American Journal of Primatology, 78(8), 883–892

Received: 30 October 2018 | Revised: 15 August 2019 | Accepted: 18 August 2019 DOI: 10.1002/ajp.23046

RESEARCH ARTICLE

Extensive variability in the gut microbiome of a highly‐specialized and critically endangered lemur species across sites

1Department of Ecology and Evolution, Stony Brook University, Stony Brook, New York Abstract 2Department of Biology, University of Deforestation continues to jeopardize Malagasy primates as viable habitats become Kentucky, Lexington, Kentucky smaller, more fragmented, and more disturbed. This deforestation can lead to 3Department of Anthropology, University of Illinois, Urbana, Illinois changes in diet, microhabitat, and gene flow between populations of endangered 4Department of Biological Anthropology, species, and it remains unclear how these changes may affect gut microbiome (GM) University of Antananarivo, Antananarivo, characteristics. The black‐and‐white ruffed lemur (Varecia variegata), which is among Madagascar ’ 5Department of Anthropology, Stony Brook Madagascar s most threatened lemur species, provides a critical model for under- University, Stony Brook, New York standing the relationships between historical and on‐going deforestation (habitat

Correspondence disturbance), feeding ecology, and GM composition and diversity. We studied four Mariah E. Donohue, Department of Biology, populations inhabiting two rainforests (relatively pristine vs. highly disturbed) in 101 Morgan Building, University of Kentucky, ‐ Lexington 40506, KY. southeastern Madagascar. We conducted full day focal animal behavioral follows and Email: [email protected] collected fecal samples opportunistically across a three‐month period. Our results

Funding information indicate that lemurs inhabiting sites characterized by habitat disturbance and low National Geographic Society, Grant/Award dietary diversity exhibited reduced gut microbial alpha diversity. We also show that Number: 9819‐15; Woese Institute for Genomic Biology; University of Illinois at these same factors were associated with high community dissimilarity using weighted Urbana Champaign; Directorate for Biological and unweighted UniFrac metrics. Finally, an indicator species analysis showed that Sciences, Grant/Award Numbers: 0820709, 0935347; Primate Conservation, Grant/ the most pristine site was characterized by an abundance of methanogenic archaea. Award Number: 74892 While it is impossible to disentangle the relative contributions of each confounding variable presented by our sampling design, these results provide crucial information about GM variability, thereby underscoring the importance of monitoring endangered species at the population‐level.

KEYWORDS dietary diversity, gut microbiome, habitat disturbance, lemur

1 | INTRODUCTION changes in the relative abundance and diversity of these microbes can trigger deleterious effects on individual health and population Elucidating the factors shaping gut microbiome (GM) patterning in viability (Brucker & Bordenstein, 2012; Dethlefsen, Eckburg, Bik, & wild populations enhances our ability to conserve and monitor Relman, 2006; Flint, Scott, Louis, & Duncan, 2012; Hooper, Littman, & endangered species. Hosts outsource many digestive and immuno- Macpherson, 2012; Sekirov, Russell, Antunes, & Finlay, 2010). logical functions to symbiotic gastrointestinal microbes (e.g., mem- Therefore, understanding variation in GM responses to ecological bers of the GM; Amato et al., 2013; Bauchop, 1971; Flint, Bayer, factors is of particular interest for species facing ongoing anthro- Rincon, Lamed, & White, 2008; Hird, 2017; Lambert, 1998), and pogenic habitat alteration (Amato et al., 2016; Stumpf et al., 2016).

Several studies show that diet plays a crucial role in shaping the 2 | METHODS GM throughout an individual’s life (Clayton et al., 2016; David et al., 2014; De Filippo et al., 2010; Trosvik, Rueness, de Muinck, 2.1 | Ethics statement Moges, & Mekonnen, 2018; Turnbaugh, Bäckhed, Fulton, & Gordon, The methods used for noninvasive fecal collections of wild primates 2008) and over evolutionary timescales (Gomez et al., 2015; were reviewed and approved by the Stony Brook University IACUC Groussin et al., 2017; Ley et al., 2008; Springer et al., 2017; Trosvik committee (IACUC #2016–2254—USDA—NF). Field data collection et al., 2018; Yildirim et al., 2010). Anthropogenic disturbance is protocols were also approved by MNP (Madagascar National Parks), expected to impact food availability in tropical habitats through the body governing research in Madagascar’s protected areas. This changes to microclimates that drive variation in species composi- research adhered to the American Society of Primatologists tion (Abbas et al., 2011; Arrigo‐Nelson, 2006; Herrera, Wright, Principles for the Ethical Treatment of Non‐Human Primates. Lauterbur, Ratovonjanahary, & Taylor, 2011). Some primate species mayactuallybenefitfromhabitat disturbance, as plant species diversity often increases and can provide a new array of food items 2.2 | Study sites (e.g. Colobus sp.:Chapman,Struhsaker,Skorupa,Snaith,&Rothman, 2010; Dunham, 2017; Cercopithecus sp.: Kaplin & Moermond, 2000; This study was conducted in the two southernmost habitats within Microcebus sp. and Cheirogaleus sp.: Crowley, Blanco, Arrigo‐Nelson, the V. variegata range: Ranomafana National Park and Manombo & Irwin, 2013). However, the ability to incorporate novel foods Special Reserve (Figure 1). Within each area, we identified a gradient likely depends on habitat, severity of disturbance, and the species’ of habitat disturbance based on local interviews, Landsat images, and degree of specialty. Accordingly, negative consequences of habitat published papers. disturbance on diet have been described in Propithecus edwardsii Ranomafana National Park (RNP; −21.27 latitude and 47.33 (Arrigo‐Nelson, 2006; Matos, 2017), Rhinopithecus bieti (Huang longitude; 800‐1,200 m), established in 1991, is a 43,500‐hectare et al., 2017), Hylobates lar and Presbytis melalophos (Johns, 1986), expanse of continuous montane rainforest in Madagascar’s Ateles geoffroyi (Chaves, Stoner, & Arroyo‐Rodríguez, 2011), and southeastern province of Fianarantsoa (Wright et al., 2011). Alouatta palliata (Asensio, Cristobal‐Azkarate, Dias, Vea, & Rodrí- Within RNP, groups were followed in a pristine site (Mangevo guez‐Luna, 2007; Dunn, Cristóbal‐Azkarate, & Veà, 2010), all of or RNP‐P) and a lightly disturbed site (Vatoharanana or RNP‐LD). which show reduced dietary diversity and increased foraging effort RNP‐PandRNP‐LD are located within RNP’ssouthern in smaller, more degraded forests. block, separated by 8.34 km of continuous primary rainforest Black‐and‐white ruffed lemurs (Varecia variegata) are a critically (Holmes et al., 2013). endangered primate species endemic to the eastern rainforests of Disturbance categories withinRNPweredesignatedbasedon Madagascar. As dietary specialists consuming primarily fruit—a rare park history. RNP‐P is a seldom visited section of RNP, containing trait in lemurs (Wright, 1999)—V. variegata are considered especially 425 km2 of pristine primary rainforest virtually unaffected by sensitive to anthropogenic disturbance (Balko, 1998; Britt, 2000; human activities such as logging and hunting (Baden, Webster, and Herrera et al., 2011; Ratsimbazafy, 2002) and are often the first Kamilar (2016); Mancini (2016); Matos, 2017). RNP‐LD was species to succumb to local extinction (White, Overdorff, Balko, & selectively logged in 1986, during which time over 1,000 trees Wright, 1995). Deforestation has fragmented and degraded much of were removed from the 3.25 km2 area and an unspecified number their range, creating extreme population differentiation with very were damaged (Lehtonen, Mustonen, Ramiarinjanahary, Niemelä, limited gene flow in southern localities (Baden et al., 2014). & Rita, 2001). While RNP‐LD is typically categorized as “pristine” In this study, we aim to understand GM variability across distinct or “relatively pristine” primary forest, here it is designated as V. variegata populations. We selected groups occupying forests with “lightly disturbed” or “recovering” because it is among the most different histories and intensities of habitat disturbance to determine disturbed sections of V. variegata’s range within RNP (Herrera if different populations exhibit distinct GM characteristics, with the et al., 2011). In addition, many of the tree species that were ultimate goal of using this data to inform conservation efforts. removed are known V. variegata food resources, including However, we acknowledge that many environmental factors—both Canarium madagascariensis (ramy), Ocotea sp. (varongy), Crypto- related to and divorced from habitat disturbance—can influence the carya sp. (tavolo), and Chrysophyllum boivinianum (rahiaka; Arrigo‐ GM, and therefore simply aim to describe GM variation within and Nelson, 2006), highlighting the disproportionate effect of habitat between populations. We sampled from two relatively pristine sites disturbance on this highly specialized species. Since 1986, RNP‐LD within Ranomafana National Park’s continuous southern rainforest has received very little human disturbance. block, and two sites connected by a narrow corridor with high rates Manombo Special Reserve (MSR; −22.95 latitude and 47.63 of human disturbance in Manombo Special Reserve. We combine longitudes; 40–120 m) is a complex of lowland rainforest and littoral feeding observations, habitat disturbance data, and 16S rRNA gene fragments located 187.21 km southeast of RNP. Forests in MSR are amplicon sequencing to understand the extent of GM variability in divided into separate administrative units with variable degrees of these four distinct V. variegata populations occupying the southern- conservation enforcement; here, we focus on Parcel I (some most extent of the species’ range. protection) and Parcel II (almost no protection). MSR was heavily DONOHUE ET AL. | 3of12

