Loss of protozoan and metazoan intestinal symbiont biodiversity in wild primates living in unprotected forests

In light of the current biodiversity crisis, investigating the human impact on non-human primate gut biology is important to understanding the ecological significance of gut community dynamics across changing habitats and its role in conservation. Using traditional coproscopic parasitological techniques, we compared the gastrointestinal protozoan and metazoan symbiont richness of two primates: the Udzungwa red colobus (Procolobus gordonorum) and the yellow baboon (Papio cynocephalus). These species live sympatrically in both protected and unprotected forests within the Udzungwa Mountains of Tanzania with distinct ecological adaptations and diets. Our results showed that terrestrial and omnivorous yellow baboons had 2 (95% CI 1.47–2.73) and 3.78 (2.62–5.46) times higher gut symbiont richness (both including and excluding rare protozoans) compared to the arboreal and leaf-eating Udzungwa red colobus in unprotected and protected forest, respectively. We also found a consistent depletion of symbiont richness in red colobus living in the unprotected forest fragment compared to the continuous protected forests [the latter having 1.97 times (95% CI 1.33–2.92) higher richness], but not in yellow baboons. Richness reduction was particularly evident in the Udzungwa red colobus monkeys, confirming the pattern we reported previously for gut bacterial communities. This study demonstrates the impact of human activities even on the microbiodiversity of the intestinal tract of this species. Against the background of rapid global change and habitat degradation, and given the health benefits of intact gut communities, the decrease in natural gut symbionts reported here is worrying. Further study of these communities should form an essential part of the conservation framework.

www.nature.com/scientificreports/ restoring bacterial diversity 17 ). Thus, ignoring the conservation of this gut component means neglecting a whole set of biological relationships essential for survival, since host-gut symbiont interactions are among the more prevalent ecological and evolutionary drivers of biological diversity and ecosystem composition 18,19 . Several authors have suggested that such gut symbionts should be recognized as meaningful conservation targets along with their hosts 20 21 .
As a result of habitat degradation and/or contraction, studies have revealed that distribution and population size of most animal hosts decreases. If we assume that environmental changes due to human impact similarly affect gut symbionts, we would expect a decrease in prevalence and richness of gut symbionts as well [22][23][24][25] . Thus, comparing gut symbiont communities in threatened animal hosts, such as non-human primates 26 living in both protected and human-impacted habitats, is of particular relevance for understanding the impact of conservation efforts and whole ecosystem health 27 .
The Udzungwa Mountains of Tanzania (hereafter referred as the Udzungwas) are one of 35 internationallyrecognized biodiversity hotspots (https ://www.conse rvati on.org) 28 ; however, intensive agriculture is currently encroaching on these relatively intact forests 29 , making non-human primates living outside protected areas particularly vulnerable 30,31 . We recently discovered that gut metazoan richness of the endemic Udzungwa red colobus monkey (Procolobus gordonorum) varies with altitude, with lower richness at lower altitudes 32 . Because human activities are concentrated at lower altitudes, we suggested that anthropogenic disturbance might play an active role in reducing gut symbiont presence and transmission. Therefore, in the present paper, we investigated gut symbiont richness of two sympatric primate species with contrasting life history traits. The endangered Udzungwa red colobus are relatively small canopy dwellers that feed predominantly on leaves 33 ; instead, the larger yellow baboons (Papio cynocephalus) of least conservation concern, live primarily on the ground and are generalist feeders, with diets including seeds, ripe fruits and animal prey. They are also frequent raiders of human food (i.e. crops and organic waste in villages 34 ). Because gut symbiont transmission are likely to depend on host habits and lifestyle, we compared the richness of gastrointestinal protozoans and metazoans of two cercopithecines with highly different ecological adaptations but living sympatrically in contrasting habitat type, predicting higher gut symbiont richness in terrestrial hosts rather than arboreal ones. Moreover, within host species, we predict gut symbiont richness to be higher in hosts living in protected rather than unprotected forest.
