The thermal environment of sperm affects offspring success: a test of the anticipatory paternal effects hypothesis in the blue mussel

Abstract

There has been an explosion of recent evidence that environments experienced by fathers or their ejaculates can influence offspring phenotypes (paternal effects). However, little is known about whether such effects are adaptive, which would have far-reaching implications for the many species facing rapidly changing environments. For example, some arguments suggest paternal effects might be a source of cross-generational plasticity, preparing offspring to face similar conditions to their father (anticipatory hypothesis). Alternatively, ejaculate-mediated effects on offspring may be non-adaptive by-products of stress. Here, we conduct an experiment to distinguish between these predictions, exposing ejaculates of the externally fertilizing mussel Mytilus galloprovincialis to ambient (19°C) and high (24°C) temperatures, then rearing offspring groups in temperatures that match and mismatch those of sperm. We find that, overall, high temperature-treated sperm induced higher rates of normal offspring development and higher success in transitioning to second-stage larvae, which may represent adaptive epigenetic changes or selection on sperm haplotypes. However, the progeny of high temperature-treated sperm did not perform better than those of ambient temperature-treated sperm when rearing temperatures were high. Overall, these findings offer little support for the anticipatory hypothesis and suggest instead that beneficial paternal effects may be eroded when offspring develop under stressful conditions.

1. Introduction

Paternal effects, the influence that paternal environments have on offspring phenotype, were long considered negligible (with the exception of paternal care in some species), owing to the traditional view that the only heritable information in sperm is the DNA sequence [1]. However, a considerable body of evidence has recently shown that components of the ejaculate beyond the gene sequence can respond to the paternal environment and transmit information to offspring; e.g. through epigenetic mechanisms such as sperm DNA methylation patterns or RNA profiles [2,3]. These factors can be transmitted when sperm release DNA and RNA into fertilized eggs, affecting patterns of gene expression in developing offspring (e.g. [4,5]) However, whether such effects are adaptive for offspring remains, for the most part, an open question. It has been argued, for example, that paternal effects might be anticipatory, transferring information about the paternal environment to offspring and preparing them for conditions that match the father's (for empirical support see [4,6,7]). Alternatively, ejaculate-mediated paternal effects could be by-products of physiological stress responses by males or their sperm, i.e. a source of non-adaptive or maladaptive noise in offspring development (see review in [2]).

The vast majority of studies on ejaculate-mediated paternal effects have focused on environmental factors experienced by fathers before mating (reviewed in [3]). However, sperm themselves also experience fluctuating biotic and abiotic conditions after ejaculation that can cause phenotypic and molecular alterations [8]. These conditions can be especially variable for species that release gametes externally, meaning sperm experience a period without homeostatic protection from either male or female reproductive tracts. Moreover, human-induced changes are amplifying the environmental stress experienced by externally shed gametes. In particular, rising mean temperatures and increased frequency of extreme events such as heatwaves can have multiple deleterious effects on the fertilization capacity of broadcast spawning marine species [9]. Recent evidence from external fertilizers also indicates that heat stress can induce molecular responses in sperm that include changes to cellular RNA profiles [10], and that the temperature experienced by sperm can influence offspring phenotype [11]. However, it has yet to be tested how offspring perform in temperatures that match and mismatch those experienced by sperm. Therefore, it remains unclear whether changes to sperm could represent a source of adaptive cross-generational plasticity in the face of climate change (e.g. [7]), or whether they will amplify the negative effects of environmental stress on offspring.

Here, we test the performance of offspring after sperm were exposed to different post-ejaculation temperatures, by rearing the offspring in temperatures that match and mismatch those experienced by their father's sperm. We use the broadcast spawning mussel, Mytilus galloprovincialis, which is an ideal and highly relevant model for testing paternal effects of post-ejaculation temperature on offspring. Populations of M. galloprovincialis across the southern Australian coastline are experiencing both high rates of average warming and an increasing frequency of marine heatwave events, where temperatures can reach 6°C above average for extended periods [12,13]. Sperm of this species are able to maintain normal motility under such temperatures [10,14]; however, we have recently reported that high post-ejaculation temperature causes changes to sperm transcript profiles, including heat shock protein (hsp) genes [10]. Intriguingly, changes in the expression of sperm hsp and other genes have also been reported during other environmental changes in M. galloprovincialis, such as hyposalinity, suggesting that control over sperm gene expression may be an important component of adaptive gamete plasticity [15]. Moreover, variation in the environment experienced by sperm can affect offspring survival in M. galloprovincialis [16]. In this study, we exploit this system to provide a novel test of whether temperature-induced changes to sperm can adaptively prepare offspring to face similarly altered environments, i.e. whether offspring perform better when their environment matches that experienced by sperm.

