Sex‐dependent infection causes nonadditive effects on kissing bug fecundity

Abstract The influence of parasites on host reproduction has been widely studied in natural and experimental conditions. Most studies, however, have evaluated the parasite impact on female hosts only, neglecting the contribution of males for host reproduction. This omission is unfortunate as sex‐dependent infection may have important implications for host–parasite associations. Here, we evaluate for the first time the independent and nonindependent effects of gender infection on host reproductive success using the kissing bug Mepraia spinolai and the protozoan Trypanosoma cruzi as model system. We set up four crossing treatments including the following: (1) both genders infected, (2) both genders uninfected, (3) males infected—females uninfected, and (4) males uninfected—females infected, using fecundity measures as response variables. Interactive effects of infection between sexes were prevalent. Uninfected females produced more and heavier eggs when crossed with uninfected than infected males. Uninfected males, in turn, sired more eggs and nymphs when crossed with uninfected than infected females. Unexpectedly, infected males sired more nymphs when crossed with infected than uninfected females. These results can be explained by the effect of parasitism on host body size. As infection reduced size in both genders, infection on one sex only creates body size mismatches and mating constraints that are not present in pairs with the same infection status. Our results indicate the fitness impact of parasitism was contingent on the infection status of genders and mediated by body size. As the fecundity impact of parasitism cannot be estimated independently for each gender, inferences based only on female host infection run the risk of providing biased estimates of parasite‐mediated impact on host reproduction.


| INTRODUCTION
The fitness impact of parasites ranges from early mortality and complete castration to slight reduction in host fecundity and even increased reproduction (Ballabeni, 1995;Minchella, 1985;Poulin, 1998).
Even though reduced host fecundity is often an effect of parasitism, there is a still little knowledge about the mechanisms involved. Most existing evidence of parasite impact on host fitness in insects comes from studies performed on female hosts without consideration of the male infection status (Hurd, 2009). This omission is unfortunate considering that resource allocation trades off between parasite defense and other components of the phenotype differ between the sexes (Tschirren, Fitze, & Richner, 2003) resulting in sex-specific strategies to avoid or tolerate parasitism. Thus, the infection status of males may affect their reproductive investment in different ways. For example, infected males may have lower energy allocation to spermatophore production, provide lower-quality ejaculates, or fail to stimulate oviposition (Lehmann & Lehmann, 2000;Polak, 1996;Simmons, 1993). In these cases, females may receive direct benefits by mating with uninfected males. However, infected males may also positively affect the reproductive performance of females. For example, females lay more eggs when mated with infected males as infection increases the value of nuptial gifts (Hurd & Ardin, 2003). Therefore, the net resource availability for host reproduction may be contingent not only on female, but also on male infection status. Despite its importance, few studies have examined the influence of both male and female infections on host reproductive success (Sheridan, Poulin, Ward, & Zuk, 2000;Zuk & McKean, 1996), and to our knowledge, no study has examined potential interactive effects of sex-dependent infection on host reproduction.
We focus on a host-parasite interaction between the hemipteran

| Study system
The kissing bug M. spinolai is a triatomine species responsible for T. cruzi transmission in mammals of arid and semiarid areas of Chile (Botto-Mahan, Ortiz, Rozas, Cattan, & Solari, 2005). This strictly hematophagous and diurnal insect species is distributed between 26° and 34°S; its main habitat includes rocky outcrops, bird nests, rock crevices, and caves (Frías-Lasserre, 2010). Mepraia spinolai requires the blood of vertebrates to complete its life cycle that includes egg, five nymph stages, and adult (Botto-Mahan, 2009). In many triatomine species, one full engorgement is sufficient for molting from one stage to the next (Kollien & Schaub, 2000).
Trypanosoma cruzi is a heteroxenous trypanosomatid with a life cycle that involves several morphologically different stages, which can be found in insect vectors and mammalian hosts (Kollien & Schaub, 2000). This trypanosomatid multiplies and differentiates in the digestive tract of the insect vector. Infection of mammal hosts occurs by contamination of mucous membranes with insect feces, which contain the infectious metacyclic trypomastigote stage of the flagellate (Kollien & Schaub, 1997.