FIGURE 1 Map of study sites: (a) Ranomafana National Park (RNP) and (b) Manombo special reserve (MSR). For reference, Mangevo is RNP‐ P, Vatoharanana is RNP‐LD, MSR Parcel I is MSR‐D, and MSR Parcel II is MSR‐HD

impacted by Cyclone Gretelle, a category 4 storm that struck in II (MSR‐HD) as more disturbed than Parcel I (MSR‐D). Generally, January of 1997. Gretelle’s wind speeds of up to 245 km/h sheared RNP sites are considered far less disturbed than MSR sites. 80% of MSR’s lowland humid forest (Ratsimbazafy, 2002; Wright, 1999). The impacts of Gretelle continue to be augmented by intense 2.3 | Field methods anthropogenic disturbance (i.e., logging, slash‐and‐burn agriculture) in all parcels (Johnson et al., 2011). This study was conducted in 2016 throughout the Malagasy Landsat images taken from 1987 to 2017 reveal a total loss of winter, which is typically characterized by limited resource 2,000 hectares over 30 years in MSR. Today, less than 5,000 severely availability, low precipitation, and cold temperatures (Wright, fragmented hectares remain forested. The rate of deforestation is high 1999). Field data were collected 5 to 6 days per week in RNP‐LD throughout MSR, with elevated rates in Parcel II and the littoral forest. from June 1 to June 24 (n =80hr),RNP‐P from June 30 to July 11 MSR is being deforested at a rate of 2.9% per year, which is (n =254hr),MSR‐D from July 23 to August 10 (n =140.5hr),and higher than the national average of 1.1% per year (Vieilledent et al., MSR‐HD from August 11 to August 17 (n =54hr).Full‐day focal 2018) and the RNP rate (which is close to 0%). Based on animal behavior was recorded every 5 min by two independent deforestation rates and government protection, we designate Parcel observers (Altmann, 1974). When feeding or foraging was 4of12 | DONOHUE ET AL. observed, we recorded the plant species consumed and estimated 2.5 | DNA sequence generation V. variegata height from ground level in meters. Here, as in Holmes, DNA from the fecal samples was extracted using the QIAamp Gordon, Louis, and Johnson (2016), we treat feeding and foraging Powerfecal DNA Kit following the manufacturer’s protocol after as a single behavior because it was difficult to distinguish as they washing samples in 500 uL 1× PBS to remove the ethanol happened far above us in the forest canopy. preservative in which they were stored. To check the quality and Because V. variegata live in fission–fusion communities size of genomic DNA, samples were amplified using polymerase chain with dynamic group membership (Baden et al., 2016; Holmes reaction (PCR), with a mix as follows: 2.00 μl genomic DNA, 12.62 μl et al., 2016), it was difficult to confirm whether we were molecular‐grade water, 2.00 μl 10× PCR buffer, 1.20 μl50mM following known sub‐groups at each site. We labeled sub‐groups MgCl2, 0.80 μl 10 mM dNTPs, 0.60 μl of each 10 μM primer, and based on membership size and location, assuming that if we 0.18 μl Platinum Taq DNA Polymerase (Life Technologies). We used repeatedly located a characteristic number of individuals in a universal 16S rRNA primers (27F and 1492R), which generated known feeding tree, they were members of a distinct sub‐group. sequences of 1500 bp. For example, in RNP‐LD, we often observed four individuals in PCR product was confirmed with agarose electrophoresis and close proximity to each other in fruit trees surrounding our quantitative DNA ladder (Hyperladder I, Bioline USA, Boston, MA). campsite. Therefore, we named them “Group 1.” Using this DNA samples were loaded onto 1% agarose gel, stained with system, we were only able to confidently identify sub‐group ethidium bromide and visualized under UV light. Negative controls affiliation 81.25% of the time in RNP‐LDand40%ofthetimein containing no samples were used to ensure no sample was RNP‐P. In MSR, however, individuals seldom traveled, which contaminated with exogenous DNA. No DNA was detected in the allowed us to identify sub‐groups 100% of the time. In total, we negative control by gel electrophoresis, and no amplification was believe we followed at least two groups in RNP‐LD (n =7 observed in any PCR reactions using this sample as a template. The individuals), at least three groups in RNP‐P(n = 12 individuals), amplified PCR products were purified with Qiaquick PCR purification two groups in MSR‐D(n = 8 individuals) and one group in MSR‐ kit (Qiagen Inc., CA) following the manufacturer’s protocol. Bulk HD (n =4 individuals). extractable genomic DNA concentrations were also measured using Fecal samples were collected opportunistically from as the Quant‐iT™ dsDNA High‐Sensitivity Assay Kit (Life Technologies many group members as possible. We only collected samples that Inc.). To control for batch effects, DNA extractions and PCRs were we were able to locate less than 30 s after deposition. After randomized across sites, groups, and fecal collection dates. identifying a deposit, sterilized tweezers were used to extract Genomic DNA samples were then sent to the Roy J. Carver approximately 2 g from the center of the feces. Each sample was Biotechnology Center at the University of Illinois at Urbana‐Champaign placed in 2 ml tubes filled with 96% ethanol, which were labeled for amplicon library synthesis on a Fluidigm Access Array™ System. The with the date and time, alongside host location, sex, group identity, output of the system resulted in pooled amplicons flanked by barcoded and approximate age. In total, we collected 139 fecal samples: adapters ready for Illumina sequencing. The site‐specific forward and 33 from RNP‐P, 37 from RNP‐LD, 53 from MSR‐D, and 16 from reverse primers modified with Fluidigm CS1 and CS2 tails respectively MSR‐HD. were used to amplify the V3‐V5 (F357: 5′CCTACGGGAGGCAGCAG‐ 3′; R926: 5′‐CCGTCAATTCMTTTRAGT‐3′) regions of the bacterial 16S rRNA gene as well as the 349F‐806R (Arch349F: 5′‐GYGCAS- 2.4 | Feeding behavior data analysis CAGKCGMGAAW‐3′; Arch806R (5′‐GGACTACVSGGGTATCTAAT‐3′) region of the archaeal 16S rRNA gene. The pooled amplicons flanked Using observational data, we calculated the number of food by barcoded Illumina linkers were then sequenced using 250 bp paired‐ species consumed at each site to assess differences in diversity. end chemistry on an Illumina Miseq. We created a rarefaction model using the R package vegan (Okansen et al., 2018) to determine expected species richness, calculate alpha diversity, and assess if our data were adequate to 2.6 | GM bioinformatic methods make statistical judgments of food species diversity. For estimates of foraging effort, we calculated the proportion of Raw sequences were separated by primer, quality filtered (average time spent feeding per group by dividing the number of data phred score = 33), and trimmed to a maximum length of 250 bp. points spent feeding by the total number of data points for each We removed 17 samples from our