Parasite preferences for large host body size can drive overdispersion in a fly-mite association

Parasite abundance (number of parasites an individual host harbours, including 0) is rarely spread evenly among hosts (Reiczigel et al., 2019). Typically, a minority of hosts harbour the majority of parasites and by extension experience the bulk of the deleterious physiological and fitness impacts of infection (Poulin, 2007). This “overdispersion” can ultimately influence the ecology of host populations by concentrating the costs of infection in a subset of the host population (Jaenike, 1996, Poulin, 2007). Larger hosts often harbour more parasites than smaller hosts (Poulin and Morand, 1997, Watkins and Blouin-Demers, 2018, Hechinger et al., 2019). This trend can be observed within and between host species. Interspecifically, both helminth and louse burdens are higher in larger host species (Poulin and George-Nascimento, 2007, Chu et al., 2019). Among perch, larger fish harbour more internal parasites (Zelmer and Arai, 1998). However, few studies have aimed to disentangle the effects of age and accumulation of parasites over time from the preferences of parasites.

Both host traits and parasite traits can contribute to heterogeneities in infection (Shaw and Dobson, 1995). For example, host immunity affects the possibility of initial infection and/or the ability of parasites to proliferate following infection. Larger hosts may also have more biomass, space, or energy to support parasite reproduction (Poulin and George-Nascimento, 2007; Hechinger et al., 2013). Furthermore, many organisms gain mass as they age, and older organisms have more time to accumulate infections. On the other hand, many mobile parasites can distinguish between potential hosts of the same species and selectively infect hosts based on sex, caste, mating status, etc. (Christe et al., 2007, Cervo et al., 2014, Horn et al., 2020). In previous experiments, ectoparasitic mites (Macrocheles subbadius) disproportionately infected female Drosophila nigrospiracula over male conspecifics in pair-wise choice tests (Horn et al., 2020). However, whether this preference plays a role in population level aggregation is currently unclear (Polak and Markow, 1995).

Since larger hosts have more resources (biomass, energy, etc) to support parasite reproduction (Poulin and George-Nascimento, 2007; Hechinger et al., 2013), parasites may have adapted to seek out larger hosts. For example, Poecilochirus mites disproportionately infect larger host beetles (Grossman and Smith, 2008). If mites can detect and preferentially infect larger hosts, the resulting heterogeneity in infection could contribute to overdispersion of parasites in the host population. Alternatively, hosts also vary intraspecifically in their behavioural resistance against ectoparasites (Polak, 2003). Insect body size can correlate with physical endurance; for example, larger worker bees have better overall flight performance and range, potentially due to higher energy reserves (Kenna et al., 2021). Behavioural resistance against mite infection in a fly host is limited by endurance and can be exhausted by energetically demanding activity (Polak, 2003, Luong et al., 2007). Therefore, larger flies may have stronger behavioural resistance and harbour fewer ectoparasites than smaller flies. Mass, therefore, could exert negative or positive effects on ectoparasite abundance. In this study we tested three hypotheses i) mites selectively infect larger hosts in pair-wise choices, ii) larger hosts have stronger resistance (using climbing endurance as a proxy measure of resistance; Horn and Luong, 2021), and iii) mite preferences for larger hosts results in heterogeneity in mite abundance.

We tested these hypotheses using the Drosophila nigrospiracula (Diptera: Drosophilidae) Patterson and Wheeler - Macrocheles subbadius (Mesostigmata: Macrochelidae) (Berlese) association. Drosophila nigrospiracula is a cactiphilic fly which feeds and reproduces on rotting Saguaro gigantea (Johnston and Heed, 1976). Macrocheles subbadius is a hemolymph-feeding facultative ectoparasite of D. nigrospiracula, that also uses flies for dispersal between rotting cactus habitats (Polak 1996). Flies use energetically demanding behavioural defenses, e.g. grooming, kicking, jumping, to fend off mites, and flies that perform better in negative geotaxis endurance assays are more resistant to mite infection (Polak, 2003, Horn and Luong, 2021). Infection negatively impacts the survival, mating, and reproductive success of flies (Polak, 1996). Natural M. subbadius populations are overdispersed, and the degree of overdispersion increases as the necrotic cactus habitat dries out (Polak and Markow, 1995). Since Macrocheles are capable of distinguishing between potential fly hosts (Campbell and Luong, 2016, Horn et al., 2020), this system provides an opportunity to test if parasite preferences can influence heterogeneity of infection, separate from age-mediated parasite accumulation.

