Effects of anthropogenic habitat disturbance and Giardia duodenalis infection on a sentinel species' gut bacteria

Abstract Habitat disturbance, a common consequence of anthropogenic land use practices, creates human–animal interfaces where humans, wildlife, and domestic species can interact. These altered habitats can influence host–microbe dynamics, leading to potential downstream effects on host physiology and health. Here, we explored the effect of ecological overlap with humans and domestic species and infection with the protozoan parasite Giardia duodenalis on the bacteria of black and gold howler monkeys (Alouatta caraya), a key sentinel species, in northeastern Argentina. Fecal samples were screened for Giardia duodenalis infection using a nested PCR reaction, and the gut bacterial community was characterized using 16S rRNA gene amplicon sequencing. Habitat type was correlated with variation in A. caraya gut bacterial community composition but did not affect gut bacterial diversity. Giardia presence did not have a universal effect on A. caraya gut bacteria across habitats, perhaps due to the high infection prevalence across all habitats. However, some bacterial taxa were found to vary with Giardia infection. While A. caraya's behavioral plasticity and dietary flexibility allow them to exploit a range of habitat conditions, habitats are generally becoming more anthropogenically disturbed and, thus, less hospitable. Alterations in gut bacterial community dynamics are one possible indicator of negative health outcomes for A. caraya in these environments, since changes in host–microbe relationships due to stressors from habitat disturbance may lead to negative repercussions for host health. These dynamics are likely relevant for understanding organism responses to environmental change in other mammals.