data set that did not meet these focal animal. Focal animals were clustered based on‐site and standards. Because the forward and reverse sequences were group identity for statistical evaluation using ANOVA tests with separated by a large amount of unsequenced data, we chose to only Bonferroni corrections. We also created a stacked bar plot, use our forward sequence reads (V3 region) for all subsequent separated by site, showing the relative proportions of each food analyses. All downstream analyses were completed in QIIME v.1.8.0. species observed at 5 or more data points (Figure S1). Finally, we (Caporsao et al., 2010). We aligned the sequences to the SILVA 16S have provided a complete list of food species consumed by the reference databases. Chimeras and singletons were removed, and site (Table S1). successfully classified sequences were clustered into operational DONOHUE ET AL. | 5of12 taxonomic units (OTUs) using a 97% similarity cutoff. From these metadata with OTU abundances and ecological variables were OTUs, we picked representative sets and assigned taxonomy using entered into the Galaxy environment (www.huttenhower.sph. the pick_open_reference_otus.py command. harvard.edu/galaxy/) using the default settings (α = 0.05 for the factorial the Kruskal–Wallis test among variables; threshold on the logarithmic LDA score for discriminative features was 2.0). Because 2.7 | Microbial relative abundance, alpha and beta the LefSe output nests lower taxonomic ranks within higher ones, we diversity separated our results into “independent” and “dependent” biomar- Bacterial alpha and beta diversity indices were calculated using the kers. Dependent biomarkers include the entire LefSe output; core_diversity_analysis.py script after rarefaction at a minimum of independent biomarkers are those that we suspect to be the lowest 1,000 reads per sample with no maximum to capture the maximum reported taxonomic rank. Some of the biomarkers are ambiguous (i.e., amount of diversity (McMurdie & Holmes, 2014). To confirm that “other”), so many of our classifications are educated guesses. sequencing depth did not bias our results, we ran a one‐way analysis of variance (ANOVA) test using the site within the forest as the 3 | RESULTS independent variable and read depth as the dependent variable. We used one‐way ANOVAs with Bonferroni corrections to test 3.1 | Feeding behavior and general food for the effect of independent variables on alpha diversity, as composition measured using the chao1 metric. Independent variables included forest, site within the forest, and level of food species diversity There was considerable variation in both food species diversity and (hypothesis testing), in addition to group identity, sex, and time‐of‐ time spent feeding/foraging across sites. V. variegata in RNP‐P spent day (possible confounding factors). Unweighted and weighted 30.79 ± 1.56% of the day feeding/foraging, RNP‐LD 36.25 ± 1.1%, UniFrac distances were calculated to plot individual samples in MSR‐D 26.96 ± 5.65%, and MSR‐HD 11.1 ± 0.15%, with individuals in ordination space using a principal coordinate analysis (PCoA) to RNP allotting more time to feeding/foraging than those in MSR visualize beta diversity across samples and populations. For UniFrac (ANOVA: df =3; F = 13.73; p = 0.01). Despite increased energy analyses, we used a subset of 101 samples for which we could assign allocated to feeding/foraging, V. variegata in RNP‐LD consumed a group identity with 100% accuracy (see field methods). Using both maximum 3 food species, while animals from the site with the weighted and unweighted UniFrac distances parses fine‐scale second‐least food species diversity (MSR‐HD) consumed more than variation driven by high‐abundance and lowabundance OTUs, four times that number of species and spent less than 12% of the respectively. UniFrac distances were assessed statistically using time foraging/feeding. This difference in food species diversity ANOSIM tests with 999 permutations. The ANOSIM R‐statistic cannot be attributed to sampling bias; our rarefaction analysis compares the mean of ranked dissimilarities between groups to the produced a species accumulation curve with a horizontal plateau mean of ranked dissimilarities within groups. An R‐value close to 1.0 (Figure 5), indicating that continued sampling would not have yielded suggests high dissimilarity between groups, while an R‐value close to additional food sources. 0 suggests an even distribution of high and low ranks within and Dietary composition varied across sites, both within and between groups (Clarke & Gorley, 2001). Cyanobacteria were not between continuous forests. RNP‐LD focal animals primarily filtered from the data, as is somewhat common in primate consumed Chrysophyllum boivinarnum (rahiaka), with occasional microbiome studies (Mallott & Amato, 2018); this phylum comprises supplementation of Protorhus sp. (sandramy) and Cryptocarya sp. more than just environmental contaminants and some of its members (tavolo). V. variegata in RNP‐P consumed a more varied diet that may be crucial components of mammalian gastrointestinal commu- also included high proportions of Cryptocarya sp. and Chrysophyllum nities (Ley et al., 2005). boivinarnum, but expanded to other foods such as Ocotea sp We used a two‐way ANOVA to test for statistically significant (varongy) and Syzyium sp (rotra). interactions between the relative abundance of the phyla Firmicutes There was very little dietary overlap between RNP and MSR and Bacteroidetes, which are commonly used in comparative primate populations. The most commonly consumed food sources in MSR microbiome studies. Using relative abundance data, we calculated the included Ficus lutea (amontana), Blotia oblongifolia (fanjavala flowers), Firmicutes/Bacteroidetes (F/B) ratio for all samples and grouped the and Dombeya sp (valotra). With the exception of Blotia oblongifolia and results according to the site. We ran a one‐way ANOVA with “site” as Mendocia flagellaris (vahy vine), all foods observed being consumed the independent variable to determine whether location statistically across RNP and MSR sites were fruits, which comprised more than impacted F/B ratio. 97% of the total V. variegata diet over this field season. We used one‐way ANOVAs, still using “site” as the independent variable, to measure the differential abundance of phyla found at 3.2 | Sequencing effort more than 1% abundance across samples. To further understand prokaryotic variation across sites, we used Linear discriminant Our rarefaction curve did not reach a plateau, which indicated that analysis of Effect Size (LefSe) to identify microbial biomarkers (OTUs the number of OTUs would continue to increase with greater with differential relative abundance) specific to each site. Tabular sequencing depth for most samples (Figure S2). We also detected 6of12 | DONOHUE ET AL.