2. Material and methods

(a) . Spawning

Mussels were collected from Woodman Point, Western Australia (32°14′ 03.6″ S, 115°76′ 25″ E) during June–September 2019 and used on the days of collection. The sea surface temperatures at Woodman Point during this period ranged between 17 and 19°C; these are typical of recent winters in this location (Giovanni online data system, available via NASA Goddard Earth Sciences Data and Information Services Center (see https://giovanni.gsfc.nasa.gov/giovanni/); see also [14]). Spawning was induced via standard procedures, by placing mussels in a heated water bath (28°C) [10,17,18]. Heat experienced by whole mussels before spawning could possibly affect gametes; however, all mussels experienced this prior to split-ejaculate application of experimental treatments (see below). Therefore, spawning temperature did not confound gamete treatments. Furthermore, spawning mussels were taken immediately from the water bath, washed with filtered seawater (FSW; see electronic supplementary material) to remove gametes that were already released and placed in jars of FSW at ambient sea surface temperature (19°C). Thus, experimental gametes were not directly exposed to elevated temperatures before applying experimental treatments. Gamete concentrations were estimated and adjusted using standard procedures (electronic supplementary material).

(b) . Sperm treatments

We conducted experimental trials across a number of logistically feasible ‘blocks’. In each block, 4–5 individual males were used for assays. Each male's ejaculate was adjusted to 5.0 × 10⁶ cells µl⁻¹ then split into two aliquots of 300 µl (a schematic of the experimental design is provided in electronic supplementary material, figure S1). These aliquots were exposed to different temperature treatments for 15 min, using water baths in a temperature-controlled room: ambient, 19°C, and high, 24°C (the latter simulating a heatwave event; [12]). We have previously found that similar temperature regimes alter patterns of sperm gene expression in M. galloprovincialis, without affecting sperm swimming behaviour [10]. After 15 min, both samples were transferred to 19°C for 5 min to ensure eggs would not experience high temperatures during fertilization trials (electronic supplementary material, figure S1).

(c) . Fertilization trials

Following temperature treatments and return to ambient temperature, each sperm aliquot within a block was used to fertilize separate 6 ml samples of pooled eggs (5.0 × 10⁴ cells µl⁻¹) prepared from 4 to 5 females. Pooling of eggs in this way enabled us to control for male × female interaction effects on fertilization and offspring viability [19]. Sperm–egg mixes were left for 2 h, then 900 µl subsamples were removed and preserved in 1% buffered formalin. These subsamples were later assessed for fertilization rates by counting a haphazard sample of 100 eggs and scoring the proportion undergoing polar body formation and/or cell division. There was no difference in fertilization rate between the sperm temperature treatments (beta-binomial generalized linear mixed-effects model (GLMM); ${\chi }_1^2=0.191$, p = 0.663).

(d) . Offspring treatments and phenotypes

At 2 h post-fertilization, we removed two separate 500 µl aliquots from each sample to rear for offspring viability assays and diluted them in large vials with FSW to a final volume of 5 ml. These two aliquots were reared in different temperature treatments using water baths in a temperature-controlled room, reflecting the ambient and high sperm treatments (electronic supplementary material, figure S1). Therefore, we had four sperm treatment × rearing treatment combinations: ambient–ambient, ambient–high, high–ambient and high–high. Under the anticipatory hypothesis, offspring reared in high temperatures would be expected to perform better if sperm had experienced high temperatures; similarly, offspring reared in ambient temperatures would be expected to perform better if sperm treatments were ambient. At 48 h post-fertilization, each offspring vial was removed from its treatment, homogenized by swirling and 20 µl aliquots taken for fitness assays. In these aliquots, we counted the total number of surviving offspring, the proportion that were second-stage (veliger) versus first-stage (trocophore) larvae (as per [16]) and the proportion of normal versus abnormal offspring (based on morphological symmetry and swimming behaviour, following [20]).

