Patterns of host–parasite association may vary across the landscape in part because of host and parasite diversity, divergence, local ecology, or interactions among these factors. In central coastal California, we quantified parasite prevalence, infection intensity, and diversity in two sister species of woodrats (Neotoma fuscipes and Neotoma macrotis) where the species co-occur (sympatry) and where each species exists alone (allopatry). In feces from 50 adults we identified seven taxa: the protozoans Eimeria, Giardia, and Cryptosporidium, the nematodes Trichuris, Aspicularis, and Eucoleus, and a cestode in the family Anoplocephalidae. Gastrointestinal parasite infection intensity and diversity were higher in males than in females, a difference that was most pronounced in the more aggressive N. fuscipes. Both species had lower infection intensity in sympatry than in allopatry and in sympatry the two species did not differ in infection intensity in total but did maintain distinct parasite communities. Taken together, our findings suggest that host evolutionary differences, including perhaps species-specific patterns of aggressive behavior, as well as local ecology, influence the likelihood of infection by these endoparasite taxa.
Identifying the ecological settings that lead to parasite diversity and diversification is central to understanding the role parasites play in the population dynamics and evolutionary potential of their host communities. One setting that may uniquely affect both parasite diversity and diversification is where closely related host species come into secondary contact after differentiation in allopatry (isolation from one another). The divergence of host lineages while in isolation may have profound effects on the divergence of their associated parasites and pathogens (Hafner and Nadler 1990; Nieberding and Olivieri 2006). When diverged host lineages come into secondary contact, there is opportunity for mixing of their respective parasite and pathogen communities. Such mixing may lead to augmented parasite and pathogen diversity in sympatric locations (areas of host overlap) vs. allopatric locations (Pavesi 2005; Nieberding and Olivieri 2006). Moreover, if local ecology overrides host divergence, there may be greater similarity in the parasite community and total infection intensity between sympatric hosts than among allopatric sites within a single host lineage (Ezenwa 2003). Alternatively, when divergent hosts come into contact, they may retain different parasite communities or susceptibilities to infection (Raeymaekers et al. 2013). As such, despite a shared ecological setting in sympatry, phylogenetic divergence of the hosts may be the predominant factor determining the parasite community and infection intensity of each host species. Although great effort is required in any system to distinguish the effects that phylogenetic history and shared host ecology have on parasite communities, zones of contact between closely related hosts are rich arenas in which to further our understanding of host genotype–parasite interactions (Ezenwa 2003; Thrall et al. 2006).
The woodrat sister species Neotoma fuscipes (the dusky-footed woodrat) and Neotoma macrotis (the large-eared woodrat) come into secondary contact and hybridize in the Salinas Valley of California (Matocq 2002a, b; Fig. 1). Woodrats are known for the large stick houses they build, either free standing or within the shelter of boulders (Linsdale and Tevis 1951; Smith et al. 2000; Shurtliff et al. 2013). Their houses, or middens, are the site of all major life activities including the hoarding of food, rearing of young, and, most likely, mating (Linsdale and Tevis 1951; M.D.M. pers. obs.). These houses provide a stable microhabitat for many vertebrates and invertebrates (Carraway and Verts 1991 and references therein) and woodrats have been implicated in maintaining pathogens on the landscape including the bacterial and protozoal parasites Borrelia spp. (Lane and Brown 1991), Anaplasma spp. (Nieto et al. 2010), Theileria spp. (Kjemtrup et al. 2001), and Ehrlichia spp. (Nicholson et al. 1999). The diversity and prevalence of gastrointestinal parasites is much less well studied in woodrats, so here, we quantify these parasite communities in both woodrat species where they occur in sympatry and from localities where each species exists in allopatry. On the basis of the prevalence, abundance (infection intensity), and diversity of parasite ova and oocysts detected in these woodrats, we addressed the following questions: First, what is the composition of the gastrointestinal communities of N. fuscipes and N. macrotis in the central coastal California? Second, are infections in sympatric populations of each species more similar to one another than either is to an allopatric population? And third, do different taxonomic groups of intestinal parasites show distinct spatial or species-specific patterns of diversity or abundance?