associated with greater ecological overlap between humans, wild animals, and domestic species, increasing the potential for disease transmission from humans and domestic species to vulnerable wildlife (Faust et al., 2018;Goldberg et al., 2008). More recently, research suggests that these factors can influence mammalian physiology via interactions with the gut microbiome (Stumpf et al., 2016). For example, variation in diet across habitats are associated with differences in the gut microbiome of wild animals (Amato et al., 2013;Benítez-Malvido & Martínez-Ramos, 2003;Greene et al., 2019). Interactions among several host species within a shared environment may also facilitate the transmission of both commensal and pathogenic microbes, thus modifying a host's microbial community structure Moeller et al., 2013;Rwego et al., 2008oldberg, 2008. For example, humans, livestock, and nonhuman primates living in fragmented forests in Uganda share similar strains of Escherichia coli, highlighting the potential for bacterial transmission in sympatric environments . Given that the gut microbiome is known to affect host nutrition, metabolism, immune function, and behavior (Flint et al., 2012;Heijtz et al., 2011;Hooper & Macpherson, 2010), changes to its structure as a result of any of these pathways are likely to have substantial impact on host physiology and ultimately, reproductive success and survival.
In the context of disturbance-induced ecological overlap, the relationship between gut bacteria and pathogens is of particular interest. Gut bacteria may affect host susceptibility to intestinal pathogen infection and influence the progress of pathogenic infection and clinical manifestation of disease (Berrilli et al., 2012;Costello et al., 2012;Koch & Schmid-Hempel, 2011;Partida-Rodríguez et al., 2017). Alternatively, parasites may alter gut bacterial community composition, which could lead to alterations in host health (Barash et al., 2017;Berrilli et al., 2012;Cantacessi et al., 2014;Peachey et al., 2017;Šlapeta et al., 2015). There is limited research on the interactions between gastrointestinal parasites and gut bacteria and, of those, many focus on mice, amphibians, or humans (Barash et al., 2017;Berrilli et al., 2012;Cantacessi et al., 2014;Cooper et al., 2013;Jani & Briggs, 2014;Kreisinger et al., 2015;Lee et al., 2014;Shu et al., 2019). However, overall, the literature suggests that host-parasite-gut bacteria interactions are system-specific. For example, increased microbial diversity is generally believed to indicate good host health (Costello et al., 2012). Cryptosporidium spp. infection in captive Coquerel's sifaka has been associated with decreased gut microbial diversity as well as bacterial taxa linked to dysbiosis (McKenney et al., 2017). In contrast, domestic cats infected with the protozoan Giardia duodenalis have higher microbial species richness compared to uninfected individuals (Šlapeta et al., 2015).
Whether changes in microbial diversity due to pathogenic infection result in changes in host health and phenotype remain to be seen.
Given that habitat disturbance often also affects parasite prevalence and abundance patterns in wild mammals Gillespie et al., 2005;Zommerset al., 2013), parasite-bacteria relationships may be a key factor for understanding mammalian health outcomes in anthropogenically disturbed habitats (Cantacessi et al., 2014;Cooper et al., 2013;Kreisinger et al., 2015;Lee et al., 2014;Zaiss & Harris, 2016). Habitat disturbance may impact interactions between hosts and their associated microbial communities, which may then lead to downstream effects on host physiology and health, including nutritional deficits, higher prevalence of pathogens, and lower gut microbial diversity (Amato et al., 2013;Barelli et al., 2015;Estrada et al., 2017).
To improve our understanding of this interplay, we used the black and gold howler monkey, Alouatta caraya, as a model for exploring host-parasite-gut bacteria interactions in response to habitat disturbance. Primates are a relevant model in which to address these questions due to their large variation in habitat use, diet ecology, and interspecies interactions. As the most abundant primate species in northeastern Argentina, A. caraya is simultaneously a sentinel of ecosystem health and a model organism (Kowalewski & Gillespie, 2008, 2018Kowalewski et al., 2011). For example, A. caraya experience high morbidity and mortality associated with yellow fever and thus serve as an early warning system prior to human outbreaks (Holzmann et al., 2010;Oklander et al., 2017).
Anthropogenic activities in Argentina, such as deforestation and selective logging, are forcing A. caraya into ecological overlap with humans and domestic animals. In such disturbed habitats, A. caraya interact with humans and domestic animals in multiple ways, including crossing terrestrially from forest patch to patch, sharing the same water sources as cattle, and engaging in altercations with domestic dogs (Kowalewski et al., 2011;Raño et al., 2016). These interactions may lead to greater susceptibility to zoonotic diseases and greater sensitivity to gut dysbiosis via horizontal microbial transmission.
We examined the effect of ecological overlap with humans and domestic species-a proxy for anthropogenic habitat disturbanceand infection by the protozoan parasite Giardia duodenalis on the gut bacteria of A. caraya. A previous study by Kowalewski and colleagues found a high prevalence of G. duodenalis in A. caraya populations across a gradient of disturbance (Kowalewski et al., 2011). As a result, we hypothesized that A. caraya in more disturbed habitats, with increased contact with humans and domestic animals, would have a higher infection prevalence of G. duodenalis.
Additionally, research across animals, including amphibians, fish, and other species of Alouatta, indicate that habitat disturbance is one factor that is associated with differences in gut bacterial community composition (Amato, 2016;Amato et al., 2013;Huang et al., 2018;Sullam et al., 2012). Therefore, we hypothesized that anthropogenic disturbance would be associated with differences in the A. caraya gut bacterial community. Due to stressors that accompany habitat disturbance, such as changes in diet or decreased home range , we predicted A. caraya in more disturbed habitats would have decreased gut bacterial diversity.
Finally, as clinical manifestations of G. duodenalis have not yet been observed for A. caraya and given reported interactions between parasites and gut bacteria (Barash et al., 2017;Berrilli et al., 2012;Cantacessi et al., 2014;Cooper et al., 2013;Jani & Briggs, 2014;Kreisinger et al., 2015;Lee et al., 2014;Shu et al., 2019), it is crucial to understand how G. duodenalis affects the A. caraya gut bacteria and potentially A. caraya health. We hypothesized that A. caraya infected with Giardia would harbor a different gut bacterial community than uninfected individuals. In particular, based on studies of other primates (McKenna et al., 2008;McKenney et al., 2017), we predicted that Giardia infection in A. caraya would be correlated with decreased bacterial diversity and changes in bacterial community composition.

| Study site
This study was conducted in the winter (June-July) of 2016 and 2017 across three sites in the Corrientes and Chaco provinces in northeastern Argentina (  (Kowalewski et al., 2011). All habitats are prone to flooding, and flash floods have been occurring more frequently in recent years, which could affect parasite prevalence and transmission. Notably, in April-May 2017, prior to our second period of sample collection, 600 mm of rainfall across 3 days was recorded, leading to severe flooding.