(e) . Statistical analyses

We analysed offspring measures using GLMMs in R v. 4.0.3 [21] using the packages ‘lme4’ [22] and ‘glmmTMB’ [23]. Total surviving offspring were analysed as count data using a negative binomial GLMM to account for overdispersion (electronic supplementary material). We fit random effects for experimental block and male, and fixed effects of sperm treatment, rearing treatment and their interaction, along with a covariate for fertilization rate. This covariate accounts for variation in offspring numbers among samples that could be owing to variation in initial fertilization rate (as per [16]). Rates of veliger versus trocophore larvae and normal versus abnormal larvae were analysed as proportion data, the former with a beta-binomial GLMM (to account for overdispersion; electronic supplementary material) and the latter with a binomial GLMM. The same predictors were used as for the offspring survival model, except without the fertilization rate covariate, because variation in fertilization rate was already accounted for in the proportion denominators (number of surviving offspring), i.e. the weights of the binomial trials. We tested the significance of fixed effects with Type III Wald χ² tests. Coefficient estimates controlling for covariates and random effects (i.e. marginal means) were visualized using the ‘effects’ package [24].

3. Results

Our analysis revealed a positive relationship between fertilization rate and offspring survival (table 1a). When controlling for variation in fertilization rate (i.e. the covariate), we found a significant effect of rearing temperature, but not sperm temperature or their interaction, on offspring survival (table 1a); fewer offspring survived in the high treatment than in the ambient treatment (figure 1a). However, for the surviving offspring, we found significant main effects of both sperm temperature and rearing temperature on both the proportion of veligers and the proportion of normally developed larvae. Specifically, we found that significantly more offspring had transitioned to the veliger stage and were normally developed, from high temperature-treated sperm (table 1b,c and figure 1b,c). There were also higher proportions of both veliger and normally developed larvae in ambient rearing temperatures (table 1b,c and figure 1b,c).

Table 1.

Wald χ² tests of the effects of sperm temperature, rearing temperature and their interaction (ST × RT) on (a) offspring survival, (b) proportion of veliger larvae and (c) proportion of normally developing larvae, fitted using GLMMs. The offspring survival model also included a covariate of fertilization rate. *Indicates p-value < 0.05.

effect	χ2	d.f.	p-value
(a) offspring survival
sperm temperature	0.873	1	0.350
rearing temperature	4.153	1	0.042*
ST × RT	0.010	1	0.921
fertilization rate (covariate)	12.471	1	<0.001*
(b) proportion veligers
sperm temperature	4.938	1	0.026*
rearing temperature	7.667	1	0.006*
ST × RT	3.707	1	0.054
(c) proportion normal
sperm temperature	4.246	1	0.039*
rearing temperature	9.205	1	0.002*
ST × RT	3.774	1	0.052

Figure 1.

Effects of sperm temperature and rearing temperature on (a) total surviving offspring, (b) proportion of veliger larvae and (c) proportion of normally developed offspring. Means (±s.e.) in each treatment combination calculated from generalized linear mixed-effects models, controlling for covariates and random effects, using the ‘effects’ package [24]. Temperature treatments: A = ambient, 19°C; H = high, 24°C.

The main effects of sperm temperature on offspring development appeared to be driven by changes in the ambient rearing group. There were marginally non-significant interactions between sperm temperature and rearing temperature on both the proportion of veligers and the proportion of normal offspring (table 1b,c). The higher proportions of veliger larvae and normal offspring from high temperature-treated sperm were apparent for offspring reared at ambient but not high temperatures (table 1b,c and figure 1b,c; electronic supplementary material, figures S2 and S3), while the marginal means for offspring reared at high temperatures showed the opposite trends (figure 1b,c; electronic supplementary material, figures S2, S3). Indeed, when the sperm treatment effect was tested using separate models for ambient temperature-reared and high temperature-reared groups of offspring, sperm treatments were significant in the ambient temperature-reared group but not the high temperature-reared group for both the proportion of veliger larvae (ambient group: ${\chi }_1^{\hspace{0.17em}2}=5.568$, p = 0.018; high group: ${\chi }_1^{\hspace{0.17em}2}=0.258$, p = 0.612) and proportion of normally developed larvae (ambient group: ${\chi }_1^{\hspace{0.17em}2}=4.224$, p = 0.040; high group: ${\chi }_1^{\hspace{0.17em}2}=0.679$, p = 0.410).