MATERIALS AND METHODS
Sample collection and identification
We trapped woodrats in three regions in central coastal California (September through November 2011; Fig. 1). Allopatric N. fuscipes were trapped in the interior Diablo Range (36°22′5.1″N, 120°49′53.3″W) and allopatric N. macrotis ~60 km west in the Santa Lucia Mountains (36°14′0.8″N, 121°29′5.1″W). Sympatric individuals of both species were trapped ~80 km south along the Nacimiento River on the Camp Roberts Military Reservation (35°46′30.8″N 120°47′36.2″W). Traps were set at all active woodrat houses along ~500-m transects for one to three nights and only adult individuals were retained. Woodrats were individually housed in standard shoebox cages (43 × 24 × 20 cm) and offered food and water ad libitum. Collection of fecal pellets (minimum ~20 cm3) began immediately postcapture from each individual to ensure that detected endoparasites would reflect those in wild woodrats. Pellets were collected from the cage floor and stored at −20 C until analysis. Animals not retained for further research were released at their point of capture. All trapping and care followed guidelines from the American Society of Mammalogists (Sikes et al. 2011), under protocols approved by the Institutional Animal Care and Use Committee of the University of Nevada Reno, and with permits issued by the California Department of Fish and Wildlife.
We processed fecal pellets from 50 adults, 25 allopatric (13 N. fuscipes, 12 N. macrotis) and 25 sympatric (12 N. fuscipes, 13 N. macrotis). To avoid bias, samples were processed randomly, blind to species and region but stratified by these parameters. Slides were prepared by suspending 1 g of feces from each individual in a hyperosmotic sugar solution using centrifugation (Foreyt 2001). We quantified parasite ova and oocysts using a modified McMaster's method, reading each slide with a Zeiss phase-contrast light microscope across a grid, one row at a time. Each slide was scanned once at 100× to identify helminth ova and once at 400× to identify protozoa. We used morphology to identify taxa to genus (Henry 1932; Chandler 1945; Voge 1946; Foreyt 2001; Haukisalmi and Rausch 2006; Zajac and Conboy 2006; Haukisalmi and Henttonen 2007). From these scans, we determined a count (per gram of feces) of each taxon present for each woodrat.
We compared the endoparasites detected by woodrat region, species, and sex using univariate and multivariate methods. Prevalence was defined as the proportion of hosts infected with a given parasite taxon within a woodrat group; Fisher's exact test was used to compare prevalences. Infection intensity was defined as the number of oocysts or ova of a given taxon per infected woodrat host (i.e., abundance). We used analysis of variance (ANOVA) to compare infection intensity and parasite taxon richness (number of species), diversity (Simpson's D, richness accounting for relative abundance of each species), and evenness (equitability of species abundance; McCune and Grace 2002). Counts of oocysts and ova were log-transformed to meet the assumptions of parametric analysis. To combine richness and abundance information, we used Simpson's index, D, a widely used and robust analysis because it is based on calculation of the variance of the species distribution (Magurran 2004). Evenness (E) was determined from D as E = (1/D)/S (Kent 2012). Full ANOVA models included all interactions, but we report reduced models (noted FR,df,df) when they provided a better fit (R2adj-reduced>R2adj-full). We report exact P in the text (and ranges in figures) and consider P<0.05 as statistically significant.
We ordinated the parasite communities using nonmetric multidimensional scaling (NMS) with the relative Sorenson distance measure. The modeling procedure, based on ranked distances, uses a search algorithm to determine the fewest dimensions that maximize data fit (minimize stress) and avoid local minima (“autopilot,” PC-ORD, v6, MjM Software, Gleneden Beach, Oregon, USA). To compare groups statistically, we used a nonparametric multiresponse permutation procedure (MRPP), appropriate because we could not achieve multivariate normality despite log transformation (McCune and Grace 2002).