FIGURE 2 Differences in bacterial abundance by site; (a) the relative proportion of microbial phyla detected in more than 1% of samples; (b) taxa recognized by LefSe as differentially abundant biomarkers. Asterix denote independent biomarkers (i.e., Candidatus Methanomethylophilus is nested within the kingdom Archaea, phylum Euryarchaeota, etc:)

variation in sequencing depth across samples; however, these 73 to 16,562 sequences per sample (Table S2). Mean read count and differences were not statistically different between sites (ANOVA, standard deviation were 9,376 ± 2,617 for RNP‐P, 8,150 ± 1,543 for df =3, F = 1.44, SumSq = 3,4030,900, p = 0.23) and variation was RNP‐LD, 9,200 ± 3,012 for MSR‐D, and 9,409 ± 2,215 for MSRHD. All distributed evenly across sites (Figure S3). Read count ranged from samples with less than 1,000 reads were removed from downstream analyses. After filtering, we retained 28 samples from RNP‐P, 35 from RNP‐LD, 39 from MSR‐D, and 19 from MSR‐HD (n = 122 fecal samples total).

3.3 | Gut microbial composition

In total, we generated 672,052 archaeal sequence reads (mean = 5,548 per sample; range = 6–16,640) and 1,059,886 bacterial sequence reads (mean = 8,886 per sample; range = 73–16,529; Table S1). We recovered 25 phyla, 69 classes, 147 orders, 297 families, and 687 genera of bacteria. The shared bacterial phyla of V. variegata across all sites include Bacteroidetes (35%), Proteobacteria (21%), Cyanobacteria (12%), and Firmicutes (10%). All successfully classified archaea OTUs belonged to an unidentified family of the class Themoplasmatales, which is ranked within the phyla Euryarchaeota.On average, this family was found in 72% of the sequence reads obtained FIGURE 3 Chao1 alpha diversity estimates (number of unique OTUs) for each site. Error bars represent the standard deviation of from RNP‐P samples, 19% of RNP‐LD, 43% of MSR‐D, and 70% of the mean MSR‐HD (ANOVA; df =3; F = 39.09; SumSq = 5.2781; p < .0001). DONOHUE ET AL. | 7of12

FIGURE 4 Weighted UniFrac plots showing beta diversity (differences in species composition) across: (a) food species diversity (note that “Low food species diversity” represents samples from RNP‐LD and “High food species diversity” represents samples from RNP‐P, MSR‐D, and MSR‐HD); (b) group identity; (c) locality within forests; and (d) forest

Because archaea‐specific primers generated low diversity, we did not on‐site, which represents the level of habitat disturbance and food include these data in subsequent diversity analyses. species diversity (Figure 2b). All of the biomarkers identified in RNP‐ Of the five bacterial phyla detected in more than 1% abundance P were classified as methanogenic archaea. Samples from RNP‐LD across samples, all but Spirochetes were found in statistically exhibited elevated levels of bacteria associated with carbohydrate different abundances between sites (Bacteroidetes: F = 34.62, fermentation (genera Succinatimonas; see Santos & Thompson, 2013), SumSq = 1.7638, p < .0001; Cyanobacteria: SumSq = 1.1425 F = in addition to bacterial taxa associated with both symbiotic and 10.06, p <.0001; Firmicutes: F = 20.199, SumSq = 0.87576, pathogenic interactions (order: Enterobacteriales, family: Enterobacter- p < 0.0001; Proteobacteria: F = 23.551, SumSq = 1.047, p < 0.0001) iaceae). The MSR‐D samples showed an increased abundance of the (Figure 2a). We also detected a significant interaction between the pathogenic family Spirochaetaceae (Cochez et al., 2014) and the relative abundance of Firmicutes and Bacteroidetes (F = 20.198, bacterial subclass Erysipelotrichia, which is associated with diets high SumSq = 3.5030, p < .0001) in all sites. The highest Firmicutes/ in fat (Greiner & Bäckhed, 2011). Finally, MSR‐HD had just two Bacteroidetes ratio was detected in MSR‐D (8.46), followed by microbial biomarkers. The first, Actinobacteria, is a phylum of Gram‐ MSR‐HD (4.02), RNP‐LD (1.83), and RNP‐P (0.78). positive bacteria found across diverse terrestrial and aquatic From the LefSe analysis, we identified 10 independent and ecosystems. The second, Gordonibacter, is associated with gastro- 22 nonindependent potential biomarkers for discrimination based intestinal disease in humans (Opstelten et al., 2016). 8of12 | DONOHUE ET AL.

described in the Chlorocebus genus (Trosvik et al., 2018), Alouatta pigra (Amato et al., 2013), Lemur catta (Bennett et al., 2016), and Procolobus gordonorum (Barelli et al., 2015). GM robustness to habitat disturbance has been reported in parrots (Strigops habroptilus; see Perry, Digby, & Taylor, 2017), but other taxa, including frogs (Fejervarya limnocharis; see Chang, Huang, Lin, Huang, & Liao, 2016), jackals (Canis mesomelas; see Menke et al., 2017), and swan geese (Anser cygnoides; see Wu et al., 2018) show that populations in disturbed habitats exhibit distinct GMs relative to pristine habitat FIGURE 5 The number of food species consumed by site, rarefied counterparts. While our findings in V. variegata generally fit this to the site with the lowest sampling effort (RNP‐LD). The vertical line narrative, our results indicate that it is not simply the level of habitat shows the point where the rarefaction comparison is made, and disturbance driving GM diversity. Notably, our PCoA plots show that horizontal lines indicate the observed species richness for each site. the least disturbed site (RNP‐P) and most disturbed site (MSR‐HD) Although the sampling effort was low at RNP‐LD, food species diversity had already reached a horizontal asymptote, indicating that have the most similar GM community compositions in PC1 and PC2, more feeding data points would not have changed the outcome of which together encompass approximately 80% of the total variation. the results This pattern cannot be explained by proximity (Figure 1), dietary diversity (Figure 5), or dietary composition (Figure S1 and Table S1). ‐ ‐ 3.4 | Gut microbial diversity Samples from RNP P and MSR HD also contained the highest abundances of archaea. Clearly, additional environmental variables Microbial α diversity varied across sites (Figure 3), with diversity shape the GM that were beyond the scope of the research presented (ᾱ = 5844) in RNP‐P, the site with the least habitat disturbance, being herein. The remainder of our discussion focuses on the effects of significantly greater than α diversity of all other sites (ANOVA, df =3; habitat disturbance and diet, but it is important to acknowledge that F = 9.007, p < .0001). We did not detect significant differences in α confounding and unexplored factors undoubtedly contribute to the diversity among the other three sites (RNP‐LD ᾱ =3,952; MSR‐D patterns observed in this study. ᾱ = 3,941; and MSR‐HD ᾱ = 3,854). Food diversity, sex, and time‐of‐day did not yield significant differences in α diversity (results not shown). 4.1 | Gut microbial diversity is a sensitive measure PCoA plots of weighted UniFrac distances clustered samples of habitat disturbance and diet according to food species diversity (Figure 4a), group identity (Figure 4b), location within forests (Figure 4c), and forest RNP‐P had significantly higher α diversity than the other three (Figure 4d). ANOSIM tests show that these clusters are statistically sampled sites, all of which had at least some amount of disturbance significant (food species diversity: R = 0.44, p = .001; group identity: associated with them. While generally this pattern was expected, we R = 0.26; p = .001; location within the forest: 0.23, p = .001; forest: found it surprising that RNP‐LD, a high‐quality site without recent R = 0.16, p = .001). While unweighted UniFrac ANOSIM tests pro- disturbance, showed measures of α diversity similar to the more duced similar pvalues and clustering patterns (Figure S4), R‐values disturbed MSR sites. Furthermore, our UniFrac plots demonstrated for site and food species diversity decreased (food species diversity: clear differentiation between RNP‐LD and the other three sites, R = 0.60; p = .001; group identity: R = 0.31, p = .001; location within indicating a significant shift in both high‐abundance and low‐ forest: R = 0.35, p = .001; forest: R = 0.23; p = .001). Sex and time‐of‐ abundance OTUs (Figure 4a and Figure S4a, respectively). This day did not yield significant differences in β diversity using UniFrac well‐defined clustering is coupled with the highest weighted UniFrac distance (results not shown). ANOSIM statistic, further underscoring the unique GM composition of RNP‐LD in comparison to nearby RNP‐P and distant MSR‐D and MSR‐HD. Dietary diversity helps explain this pattern. 4 | DISCUSSION Individuals from the most pristine site (RNP‐P) and the two most disturbed sites (MSR‐DandMSR‐HD) were observed This study shows considerable GM variation across populations of consuming approximately the same number of food species V. variegata within and between forests. Each of the sites selected (>15), whereas those in RNP‐LD were observed consuming just varied in degree of historical and ongoing habitat disturbance. While three. We argue that this difference is not driven by plant we cannot definitively attribute GM variation to disturbance itself, phenology alone, as samples were collected from RNP‐P during the we show that these populations have indeed diverged in crucial same year and season, weeks apart. We also argue that this GM characteristics, including diversity and species composition. In difference is not the result of environmental stochasticity, general, we found that microbial OTUs clustered according to as unpublished data collected in RNP‐LD in 2017 and 2018, location and that prokaryotic α diversity was highest in the most during the same season (June–July), show that V. variegata pristine site and significantly lower in sites with a history of and other lemur species (Eulemur rufifrons and Eulemur rubriventer) disturbance. Within the primate clade, similar patterns have been were observed eating the same three food speciesexclusively DONOHUE ET AL. | 9of12

