SPECIAL ISSUE : FEMALE MATING FAILURES
Sperm as a limiting factor in mating success in Hymenoptera parasitoids
Correspondence: E-mail: email@example.com
The paradigm that females produce few costly eggs, whereas males produce an unlimited quantity of sperm holds true when matings are dispersed over time and males can replenish their sperm supply. In such a system, the best strategy for males is to mate with as many females as possible, as more is always better. However, in parasitoid species, mating often occurs at the emergence site and males may have to mate successively several females in a short period of time. This can lead to sperm depletion that can be temporary in synspermatogenic species whose males produce sperm throughout their adult life, or definitive in prospermatogenic species whose males emerge with their full sperm complement and do not produce more during their adult life. Both the spermatogeny index and the mating structure of a species will therefore influence the probability and intensity of sperm depletion in males. Sperm-limited males can court and mate females even in prospermatogenic species. These males still gain some inclusive fitness by preventing competitor males to fully inseminate females, therefore increasing the representation of their daughters in the following generation. Females that mate with partially or completely sperm-depleted males produce a constrained sex ratio that decreases their lifetime fitness. In species where sperm-depleted males occur, female choosiness based on sperm supply is predicted, as the cost of mating sperm-depleted males can be high. In addition, males are also expected to be choosy in prospermatogenic species as mating with sub-optimal females can be costly when sperm is limited.
Most hymenopteran parasitoids reproduce by arrhenotokous parthenogenesis, the dominant mode of sex determination in the Hymenoptera, where an unfertilized egg develops into a haploid male and a fertilized egg into a diploid female (Heimpel & de Boer, 2008). Some parasitoid species also reproduce through thelytokous parthenogenesis, where only females are produced and both genetic and endosymbiont-induced forms of thelytoky are known in Hymenoptera (Heimpel & de Boer, 2008). When mated, females of arrhenotokous parasitoid species can control the sex ratio of their progeny by choosing to release sperm or not at the time of oviposition. Virgin females can therefore produce only sons.
How mated females should allocate the sex ratio of their progeny is predicted by optimality models, such as the local mate competition (LMC) model (Hamilton, 1967) or the host quality (HQ) model (Charnov et al., 1981). The LMC model predicts that when mating is within a structured population, and therefore violates the panmixis assumption of Fisherian equal sex ratio (Fisher, 1930), females, when alone, should allocate to a host patch only enough sons to mate with all the daughters on the patch. When the female perceives competition from other females in that patch, she should gradually increase the sex ratio deposited, eventually approaching a 1:1 sex ratio. The LMC model thus predicts how females ovipositing in a patch respond to the risk of competition among their sons.
The HQ model predicts how host quality should influence the sex ratio deposited (Charnov et al., 1981). If sons and daughters respond differently to the quality of the host on or in which they develop, the female should allocate to the low-quality hosts, the sex that suffers less. Females generally suffer more when they develop in low-quality hosts and therefore it is observed that males are often allocated to low-quality hosts (King, 1987; King & Napoleon, 2006).
These models always assume that females can produce an optimal sex ratio and are not constrained in their sex ratio allocation, but in reality, females are sometimes constrained to produce a sex ratio that is not optimal. In solitary parasitoids that develop in solitary hosts, females must disperse to find a mate. They are thus virgin until they encounter a male and all progeny produced until then will be male. Virgin females are present in parasitoid populations with proportions varying from <1 to 15–20% (Strand, 1989; Godfray, 1990; Gu & Dorn, 2003; Martel & Boivin, 2004, 2007), but reaching 50% in Cephalonomia waterstoni Gahan (Hardy et al., 1998). Other causes of virginity include post-copulation constraints, such as in species where there is a period after mating during which the sperm is not available, sperm incompatibility, or when too many matings occurred (Godfray, 1990). The presence of virgin females producing a constrained sex ratio will also influence the sex ratio produced by unconstrained females.