| Sample collection
Fecal samples were collected from 61 individuals (remote = eight individuals, rural = 29 individuals, village = 24 individuals) noninvasively and opportunistically during the 2016 and 2017 winter seasons, following previous protocols (Gillespie, 2006). Sex, age class (infant, juvenile, subadult, or adult), social group, and habitat type of each individual were noted (Table S1). For the most part, social groups, and thus individuals, were sampled only once during the study. We sampled between five and seven social groups per site, and almost all individuals from each social group were sampled and sequenced. Individuals were sampled for both Giardia presence and gut bacterial community characterization. For each sample, one gram of fecal matter was immediately homogenized in sterile cryovials in RNAlater nucleic acid stabilizing buffer (Ambion, Life Technologies, Grand Island, NY) for Giardia analysis, and another gram was immediately homogenized in sterile fecal vials with 95% ethanol for microbiome analysis. All samples were stored at 4°C until transport to the USA for processing, and all research procedures were approved by Emory University and Northwestern University (Northwestern IACUC Field Research 2019-001) and complied with applicable laws in Argentina. Import and export permits were obtained from the Centers for Disease Control (CDC) and Argentina's Ministerio de Ambiente y Desarrollo Sostenible, respectively.

| Giardia duodenalis analyses
All Giardia analyses were conducted at Emory University, Atlanta, Georgia. The CDC provided technical assistance with all Giardia protocols. Due to the heterogeneity of G. duodenalis, a multilocus approach was utilized to target three genes: glutamate dehydrogenase (gdh), beta-giardin (bg), and triosephosphate isomerase (tpi). DNA was extracted from the RNAlater-preserved fecal samples using the FastDNA Spin Kit for Soil (MP Biomedicals LLC), and the multilocus genes were amplified using a nested PCR protocol adapted from Roellig et al. (2015). Briefly, all PCR reactions were prepared in a final volume of 25 μl containing 1× Taq PCR Master Mix (Qiagen), 400 ng/ μl BSA, 500 nM of each primer, nuclease-free water, and genomic DNA (2 μl in first PCR reaction and 2 μl of first reaction product in the second PCR reaction) (amplification conditions listed in Table S2). Positive and negative controls were included in each reaction. The presence of Giardia infection was verified by running 5 μl of PCR product on a 1% agarose gel.

| Microbiome analyses
DNA was extracted from the ethanol-preserved fecal samples using the Qiagen Powersoil Kit at Northwestern University, Evanston, Illinois. The V4 region of the 16S ribosomal RNA gene was amplified using a modified version of the Earth Microbiome Project protocol (Mallott & Amato, 2018;Thompson et al., 2017) and the 515Fa/926R primer set (Walters et al., 2016). Extraction and negative controls were both included in the sequencing run. We barcoded and pooled all amplicons in equimolar concentrations for sequencing on the Paired-end sequences were joined and processed using QIIME2 v2019.7 (Bolyen et al., 2019). Quality filtering and the removal of chloroplast and mitochondria sequences resulted in a total of 856,786 reads with an average of 13,387 reads per samples. The DADA2 algorithm was used to cluster amplicon sequence variants (ASVs), and taxonomy was assigned by comparing ASVs to the GreenGenes 13.8 reference database. When we compared taxonomic assignments between Greengenes and SILVA, we found that the Greengenes database assigned more features (2,568) than SILVA (2,559) did for this particular dataset. All samples were rarefied to 8,000 reads per sample based on alpha rarefaction curves (Figure S1a-c). Alpha diversity (Faith's phylogenetic distance, Shannon diversity index, and number of observed ASVs) and beta diversity (unweighted and weighted UniFrac distances) were calculated in QIIME2.  these years compared to our data from 2017 ( Figure 1) (Kowalewski et al., 2011;Kuthyar, unpublished data).

| Habitat disturbance and gut bacteria
Bacterial community composition varied across habitats (Figure 2;

| Habitat disturbance and Giardia duodenalis
Habitat disturbance leads to increased contact between wildlife, livestock, and humans, which increases the risk for zoonotic disease

| Habitat disturbance and gut bacteria
Habitat type was also associated with variation in the A. caraya gut bacterial community, where different habitats were associated with different gut bacterial community compositions. Multiple factors may contribute to these patterns, including diet, intra-and interspecies contact, and the physical environment (Mccord et al., 2014).