(Chrysophyllum boivinarnum, Protorhus sp., and Cryptocarya sp.). remains unclear, especially because so few primate microbiome Therefore, we argue that habitat disturbance and dietary diversity studies report information about archaea. Because archaea are harbor statistically similar signals on V. variegata gut microbial α known to thrive in more acidic environments (Schäfer, 1998), and diversity and that these effects may manifest independently, not diet/ecology impact stomach acidity (Beasley, Koltz, Lambert, Fierer, additively as we had expected. It is also worth noting that RNP‐LD & Dunn, 2015), it is possible that differential abundance of archaea is the only site accessible to tourists, so it is possible that higher reflects variation in gut pH across disturbance regimes. This could visitor presence and potentially increased stress also contribute to have major consequences for individual health and species conserva- RNP‐LD’s distinct microbial patterning. tion, as high stomach acidity is thought to protect against pathogenic invasion (Beasley et al., 2015). 4.2 | Approximations of gut health

Reduced α diversity is expected to jeopardize immunological 5 | Limitations function, host health, and population viability (Amato et al., 2013). Therefore, we expected that OTUs associated with poor gastro- Geographic location and habitat disturbance were confounded in intestinal health would be disproportionately represented in sites this study, and as such, it was not possible to separate these two with low α diversity. factors. Therefore, while we argue that many of the patterns Although identifying pathogens was beyond the scope of this described herein are connected with the level of disturbance, it is study, an indicator species analysis (LefSe) confirmed that the two important to consider other factors that may drive GM composition most disturbed sites (MSR‐D and MSR‐HD) were characterized by and diversity, including host relatedness, altitude, and distance microbes associated with the disease, including Gordonibacter and between populations. Furthermore, we were not able to determine Spirochaetae. However, despite having similar α diversity, samples how many individuals were sampled repeatedly—as a result, site‐ from RNP‐LD showed an abundance of bacteria associated with specific effects may be amplified. In addition, future studies may carbohydrate fermentation, a hallmark of a healthy frugivore GM, also include nutritional analyses of food items, as feeding observa- with no observed increase in disease‐associated microbes. tions provide an admittedly limited picture of this dynamic and We also found that samples from MSR had significantly higher complicated system. F/B ratios than those from RNP. The F/B ratio has been correlated with energy harvest potential (Ley, Turnbaugh, Klein, & Gordon, 5.1 | Significance and future directions 2006), with increased Firmicutes associated with obesity (Koliada et al., 2017; Ley et al., 2005, 2006; Turnbaugh et al., 2006) and At its core, this study represents a first step toward understanding consumption of low‐quality, fibrous foods (Clayton et al., 2018; population‐level variation in the V. variegata GM. However, we hope Gomez et al., 2015). Increased Bacteriodetes has been demonstrated that our findings can be extrapolated and applied across taxa to in individuals with increased sugar consumption (De Filippo et al., understand global microbiome patterning in response to human 2010; Ley, 2010; Wu et al., 2011). These findings align with the disturbance. Currently, there are few published studies examining observation that while V. variegata in MSR were mostly frugivorous, these dynamics, despite their clear utility in conservation biology and fruits (which are high in sugar) were occasionally supplemented with potential for expanding our understanding of nature’s rules in the leaves and flowers (high in fiber). It is plausible that, as a result of Anthropocene. Most importantly, the GM—as it is dynamic across an consuming a more varied, less nutritious diet, maintaining a GM individual’s lifetime—can and should be used to monitor the rate and characterized by more efficient energy harvest potential (i.e., higher severity of human impacts on endangered populations. The data F/B ratio) is key to V. variegata survival in MSR. presented herein represent a baseline GM for the MSR V. variegata population, which can be tracked over time as a component of efforts to conserve this unique and shrinking forest. 4.3 | The mysterious role of methanogenic archaea