| Infected and uninfected adults
Individuals of M. spinolai used in this study were obtained from the first generation of a cohort of field-collected insects. During their development, insects (from the first instar nymph to adult) were reared individually in plastic containers maintained in a growth chamber at 26 ± 0.5°C, 65%-70% relative humidity, and 14 hr:10 hr light:dark cycle. Adults infected with T. cruzi were obtained by allowing insects to feed on infected laboratory rodents during their five nymphal stages. Only infected rodents in good condition and within the first 5 weeks of infection were used for feeding purposes (Wallace et al., 2001). Uninfected adults were obtained by allowing nymphs to feed on uninfected laboratory rodents.

| Experimental design and reproductive output
Once reached the adult stage, combinations of infected and uninfected virgin insects were assigned to four mating treatments: infected males and females (n = 19), infected males and uninfected females (n = 16), uninfected males and infected females (n = 25), and uninfected males and females (n = 24). Each pair was in a 7-cm height, × 6-cm diameter plastic container with a meshed lid. Every container was provided with folded piece of paper as refuge. Laboratory conditions were as described above. All pairs fed to engorgement every 3 weeks on uninfected rodents. Pair survivorship and sexual activity (e.g., males mounting or trying to mount females) were recorded daily prior to removal of dead adults. While female adults were alive, eggs were collected from parental containers daily, counted, individually weighed (±0.05 mg), and placed in new containers. Eggs were classified as yolky or yolkless eggs considering deformations of surface and color. Eclosion of first instar nymphs was recorded daily until 1 month after the female parent's death.

| Statistical methods
To examine the effect of the infection status of males, females, and their interaction, two-way ANCOVAs were used. The infection status of males and females was the main factor. The total time spent together (i.e., total time the pair was together) and adult female survivorship (i.e., time elapsed from the first day of mating to female death) were the covariates. Dependent variables consisted of the following: (1) production of yolky eggs, (2) weight of yolky eggs, (3) production of yolkless eggs, (4) reproductive investment (number × mean weight of yolky eggs), (5) the day by which the female laid 50% of her eggs (E50, hereafter), and (6) the number of first-stage nymphs. All dependent variables were checked for homogeneity of variance and normality and transformed when needed. For the number of nymphs, we used GLM with Poisson distribution errors and log link to compare treatments.
When the interaction between female infection status and male infection status was significant, we examined the significance of each main effect by comparing one specific factor at a variable level of another using interaction slices (Schabenberger, Gregoire, & Kong, 2000). All analyses were performed in JMP version 8.0.2.
Because differences in reproductive output can be attributable not only to an effect of T. cruzi but also to variation in the volume of ingested blood, we compared the volume of blood ingested by females in a one-way ANCOVA, using body size as covariate and female infection status as main effect. Body size was estimated as a factor from the linear combination of the equations for body length (mm), abdomen width (mm), and body weight (mg), which together accounted for 91.7% of the variance. In addition, we compared survivorship between infected and uninfected adult males and females, calculated as the number of days elapsed between the first day of mating and its death, using Student's t tests (Sokal & Rohlf, 1995).  (Table 1). Interaction slices revealed that uninfected females produced 43.1% less yolky eggs when crossed with infected than uninfected males (p = .008; Figure 2a, Table 2). Likewise, when crossed with uninfected males, egg production of infected females decreased 53.4% in comparison with crossings with uninfected females (p = .006). For yolky egg weight (Figure 2b, Table 2), the only significant slice indicates that crossings between uninfected females × infected males produce eggs 7.3% lighter than eggs from crossings between uninfected females × uninfected males (p = .041).

| RESULTS
The reproductive investment (Figure 2c, Table 2) followed a similar but less strong pattern than that observed for the number of yolky eggs, suggesting the overall investment of females in reproduction depends on the infection status of males and females altogether.