Fly and mite cultures are fully described in Horn and Luong (2021). Because mites preferentially infect mated and female flies, we conducted assays using mated females (Horn et al., 2020). Adult flies were collected from stock bottles (Instant potato and formula 4–24 Instant Drosophila Medium; Carolina Biological Supply Company, NC USA) and moved to agar vials at a 1:1 ratio of males to females for at least 72 h before experiments. Adult female mites, the infectious stage, were collected from cultures using Berlese funnels for use in the experiments below. Flies were raised separately from mites; thus, they did not have an opportunity to acquire infections with age. Data from the following experiments and associated R scripts are available via Open Science Foundation (doi: 10.17605/OSF.IO/29F85). Statistical analysis was performed in R Version 4.2.1 (R Core Team 2022, R-project.org/). Fig. 1 was made in Microsoft excel and Fig. 2, Fig. 3 were made with the ggplot2 and the ggeffects packages (Wickham, 2016, Lüdecke, 2018).

The first experiment tested if mites preferentially infect flies based on host size while eliminating differences in resistance. Female D. nigrospiracula were weighed (±0.005 mg; Mettler Toledo XP105 balance, OH, USA), and pairs were made by matching flies in the bottom third of flies by mass with the top third of flies. Flies were tethered to cotton with Elmer’s Rubber Cement (Elmer’s Brand, USA), eliminating differences in resistance (Campbell and Luong, 2016). Fly pairs were placed into Y-mazes, then a single adult female mite was introduced to each Y-maze, and the ends were sealed with cotton. After 1 h under an opaque box, we recorded which fly the mite selected, if either. A binomial test on the Y-mazes in which infection occurred was used to test if mites preferred high or low mass flies (binom.test, Stats, Ho: proportion=0.5).

In 21 of 43 (49%) pairwise choice tests, the mite infected a fly. Among trials where infection occurred (n=21), 18 (86%) mites infected the heavier fly and three (14%) infected the lighter fly (Fig. 1). The heavy fly was significantly more likely to be infected (binom.test, P=0.0015). The mean mass of the heavy fly was 3.00±0.04 mg (mean ± S.E.M.), and the mean mass of the light fly was 2.22±0.05 mg. In absolute terms the average difference between the heavy and light flies was 0.78±0.06 mg, and the average percent difference was 30%. The smallest difference was 0.44 mg or 16%. The largest difference was 1.36 mg or 54%. Differences in mass and mite choice for each pair of flies with an infection is visualized in Fig. 1.

Second, we tested if larger flies had stronger endurance than smaller flies using negative geotaxis assays. Fly endurance was defined as the number of times a fly could ascend a vial following knockdown and is a proxy measure of ectoparasite resistance that avoids the confounding variable of mite preference (Horn and Luong, 2021). Two fly cohorts were used: in one cohort flies were aged 30-32 days post-eclosion and in the other they were aged 10-12 days. These times represent the approximate period of peak reproduction and post-peak reproduction (Luong and Polak, 2007). After aging, individual flies were weighed then placed into individual vials and given 1 h to acclimatise to the new environment. Vials were marked 5 cm above the base. Flies were induced to climb by tapping the vial to the table, knocking the fly down. When flies climbed to the 5 cm mark the knockdown was repeated. A fly was considered exhausted when it did not ascend the vial within 15 s of being knocked down; the number of times a fly reached the mark was recorded (cycles climbed). Generalised linear regression was used to test if there was a significant relationship between mass and the number of cycles climbed, if fly age affected climbing endurance, as well as to test for an interaction between mass and age (glm.nb function, link=log, Ho: b=0).

The mean fly mass was 2.99±0.05 mg and flies climbed a mean 16.2 (12.0-20.8, 95% bootstrapped confidence interval (CI)) cycles (n=96); this excludes two flies from the 10–12 days old cohort heavier than 4 mg which had substantially deformed abdomens. The interaction between mass and age was not a significant predictor of climbing endurance (cycles climbed) (z=0.66, P=0.51), the regression coefficient of the interaction was 0.020 (95% CI: -0.038-0.077) (summary, confint). Mass was not a significant predictor of cycles climbed in a model including the mass*age interaction (z=-0.46, P=0.65), regression coefficient = -0.36 (95% CI: -1.90-1.18); nor was mass a significant predictor of cycles climbed in a model without the mass*age interaction, i.e. the simplified model Cycles∼Mass+Age (z=0.49, P=0.63), regression coefficient = 0.14 (95% CI: -0.41-0.66). Hence endurance, a proxy measure of fly resistance, was not affected by fly mass (Fig. 2). Age was also not a significant predictor of cycles climbed in the interaction model (z=-0.62, P=0.54), nor the simple model (z=0.24, P=0.81); the regression coefficients were -0.055 (95% CI: -0.23-0.12) and 0.0032 (95% CI: -0.023-0.029), respectively. However, this may be a consequence of inter-cohort differences masking effects of aging.