Morphology of parasites detected
We detected parasites of seven taxa: three protozoan genera (Eimeria, Giardia, and Cryptosporidium), one cestode family (Anoplocephalidae), and three nematode genera (Trichuris, Aspicularis, and Eucoleus). Of the protozoans, the Eimeria oocysts were ellipsoidal (Fig. 2), averaging 21 × 30 µm, with a smooth, transparent cyst wall. Four sporocysts filled the interior of the sporulated oocysts without overlap. Their morphology and host origin are most consistent with E. neotomae (Henry 1932). The Giardia cysts (Fig. 2) were ellipsoidal, averaging 16 × 20 µm, slightly larger than those infecting other rodents (Adam 2001). Cysts had thin walls, contained two or four nuclei, and had one to two slender, linear, intracytoplasmic flagella. Cryptosporidium were rare and had the smallest oocysts, averaging 6 × 6 µm.
The cestode ova were ovoid, averaging 60 µm in diameter, with a visible pyriform apparatus (Fig. 2). Morphology, and host identity and origin best match the Anoplocephalidae, common in North American rodents (Haukisalmi and Rausch 2006). We did not sample adult tapeworms from woodrats, which is required for accurate cestode species identification.
Trichuris (Fig. 2) were the most abundant nematode ova found. These were narrow, averaging 36 × 85 µm, with conspicuous bipolar plugs. Their morphology and host origin best match T. neotomae (Chandler 1945). The remaining nematodes, Aspicularis and Eucoleus, were much rarer in the woodrats. The Aspicularis ova were ellipsoid, averaging 34 × 87 µm, with a double shell wall containing an undifferentiated embryo. The Eucoleus ova were asymmetric, averaging 37 × 75 µm, with bipolar plugs and a multicelled embryo (Zajak and Conboy 2006).
Prevalence and infection intensity by region, species, and sex
Protozoan endoparasites were the most prevalent, with infection rates by Eimeria and Giardia ≥40% in both species followed by the cestodes and Trichuris nematodes at 34% and 20%, respectively. Cryptosporidium, Aspicularis, and Eucoleus were rare, found in one woodrat each. Males had a higher prevalence of nematodes and cestodes than females, the only significant differences in prevalence by group (Fig. 3). We did not detect any parasites in five animals (three N. macrotis, two N. fuscipes). We calculated each woodrat's pooled prevalence as the average of its presence–absence values across the seven parasite taxa. This value was not significantly higher in allopatric than in sympatric woodrats (28% vs. 23%; FR,1,46 = 1.5, P = 0.221), marginally higher in N. fuscipes than N. macrotis (30% vs. 22%; FR,1,,46 = 3.6, P = 0.066), and higher in males than in females (32% vs. 19%; FR,1,46 = 8.8, P = 0.005). No interactions among region, species, and sex were detected (F1,42≤ 2.1, P≥0.151).
Infection intensity by all parasites was higher in allopatric than in sympatric woodrats (FR,1,46 = 6.4, P = 0.015), marginally higher in N. fuscipes than in N. macrotis (FR,1,46 = 2.7, P = 0.109), and higher in males than in females (FR,1,46 = 7.9, P = 0.007, Fig. 4). There were no interactions among region, species, and sex (F1,42 ≤1.8, P≥0.188).
When the infection intensity of protozoans, cestodes, and nematodes was considered separately, the pattern was more complex. Infection intensity differed by region in protozoans (allopatry>sympatry), differed by species in nematodes (N. fuscipes>N. macrotis), and differed by sex in nematodes and cestodes (males>females; Fig. 4). In protozoans, infection intensity was marginally higher in allopatric than in sympatric woodrats (F1,42 = 3.6, P = 0.065), but there was no difference by species or sex (F1,42≤0.2, P≥0.648). In nematodes, intensity did not differ by region (F1,42 = 0.0, P = 0.938), but was higher in N. fuscipes than in N. macrotis (F1,42 = 4.4, P = 0.042) and much higher in males than females (F1,42 = 16.0, P<0.001). A sex-by-species interaction (F1,42 = 6.9, P = 0.012) showed that the sex-difference in nematode infections was more pronounced in N. fuscipes (male vs. female: 12.9 vs. 0.0) than in N. macrotis (1.1 vs. 0.2). For the cestodes, there was no difference by region (FR,1,46 = 1.5, P = 0.229) or species (FR,1,46 = 0.0, P = 0.953), but males had a higher average intensity of infection than females (FR,1,46 = 4.0, P = 0.052).
