terms of physiology, experimentally generated hypomorphic and null SFP alleles have been used to reveal that certain SFPs have a variety of effects on post-mating outcomes (Wolfner 2002; Chapman et al. 2003; Chapman and Davies 2004), including effects on female gene expression (McGraw et al. 2008; Gioti et al. 2012). Experimental evolution work that eliminated female coevolution (Rice 1996) or sexual selection (Hollis et al. 2019) encouraged the interpretation that these SFP–female interactions could translate into a broader antagonistic evolutionary dynamic, based upon the concurrence in these experiments of male fitness improvements via sperm competition advantages, possible female fitness reductions via shortened lifespans and differences across reproductive transcriptomes in both sexes (Hollis et al. 2014; Hollis et al. 2016). Considering natural evolution among drosophilid flies, gene expression in the female reproductive tract is perturbed by mating to a heterospecific versus conspecific partner (Bono et al. 2011; Ahmed-Braimah et al. 2021), which may factor into reproductive isolation, and must begin as within-species population differentiation in male–female genetic interactions (Alipaz et al. 2001; Garlovsky et al. 2020; Lollar et al. 2023), such as SFP–female interactions (Wensing and Fricke 2018). Attempts to associate natural genotypic variation in male melanogaster with whole-female post-mating transcriptome variation have yielded mixed findings (McGraw et al. 2009; Delbare et al. 2017), leaving both the potential coevolutionary interactions and the identities of prospective female-expressed, counter-evolving loci still unclear. An approach using strategically selected male genotypes associated with both SFP dosage variation and adaptive evolutionary context, and focusing more sensitively on select female organ transcriptomes, could be informative.

Here, we explicitly test for coevolution of reproductive tract gene expression using natural D. melanogaster populations from the well-studied North American latitudinal cline (Adrion et al. 2015). In the male accessory gland, the primary source of seminal fluid, Panama males exhibit highly diverged gene expression from males from Maine, with a significant enrichment for differentially expressed SFP genes (Fig. 1a) (Cridland et al. 2023). Moreover, this SFP gene expression differentiation is starkly asymmetric, with Panama males exhibiting naturally hypomorphic expression of 48 SFPs relative to Maine males (∼16% of D. melanogaster SFPs). Cridland et al. used population branch length analysis by comparisons to the ancestral range Zambian population (Pool et al. 2012) and D. simulans to determine that much of this divergence results from evolution in the Panama population. Further contextualizing their observations, work over many decades has determined that phenotypic clines in D. melanogaster result from recent, spatially varying selection (Schmidt and Paaby 2008; Turner et al. 2008; Reinhardt et al. 2014; Adrion et al. 2015; Svetec et al. 2016). Together, these patterns of asymmetric, SFP-enriched, recent transcriptome evolution that are specific to the Panama population strongly suggest a role for selection in driving the accessory gland expression divergence. RNAi experiments on several genes connect reduced SFP expression to male fitness (Chapman 2001), consistent with the hypothesis that latitudinal SFP expression differentiation may be a target of selection, and mirroring the naturally hypomorphic expression of many SFPs in Panama. We hypothesized that under the sexual conflict model, recent, adaptive male gene expression differentiation would correspond to diverged male–female interactions. For example, Maine females mated to Panama males might enact a more Panama-like transcriptional response, or more generally an aberrant response, consistent with the idea that recent selection on accessory gland transcriptomes has changed the composition of seminal fluid and its influences on female transcriptomes. The disparate effects of conspecific versus heterospecific ejaculates on the female reproductive tract transcriptome in related species (Bono et al. 2011; Ahmed-Braimah et al. 2021) imply that such incompatibilities must arise as within-species population divergence. Here, we test this prediction using these recently diverged populations of D. melanogaster.

Fig. 1.

Reproductive transcriptome population differentiation between Panama, Maine, and Zambia. a to c) Male accessory gland transcriptomes; differentially expressed SFPs are colored according to the higher-expressing population: red for Zambia, yellow for Panama, blue for Maine, gray for non-SFP genes, with dotted lines showing the thresholds for differential expression. a) Maine versus Panama accessory gland. b) Zambia versus Maine accessory gland. c) Zambia versus Panama accessory gland. d to f) Lower reproductive tract transcriptomes in females mated to a same-population mate. Differentially expressed genes are colored according to the higher-expressing population: red for Zambia, yellow for Panama, blue for Maine, gray for non-differentially expressed genes. d) Maine versus Panama FRT. e) Zambia versus Maine FRT. f) Zambia versus Panama FRT.

To test for male effects on female transcriptome variation, we measured female post-mating gene expression following mating to either a co-evolved same-population male, or a diverged male (Fig. 2). We measured gene expression in the somatic female reproductive organs that directly interact with semen (i.e. the uterus, sperm storage organs, reproductive glands, and common oviduct). Gene expression in these organs is strongly mating responsive over a well-defined time course (Mack et al. 2006; McDonough-Goldstein et al. 2021a), providing a high-dimensional assay to quantify the effects of adaptive male transcriptome evolution on females and simultaneously identify interacting female genetic pathways.