| DISCUSSION
The interaction between infection and gender influenced the most reproductive variables (Table 1, Figure 2  Number of nymphs χ 2 1 = 46.9 <.001 χ 2 1 = 104.7 <.001 χ 2 1 = 84.7 <.001 All analyses included female survival as covariate. 2008). In this study, infected females reached smaller size at maturity compared to those uninfected, suggesting that like other parasite species, T. cruzi probably curtails essential nutrients involved in host growth (Hurd, 1990;Thompson, 1983). In this vein, Hurd, Hogg, and Renshaw (1995) suggested that female body size could affect reproduction in two ways. First, fecundity may be limited by the number of ovarioles present in each ovary, which is function of female body size.
Second, body weight reduction may negatively affect blood feeding and blood meal utilization for egg production. As in previous studies, M. spinolai individuals were exposed to T. cruzi from the first nymph stage on, hence increasing the chance of parasite-insect competition.
In consequence, it is likely that final insect size results from a tradeoff involving a higher energy allocation to insect survival rather than reproduction (Botto-Mahan, 2009).
One of the most frequently observed patterns in insect reproduction is size-assortative mating, that is, the preferential mating between similar sized individuals. Most explanations to this pattern base on mate choice through sexual selection (Arnqvist, 2011;Arnqvist & Rowe, 2005;Baldauf, Kullmann, Schroth, Thünken, & Bakker, 2009;Gagnon & Turgeon, 2011). However, size-assortative mating may also occur due to mating constraints when males and females differ sufficiently in body size (Crespi, 1989;Han, Jablonski, Kim, & Park, 2010). In the study system, T. cruzi reduces the body size of male and female kissing bugs (Figure 1; Botto-Mahan, 2009).
In this way, infected-uninfected pairs had body size mismatches that probably translated into poor physical contact and limited sperm transfer during copulation. Infected-infected pairs, like uninfecteduninfected pairs, may not experience size-related mating constraints.
This may explain (1) why infected males sired more nymphs when crossed with infected females than when crossed with uninfected females ( Figure 2d) and (2) why uninfected males sired more nymphs when crossed with uninfected females than they did when crossed with infected females (Figure 2d). The body size hypothesis may also relate to the ability of females to detect and avoid parasites by directly assessing the male infection status through visual, tactile, or olfactory detection, or indirectly through detection of a signal sensitive to infection, such as body size (David & Heeb, 2009). In this study, uninfected females had low egg production and reproductive investment when crossed with infected males (Figure 2). If male body size indicates quality, females lose out by copulating with small-sized males with high parasite loads, incompatible genotypes, or lacking direct resources to offer.
We have presented evidence that T. cruzi influences reproduction through both genders. Even though our study was not designed to inquire into the mechanisms involved in parasite-induced fitness impact, it is likely that parasites impose a direct cost on female reproduction by reducing resource allocation to reproduction. The effect of T. cruzi on males, however, is less clear. There is some evidence for hemipterans that the quantity and quality of the seminal fluid depend on male environment (Kaldun & Otti, 2016). to the understanding of size-assortative mating in insects. A variety of factors have been suggested to cause size-assortative mating, including mate preferences, mate availability, and mating constraints (Crespi, 1989;Han et al., 2010;Nuismer, Otto, & Blanquart, 2008;Thomas et al., 1995). Results of this study suggest that we should add infection status to this list of variables. The extent to which parasite-induced size reduction affects size-assortative mating in natural populations needs to be assessed in future studies.

CONFLICT OF INTERESTS
We have no conflict of interests.

AUTHOR CONTRIBUTIONS
C.B.M. and R.M. designed the study; C.B.M. and V.C. collected data; C.B.M., V.C., and R.M. analyzed data and participated in manuscript writing. All authors gave final approval for publication.

DATA AVAILABILITY
Raw data are available on the following repository: https://figshare. com/s/5d2e40f59316d1679e1f.