Parasite richness, diversity, and evenness by region, species, and sex
Endoparasite richness was higher in males than in females (FR,1,46 = 8.8, P = 0.005) and marginally higher in N. fuscipes than in N. macrotis (FR,1,46 = 3.6, P = 0.066), but did not differ in woodrats by region (FR,1,46 = 1.5, P = 0.221, Fig. 5A). Region, species, and sex did not interact to affect richness (F1,42≤2.1, P≥0.151). The woodrats did not differ in parasite diversity (Simpson's D) or evenness (E) by region, species, sex, or the interactions between these effects (F1,32≤1.5, P≥0.227, Fig. 5B, C).
The NMS ordination—on the basis of 44 woodrats harboring the four moderately abundant parasites—did not show clear separation in the endoparasite communities by region or species, but males and females had distinct communities (Fig. 6). Three dimensions captured most of the variance in the parasite communities, with higher dimensions offering no improvement. When the parasite communities were compared by permutation (MRPP with T-test statistic), results supported the ordination. There was no difference by region (T = 0.94, P = 0.863) or species (T = 0.90, P = 0.841), but males and females harbored distinct parasite communities (T = 3.1, P = 0.014). Females had less heterogeneity than males (Sorenson's distance among females = 0.63 vs. among males = 0.71). Within regions, the two species also did not differ in parasite communites (allopatry, N. fuscipes vs. N. macrotis: P = 0.757; sympatry, N. fuscipes vs. N. macrotis: P = 0.931).
We used Sorenson distance (1 − Sorenson similarity) to qualitatively compare the parasite communites by region and species (Table 1). The parasites in N. macrotis were most similar: 22% distant from allopatry to sympatry. In contrast, N. fuscipes communities were more distinct: 53–73% distant from other N. fuscipes or N. macrotis. Sympatric N. fuscipes had the most unusual parasites by this measure, 70–73% distant from allopatric and sympatric N. macrotis.
Neotoma fuscipes and N. macrotis of the central coastal California maintain a diverse community of parasites. Protozoan parasites of the genera Eimeria and Giardia were the most common in the sampled woodrats, and both Giardia, which was widespread, and Crytosporidium, which was rare, had not been previously documented in these rodents. Although the parasite communities of the two species were similar (Fig. 6B), infection intensities were higher in allopatry than in sympatry and in males than in females. The parasite communities harbored by these woodrats appear to be the result of deep historical associations between the hosts and parasites, influenced by the current ecological setting of each population and sex-specific factors.
Overall parasite prevalence and diversity
The most common parasites in both woodrat species by fecal analysis were coccidia (Eimeriidae). Coccidian parasites including the cyst-forming species (Sarcocystis neotomafelis and Toxoplasma gondii) have been observed in N. micropus from the southern US and Mexico (Galaviz-Silva et al. 1991; Charles et al. 2012). However, coccidian parasites in N. fuscipes and N. macrotis have not been widely reported outside of the initial description of E. neotomae in N. fuscipes from Northern California (Henry 1932).
Giardia was the most prevalent parasite that is associated with zoonotic disease, and the second-most common organism found in our samples. Cryptosporidium, also potentially zoonotic, was rare. Waterborne parasites have been documented in many terrestrial mammals (Applebee et al. 2005). Considering the cosmopolitan host range of Giardia species and many coccidia, it is not surprising that they would infect woodrats.
In addition to protozoans, anoplocephalid cestodes and nematodes (Trichuris) were prevalent. Anoplocephalidae utilize rodents as definitive hosts in North America and two species have been reported to parasitize Neotoma in the western US (Miller and Schmidt 1982; Haukisalmi and Rausch 2006). Andrya neotomae was first described from N. fuscipes collected in Monterey County, California (now N. macrotis sensu Matocq 2002a, b; Voge 1946). The second, Paranoplocephala primordialis (originally A. primordialis), has been found in the small intestines of N. fuscipes (now N. macrotis sensu Matocq 2002a, b; Voge 1955). Although P. primordialis is a generalist parasite of arvicoline and sciurid rodents in North America, neither of these anoplocephalids has been described as parasitizing N. macrotis definitively (Rausch and Schiller 1949; Voge 1955). In Northern California, Chandler (1945) found N. fuscipes infected by Trichuris nematodes. We are unaware, however, of documented infection by this nematode farther south in N. fuscipes or within the range of N. macrotis.