Several bacterial taxa varied between individuals living in rural
and village habitats, including Lachnospiraceae, Erysipelotrichaceae, Ruminococcaceae, Bacteroides uniformis, and Prevotella copri-a pattern that may be associated with diet. Previous studies have shown that differences in diet composition, and thus nutrient availability, across habitats influence the types of gut bacteria residing in animal hosts (Amato et al., 2013;Barelli et al., 2015;Trosvik et al., 2018).
The same bacterial taxa in which we detected variation in our study showed a response to changes in diet and/or habitat in other studies.
High abundances of Prevotella have been found in vervets and macaques eating a Western diet (Amato et al., 2015;Ma et al., 2014), F I G U R E 4 Significant differences in bacterial richness (observed ASVs) between Giardia-infected and -uninfected Alouatta caraya individuals red colobus monkeys in protected forest in Tanzania (Barelli et al., 2015), and black howler monkeys consuming a leafy diet (Amato et al., 2014(Amato et al., , 2019. Ruminococcaceae relative abundances are reported to be higher in red colobus monkeys in disturbed habitats (Barelli et al., 2015), in black howler monkeys during seasons of low energy intake (Amato et al., 2014), and in frogs with low dietary diversity in farmland habitats (Chang et al., 2016). Similar dynamics between these bacterial taxa and diet may be operating in our system. In A. caraya, there are differences in overall seasonal patterns of food availability as well as the availability of new and mature leaves across habitats (Kowalewski & Zunino, 2004), and lower food abundances have been recorded in rural and village habitats compared to remote habitats (Zunino et al., 2001). However, since we did not collect diet data in this study, it is unclear whether previously documented differences in food availability correlate to differences in actual food consumption across sites. Future studies should incorporate data on A. caraya diet to see which dietary factors (i.e., leaf and fruit abundance, rate human food scrap consumption, proportion of the diet that is fruit, or amounts and ratios of specific nutrients consumed) drive these changes in bacterial taxa across habitats.
Interestingly, there was no difference in bacterial richness or diversity across habitat types in A. caraya. This finding is in contrast to other studies in a range of animals. For example across various species of primates, including another species of howler monkey, habitat fragmentation has a modest but significant effect on gut bacterial diversity (Amato et al., 2013;Barelli et al., 2015;Mccord et al., 2014). Additionally, house sparrows in urban environments were characterized by decreased microbial diversity and fewer metabolic functions (Teyssier et al., 2018). These patterns are often attributed to variation in diet since diet diversity is often linked to gut microbiome diversity (Heiman & Greenway, 2016), and habitat disturbance often alters diet diversity as a function of a decline in plant species diversity in the habitat and food availability. Indeed, reduced dietary diversity is associated with reduced bacterial diversity in both red colobus monkeys and howler monkeys in fragmented and/or disturbed habitats (Amato et al., 2013;Barelli et al., 2015). However, habitat disturbance and fragmentation may not impact the interaction between diet and gut bacterial structure in the same way across all species. For example, vampire bats exhibit less diverse diets in fragmented habitats but show no changes in gut bacterial diversity between habitats (Ingala et al., 2019). Furthermore, in the disturbed rural and village habitats of this study, A. caraya diet composition may be more diverse compared to diets observed in remote habitats (M. Kowalewski, personal observation) as a result of edge effects and secondary forest succession after selective deforestation (Kowalewski & Zunino, 1999). Future research should include data on actual food consumption to further explore the interaction between diet diversity and gut microbial diversity.
Beyond diet, patterns of microbial transmission may also contribute to the cross-habitat differences in bacterial community composition identified in this study. Evidence of these dynamics has been reported previously. For example, gorillas that lived in ecological overlap with humans and livestock in Rwanda harbored similar strains of E. coli to those of humans and livestock compared to gorillas that did not overlap . Further, a study in Uganda found that Cryptosporidium spp. isolates looked genetically identical regardless if they came from a human, nonhuman primate, or livestock source (Salyer et al., 2012). Here, as high ecological overlap exists among A. caraya, humans, and domestic animals in the anthropogenically disturbed habitats, A. caraya in fragmented rural and village habitats may be more exposed to cross-species bacterial transmission, thus altering their bacterial community structure, than at remote habitats, where they are mostly isolated from humans and livestock.
Finally, the physical environment and the resulting exposure to environmental bacterial pools may shape a host's gut bacteria across habitats. In a previous study, soil was found to best predict the gut bacteria of baboons in Kenya (Grieneisen et al., 2019