V. variegata GMs in our study were similar to other wild primate ACKNOWLEDGEMENTS species, with the expected predominance of Bacteroidetes, Proteo- bacteria, and Firmicutes at the phylum‐level (Yildirim et al., 2010). The authors thank the Madagascar Ministry of Environment, Forests, Despite a convergence of common prokaryotes at the highest and Ecology, the Madagascar National Parks in Ranomafana National taxonomic levels, we discerned significant variation in relative Park and Manombo Special Reserve for allowing us to conduct this abundances across sites. Perhaps the most striking aspect of this field research, and Center ValBio Research station and MICET/ICTE prokaryotic diversity was detected in RNP‐P, where samples were for logistical support. The authors thank Feno, De l’Or Boto, characterized by an elevated abundance of methanogenic archaea. Velontsara Baptiste, and Elizabeth Wallace for their help with data Interestingly, Trosvik et al. (2018) detected a similar pattern in collection, in addition to Katie Everson, Derek Filipek, Levi Gray, Kara Chlorocebus monkeys occupying pristine habitats. Whether this Jones, and Thomas Maigret for their role in editing this manuscript. pattern is coincidence or indicative of underlying biological processes Finally, we thank Dr. Robert Thacker for all of his advice and 10 of 12 | DONOHUE ET AL. guidance. This study was funded by the National Geographic Young Beasley, D. E., Koltz, A. M., Lambert, J. E., Fierer, N., & Dunn, R. R. (2015). Explorer’s Grant #9819‐15, Primate Conservation Inc Grant #74892, The evolution of stomach acidity and its relevance to the human microbiome. PLOS One, 10(7), e0134116. https://doi.org/10.1371/ NSF BCS 0935347 and 0820709, and the Woese Institute for journal.pone.0134116 Genomic Biology and the University of Illinois at Urbana Champaign. Bennett,G.,Malone,M.,Sauther,M.L.,Cuozzo,F.P.,White,B.,Nelson,K.E.,… This study was approved by the Stony Brook University IACUC and Amato, K. R. (2016). Host age, social group, and habitat type influence the complied with Malagasy laws. gut microbiota of wild ring‐tailed lemurs (Lemur catta): Ring‐tailed lemur gut microbiota. American Journal of Primatology, 78(8), 883–892. https:// doi.org/10.1002/ajp.22555 ORCID Britt, A. (2000). Diet and feeding behaviour of the black‐and‐white ruffed lemur (Varecia variegata variegata) in the Betampona Reserve, Eastern Mariah E. Donohue http://orcid.org/0000-0002-4311-453X Madagascar. Folia Primatologica, 71, 133–141. https://doi.org/10.1159/ Abigail E. Asangba http://orcid.org/0000-0003-4223-6903 000021741 Brucker, R. M., & Bordenstein, S. R. (2012). Speciation by symbiosis. Trends in Ecology & Evolution, 27, 443–451. https://doi.org/10.1016/j.tree. 2012.03.011 REFERENCES Caporaso, J. G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F. D., Costello, E. K., … Knight, R. (2010). QIIME allows analysis of high‐ Abbas, F., Morellet, N., Hewison, A. J. M., Merlet, J., Cargnelutti, B., throughput community sequencing data. Nature Methods, 7, 335–336. Lourtet, B., … Verheyden, H. (2011). Landscape fragmentation https://doi.org/10.1038/nmeth.f.303 generates spatial variation of diet composition and quality in a Chang, C. W., Huang, B. H., Lin, S. M., Huang, C. L., & Liao, P. C. (2016). generalist herbivore. Oecologia, 167(2), 401–411. https://doi.org/10. Changes of diet and dominant intestinal microbes in farmland frogs. 1007/s00442‐011‐1994‐0 BMC Microbiology, 16, 33. https://doi.org/10.1186/s12866‐016‐ Altmann, J. (1974). Observational study of behavior: Sampling 0660‐4 methods. Behaviour, 49(3), 227–266. https://doi.org/10.1163/ Chapman, C. A., Struhsaker, T. T., Skorupa, J. P., Snaith, T. V., & Rothman, J. M. 156853974X00534 (2010). Understanding long‐term primate community dynamics: Implica- Amato, K. R., Metcalf, J. L., Song, S. J., Hale, V. L., Clayton, J., Ackermann, tions of forest change. Ecological Applications, 20(1), 179–191. https://doi. G., … Braun, J. (2016). Using the gut microbiota as a novel tool for org/10.1890/09‐0128.1 examining colobine primate GI health. Global Ecology and Conservation, Chaves, Ó. M., Stoner, K. E., & Arroyo‐Rodríguez, V. (2011). Differences in 7, 225–237. https://doi.org/10.1016/j.gecco.2016.06.004 diet between spider monkey groups living in forest fragments and Amato, K. R., Yeoman, C. J., Kent, A., Righini, N., Carbonero, F., Estrada, A., continuous forest in Mexico. Biotropica, 44(1), 105–113. https://doi. … Leigh, S. R. (2013). Habitat degradation impacts black howler org/10.1111/j.1744‐7429.2011.00766.x monkey (Alouatta pigra) gastrointestinal microbiomes. The ISME Clarke, K.R., & Gorley, R.N. (2001). Primer v5: user manual/tutorial. journal, 7, 1344–1353. https://doi.org/10.1038/ismej.2013.16 Primer‐E Ltd: Plymouth. Arrigo‐Nelson, S.J. (2006). The impact of habitat disturbance on the feeding Clayton, J. B., Vangay, P., Huang, H., Ward, T., Hillmann, B. M., Al‐Ghalith, ecology of the Milne‐Edwards’ sifaka (Propithecus edwardsi) in Ranoma- G. A., … Knights, D. (2016). Captivity humanizes the primate fana National Park, Madagascar (PhD thesis). State University of New microbiome. Proceedings of the National Academy of Sciences, 113, York, Stony Brook, NY. 10376–10381. https://doi.org/10.1073/pnas.1521835113 Asensio, N., Cristobal‐Azkarate, J., Dias, P. A. D., Vea, J. J., & Rodríguez‐ Clayton, J. B., Gomez, A., Amato, K., Knights, D., Travis, D. A., Blekhman, R. Luna, E. (2007). Foraging Habits of Alouatta palliata mexicana in three … Johnson, T. J. (2018). The gut microbiome of nonhuman primates: forest fragments. Folia Primatologica, 78, 141–153. https://doi.org/10. Lessons in ecology and evolution. American Journal of Primatology, 1159/000099136. e22867. https://doi.org/10.1002/ajp.22867 Baden, A. L., Holmes, S. M., Johnson, S. E., Engberg, S. E., Louis, E. E., Jr, & Cochez, C., Heyman, P., Heylen, D., Fonville, M., Hengeveld, P., Takken, W., Bradley, B. J. (2014). Species‐level view of population structure and … Sprong, H. (2014). The presence of Borrelia miyamotoi, a relapsing gene flow for a critically endangered primate (Varecia variegata). fever Spirochaete, in Questing Ixodes ricinus in Belgium and in the Ecology and Evolution, 4(13), 2675–2692. https://doi.org/10.1002/ Netherlands. Zoonoses and Public Health, 62(5), 331–333. https://doi. ece3.1119 org/10.1111/zph.12154 Baden, A. L., Webster, T. H., & Kamilar, J. M. (2016). Resource seasonality Crowley, B. E., Blanco, M. B., Arrigo‐Nelson, S. J., & Irwin, M. T. (2013). and reproduction predict fission‐fusion dynamics in black‐and‐white Stable isotopes document resource partitioning and effects of forest ruffed lemurs (Varecia variegata): Ruffed lemur fission‐fusion dy- disturbance on sympatric cheirogaleid lemurs. Naturwissenschaften, namics. American Journal of Primatology, 78(2), 256–279. https://doi. 100, 943–956. https://doi.org/10.1007/s00114‐013‐1094‐6 org/10.1002/ajp.22507 David, L. A., Maurice, C. F., Carmody, R. N., Gootenberg, D. B., Button, J. E., Balko, E. (1998). A behaviorally plastic response to forest composition and Wolfe, B. E., … Turnbaugh, P. J. (2014). Diet rapidly and reproducibly logging disturbance by Varecia variegata variegata in Ranomafana alters the human gut microbiome. Nature, 505, 559–563. https://doi. National Park, Madagascar (PhD thesis). State University of New York, org/10.1038/nature12820 College of Environmental Science and Forestry, New York, NY. Dethlefsen, L., Eckburg, P. B., Bik, E. M., & Relman, D. A. (2006). Assembly Retrieved from https://doi.org/10.13140/2.1.4927.0401 of the human intestinal microbiota. Trends in Ecology & Evolution, 21(9), Barelli, C., Albanese, D., Donati, C., Pindo, M., Dallago, C., Rovero, F., … De 517–523. https://doi.org/10.1016/j.tree.2006.06.013 Filippo, C. (2015). Habitat fragmentation is associated to gut Dunham, N. T. (2017). Feeding ecology and dietary flexibility of Colobus microbiota diversity of an endangered primate: Implications for angolensis palliatus in relation to habitat disturbance. International conservation. Scientific Reports, 5,1–12. https://doi.org/10.1038/ Journal of Primatology, 38(3), 553–571. https://doi.org/10.1007/ srep14862 s10764‐017‐9965‐x Bauchop, T. (1971). Stomach microbiology of primates. Annual Review of Dunn, J. C., Cristóbal‐Azkarate, J., & Veà, J. J. (2010). Seasonal variations Microbiology, 25(1), 429–436. in the diet and feeding effort of two groups of howlers in different DONOHUE ET AL. | 11 of 12