Fig. 2.

Experimental strategy using crosses among 3 populations to distinguish interpopulation differentiation and male genotype effects.

Results

Following mating to a same-population male, female reproductive tract transcriptomes were modestly differentiated between Maine and Panama (21 differentially expressed genes, Fig. 1d). Regarding coevolutionary dynamics, the key result is that male geographic origin had little detectable effect on FRT transcriptomes. Among >8,100 FRT-expressed genes, only 3 genes were differentially expressed between females mated with foreign versus same-population males (Fig. 3a and b) which is no more than expected by chance given multiple hypothesis testing. Consistent with the conservative interpretation that they are false positives, these 3 genes exhibited modest log fold changes.

Fig. 3.

Transcriptomes from interpopulation test crosses. a to d) Mean log counts per million in females with a matched-population mate (x-axis) versus a diverged mate (y-axis). Data points for significantly differentially expressed genes are large red points with the gene name labeled. a) FRT in Maine females with a Maine versus a Panama mate. b) FRT in Panama females with a Panama versus Maine mate. c) FRT in Zambia females with a Zambian versus Panama mate. d) Head of Zambia females with a Zambian versus Panama mate.

To explore whether North American population differentiation in the accessory gland is too recent or subtle to reveal real coevolutionary interactions in the species, we conducted a second set of experiments incorporating the Zambian ancestral range population (Fig. 2). Zambia has the highest genetic diversity in the species (Pool et al. 2012) and is far more diverged relative to the N. American populations than they are to each other. In particular, Panama males show extreme accessory gland transcriptome divergence relative to Zambia males (Cridland et al. 2023): >2,300 genes are differentially expressed in the accessory gland between Panama and Zambia, including 104 SFP genes (i.e. 27% of all expressed genes and 38% of SFPs (Wigby et al. 2020) are differentially expressed, Fig. 1c), versus 58 SFPs among 1,176 differentially expressed genes between Maine and Zambia (Fig. 1b). Accordingly, we collected additional transcriptomes to determine whether Zambia females respond differentially to Panama males. From these new crosses, we also collected female head transcriptomes, because reception of certain SFPs induces various female behaviors (Chen et al. 1988; Isaac et al. 2010; Bath et al. 2017; Scheunemann et al. 2019) and because a prior study comparing whole-fly, mated-state female transcriptome divergence highlighted some female head-expressed genes (Delbare et al. 2017). We find that mating to a Panama rather than a Zambian male has no effect on Zambia female transcriptomes in the lower reproductive tract (Fig. 3c) or head (Fig. 3d). This negative result is not because female transcriptomes are geographically invariant: Zambia females differentially express 1,182 and 884 reproductive tract genes relative to Maine and Panama, respectively (Fig. 1e and f). Together, these findings indicate that female genotype is substantially more important to interpopulation differences in the mated-state FRT transcriptome than either male genotype or male–female genotypic interactions, which appear negligible.

Discussion

Using populations with pronounced divergence in male accessory gland and SFP gene expression, we find no evidence that this divergence affects the reproductive tract or head transcriptomes of their mates. Previous analyses supported the inference that male accessory gland transcriptome divergence between the 2 American populations was influenced by selection. Thus, our results suggest 2 possible interpretations: (i) that these female transcriptomes are indifferent to recent male accessory gland adaptations or (ii) that the extensive accessory gland and SFP variation among these 3 populations is not the result of selection. Notably, beyond the geographic SFP expression differentiation, there are 1,927 nonsynonymous SNPs in SFP genes segregating in North America, including at least 1 nonsynonymous SNP in 219 of the 273 SFP genes (Svetec et al. 2016; Cridland et al. 2023). Many of these SNPs are observed at different frequencies between the American populations, with allele frequency differences for some SFP nsSNPs exceeding 60%, indicating protein sequence variation may contribute to diverged seminal fluid composition between the populations. Given the differential male expression of many non-SFP genes (Fig. 1a to c), there might be interpopulation variation in the concentrations of amines, lipids, or sugars in seminal fluid as well (Baer et al. 2001; Poiani 2006; Kim et al. 2024). Our negative results indicate that none of these possible dimensions of ejaculate variation has any detectable effect on the female reproductive tract transcriptomes. Thus, across 2 sets of organs with hundreds of mating-responsive genes (Dalton et al. 2010; McDonough-Goldstein et al. 2021a)—including the 5 reproductive organs that directly contact sperm and semen, and the head tissues that mediate wide-ranging behavioral post-mating responses—females exhibit effectively identical post-mating transcriptomes regardless of substantial accessory gland divergence of their mates.