The four most abundant parasites in our sample (Fig. 2) were found in both woodrat species. The pervasiveness of these four genera suggests that woodrats are likely susceptible throughout their ranges. Furthermore, mixed infections with protozoans such as Giardia spp. and helminths, especially Trichuris species, are also common in rodents worldwide (Fuehrer et al. 2012 and references therein). Although common in small mammals, Trichuris spp. are often host specific and resistant to transmission across genera. For example, the transmission and persistence of T. arvicolae is associated with communal, as opposed to solitary, nesting rodent species (Sanchez et al. 2011). Although woodrats typically share their houses only with their own young, nests are maintained across generations, allowing for parasite ova to be concentrated. In our study, both N. fuscipes and N. macrotis appeared to be infected with the same species of helminth, T. neotomae, although identification of the adult whipworms is required for confirmation. Indeed, more precise identification of all parasites described herein will be possible through further morphologic and DNA-based analyses. Identifying any health or fitness consequences of these infections also requires further investigation.
Male-biased patterns of infection
Male woodrats had higher prevalence and infection intensity of nematodes and cestodes than did females (Figs. 3, 4, and 6). Males also supported a richer, more even (lower E) parasite community than did females (Fig. 5). Consistent with this pattern, among individuals, males harbored more heterogeneous communities than did females (i.e., interindividual distances were higher). Sex-biased infestations have been repeatedly reported for a variety of parasites of vertebrates (e.g., Poulin 1996; Zuk and McKean 1996; Krasnov et al. 2005; Krasnov and Matthee 2010). In the majority of mammals, parasite prevalence, infection intensity, and species richness are higher in males than in females (Soliman et al. 2001; Rossin and Malizia 2002; Hillegrass et al. 2008). Some mammalian hosts do not exhibit this pattern, but where data are available in rodents, males consistently have higher infection intensities than females (Wirsing et al. 2007; Hillegrass et al. 2008; Krasnov and Matthee 2010).
One hypothesis explaining male-biased parasitism correlates differences in infection intensity to sex differences in movement patterns. Males may defend larger territories than females or have higher contact rates with other hosts during reproductive activity. Both factors may increase contact with infective stages of parasites and thereby facilitate transmission (Brown et al. 1994). Likewise, males with larger home ranges may encounter more parasite-dense areas (Ims 1987; Brei and Fish 2003; Nunn and Dokey 2006). Higher host–parasite encounter rates would contribute to both higher infection intensity and greater among-individual heterogeneity, as we have documented. Perhaps also contributing to greater heterogeneity among males is that for a given locality, males are likely drawn from a larger and more diverse pool of regional populations and habitats than females because males typically disperse farther than females (Matocq 2004; Matocq and Lacey 2004; McEachern et al. 2009). Interestingly, the relative philopatry of females also means that locally, females are more genetically homogeneous than males (Matocq and Lacey 2004). As such, the lower interindividual genetic distance among females compared with males parallels our observed pattern of greater similarity among females than males in parasite communities.
Another source of sex-biased patterns of parasitism may be androgen hormones. Testosterone can act as a “double-edged sword,” favoring large body size and secondary sexual characteristics but also having an immunosuppressive effect (Ezenwa et al. 2012). The “immunocompetence handicap” hypothesis argues that male sexual traits lower ability to resist infection through steroid suppression of the immune system (Folstad and Karter 1992).