The third experiment tested the relationship between host body size and mite abundance (number of mites an individual fly harbours, including 0). Female flies were exposed to mites in infection microcosms (120 mL glass jars). These infection jars simulate semi-natural conditions whereby flies are surrounded by mites in an infectious environment but are free to groom and defend themselves. Flies were aged to 28+ days prior to infection assays. Jars (n =13) were filled with mite media, and a vertical depression was formed using a tongue depressor. Flies were then placed into the depression and the jar was sealed with 000 mesh that allows air flow but not mites or flies to cross. The mite cultures in our laboratory generally contain ∼30-60 mites/75 mL of media based on Berlese funnel estimates (personal observations). Thus, these volumes meant there were 2-4 times more mites than flies in each jar. Infection jars were incubated for 24 h, after which flies were retrieved from the jars via aspiration (total of 106 flies recovered). Flies were scored (number of mites) and immediately weighed. Mites were not removed prior to weighing, to limit damage to the fly body, but previous research found the larger mite Macrocheles muscaedomesticae (which is nearly twice the size of M. subbadius) is ∼20 times smaller than flies by mass (Durkin et al., 2019). Parasite abundance was modelled using generalised linear mixed effect models (glmmTMB, family=nbinom2, link=log Ho: b=0), with mass as a fixed variable as well as mite line and infection jar as random effects. We also tested if heavier flies were more likely to be infected by modelling infection status (infected=1 or uninfected=0) with mass as a fixed variable as well as mite line and infection jar as random effects (glmmTMB, family= binomial, link=logit, Ho: b=0).

The mean mass of all flies recovered from infection jar experiments was 2.61±0.05 mg (n=106), and 43 out of 106 (41%) flies recovered from the jars were infected with at least one mite. The mean mass of uninfected flies was 2.56±0.06 mg (n=63), whereas the mean mass of infected flies was 2.68±0.08 mg (n=43). Mean mite abundance across all flies was 1.02 (0.68-1.4, 95% bootstrapped CI) mites. Mass was a significant predictor of mite abundance (z=5.3, P<0.0001, Fig. 3a); estimated regression coefficient was 1.66 (95% CI: 1.04-2.27). We also tested if heavier flies were more likely to be infected with parasites. Mass was also a significant predictor of infection status (z=2.07, P=0.038, Fig. 3B); the estimated regression coefficient was 1.23 (95% CI: 0.07-2.39). Heavier flies were more likely to be infected and harboured more mites on average.

We set out to test three hypotheses about the relationship between host body mass and ectoparasite infection outcomes: i) mites selectively infect larger hosts in pairwise choices; ii) larger hosts have stronger resistance against ectoparasites; and iii) mite preferences for larger hosts generates heterogeneity in mite abundance. Pair-wise choice experiments showed mites selectively infected heavier female flies over lighter females (Fig. 1). Furthermore, mites disproportionately infected larger hosts even when flies were able to groom/resist infection in mesocosms (Fig. 3). Our results are consistent with the hypothesis that mites preferentially infect larger hosts, and the hypothesis that mite preferences generate variation in mite abundance.

When flies were restrained to eliminate differences in resistance, M. subbadius infected the larger fly in 86% of Y-mazes (Fig. 1). This preference appears stronger than previously reported preferences of M. subbadius for female D. nigrospiracula over male flies (71%) as well as mated females over unmated females (65%) (Horn et al., 2020). Because female D. nigrospiracula are larger than conspecific males, the strength of the preference observed here suggests that the previously reported preference of M. subbadius for female D. nigrospiracula over males was driven by sexual dimorphism in size (Horn et al., 2020). Our results here and reported previously (Fig. 1; Horn et al., 2020) are consistent with observations showing wild caught females generally harbour more mite infections than wild caught males (Polak and Markow, 1995). It may be fruitful for future research to simulate fly and mite populations, accounting for sex and/or size differences, to evaluate which preference best explains the overdispersion observed in field data (Polak and Markow, 1995).