sized forest fragments. International Journal of Primatology, 31, 50, 227–246. https://doi.org/10.1002/(SICI)1098‐2345(200004) 887–903. https://doi.org/10.1007/s10764‐010‐9436‐0 50<227:AID‐AJP1>3.0.CO;2‐S De Filippo, C., Cavalieri, D., Di Paola, M., Ramazzotti, M., Poullet, J. B., Koliada, A., Syzenko, G., Moseiko, V., Budovska, L., Puchkov, K., Perederiy, Massart, S., … Lionetti, P. (2010). Impact of diet in shaping V., … Vaiserman, A. (2017). Association between body mass index and gut microbiota revealed by a comparative study in children from Firmicutes/Bacteroidetes ratio in an adult Ukrainian population. BMC Europe and rural Africa. Proceedings of the National Academy of Microbiology, 17, 120. https://doi.org/10.1186/s12866‐017‐1027‐1 Sciences, 107(33), 14691–14696. https://doi.org/10.1073/pnas.100 Lambert, J. E. (1998). Primate digestion: Interactions among anatomy, 5963107 physiology, and feeding ecology. Evolutionary Anthropology: Issues, Flint, H. J., Bayer, E. A., Rincon, M. T., Lamed, R., & White, B. A. (2008). News, and Reviews, 7(1), 8–20. Polysaccharide utilization by gut bacteria: Potential for new insights Lehtonen, J. T., Mustonen, O., Ramiarinjanahary, H., Niemelä, J., & Rita, H. from genomic analysis. Nature Reviews Microbiology, 6, 121–131. (2001). Habitat use by endemic and introduced rodents along a gradient https://doi.org/10.1038/nrmicro1817 of forest disturbance in Madagascar. Biodiversity and Conservation, 10, Flint, H. J., Scott, K. P., Louis, P., & Duncan, S. H. (2012). The role of the gut 1185–1202. https://doi.org/10.1023/A:1016687608020 microbiota in nutrition and health. Nature Reviews Gastroenterology & Ley, R. E. (2010). Obesity and the human microbiome:. Current Opinion in Hepatology. 9, 577–589. https://doi.org/10.1038/nrgastro.2012.156 Gastroenterology, 26(1), 5–11. https://doi.org/10.1097/MOG.0b013e328 Gomez, A., Petrzelkova, K., Yeoman, C. J., Vlckova, K., Mrazek, J., Koppova, 333d751 I. … Leigh, S. R. (2015). Gut microbiome composition and metabolomic Ley, R. E., Backhed, F., Turnbaugh, P., Lozupone, C. A., Knight, R. D., & profiles of wild western lowland gorillas (Gorilla gorilla gorilla) reflect Gordon, J. I. (2005). Obesity alters gut microbial ecology. Proceedings host ecology. Molecular Ecology, 24, 2551–2565. https://doi.org/10. of the National Academy of Sciences, 102(32), 11070–11075. https:// 1111/mec.13181 doi.org/10.1073/pnas.0504978102 Greiner, T., & Bäckhed, F. (2011). Effects of the gut microbiota on obesity Ley, R. E., Hamady, M., Lozupone, C., Turnbaugh, P., Ramey, R. R., Bircher, and glucose homeostasis. Trends in endocrinology and metabolism: TEM, S. … Gordon, J. I. (2008). Evolution of mammals and their gut 22(4), 117–123. https://doi.org/10.1016/j.tem.2011.01.002 microbiomes. Science, 320(5883), 1647–1651. https://doi.org/10. Groussin, M., Mazel, F., Sanders, J. G., Smillie, C. S., Lavergne, S., Thuiller, 1126/science.1155725 W., & Alm, E. J. (2017). Unraveling the processes shaping mammalian Ley, R. E., Turnbaugh, P. J., Klein, S., & Gordon, J. I. (2006). Human gut gut microbiomes over evolutionary time. Nature Communications, 8, microbes associated with obesity. Nature, 444(21/28), https://doi.org/ 14319. https://doi.org/10.1038/ncomms14319 10.1038/nature4441022a Herrera, J. P., Wright, P. C., Lauterbur, E., Ratovonjanahary, L., & Taylor, L. Mallott, E. K., & Amato, K. R. (2018). The microbial reproductive ecology L. (2011). The effects of habitat disturbance on lemurs at Ranomafana of white‐faced capuchins (Cebus capucinus). American Journal of National Park. International Journal of Primatology, 32, 1091–1108. Primatology, 80(8):e22896. https://doi.org/10.1002/ajp.22896 https://doi.org/10.1007/s10764‐011‐9525‐8 Mancini, A. (2016). Population genetic analysis of the critically endan- Hird, S. M. (2017). Evolutionary biology needs wild microbiomes. Frontiers gered black‐and‐white ruffed lemur (Vareica variegata) in Ranomafana in Microbiology, 8(725), https://doi.org/10.3389/fmicb.2017.00725 National Park, Madagascar. MA thesis, The City University of New Holmes, S. M., Baden, A. L., Brenneman, R. A., Engberg, S. E., Louis, E. E., Jr, York, Hunter College. & Johnson, S. E. (2013). Patch size and isolation influence genetic Matos, M. D. P. (2017). Comparing the feeding ecology of Propithecus patterns in black‐and‐white ruffed lemur (Varecia variegata) popula- edwardsi in disturbed and undisturbed forest in Ranomafana National tions. Conservation Genetics, 14, 615–624. https://doi.org/10.1007/ Park, Madagascar (MA dissertation). Universidade de ÉVORA, Évora, s10592‐0130455‐1 Portugal. Holmes, S. M., Gordon, A. D., Louis, E. E., Jr, & Johnson, S. E. (2016). McMurdie, P. J., & Holmes, S. (2014). Waste not want not: Why rarefying Fission‐fusion dynamics in black‐and‐white ruffed lemurs may microbiome data is inadmissible. PLOS Computational Biology, facilitate both feeding strategies and communal care of infants in a 10(4):e1003531. https://doi.org/10.1371 spatially and temporally variable environment. Behavioral Ecology and Menke, S., Meier, M., Mfune, J. K. E., Melzheimer, J., Wachter, B., & Sommer, Sociobiology, 70(11), 1949–1960. https://doi.org/10.1007/s00265‐ S. (2017). Effects of host traits and land‐use changes on the gut 016‐2201‐4 microbiota of the Namibian black‐backed jackal (Canis mesomelas). FEMS Hooper, L. V., Littman, D. R., & Macpherson, A. J. (2012). Interactions Microbiology Ecology, 93(11), https://doi.org/10.1093/femsec/fix123 between the microbiota and the immune system. Science, 336(6086), Oksanen, J., Guillaume, F., Friendly, M., Kindt, R., Legendre, P., McGlinn, D., … 1268–1273. https://doi.org/10.1126/science.1223490 Wagner, H. (2009). vegan: community ecology package. R package Huang, Z. P., Scott, M. B., Li, Y. P., Ren, G. P., Xiang, Z. F., Cui, L. W., & Xiao, W. version 1.15.4. (http://CRAN.Rproject.org/package=vegan) (2017). Blackand‐white snub‐nosed monkey (Rhinopithecus bieti) Oksanen, J., Blanchet, F. G., Friendly, M., Kindt, R., Legendre, P., McGlinn, feeding behavior in a degraded forest fragment: Clues to a stressed D., … Wagner, H. (2018). vegan: Community Ecology Package, R population. Primates, 58,517–524. https://doi.org/10.1007/s10329‐017‐ package version 2.5‐5. 06187 Opstelten, J. L., Plassais, J., van Mil, S. W. C., Achouri, E., Pichaud, M., Johns, A. D. (1986). Effects of selective logging on the behavioral ecology Siersema, P. D., … Cervino, A. C. L. (2016). Gut microbial diversity is of west Malaysian primates. Ecology, 67, 684–694. https://doi.org/10. reduced in smokers with Crohnʼs disease:. Inflammatory Bowel 2307/1937692 Diseases, 22(9), 2070–2077. https://doi.org/10.1097/MIB.0000000 Johnson, S. E., Ingraldi, C., Ralainasolo, F. B., Andriamaharoa, H. E., 000000875 Ludovic, R., Birkinshaw, C. R., … Ratsimbazafy, J. H. (2011). Gray‐ Perry, E. K., Digby, A., & Taylor, M. W. (2017). The low‐diversity fecal headed lemur (Eulemur cinereiceps) abundance and forest structure microbiota of the critically endangered Kākāpō is robust to anthro- dynamics at Manombo, Madagascar: Gray‐headed lemur abundance pogenic dietary and geographic influences. Frontiers in Microbiology, 8, and forest structure. Biotropica, 43(3), 371–379. https://doi.org/10. 2033. https://doi.org/10.3389/fmicb.2017.02033 1111/j.1744‐7429.2010.00705.x Ratsimbazafy, J. H. (2002). On the brink of extinction and the process of Kaplin, B. A., & Moermond, T. C. (2000). Foraging ecology of the mountain recovery: reponses of the black‐and‐white ruffed lemurs (Varecia variegata monkey (Cercopithecus l'hoesti): Implications for its evolutionary variegata) to disturbance in Manombo (PhD thesis). State University of history and use of disturbed forest. American Journal of Primatology, New York, Stony Brook, NY. 12 of 12 | DONOHUE ET AL.