Allopatry vs. sympatry
In comparisons of allopatric with sympatric sites by parasite taxon, woodrat hosts from sympatric sites had a lower infection intensity of protozoal parasites. It is possible that the ecology of our sympatric site somehow lessens the opportunity for infection by protozoans, especially with pathogens such as Giardia, which can be transmitted through contaminated water or terrestrially (Sulaiman et al. 2003). At the sympatric site, the relatively large and fast-flowing Nacimiento River may be a more dilute source of waterborne pathogens than the small, intermittent streams that characterize the allopatric sites.
A second, nonexclusive possibility is that lower overall infection intensities in sympatry may result in part from altered contact rates between conspecific hosts because of the presence of heterospecifics, who may compete for similar foraging areas, denning sites, and mates. Finally, we may be missing some species-specific patterns between woodrat hosts and their parasites in sympatry because of the relatively coarse taxonomic scale of parasite identification that we used.
Despite their being evolutionarily distinct lineages, we found few differences in parasite prevalence between N. fuscipes and N. macrotis. The only exception was that N. fuscipes had higher nematode infection intensity than N. macrotis. This could result in part from higher levels of aggression in N. fuscipes than in N. macrotis, a behavioral pattern we have observed in the field and laboratory, particularly in males (P.J.M., M.D.M., unpubl.). A greater tendency for aggression in N. fuscipes, especially males, may lead to more contact among individuals and parasite transmission. Asymmetries in aggressive behavior between sympatric woodrats have been documented between other species (Dial 1988; Shurtliff et al. 2013). A significant sex-by-species interaction demonstrates a greater difference between male and female nematode infections in N. fuscipes than in N. macrotis, which is consistent with the pattern of aggressive behavior we have observed in these woodrats (P.J.M., M.D.M., unpubl.).
Beyond the difference in nematode infection intensity, the similarity in parasite communities in the two woodrat species and among populations is striking. On the basis of species identity and abundance the two sampled N. macrotis populations are highly similar in their endoparasites, with a Sorenson similarity index of 0.78 (identical = 1). Despite the distance (~80 km) and likely microhabitat differences between these two populations, their similar parasite communities suggest that N. macrotis interacts with infective parasites in a consistent way. However, if site-specific ecology was the only force driving infection, we would expect N. fuscipes in sympatry with N. macrotis at Camp Roberts to harbor a similar parasite community. However, the parasite communities infecting N. macrotis and N. fuscipes at Camp Roberts were quite distinct (0.30 similarity). The parasite community hosted by N. macrotis at Camp Roberts was similarly distinct from that infecting N. fuscipes in the Diablo Range, ~80 km away (0.34 similarity). Habitat differences likely play a role, because the community found in N. fuscipes from Camp Roberts is also quite dissimilar from that in N. fuscipes in the xeric Diablo Range (0.37 similarity). We suspect there might be substantial differences in the parasite communities between the relatively mesic coastal and xeric interior ranges. Nonetheless, even in sympatry, N. macrotis and N. fuscipes interact quite differently with the same resident parasites.
In our preliminary investigation of host–pathogen/parasite patterns in this woodrat system, we find evidence that suggests that these patterns are determined by both evolutionary differences between the hosts and site-specific ecology and behavior. In sympatry, the two closely related hosts interact with the local parasites distinctly, in contrast to the relative constancy of the host–parasite signal we detect across a large spatial expanse in one of the two species (N. macrotis across the Santa Lucia Mountains). Sex-biased patterns of infection are also important here, which suggests a critical role of behavior in parasite maintenance and transmission. Clarifying the roles that evolutionary history and local ecology play in parasite prevalence and infection intensity will require increased spatial and genetic sampling of the parasite fauna. This is especially true in the context of potential hybridization between these woodrat species where they meet on the landscape.
For assistance in the field we thank J. Bender, A. Hollingsworth, C. Feldman, and M. Gritts. For assistance in the laboratory we thank M. Gritts, J. Hsueh, and E. Tobin. We thank the staff and administration of the Camp Roberts Military Reservation for access to their site and the Hollister office of the Bureau of Land Management for access to the Clear Creek Management area. Funding for this research was provided in part by a National Science Foundation grant (M.D.M., DEB-0952946) and the College of Agriculture, Biotechnology and Natural Resources at the University of Nevada, Reno.