In species with panmictic mating structure, and therefore with a 1:1 sex ratio, virgin females constrained to produce only sons are not heavily penalized as the value of sons and daughters is equal. As the proportion of constrained females producing only males increases, unconstrained females should increase the proportion of females produced as the value of sons decreases because of their excess production by constrained females. In practice, there are probably small advantages to produce daughters over males, so females are unlikely to refuse mating with a male, but they are unlikely to refuse an oviposition opportunity to wait for a mate (Ode et al., 1997). This equilibrium, however, is modified for species that are egg-limited or when oviposition incurs mortality risks. The effect of constrained oviposition is different in species with structured populations. In such case, the optimal sex ratio is female-biased and the sex ratio strategy of unconstrained females is only slightly affected by the presence of constrained females. The sex ratio of the populations should remain female-biased, although slightly less so than in the absence of constrained females. It is then predicted that, in these species, constrained females should invest in the removal of the constraint, i.e., favor mating over oviposition opportunities (Godfray, 1990).
Constrained sex ratio after mating
Even mated females can produce a constrained sex ratio depending on the quantity or quality of sperm transferred by the male during copulation (Henter, 2004). The dominant paradigm in reproductive biology has long been that females produce few costly gametes, the oocytes, whereas males produce an unlimited supply of cheap gametes, the sperm (Dewsbury, 1982; Hardy et al., 1998). Males' fitness should then be limited only by their capacity to acquire mates, and for males, the best strategy is to inseminate as many females as possible (Bateman, 1948). However, in numerous species, the production of sperm by males incurs non-trivial costs and males should therefore use their sperm with parsimony (Simmons, 2001). Males can optimize sperm use through mate choice or by adjusting sperm numbers according to proximal cues of female quality or the probability of sperm competition (Simmons, 2001; Martel et al., 2008a). Although the dominant view was that of indiscriminate males competing for females, males should choose females with the best reproductive value when sperm is limited or costly and reject low-quality females in the hope of encountering a female with a higher value. Such male mate choice is more likely to evolve in situations where females are encountered simultaneously and where males are unlikely to be able to mate them all (Edward & Chapman, 2011), a situation often found in parasitoids that are gregarious or quasi-gregarious (solitary parasitoids that attack aggregated hosts). Many haplodiploid parasitoids exhibit a female-biased sex ratio resulting in a female-biased operational sex ratio, a condition favoring male mate choice (Edward & Chapman, 2011).
The temporal distribution of sperm production is thus of interest when it comes to understand mate choice, sperm allocation, and the risk for females of not getting a full complement of sperm during mating. In females, the ovigeny index classifies species based on the proportion of oocytes present at emergence in female parasitoids (Jervis et al., 2001). In a similar manner, species can be classified according to a spermatogeny index. Males are prospermatogenic when they emerge with all their sperm stock and cannot produce more during their adult life, or synspermatogenic if they emerge with no sperm and produce sperm during their adult life (Boivin et al., 2005). Both pro- and synspermatogenic males may become sperm-depleted when all their stock has been transferred following several successive matings. Sperm depletion is permanent in prospermatogenic species and can be temporary in synspermatogenic species. When successive matings result in a gradual decrease in the quality of the transferred ejaculate, additional matings, although still advantageous for males, do not result in a linear increase in the male reproductive success. Evaluation of the best reproductive strategy for males should then be estimated through optimality models based on number and quality of sperm transferred rather than the number of matings performed by a male.
The temporal distribution of sperm production, the management of sperm by males, and its impact on female insemination have been studied in detail in the egg parasitoid Trichogramma euproctidis (Girault) (Hymenoptera: Trichogrammatidae).
Sperm management in Trichogramma euproctidis
Males of the egg parasitoid T. euproctidis have a short life span (Boivin & Lagacé, 1999) and emerge a few minutes before females (Doyon & Boivin, 2006). They remain at the emergence site (Pompanon et al., 1995; Martel & Boivin, 2004) and mate with emerging females (van den Assem et al., 1980; Hardy, 1994). Males that develop singly in eggs of Ephestia kuehniella Zeller emerge with 800 sperm per seminal vesicle, which are swollen regions of the vas deferens (Boivin, 2010a), for a total of 1 600 sperm (Damiens & Boivin, 2005). This species is strictly prospermatogenic and therefore males do not produce sperm during their adult life. In the related species Trichogramma brassicae Bezdenko, spermatogenesis ceases in the late pupal stage (Chihrane & Laugé, 1994), after which the testes start to degenerate. At each mating, males ejaculate around 100 sperm and the female, also emerging singly from E. kuehniella, stores around 50 of these sperm in her spermatheca. The male ejaculates 100 sperm per mating for the first 10 females mated, then the quantity of sperm transferred gradually decreases (Damiens & Boivin, 2005).