Of course, it remains possible that accessory gland adaptations influence unmeasured female phenotypes while leaving no signature on reproductive tract and head transcriptomes in our pairwise tests. For example, perhaps the FRT's proteomic post-mating response is more sensitive to coevolution (McDonough-Goldstein et al. 2021b). Even so, our results suggest it is unlikely that male variation regulates many granular aspects of female physiological variation via many individual components of semen. The fact that diverse aspects of female post-mating biology are altered by male knockdown of certain individual SFPs may indicate only that these specific SFPs—primarily Sex Peptide—are cues D. melanogaster females use to recognize that mating has occurred. Upon receiving these minimal cues, females hang their own regulatory scaffolding for enacting their post-mating physiology, including its transcriptomic components. Females may often use minimal cues to initiate reproductive state transitions. For example, in mice, mechanical stimulation is sufficient to induce pseudopregnancy (Stone and Srodulski 2023), which can similarly be described as a collection of many physiological and behavioral changes; this does not indicate that male mice use physical stimulation to regulate each discrete component of pseudopregnancy. Similarly in flies, it may be mischaracterizing the balance of control—and by extension, the evolutionary implications—to describe Sex Peptide and other SFPs as vehicles by which males regulate female immunity, behavior, gene expression, and metabolism.

Integrating our findings with the literature on SFP function and evolution, we deduce that many SFPs are likely under recent, locally variable selection to optimize functions unrelated to female gene expression. Latitudinal interpopulation accessory gland differentiation (Cridland et al. 2023) makes environmental adaptation an obvious possibility, particularly in response to latitudinally varying environmental factors that could influence sperm viability, storage, or competition. SFPs are components of fluids that are, effectively, cell culture media for spermatozoa, both in seminal fluid and in the female storage organs (Thayer et al. 2024); perhaps these “recipes” need optimization across different temperatures (Scossiroli 1954; van Heerwaarden and Sgrò 2021; Gandara and Drummond-Barbosa 2023). Variation in seasonality (Fan et al. 2025) and ecological community composition may also be relevant. Sperm competition among males is another prospective source of selection on SFP sequence and expression (Fiumera et al. 2005; Fiumera et al. 2007; Patlar and Civetta 2022). Some experiments associate sperm competition advantages with collateral damage to females (Rice 1996; Civetta and Clark 2000; Hollis et al. 2019), implying that sperm competition adaptations are inherently also sexually antagonistic and should still generate female counter-adaptations. Because we find no evidence of antagonistic coevolution in reproductive tract transcriptomes, our data are more easily compatible with a model wherein any possible sperm competition adaptations among these populations are not strongly deleterious to female fitness (Jiang et al. 2011; Castrezana et al. 2017), either because the laboratory-based findings do not reflect phenomena in natural populations or because female fitness and lifespan are not strongly correlated. While male × male × female interaction effects on sperm competition evolution have been modeled (Clark 2002), tradeoffs between sperm competition and environmental adaptation are relatively less explored.

Materials and Methods

Fly Samples

Flies were sampled from 30 isofemale lines that were established from inseminated females collected in Fairfield, Maine (September 2011) and Panama City, Panama (January 2012) (Zhao et al. 2015). We also sampled flies from 10 isofemale lines from Zambia (Pool et al. 2012). Three to 7-d-old, virgin females were introduced to 3 males in a vial and watched to confirm copulation. Males of mixed ages were held in vials in same-genotype groups of 5 for 5 d prior to the experiment, ensuring that none had recently mated. After mating, males were removed and females were rested for 4 to 6 h before dissection, with the same range of rest times evenly represented among replicates and treatments. Flies were housed on standard yeast–cornmeal–agar food at 25 °C on a 12 h light/dark cycle. We sampled 15 flies from different isofemale lines and pooled for each North American population (i.e. 15 flies per population sample, with 15 genotypes equally represented). For Zambia, we used 1 fly from each of 10 isofemale lines per sample. We used 3 replicates per treatment; however, because Zambia line 110 females would not readily mate with Panama males, the Zambia × Panama treatments had only 9 flies in each replicate.

RNA Isolation and Sequencing

The lower reproductive tract of mated females was dissected from the common oviduct to the posterior uterus, inclusive of the sperm storage organs, female accessory glands, and reproductive-associated fat body. RNA extraction and sequencing methods followed (Cridland et al. 2023). Paired-end, 150-base pair reads were generated on an Illumina NovaSeq 6000. One Zambia × Zambia head replicate failed RNA quality screening and was dropped from further analysis.

Analysis

Differential expression analysis followed (Cridland et al. 2023). Reads were aligned to genome v6.41 (downloaded 2021 August 9 from Flybase (Jenkins et al. 2022) with hisat2 v2.1 (Kim et al. 2015). Read counts were generated using featureCounts v2.0.3 (Liao et al. 2014) and differential expression was called using limma 3.50.3 (Ritchie et al. 2015) within R 4.1.2. To be considered expressed, genes required a median TPM ≥ 1 in at least one treatment per pairwise comparison. To be called differentially expressed, we required a logFC with an absolute value >1 and adjusted P value <0.05.

Funding

This work was supported by the National Institute of General Medical Sciences via R35GM134930, R35GM156525, and F32GM146419, and by the National Institute of Child Health and Human Development via K99HD115833.

Data Availability

Sequences are hosted at NCBI SRA under PRJNA1257015.