Can paternal effects via seminal fluid contribute to the evolution of polyandry?

Abstract

Genetic benefits from mating with multiple males are thought to favour the evolution of polyandry. However, recent evidence suggests that non-genetic paternal effects via seminal fluid might contribute to the observed effects of polyandry on offspring performance. Here, we test this hypothesis using the field cricket Teleogryllus oceanicus. Using interference RNA, we first show that at least one seminal fluid protein is essential for embryo survival. We then show that polyandrous females mated to three different males produced embryos with higher pre-hatching viability than did monandrous females mated with the same male three times. Pseudo-polyandrous females that obtained sperm and seminal fluid from a single male and seminal fluid from two additional males had embryos with viabilities intermediate between monandrous and polyandrous females. Our results suggest either that ejaculate mediated paternal effects on embryo viability have both genetic and non-genetic components, or that seminal fluids transferred by castrated males provide only a subset of proteins contained within the normal ejaculate, and are unable to exert their full effect on embryo viability.

1. Introduction

Mating by females with multiple males (polyandry) is taxonomically widespread [1]. The evolution of polyandry has been an enduring puzzle, because mating is known to be costly for females [2]. Where females can acquire resources from males to offset these costs, multiple mating makes adaptive sense [3]. However, an explanation for polyandry in species where males provide no obvious direct benefits remains the subject of debate. Females have been proposed to gain genetic benefits from polyandry because postcopulatory sexual selection can increase the probability that their ova are fertilized by genetically superior males [4]. Evidence of genetic benefits have been sought using experiments in which female mating frequency is held constant while the number of males mated is varied [5]. Such studies have provided general, albeit weak support for increased offspring fitness resulting from polyandry [6]. For example, in the Australian field cricket Teleogryllus oceanicus, polyandry elevates embryo survival [7] and the fertility of sons [8]. While findings such as these are taken as evidence for genetic benefits from polyandry, conceptually the problems with genetic benefit arguments for the evolution of polyandry are the same as those for the evolution of premating female choice; genetic variance in breeding values for offspring performance should be lost in the face of selection [9], eliminating the genetic benefits of polyandry. Non-genetic paternal effects offer a resolution to this paradox [10,11].

There is now considerable evidence that environmentally induced changes in the behaviour and physiology of males can be transmitted to offspring [12], and that non-genetic inheritance can be mediated via non-sperm components of the ejaculate [13]. Seminal fluid contains a cocktail of biologically active molecules, including proteins, lipids and RNAs, that have effects on the fertilization capacity of sperm, the development and viability of embryos and the phenotypes of adult offspring [12,14–16]. Because variation in non-genetic paternal effects can be environmentally induced in each generation, the benefits of multiple mating for females will be maintained in the population, providing a mechanism for the evolutionary maintenance of both mate choice and polyandry [10,11,17].

In T. oceanicus, seminal fluid proteins have been found to mediate non-genetic paternal effects on offspring viability, either directly and/or indirectly through their effects on maternal provisioning [18,19]. Here, we examine the effects of seven seminal fluid proteins on offspring viability using interference RNA. We then test empirically whether polyandry is beneficial for female T. oceanicus because of the non-genetic paternal effects of seminal fluid on offspring performance. We adopted the widely used protocol of allowing females to mate multiply with three different males (polyandrous) or three times with the same male (monandrous) [5,6]. However, in a third group, females received sperm and seminal fluid from one male but only the seminal fluid fraction of the ejaculate from two castrated males. Thus, ‘pseudo-polyandrous' females received seminal fluid from the same number of males as the polyandrous group but remained genetically monandrous. If non-genetic paternal effects mediated via seminal fluid favour the evolutionary maintenance of polyandry, we would expect both pseudo-polyandrous and polyandrous females to have greater offspring viability than monandrous females.

2. Methods

Crickets were taken from a large outbred population originally collected from Carnarvon, Western Australia. Male and female crickets were separated at the final instar, housed in single-sex cultures and checked daily for adult emergence.

(a). Effect of seminal fluid proteins on embryo viability

We used RNAi to explore the effects on embryo survival of seven seminal fluid proteins that have been shown to affect ejaculate quality [20,21]; ToSfp001 (isotig01262), ToSfp005 (isotig01832), ToSfp011 (isotig01709), ToSfp017 (isotig05129), ToSfp022 (isotig00444) and ToSfp23 (isotig00811) [20,22]. Detailed methods for RNAi have been published elsewhere [21] and are reproduced in the electronic supplementary material. Ten males were used for each gene, and for a control group in which males were injected with the carrier vehicle for the dsRNA. Five males died before they could be mated, one ToSfp001, one ToSfp005 and three ToSfp022 knockdowns, reducing the sample sizes for these genes. Females were aged 10 days post-adult eclosion when mated to knockdown males. Following mating, females were provided with a dish of moist sand and left to oviposit for 7 days. Eggs were then rinsed from the sand and two batches of 50 (where available) eggs that showed evidence of embryonic development (see [19]) were counted and placed onto damp cotton wool in a small container (7 × 7×5 cm) and kept at 26°C. After two weeks, hatching was monitored daily and the nymphs counted and removed until no further hatching occurred. Embryo viability was calculated as the proportion of developing eggs that hatched.