This pattern of sperm depletion is explained by the structure of the seminal vesicle of T. euproctidis. The seminal vesicle is divided into an anterior chamber and a smaller vesicular pocket. The valves between the two chambers of the seminal vesicles stay open between copulations, allowing continuous movement of sperm between the chambers. At mating, the valve closes and the content of the vesicular pocket is ejaculated. Each vesicular pocket represents around 9% of the total volume of the seminal vesicle, but it can contain a maximum of around 50 sperm (thus, the total ejaculate of 100 sperm for the two vesicular pockets). The number of sperm transferred is initially constant as it is limited by the size of the vesicular pocket. However, when 9% of the total number of sperm in a seminal vesicle falls below 50, the vesicular pocket contains fewer and fewer sperm, explaining the gradual decrease in the number of sperm transferred during mating (Damiens & Boivin, 2005). After 20 matings, a male transfers few sperm and when females mated by such sperm-depleted males are tested, they never produce daughters.
Sperm depletion in males
When sperm production is costly, multiple mating leads to diminishing reproductive returns for males. This decline may be apparent in a decrease in sperm number, sperm size, stimulation of female egg production rate, and inhibition of female remating (Lewis, 2004). Male mating history is thus an important aspect in a female's reproductive success. The number of matings and the rate of matings have been shown to influence the sex ratio produced by females. Females mated late in a succession of matings produce a constrained sex ratio, in addition to T. euproctidis, in the pteromalids Pachycrepoideus vindemiae (Rondani), Lariophagus distinguendus (Förster), and Dinarmus basalis (Rondani) (Nadel & Luck, 1985; Bressac et al., 2008; Steiner et al., 2008), the braconid Bracon hebetor (Say) (Ode et al., 1996), and the bethylid Cephalonomia hyalinipennis Ashmead (Pérez-Lachaud, 2010).
Sperm-depleted males are known to continue mating with females (Simmonds, 1953; Laing & Caltagirone, 1969; Ramadan et al., 1991; Ode et al., 1997; Damiens & Boivin, 2006; Steiner et al., 2008) and females mated by such males remain pseudo-virgin. In T. euproctidis, females mated first by a sperm-depleted male then by a virgin male obtain only about 15 sperm, one-third of the normal sperm complement after mating with a virgin male. After mating with a sperm-depleted male, a female has to mate with three virgin males to obtain her full sperm complement (Damiens & Boivin, 2006). This suggests that sperm-depleted males transfer seminal fluids during mating and that the second male has to move some fluid out of the spermatheca in order for his sperm to enter. In monoandrous species, pseudo-virgin females are constrained for life to produce only sons, but even in polyandrous species, females that mated first with a sperm-depleted male pay a cost, as they have to mate with several non-depleted males to obtain their full sperm complement.
Sperm depletion in prospermatogenic males is permanent and at that point, prospermatogenic males have reached the end of their reproductive life. The presence of post-reproductive life in numerous species remains an interesting question in evolutionary ecology as the selection for living beyond the end of the reproductive ability is at best weak (Tully & Lambert, 2011). When parental care is important for progeny survival, post-reproductive survival is expected, but it is also observed in species without parental care (Damiens & Boivin, 2006; Tully & Lambert, 2011). An increase in inclusive fitness can be a selective force acting on post-reproductive life span. In T. euproctidis, sperm-depleted males that continue to mate prevent competitor males to fully inseminate females, therefore increasing the representation of their daughters in the following generation.