4. Discussion

We show that the exposure of sperm to different temperatures alters the early development of larvae, affecting the rate of transition to second-stage veliger larvae and the proportion of normally developed offspring. The overall effect of treating sperm with high temperatures was to increase the proportion of both veliger larvae and normal development. Both the success in reaching the veliger stage and morphological normality at 48 h are important predictors of later survival in mussels [25–27]. These results lend credence to recent suggestions that the apparent phenotypic stability of external gametes and fertilization under high temperatures [9] could mask molecular changes that have consequences for resulting offspring [10]. Further, although the overall sperm treatment effect appears beneficial for offspring, differences in marginal means within the ambient rearing temperature appeared to primarily drive the overall sperm effects, with no apparent increases in the proportion of veligers or normal larvae across sperm treatments within the high rearing temperature. Indeed, in marked contrast with the pattern expected under the anticipatory paternal effects hypothesis, there was a trend towards lower proportions of veligers and normal larvae from high temperature-treated sperm in high rearing temperatures. Our findings therefore suggest that sperm-mediated cross-generational plasticity is unlikely to buffer direct detrimental effects of heat stress on offspring.

Several potential mechanisms could drive the overall sperm treatment effects. One possibility is that sperm experiencing high temperatures undergo epigenetic changes, e.g. to DNA methylation or small RNA expression, which affect gene expression in early embryos [2,28]. Such changes might benefit offspring by increasing development rate during periods of stress, so offspring spend less time in vulnerable, early larval stages. An alternative (although non-mutually exclusive) explanation to nongenetic changes could be that environmental stress, such as high temperature, imposes selection among genetically different haploid sperm within an ejaculate, such that certain sperm haplotypes are more successful and transmit genes that are beneficial for normal offspring development [29]. For example, recent elegant experiments in zebrafish have shown that phenotypic variation within an ejaculate can reflect haploid genetic variation, and selection among phenotypes can affect offspring traits [30,31]. Given previous evidence that heat stress alters sperm RNA content in M. galloprovincialis [10] while the proportion of motile sperm in ejaculates remains constant, we suggest that epigenetic changes are the more likely mechanism underlying paternal effects reported here. However, we consider future molecular studies of sperm genetic and epigenetic variation to be a priority for this system, to reveal the full evolutionary implications of the reported paternal effects.

Regardless of the underlying proximate mechanisms, the sperm treatment effects appeared greatly reduced when offspring were reared at high temperatures themselves. A possible explanation for this could relate to the stress-coping mechanisms of sperm themselves; we have recently reported in M. galloprovincialis that after exposure to high temperatures, sperm contain lower abundance of transcripts related to stress responses (e.g. the glycolytic protein gene gapdh and the heat shock protein gene hsp90) [10]. Sperm might translate such RNAs to maintain normal function at high temperatures, then lose them to co-translational degradation [32]. If sperm transcripts transferred during fertilization are used by embryos [33], this could mean that fewer stress response transcripts are available to early embryos before activation of the zygotic genome. While this conclusion remains speculative, our results do suggest that the apparent beneficial effects of sperm treatment to offspring are offset when high temperatures persist through early larval stages.

In conclusion, we provide a rare test of whether post-ejaculation environments can influence offspring phenotype and indeed find evidence that early larval stages are sensitive to the temperature experienced by sperm. However, contrary to recent theoretical predictions [3], our results do not support an anticipatory or ‘environment-matching effect’, whereby environmental changes to the ejaculate prepare offspring to perform best in similar conditions. Instead, average beneficial ejaculate-mediated paternal effects might in some cases be dampened or reversed by other environmental interactions when offspring themselves develop in stressful conditions. This highlights that the evolutionary implications of paternal effects might be more complex than current predictions, and that empirical tests of such complexities will be crucial to understanding the cross-generational effects of environmental change.

Data accessibility

Data associated with this manuscript are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.9ghx3ffhc [34].

Authors' contributions

R.A.L., J.P.E. and W.J.K. conceived of the study and designed the experiments. R.A.L. conducted the experiments, data collection and statistical analyses. R.A.L. wrote the first draft of the manuscript, and all authors contributed to the final version. All authors gave final approval for publication and agreed to be held accountable for the content therein.

Competing interests

We declare we have no competing interests.

Funding

This work was supported by an Australian Research Council grant awarded to J.P.E. and W.J.K. (grant no. DP170103290).