(b). Effect of polyandry on embryo viability

On emergence females were housed in groups of 20 in 5 l containers while males were housed in individual containers (7 × 7 × 5 cm). Both were provided with cat chow and water ad libitum. Each male was assigned for use in one of three treatments; monandrous, polyandrous or pseudo-polyandrous. The males to be used in the pseudo-polyandrous treatment were castrated on the day of their emergence. These males were kept at 4°C for 30 min and then at −20°C for 2 min after which each cricket was placed under a dissecting microscope. With the wings held laterally, an incision was made through the dorsal surface between the second and third abdominal segments. Each testis was removed with forceps and the wound closed with cyanoacrylate glue. Following surgery, the cricket was placed back in its individual container and kept at 26°C on a 12 : 12 light : dark cycle. When 13 days of adult age, all males were mated to a random female from the stock population to ensure they were capable of mating. For the castrated males, once mating had occurred, the spermatophore was collected from the female and its contents evacuated into 20 µl phosphate-buffered saline (PBS, Astral Scientific) on a microscope slide for 10 min and checked for sperm. If sperm were detected the cricket was discarded.

Crickets were mated when 14 days of adult age. For mating, individual females were transferred to the box of a male in their allocated treatment. Once mating had occurred the pair was observed to ensure the spermatophore remained attached for a period in excess of 40 min, the time required for full ejaculate transfer [23]. The pair was then separated. The female was placed into her own individual contained with food and water provided ad libitum, while the male remained in his original container. For the monandrous treatment, the female was remated to the same male on days 15 and 16 following the same protocol. For the polyandrous treatment, the female was mated to different males on each subsequent day. For the pseudo-polyandrous treatment, the female mated with an intact male on day 14, the same intact male on day 15 followed by a castrated male, and the same intact male followed by a second castrated male on day 16. Thus, monandrous females (n = 79) received three quantities of sperm and seminal fluid from the same male, polyandrous females (n = 86) received three quantities of sperm and seminal fluid, one from each of three different males, and pseudo-polyandrous females (n = 73) received three quantities of sperm and seminal fluid from the same male plus two quantities of seminal fluid, one from each of two different males. Following the final mating, females were placed back into their individual boxes with a small petri dish of washed sand to lay eggs for 7 days. Hatching success was assessed as described above.

Embryo viability was analysed using generalised linear models with the number of hatched nymphs as the numerator, the number of developing eggs sampled as the denominator, a quasibinomial distribution to account for overdispersion and logit link function. Analyses were conducted in R [24] using the lme4 package.

3. Results

The proportion of eggs that hatched varied significantly among females depending on which of the seven sfp genes had been knocked down in the males with whom they were mated (χ² = 35.55, d.f. 6, p < 0.001). Planned contrasts comparing each gene knockdown to the control revealed that knockdown of ToSfp011 resulted in complete hatching failure (χ² = 43.2, d.f. 1, p < 0.001) while embryo viability was greatly reduced after knockdown of ToSfp022 (χ² = 4.92, d.f. 1, p = 0.027) (figure 1). Only ToSfp011 survived the Benjamini–Hochberg procedure with a false discovery rate of 5% (P_crit = 0.014; electronic supplementary material, table S1).

Figure 1.

The proportion of developing embryos that hatched when females mated either with males having one of their seminal fluid protein genes knocked down using interference RNA, or a control group where males were injected with the carrier vehicle used for double-stranded RNA.

There was significant variation in embryo viability due to mating treatment (χ² = 6.17, d.f. 2, p = 0.046). While monandrous females had a significantly lower proportion of embryos hatching than polyandrous females (effect estimate −0.381 ± 0.161, z = 2.37, p = 0.046), the hatching success for pseudo-polyandrous females was intermediate: pseudo-polyandrous females did not have a significantly higher hatching success than monandrous females (effect estimate 0.101 ± 0.170, z = 0.59, p = 0.825) but neither did they have a significantly lower hatching success than polyandrous females (effect estimate −0.280 ± 0.163, z = 1.725, p = 0.196). Violin plots show that the distribution of hatching success values shifted toward increasing values, from monandrous to pseudo-polyandrous to polyandrous females (figure 2).

Figure 2.

The effect of mating treatment on the proportion of developing embryos that hatched. Violin plots show the probability density of embryo viability values. The internal box plots show the median and interquartile ranges.