Determining the representation in natural populations of pseudo-virgin females that mated with a sperm-depleted male is difficult. Sampling females in the field and checking for the absence of sperm would not differentiate among truly virgin females, females that exhausted their sperm supply, and pseudo-virgin females. Some laboratory data support the presence of sperm-depleted males on and outside the emergence patch and their importance in mating females. In T. euproctidis, the number of matings is not distributed equally among males present on the emergence patch. Some males mate more often and deplete their sperm supply before leaving the emergence patch (Martel & Boivin, 2007).
When partially or completely sperm-depleted males occur in a population, multiple mating with different males by females (polyandry) could reduce the risk of producing a constrained sex ratio. Polyandry is found in most groups of insects despite the fact that females can often obtain all the sperm supply they need from a single mating and that mating with several males is costly. These costs include time and energy for additional mating, increased risk of predation, increased risk of infection of horizontally transmitted disease during copulation, and caustic seminal fluid that reduce fitness (Yasui, 1998; Fedorka & Mousseau, 2002). These costs decrease both female life span and egg production rate (Arnqvist & Nilsson, 2000). In Hymenoptera parasitoids, about 20% of the species are polyandrous, this percentage rising to 60% for gregarious species (Ridley, 1993; Pérez-Lachaud, 2010) and polyandry has been shown to occur under field conditions (Allen et al., 1994). Polyandry is not unique to Hymenoptera parasitoids. In Aleochara bilineata Gyllenhal, a Coleoptera ectoparasitoid of Diptera pupae whose first instars express several behaviors similar to those expressed by Hymenoptera parasitoids (Royer & Boivin, 1999; Royer et al., 1999), females increase their pre-mating latency after one mating, suggesting that it is a behavior to decrease the probability of multiple mating with the same male, and therefore to increase the chance of polyandry with another male (Lizé et al., 2009).
Polyandry can benefit females through sperm competition allowing the most competitive sperm (possibly carrying the best genes) to fertilize the egg (Sivinski, 1984; Simmons, 2001), or through increased genetic diversity (Hosken & Stockley, 2003; Ivy & Sakaluk, 2005), because the off-patch mates are unlikely to be related to the female. Although in arrhenotokous species like T. euproctidis, such benefits are lower because of the purge of deleterious alleles in haploid males, genetic diversity might be more important in species with complementary sex determination (individuals with heterozygous alleles developing in a female and individuals with hemi- or homozygous alleles developing in a male), such as Cotesia glomerata (L.) (Zhou et al., 2006) or B. hebetor, where the females rarely mate more than once (Whiting, 1943; Ode et al., 1995). Material benefits have also been proposed as an explanation for the presence of polyandry in female parasitoids. Although nuptial gifts have rarely been reported in Hymenoptera (Vahed, 1998), females may still gain from multiple mating either from accessory substances transferred at mating or through supplementary sperm supply. Polyandry can also be a strategy that females use to reduce costs incurred due to male harassment. Females then remate because the cost of resisting mating attempts by males is higher than the cost of additional mating. Polyandry brings no benefit of additional fecundity or longevity in T. euproctidis (Jacob & Boivin, 2005) and convenience mating or avoidance of sperm-depleted males are the most likely hypotheses.
Several related bet-hedging hypotheses could explain the evolution of polyandry when genetic benefits increase the geometric mean fitness of females (Zeh & Zeh, 2003). In the ‘intrinsic male quality hypothesis’, polyandry could appear in a stable environment if females reduce the risk that all their progeny is sired by a low-quality male by mating with several males. This hypothesis can only apply to species where females are unable to evaluate precisely male's quality or can only distinguish broad categories. A recent meta-analysis suggests that polyandry does bring genetic benefits in insects by increasing hatching success, clutch production, and fertility (Slatyer et al., 2012). These effects, however, are small but when a composite effect size is used, that integrates measures of offspring fitness, a small but significant positive effect of polyandry is present.
Polyandry could thus be a strategy expressed by females to decrease the probability of being mated by a sperm-depleted male or to replenish their sperm supply if mated by a partially sperm-depleted male (Pérez-Lachaud, 2010). By doing so, they decrease the risk of producing male-only progeny in exchange for the costs of mating several times.