| Gut bacteria and Giardia duodenalis
As hypothesized, G. duodenalis infection was associated with differences in gut bacterial community membership. Past studies in frogs found that parasitic infection led to altered skin bacterial communities and even increased the pathogen load in some cases (Jani & Briggs, 2014;Shu et al., 2019). Certain bacterial taxa, such as Lachnospiraceae, Anaeroplasmataceae, and Ruminococcaceae, that varied with Giardia infection are all associated with providing crucial metabolic services to the host (Biddle et al., 2013;Petzel et al., 1989), and changes in their abundance could have impacts on howler health beyond susceptibility to infection and symptoms of disease. The modified bacterial community structure, along with living in a disturbed habitat and harboring a parasite, could further alter the nutritional, and thus overall health, status of A. caraya in these contexts.
Bacterial richness was also significantly associated with infection, where infected individuals had fewer observed ASVs on average compared to uninfected individuals. It is possible these reductions in diversity could enhance the potential for colonization by potentially pathogenic microbes, such as Enterococcus sp. (Iebba et al., 2016). However, more research is necessary. Previously, parasite infection was associated with decreased microbial diversity in birds in the Galapagos, where initial infection led to changes in host behavior, leading to further susceptibility to future infection (Knutie, 2018). However, infection with chytrid fungus in frogs was associated with increased bacterial diversity (Becker et al., 2017). In past studies, Giardia has been associated with bacterial overgrowth as well as changes in the relative abundance of certain taxa, including increased abundance of Proteobacteria and decreased abundance of Firmicutes and Bacteroidetes (Barash et al., 2017;Halliez & Buret, 2013;Müller & Von Allmen, 2005).
Additionally, dogs infected with Giardia had significant differences in their bacterial community structure compared to uninfected dogs (Šlapeta et al., 2015). Overall, this study indicates that anthropogenic habitat disturbance influences multiple groups of mammalian gut microbes and that these microbes may also affect each other. Understanding the influences these interactions have on host health outcomes in disturbed habitats will be important for conservation efforts moving forward.
Habitat disturbance may not only affect host health through shifts in food availability and quality, but also indirectly through changes to host-associated bacterial communities and susceptibility to parasite infection. Further, integration of the gut microbiome into a disease ecology framework will be crucial in understanding the intrinsic and extrinsic factors that impact host survival and reproduction.
As a sentinel species of this semideciduous ecosystem, A. caraya plays a crucial role in advising wildlife health surveillance of increasingly fragmented and disturbed habitats. Their behavioral plasticity and dietary flexibility allow them to exploit a large range of habitat conditions, resulting in resilience in fragmented and disturbed habitats (Kowalewski et al., 2011;Miner et al., 2005). As habitats are becoming harsher due to anthropogenic pressures, however, previously resilient wild animals may be reaching their limits of plasticity.
In addition to more obvious factors, the understudied interactions between habitat disturbance, pathogen transmission, and a host's gut bacteria could negatively influence host health, reproductive output, and even survival. As such, monitoring both the host's gut bacterial community and pathogen load could be an important noninvasive method to improve understanding of the effects of anthropogenic disturbance on wild mammal health and inform conservation strategies. Recursos Naturales, and the Dirección de Parques y Reservas de la Provincia de Corrientes for logistical support and permission to conduct this investigation. The authors thank L. Ragazzo and J. Deere for assistance with laboratory and statistical analyses. K.R.A is supported as a CIFAR Azrieli Global Scholar.

DATA AVA I L A B I L I T Y S TAT E M E N T
Scripts for QIIME2 and R can be found at https://github.com/Krama to-lab/kuthy ar_giardia. The datasets used for analyses can be accessed via Dryad at https://doi.org/10.5061/dryad.wm37p vmj5. All sequences have been uploaded to SRA and will be publicly accessible with publication (PRJNA607274).