A previous study found M. subbadius preferentially infect D. nigrospiracula which already harbour conspecific mites (Brophy and Luong, 2022). This preference for already infected flies increases the importance of the first mite choosing between naïve hosts, since it may lead to “runaway infections”. This bias may contribute to the uneven infection abundances observed in our infection jar experiments (Fig. 3a). Mites are disproportionately likely to infect large naïve flies (Fig. 1), which then also become targeted by additional mites selecting infected hosts.

Why Macrocheles mites prefer larger flies over smaller flies when selecting a host is not currently known. Both M. subbadius and M. muscaedomesticae feed on fly tissue, specifically hemolymph (Polak, 1996, Campbell and Luong, 2016, Durkin et al., 2019). Although these mites are facultative parasites, M. subbadius females that feed on a fly host lay more eggs than free-living conspecifics (Luong and Subasinghe, 2017). Larger flies may have more available resources (e.g. nutrients in the hemolymph), and mites may gain a fitness benefit from choosing higher mass hosts (Hechinger et al., 2019). Future research could test if mites that feed on larger hosts produce more eggs than mites that feed on smaller hosts. Similarly, larger flies may be more likely to successfully disperse from cactus rot habitats even when infected; a possibility that could be tested using flight mill experiments (Luong et al., 2007). Alternatively, larger hosts may give off more cues (respiratory, pheromones, etc) that attract mites (Grossman and Smith, 2008, Cervo et al., 2014). Olfactory experiments could determine if mites are more attracted to larger hosts even when they cannot make direct contact.

Body size in insects, including Drosophila, is generally correlated with survival and reproductive success (Neems et al., 1990, Honěk, 1993, Polak, 1996, Beukeboom, 2018). Furthermore, there is substantial, heritable, variation in Drosophila mass (Prout and Baker, 1989; Liebowitz and Fontdevila, 1995), as well as Drosophila ectoparasite resistance (Polak, 2003). Why variation persists despite the reproductive benefits of body size could be explained by an infection-size trade-off. Larger flies may be disproportionately infected due to parasite preferences for higher mass hosts (Fig. 1). Heavy mite infection (3+ infections) can reduce the fecundity and survival of female D. nigrospiracula by over 50% (Polak, 1996), and increased mite burdens with mass may trade off with beneficial effects of body size (Fig. 3A). The strength of this trade-off may be dependent on the environment, specifically the size of the mite population. Experimental evolution could test if mite exposure or removal of mites over generations affects the average body size of fly populations.

Although body size and overall fitness (in terms of reproduction and overall survival) are positively correlated in insects, the effect of body size on mobility (e.g. climbing, flight, and endurance generally) is more ambiguous (Luong et al., 2007, Brown et al., 2017, Kenna et al., 2021). Flight endurance in bees was positively correlated with body size (Kenna et al. ,2021). Contrarily, smaller Batocera rufomaculata fly longer than larger conspecifics, although this may be adaptive phenotypic plasticity (Brown et al., 2017). Early emergence in response to environmental stress leads to smaller overall beetle size but also disproportionate investment in flight muscles which may help escape deteriorating environments (Brown et al., 2017). Luong et al. (2007) found that size was not a significant predictor of hovering endurance in young D. nigrospiracula, and here we extend this observation later in the fly life span. Polak (1996) found larger D. nigrospiracula females lived longer, thus we anticipated larger flies to be in better health at older ages; however, we did not detect a significant interaction between mass and fly age on cycles climbed. The differences between previous literature as well as our data here (Brown et al., 2017, Kenna et al., 2021) suggest that the relationship between mass and endurance may be taxon-dependent, which could be the subject of a future meta-analysis.

Our results are consistent with the general trend of size-biased infection (Zelmer and Arai, 1998, Poulin, 2007); i.e. larger hosts harbour more parasites on average. In our infection jar mesocosms, larger flies acquired more infections despite having no opportunities to acquire parasites over their lifespan. Since body size did not affect fly endurance positively or negatively, size differences in ectoparasite resistance were also not a contributor. Thus, parasite preference was the primary driver of variation in parasite abundance observed here (Fig. 3). Understanding the role parasite preferences play in shaping heterogeneity in infection may help explain population properties such as overdispersion and parasite-mediated population suppression. In turn, this may extend our understanding of size-biased infections to diverse host-parasite associations.

留言 (0)

沒有登入
gif