The presence of sperm-depleted males on the emergence patch poses a problem for females. Males that are successful at mating with several females do so because they are of higher quality, or they are better at advertising it. As time passes after emergence, these high-quality males are also those most susceptible to become sperm-depleted. The optimal strategy for females would then be to mate only with high-quality males soon after emergence, but gradually accept lower-quality males as these are less likely to be sperm-depleted. This implies that females can discriminate males based on their sperm reserve or on the number of matings they have done. Female T. euproctidis prefer large to small males (Boivin & Lagacé, 1999), but when offered virgin males and males that have mated several times, they did not show preference for virgin males (Jacob, 2004). This lack of discriminating capacity suggests that females mate several times on the emergence patch as an insurance against sperm-depleted males (Martel & Boivin, 2007). On the other hand, male T. euproctidis discriminates between virgin and mated females (Martel et al., 2008b) and prefers large to small ones (Boivin & Lagacé, 1999).
Even in gregarious or quasi-gregarious species such as T. euproctidis, mating is not exclusively done on the emergence patch. In a greenhouse experiment, Martel et al. (2010) used two isofemale lines of T. euproctidis to show that over 40% of all mating occurs outside the emergence patch and that this percentage depended on the sex ratio on the emergence patch. The higher the sex ratio, the more likely that mating occurred on the emergence patch. Even when mating occurs on patch, unsynchronized emergences or differences in emergence location can lead to non-random mating between kin and non-kin on the patch, a model termed ‘asymmetrical local mate competition’ (Shuker et al., 2005, 2006, 2007).
Phenotypic plasticity and reproductive success
The probability of a female parasitoid to produce a constrained sex ratio is thus based on the relation between the fecundity of the female and the number and quality of the sperm she acquired during copulation. An additional complexity has to be added as both the number of gametes (oocytes and sperm) and their size are part of the phenotypic plasticity of parasitoids. The size of parasitoid males and females varies depending on host size, quality, sex, and species (Boivin, 2010b), and both sexes allocate resources differently to survival and reproduction according to environmental conditions or food availability (Rivero & West, 2002). In insects, large males can produce more sperm (Wiernasz et al., 2001; Bangham et al., 2002) and inseminate more females (Boivin & Lagacé, 1999), and large females are known to produce more eggs and are more fecund (Honek, 1993; Rivero & West, 2002; Durocher-Granger et al., 2011). These differences can be important. In T. euproctidis, males that developed singly in a small host egg, Plutella xylostella L., a medium-sized host egg, E. kuehniella, or a large host egg, Trichoplusia ni (Hübner), had 787, 1 618, and 3 331 sperm in their seminal vesicles, respectively (Martel et al., 2011). Reared in the same hosts, female T. euproctidis had spermatheca of 4 882, 6 213, and 12 288 μm3 and these spermathecae contained 23, 59, and 117 sperm, respectively (Martel et al., 2011). Whereas the size of the spermatheca of a female is more or less in proportion to the number of oocytes available to her, small males would be unable to fill the spermatheca of a large female. In addition, these prospermatogenic small males would exhaust their sperm supply faster than large males. These differences in size translate into fewer and smaller gametes both when the difference in size is due to the size of the host (Martel et al., 2011) or to gregarious development in a large host (Durocher-Granger et al., 2011). A decrease in the host population density that results in an increase in superparasitism or a shift to a smaller host during the season could thus change the probability that females do not receive their full sperm complement, and therefore produce a constrained sex ratio. The presence of such variability in the reproductive capacity in both males and females is also a basis for the presence of mate choice by both males and females (Edward & Chapman, 2011).
Finding a mate is only the first step for a female to produce an optimal sex ratio throughout her life. The probability of finding a proper mate is influenced by several population dynamics issues and by the biology of a species. Whereas, the probability of finding a mate is high for gregarious and quasi-gregarious parasitoid species, solitary species must allocate time and energy to mate searching with the result that sometimes early progeny production is only males.
Finding a mate does not always guarantee that the female will be able to produce an optimal sex ratio. The quality of both the male and his sperm will affect the capacity of the female to produce daughters at an optimal rate. Details on sperm production and management by male parasitoids are available for only a few species and more data are needed to evaluate the impact of these factors in the population dynamics of parasitoids.