Gut symbiont richness across primate species. In both protected Mwanihana (MW) and unprotected Magombera (MA) forests, yellow baboons had higher mean symbiont richness compared to Udzungwa red colobus, both including and excluding rare protozoans (GLM: Z value: 7.081, df = 136, p value < 0.001; Fig. 3, Table 1; SI Fig. S1, SI Table S2). Based on Binomial GLM analysis, the prevalence of Strongyloides sp., strongylid and spirurid eggs, as well as cysts of Balantioides sp. were all significantly higher in baboons than in red colobus as reported in Table 2. Instead, Trichuris sp. was found in a higher proportion of red colobus samples compared to those of baboons (Binomial GLM: Z value = -1.992, df = 163, p value = 0.0463; Table 2). Shedding intensities revealed that baboons had higher mean eggs/cysts per gram than red colobus (SI Table S3 46] times more symbionts than Udzungwa red colobus monkeys sharing the same habitat type; while in MW, yellow baboons had about 2 (95% CI 1.47-2.7) times more gut symbionts than red colobus. Moreover, gut symbiont richness of red colobus was on average 1.97 times (95% CI 1.33-2.92) higher in individuals living in MW than in those inhabiting MA. In contrast, the same pattern was not observed in yellow baboons where the difference in symbiont richness between forests was not statistically significant (p value > 0.05; Fig. 4, SI Fig. S2). Gut symbiont distribution and prevalence varied with the taxon investigated. For example, in red colobus, a higher prevalence of Strongyloides sp. and Trichuris sp. were observed for individuals living in MW compared to MA, but not for the other three taxa (Fig. 2, Table 2; SI Table S1). In baboons, only a higher prevalence of Trichuris sp. (Fig. 2 www.nature.com/scientificreports/ Finally, modeling showed that overall, altitude influenced neither gut symbiont prevalence nor richness. Only the egg shedding intensity of Strongyloides sp. appeared to be influenced by this parameter (SI Table S4). www.nature.com/scientificreports/

Discussion
In this study, we investigated whether habitat degradation impact on protozoan and metazoan richness and prevalence of two wild primate species having different ecological adaptations (terrestrial or arboreal), dietary strategies (omnivorous or folivorous) and living in protected and unprotected forests within an Afrotropical biodiversity hotspot. Through the analysis of non-invasively collected fecal samples, hereafter we compare and discuss the possible reasons why gut symbiont richness and prevalence differ between and within primate hosts.
To avoid potential effects of seasonality [35][36][37] and altitudinal gradients 32 , fecal samples were collected during a very narrow temporal window from social groups living at similar altitudes; thus, any intraspecific differences should reflect the effects of habitat type. Despite the expected difference between species (with yellow baboons harboring a greater gut symbiont richness compared to the Udzungwa red colobus monkeys), within species, Udzungwa red colobus living in protected forest had a higher level of gut symbiont richness than those living in unprotected forest; this intraspecific difference confirms our previous results for gut bacteria 4,38 .
Although they have been evolving with their hosts for millions of years, until recently protozoan and metazoan gut symbionts have been mainly perceived as pathogens, considered harmful rather than integral to gut homeostasis and host health. However, since gut communities have now been recognized as one of the keystones of the human biome and beneficial against inflammatory mechanisms by manipulating the host immune system 39,40 , they are being considered relevant to the conservation status of their hosts 20 . This paper contributes to the growing body of knowledge revealing micro biodiversity loss in degraded versus intact habitats 4,9,38 , as well as the complex interplay between hosts, symbionts and/or microbiota 21,[41][42][43] .
None of the observed taxa of protozoans and metazoans found in this study were new to non-human primates 44 , having been noted previously for other red colobus [Ugandan red colobus, Piliocolobus tephrosceles 45,46 ; eastern black-and-white colobus Colobus guereza and Angolan black-and-white colobus C. angolensis 47 ] or other baboons [olive baboons, Papio anubis 48 ; guinea baboons, Papio hamadryas papio 49 ; yellow baboons, Papio cynocephalus 50 ]. We found yellow baboons had a higher gut symbiont richness and prevalence than Udzungwa red colobus. Although highly speculative because accurate data are not available, we believe that group density [i.e. number of individuals per group 51,52 ] is unlikely to be responsible for this result, as predicted in other animal hosts 53,54 . In fact, yellow baboons and Udzungwa red colobus populations in this study area are known to have comparable group sizes (approximately 40-60 individuals per social group). Body size has also been purported to explain gut symbiont richness and prevalence 55,56 , and since yellow baboons are larger than red colobus, this morphological difference could partially explain the higher richness in baboons. However, the two primate