4. Discussion

Females mated to three different males produced offspring with greater pre-hatching embryo viability than did females mating three times with the same male. This finding replicates that of a previous study in which females were allowed to mate four times with the same male or with four different males [7]. Thus, it is mating with different males, rather than the number of times a female is exposed to males and receives an ejaculate, that is beneficial for female T. oceanicus in that polyandry increases the early embryonic viability of offspring. Similar findings have been reported for a number of species [6]. We found mixed support for the role of seminal fluid in this polyandry effect.

Using interference RNA, we found that at least one seminal fluid protein, ToSfp011, was essential for the development of embryos to hatching. In a previous study, we documented significant variation in embryo viability among genetically monogamous females mated to castrated seminal fluid donors [19]. In that study, the seminal fluid of some donors was found to enhance embryo viability while the seminal fluid of others reduced embryo viability. Importantly, increasing expression of ToSfp011 in the accessory glands of seminal fluid donors was associated with reduced embryo survival [19]. Taken together with the findings of our interference RNA, the data suggest that while ToSfp011 is essential for embryo development it may be harmful when abundant. This pattern of biphasic dose dependence, known as hormesis, is a widespread stimulatory response found across biological systems [25,26]. For males, increased expression of ToSfp011 increases sperm viability [21] which in turn promotes a male's competitive fertilization success [27]. Polyandry then may represent a bet-hedging strategy among females seeking to ameliorate the harmful effects of male seminal fluid proteins that function in sperm competition. In support of this suggestion, Garcia-Gonzalez & Simmons [28] used a white-eye mutation that allowed them to monitor the pre-hatching viability of embryos sired by two different males mated to the same female. They found that a male that imparted high viability to his own embryos did so for the embryos sired by his sperm competition rival. The precise mechanism by which ToSfp011 affects embryo viability remains uncertain; candidate mechanisms include a direct effect on embryo development if it were able to enter the egg, perhaps bound to sperm [19], or indirect through its effects on the genetic integrity of sperm themselves.

By mating genetically monogamous females to additional castrated males, we asked whether paternal effects mediated by seminal fluids from multiple males might contribute to the enhanced embryo viability enjoyed by polyandrous females. Pseudo-polyandrous females, those receiving sperm from a single male but seminal fluid from multiple males, did not have a significantly higher embryo viability than monandrous females. The viability of their embryos fell intermediate between that of monandrous and genetically polyandrous females. We offer two potential explanations for this finding.

Offspring viability may depend on both genetic and nongenetic paternal effects. Polyandry can in theory allow a female to bias fertilization toward males of superior genetic quality so that the benefits of genetic and non-genetic paternal effects may be additive, explaining the significantly greater embryo viability only for polyandrous females in our study. Indeed, paternal effects on embryo viability have been reported to harbour significant additive genetic variance in this species [29]. However, these paternal effects are genetically correlated with a male's expenditure on his seminal fluid producing accessory glands [29], and recent work suggests that nongenetic paternal effects of seminal fluid can largely account for the apparent genetic effects on embryo viability [19].

Alternatively it may be that castrated males were unable to deliver the full complement of seminal fluid proteins (sfps) contained within a normal ejaculate, so that females mated to castrated males did not accrue the full benefits of seminal fluid doses from multiple males. There are at least 21 different proteins that have been identified from the seminal fluid of T. oceanicus [22]. While most of these sfps have transcripts originating from the accessory glands, four proteins (Tosfp006, Tosfp012, Tosfp015 and Tosfp017) have transcripts originating from the testes [22] suggesting that the testes contribute both sperm and sfps to the ejaculate. The metabolic functioning of seminal fluid has been found to depend on networks of interacting proteins [30,31]. Thus, the intermediate levels of embryo viability seen in pseudo-polyandrous females may have arisen because one or more components of the seminal fluid protein network of castrated males were absent, reducing the functionality of their seminal fluid.

In summary, we replicate previous work that shows how female T. oceanicus benefit from polyandry through increased early embryonic viability. Our data offer mixed support for the notion that seminal fluid proteins might contribute to this benefit of polyandry. Unlike genetic benefits that are expected to erode in the face of selection, variation in seminal fluid quality can vary between and within generations, for example, through variation in nutrition [32] or social environment [33], so that non-genetic paternal effects mediated via seminal fluid have the potential to contribute to the evolutionary maintenance of polyandry [34].

Ethics

This work was conducted on insects for which ethical approvals were not required.

Data accessibility

Data are available from Dryad: https://doi.org/10.5061/dryad.9zw3r22br [35].

Authors' contributions

L.W.S. conceived the study and its design, analysed the data and drafted the manuscript. M.L. conducted the experiment, recorded the data, drafted the methods and edited the final version of the manuscript. Both authors agree to be held accountable for the content therein and approve the final version of the manuscript.

Competing interests

The authors declare no competing interests.

Funding

This research was supported by the ARC (DP160100797).