Santos, E. D. O., & Thompson, F. (2013). The family Succinivibrionaceae. In National Park, Madagascar. Folia Primatologica, 64, 124–131. https:// Rosenberg, E., DeLong, E. F., Lory, S., & Thompson, F. (Eds.), The doi.org/10.1159/000156842 prokaryotes: Gammaproteobacteria. Berlin Heidelberg, Germany: Wright, P. C. (1999). Lemur traits and Madagascar ecology: Coping with Springer‐Verlag Berlin Heidelberg. an island environment. American Journal of Physical Anthropology, 110, Schäfer, G. (1998). How can archaea cope with extreme acidity? In Poole, 31–72. https://doi.org/10.1002/(SICI)1096‐8644(1999)110:29+% R. (Ed.), Mechanisms by which bacteria respond to pH (pp. 131–151). 3C31::AID‐AJPA3%3E3.0.CO;2‐0 London: John Wileyand Sons. Wright, P. C., Tecot, S. R., Erhart, E. M., Baden, A. L., King, S. J., & Grassi, C. Sekirov, I., Russell, S. L., Antunes, L. C. M., & Finlay, B. B. (2010). Gut (2011). Frugivory in four sympatric lemurs: Implications for the future microbiota in health and disease. Physiological Reviews, 90, 859–904. of Madagascar's forests. American Journal of Primatology, 73, 585–602. https://doi.org/10.1152/physrev.00045.2009 https://doi.org/10.1002/ajp.20936 Springer, A., Fichtel, C., Al‐Ghalith, G. A., Koch, F., Amato, K. R., Clayton, J. Wu, G. D., Chen, J., Hoffmann, C., Bittinger, K., Chen, Y. Y., Keilbaugh, S. A., B., … Kappeler, P. M. (2017). Patterns of seasonality and group … Lewis, J. D. (2011). Linking long‐term dietary patterns with gut membership characterize the gut microbiota in a longitudinal study of microbial enterotypes. Science, 334(6052), 105–108. https://doi.org/ wild Verreaux's sifakas (Propithecus verreauxi). Ecology and Evolution, 10.1126/science.1208344 7(15), 5732–5745. https://doi.org/10.1002/ece3.3148 Wu, Y., Yang, Y., Cao, L., Yin, H., Xu, M., Wang, Z., … Deng, Y. (2018). Stumpf, R. M., Gomez, A., Amato, K. R., Yeoman, C. J., Polk, J. D., Wilson, B. Habitat environments impacted the gut microbiome of long‐distance A., … Leigh, S. R. (2016). Microbiomes, metagenomics, and primate migratory swan geese but central species conserved. Scientific Reports, conservation: New strategies, tools, and applications. Biological Con- 8, 13314. https://doi.org/10.1038/s41598‐018‐31731‐9 servation, 199,56–66. https://doi.org/10.1016/j.biocon.2016.03.035 Yildirim, S., Yeoman, C. J., Sipos, M., Torralba, M., Wilson, B. A., Goldberg, Trosvik, P., Rueness, E. K., de Muinck, E. J., Moges, A., & Mekonnen, A. T. L., … Nelson, K. E. (2010). Characterization of the fecal microbiome (2018). Ecological plasticity in the gastrointestinal microbiomes of from non‐human wild primates reveals species specific microbial Ethiopian Chlorocebus monkeys. Scientific Reports, 8, 20. https://doi. communities. PLOS One, 5(11), e13963. https://doi.org/10.1371/ org/10.1038/s41598‐017‐18435‐2 journal.pone.0013963 Turnbaugh, P. J., Bäckhed, F., Fulton, L., & Gordon, J. I. (2008). Diet‐ induced obesity is linked to marked but reversible alterations in the mouse distal gut microbiome. Cell Host & Microbe, 3, 213–223. https:// SUPPORTING INFORMATION doi.org/10.1016/j.chom.2008.02.015 Turnbaugh, P. J., Ley, R. E., Mahowald, M. A., Magrini, V., Mardis, E. R., & Additional supporting information may be found online in the Gordon, J. I. (2006). An obesity‐associated gut microbiome with Supporting Information section. increased capacity for energy harvest. Nature, 444(21/28), 1027–1031. https://doi.org/10.1038/nature05414 Vieilledent, G., Grinand, C., Rakotomalala, F. A., Ranaivosoa, R., Rakotoar- ijaona, J. R., Allnutt, T. F., & Achard, F. (2018). Combining global tree How to cite this article: Donohue ME, Asangba AE, cover loss data with historical national forest cover maps to look at Ralainirina J, Weisrock DW, Stumpf RM, Wright PC. Extensive six decades of deforestation and forest fragmentation in Madagascar. variability in the gut microbiome of a highly‐specialized and Biological Conservation, 222, 189–197. https://doi.org/10.1016/j. critically endangered lemur species across sites. Am J Primatol. biocon.2018.04.008 White, F. J., Overdorff, D. J., Balko, E. A., & Wright, P. C. (1995). 2019;e23046. https://doi.org/10.1002/ajp.23046 Distribution of Ruffed Lemurs (Varecia variegata) in Ranomafana