www.nature.com/scientificreports/ species also have very different behavioral traits that could account for the differences in gut symbiont diversities. The most likely driving forces here are diet and time spent on the ground, given the differences in specific gut symbionts' prevalence. The Udzungwa red colobus are predominantly arboreal, occasionally descending to the ground, while yellow baboons spend much of their time travelling on the forest floor and between nearby villages, where they feed on crops and/or on human food waste. These lifestyle differences would explain the higher prevalence of trophically transmitted gut symbionts like spirurid nematodes and dicrocoeliid trematodes in yellow baboons compared to red colobus: both these symbionts are transmitted indirectly through the ingestion of an intermediate host (often arthropods) 44 , and baboons consume more terrestrial arthropods in their diet than the mainly leaf-eating Udzungwa red colobus monkeys 51,52 . Similarly, fecal-orally transmitted protozoans from contaminated soil (e.g., Entamoeba coli, Entamoeba sp., Iodamoeba buetchlii), and associated  www.nature.com/scientificreports/  www.nature.com/scientificreports/ with domestic animals (Blastocystis sp.) were only found in yellow baboons. Another protozoan transmitted through contaminated water (Balantioides sp.), as well as two nematode taxa with soil-dwelling free-living stages acquired through the skin (Strongyloides sp. and strongylids) were also found significantly more frequently in yellow baboons than red colobus. That these taxa are more likely to be found in terrestrial rather than arboreal hosts has also been observed intraspecifically for chimpanzees Pan troglodytes; i.e. individuals that used the ground more frequently for locomotion were more infected 57 . The only difference in gut symbiont prevalence noted here that is not explained by the arboreal-terrestrial lifestyle was for Trichuris sp. Although this taxon is also soil-transmitted, this genus was more prevalent in arboreal Udzungwa red colobus rather than baboons. Another larger study specifically focused on on Trichuris sp. in non-human primates was also unable to identify the factors responsible for prevalence in particular primate species, but terrestriality was not one of them 58 . However, Trichuris sp. do appear to prefer dense shade which may explain the discrepancy 59 , since red colobus tend to live in deeper shader forest, while baboons spend much of their time on forest edges. The same peculiarity of Trichuris sp. may also explain why in both species, this genus was more prevalent in MW than MA.
Studies on the effect of human disturbance on intraspecific variation of gut symbiont richness and prevalence in wild animals still present very mixed results 60 . For example, several studies conducted on primates have found lower gut symbiont richness in intact, undisturbed habitats compared to disturbed and degraded ones (Tana River red colobus, Procolobus rufomitratus and mangabey, Cercocebus galeritus galeritus 54 ; lion-tailed macaque, Macaca silenus 61 ; redtail guenon, Cercopithecus ascanius 62 ). However, the same studies were unable to reveal correlations between richness and habitat type on other primate populations (red colobus, Piliocolobus tephrosceles and black-and-white colobus, Colobus guereza 62 ). Similarly, no correlation was found between prevalence and habitat type (lion-tailed macaque, Macaca silenus 61 ; Tana River red colobus, Procolobus rufomitratu 54 ; red colobus, Piliocolobus tephrosceles and black-and-white colobus, Colobus guereza 62 ). On the other hand, similar to our results, in the black howler monkey (Alouatta pigra) gut symbiont reduction was clearly associated with habitat degradation 63 , as well as in other mammals such as Australian skinks Lampropholis guichenoti 64 and several sigmodontinae rodent species (Akodon cursor, A. montensis and Oligoryzomys) from Brazil 65 .
The discrepancy in finding correlations between gut symbiont richness and habitat type within each primate species investigated might be explained by the different ecological adaptations of each species. It is likely that host species dependent on forest integrity for its diet like the leaf-eating red colobus are influenced differently from human disturbance. In fact, human presence and intensified activities in the unprotected MA forest may have played a critical role in shaping habitat features over time and consequently reducing gut symbiont presence and potential transmission. For example, logging may have changed the forest structure by increasing forest edges and fragmentation of canopy cover may have promoted increased exposure to wind and sun radiation facilitating drier conditions and reducing survival of free-living stages of gut symbionts, resulting in lower infections of individuals living in such altered habitats. Although speculative, the lower contamination of gut symbionts in degraded habitats, such as the small lowland forest fragment of Magombera, could also potentially be due to the proximity of sugar cane plantations 66 and cultivated fields which are often treated with pesticides, fertilizers and anti-helminthics.
Although not true for all animal species, the endangered Udzungwa red colobus monkeys have now been shown to suffer biodiversity loss at multiple levels in degraded, unprotected habitats. Not only are there lower population densities of this species in MA 30,67 , but lower taxonomic and functional diversity of gut bacteria were found in individuals living in this unprotected forest 4,38 . Furthermore, the present study confirms that gut symbiont richness is reduced in red colobus inhabiting MA. If such associations are confirmed in other specialist feeders, it is plausible that gut communities could be used as a tool for evaluating individual health and the conservation status of a species, especially in the case of threatened taxa where non-invasive fecal sampling is necessary 68 . However, further studies are required to identify potential interactions between gut components and between these components and host health before these relationships can be used as biomarkers for the wellbeing of individuals and species.  Table 3; see 69 for details). In the past 50-60 years, these two forests have been gradually separated from each other by at least 6 km of fields and villages 70 . Troops of Procolobus gordonorum, the endemic Udzungwa red colobus monkeys, and Papio cynocephalus, the yellow baboon live sympatrically in both forests. The Udzungwa red colobus are arboreal and mainly leaf-eating, whereas yellow baboons are terrestrial and omnivorous, feeding regularly on sugar cane as well as human organic waste 51  www.nature.com/scientificreports/ sively during a single defecation event from each group. Social groups sampled during the same day were located at least 2 km apart (more details for sampling procedures in 4,70,71 ). All samples were collected in 2016 during the same 4-week temporal window to avoid potential seasonal effects 36,37 . An aliquot of 2 g of fresh feces was preserved in 10 ml of 10% neutral buffered formalin and stored at ambient temperature (20-25 °C) until transport to the Fondazione E. Mach in Italy where they were kept at 4 °C before definitive shipment of aliquots to the University of Veterinary and Pharmaceutical Sciences in Brno, Czech Republic for parasitological analysis.

Methods
protozoan and metazoan symbionts examination. Host blood, tapeworm proglottids and adult nematodes that were clearly visible by eye in fecal pellets were noted in the field prior preserving the 167 samples. As recommended for parasitological analyses of non-human primates 44,72 , we used both modified Sheather's flotation and fecal sedimentation techniques to identify symbionts. First we extracted and weighed sediment  www.nature.com/scientificreports/ from the whole sample, then examined 2 ml of fecal suspension with optical microscopy at 400× magnification (Olympus CX40) for gut symbiont identification. Based on morphological characteristics 44 eggs and cysts were quantified for each symbiont taxon and the number of eggs or cysts per gram of sediment were calculated according to the following formula: n = N/m, where n = number of eggs/weight of sediment (g), N = number of eggs in examined amount of sediment and m = weight of examined sediment (g). Lastly, using what of the fecal pellet after extraction of sediment, adult and larval nematodes were collected using the 'gauze-washing' method 73 and observed via stereo microscopy at 8-40× magnification (OLYMPUS SZ51).

Data analyses.
Three standard indices of gut symbiont infection were estimated and used as dependent variables in the subsequent regression models: 1. gut symbiont prevalence (i.e. number of infected samples out of the total number of samples examined); 2. gut symbiont richness (i.e. number of different symbiont taxa per sample); 3. gut symbiont egg or cyst shedding intensity (i.e. number of eggs or cysts per gram of fecal sediment for Strongyloides sp. and Balantioides sp. only).
To assess if the prevalence of each symbiont varied with host species, forest type or the interaction of these two factors, a Generalized Linear Model (GLM) with Binomial distribution and logit link was carried out. Similarly, the association of symbiont richness with primate species, forest type or their interaction was studied using a GLM with Poisson distribution and log link. Because some protozoans (e.g. amoebas and Blastocystis) are difficult to detect with optical microscopy 74,75 , the models including symbiont richness were run twice: once with all protozoan taxa, and a second time excluding rare taxa, i.e. where rare is defined as being present in fewer than 10 samples (Iodamoeba buetschlii, Entamoeba coli, Entamoeba spp., Blastocystis sp.). Since results from a previous study on the same population of red colobus monkeys revealed that altitude may affect gut symbiont richness in this species 32 , the sampling design was limited to social groups of both species at low altitude (less than 450 m a.s.l.). The altitudinal range in the present study was much narrower (216-441 m) than in the previous investigation (216-1535 m 32 ). Therefore, we consider inclusion of altitude in the models by computing the Akaike Information Criteria (AIC) values for each model with or without altitude and decided not to include it (results in Supplementary Information). Gut symbiont diversity measures such as prevalence and richness may depend on infection shedding 76 . Thus, in infected fecal samples, a linear model (LM) was used to assess if the log-transformed egg or cyst shedding intensity for each individual host was associated with primate species or forest type (and their interaction), or altitude. Altitude was standardized (subtracted by its mean and divided by its standard deviation, so to have mean 0 and standard deviation 1). Assessment of the appropriateness of statistical assumptions for each model was carried out by computing overdispersion for GLM and graphically inspecting model residuals 77