High diversity and unique composition of gut microbiomes in pygmy (Kogia breviceps) and dwarf (K. sima) sperm whales

Mammals host diverse bacterial and archaeal symbiont communities (i.e. microbiomes) that play important roles in digestive and immune system functioning, yet cetacean microbiomes remain largely unexplored, in part due to sample collection difficulties. Here, fecal samples from stranded pygmy (Kogia breviceps) and dwarf (K. sima) sperm whales were used to characterize the gut microbiomes of two closely-related species with similar diets. 16S rRNA gene sequencing revealed diverse microbial communities in kogiid whales dominated by Firmicutes and Bacteroidetes. Core symbiont taxa were affiliated with phylogenetic lineages capable of fermentative metabolism and sulfate respiration, indicating potential symbiont contributions to energy acquisition during prey digestion. The diversity and phylum-level composition of kogiid microbiomes differed from those previously reported in toothed whales, which exhibited low diversity communities dominated by Proteobacteria and Actinobacteria. Community structure analyses revealed distinct gut microbiomes in K. breviceps and K. sima, driven by differential relative abundances of shared taxa, and unique microbiomes in kogiid hosts compared to other toothed and baleen whales, driven by differences in symbiont membership. These results provide insight into the diversity, composition and structure of kogiid gut microbiomes and indicate that host identity plays an important role in structuring cetacean microbiomes, even at fine-scale taxonomic levels.


Results
Composition of cetacean gut microbiomes. A total of 1,720 symbiont OTUs were recovered from K. breviceps (n = 1,368) and K. sima (n = 890), representing 11 bacterial phyla and one archaeal lineage (Fig. 1, Table S1). The gut microbiomes of K. breviceps and K. sima were dominated by the bacterial phyla Firmicutes and Bacteroidetes, together accounting for >68% of the total gut microbial communities in each kogiid host. Actinobacteria, Proteobacteria, Synergistetes and Verrucomicrobia were also common members of kogiid gut microbiomes ( Fig. 1), with six additional rare phyla (<1% relative abundance) detected in both K. breviceps and K. sima (Table S1). Notably, some differentiation of gut microbiomes in K. breviceps and K. sima was observed at the phylum level. In total, five phyla exhibited differential abundances between hosts, including significantly (P < 0.05) higher relative abundances of Bacteroidetes, Verrucomicrobia and Lentisphaerae in K. breviceps, and Actinobacteria and Cyanobacteria in K. sima (Table S1).
Comparative analysis with other toothed and baleen whale species revealed that the dominance of the bacterial phyla Firmicutes and Bacteroidetes in kogiid microbiomes was more characteristic of baleen whales than those of other toothed whales (Fig. 1). The vast majority (84 to 91%) of the gut microbiomes in baleen whale hosts (humpback whales, right whales, and sei whales) were comprised of Firmicutes and Bacteroidetes, whereas non-kogiid toothed whale hosts were dominated by Proteobacteria (60% in bottlenose dolphins) and Actinobacteria (66% in beluga whales, Fig. 1). Despite these shared dominant phyla, several bacterial phyla distinguished kogiid microbiomes from those in baleen whales: Actinobacteria and Synergistetes were common in kogiid hosts (6-15 and 2-4%, respectively) while rare in baleen hosts (<2 and <0.2%, respectively), while Spirochaetes were common members of baleen whale microbiomes (3-8%) and extremely rare in kogiid hosts (<0.03%).
Diversity of cetacean gut microbiomes. Gut microbiomes in kogiid whales exhibited high OTU-level diversity, averaging 432 (±7 SE) and 416 (±18 SE) symbiont OTUs per host in K. breviceps and K. sima, respectively. No significant differences in richness (S), diversity (1/D), evenness (E 1/D ) and dominance (d) were observed in gut microbial communities between kogiid hosts (post hoc Tukey's HSD tests, Fig. 2), indicating similar richness and evenness of gut microbiomes in these closely related species. Broader comparisons of diversity including previously characterized cetacean gut microbiomes revealed that kogiid whales hosted microbial communities with significantly (P < 0.05) higher diversity than those occurring in other toothed whales (bottlenose dolphins = 52 ± 18 OTUs, beluga whales = 84 ± 8, post hoc Tukey's HSD tests, Fig. 2). Similarly, a single OTU dominated the microbial communities of bottlenose dolphins (67%) and beluga whales (74%), while the most abundant OTU accounted for less than a quarter of the microbiome in K. breviceps (23%) and K. sima (15%). In contrast, kogiid gut microbiomes did not differ significantly in richness, diversity and dominance compared to the gut microbial communities inhabiting three baleen whale species (post hoc Tukey's HSD tests, Fig. 2).
Community structure of cetacean gut microbiomes. Despite similarities in community diversity, the gut microbiomes in K. breviceps and K. sima exhibited significant differences in community structure according to both OTU-based and phylogenetic beta-diversity metrics (PERMANOVA, P < 0.01, Table 1), indicating that distinct microbial communities inhabit each host species. These differences were visible in cluster plots based on microbiome similarity, which grouped kogiid individuals by host species (Fig. 3). Importantly, these structural differences were driven by differences in the relative abundance of shared symbiont taxa, as analyses based on symbiont membership revealed no significant differences (PERMANOVA, P > 0.11) in microbiome community structure between kogiid hosts for either beta-diversity metric ( Table 1, Fig. S1) and no host-specific clustering of microbiomes in cluster plots (Fig. S1). Indeed, the 538 OTUs shared between K. breviceps and K. sima accounted for >99% of the recovered sequences, while the unique OTUs in each hosts' microbiome (n = 830 in K. breviceps, n = 352 in K. sima) were extremely rare (in total, representing <1% of all sequences). Significant differences in dispersion were detected between the microbiomes of K. breviceps and K. sima (PERMDISP, P < 0.01, Table S2), indicating greater intra-specific variability in microbial communities among K. sima individuals. Again, these differences were driven by variability in the relative abundance of shared symbionts, as analyses based on symbiont membership revealed no significant differences in dispersion (PERMDISP, P > 0.05, Table S2). No significant differences in the structure of kogiid gut microbiomes were detected based on sex (ANOSIM, P > 0.18) or carcass condition (P > 0.36, Table S3).
The high degree of host-specificity exhibited by kogiid gut microbiomes was also observed across more distantly related host species, including other toothed whales and baleen whales, as cluster plots based on microbial community similarity grouped all individuals by host species (Fig. 3). Host species accounted for the majority (69%) of variation observed between samples and significant differences between hosts were observed for most pairwise comparisons (Table 1), with exceptions generally involving pairwise comparisons with low statistical power (i.e. hosts with low replication, T. truncatus and D. leucas). Differences between the gut microbiomes of these more distantly related cetacean hosts were driven by symbiont membership, rather than relative abundance of shared taxa alone, as analyses of presence-absence data revealed a higher proportion of variation (83%) explained by host species and more distinct clustering of host species in cluster plots (Fig. S1).
Dominant symbiont OTUs in kogiid gut microbiomes. Core microbiomes in kogiid whales were identified as symbiont OTUs detected in all samples of each host species, thereby excluding potentially transient members of the recovered communities. Core microbiomes were comprised of 147 and 155 OTUs in K. breviceps and K. sima, respectively, and represented a small fraction of overall symbiont diversity (10.7% in K. breviceps, 17.4% in K. sima). However, core OTUs also represented dominant symbiont taxa and accounted for the vast majority of overall symbiont communities (96.2 to 97.4%). A high degree of overlap in symbiont membership was observed when comparing the core communities in K. breviceps and K. sima. The shared component of core microbiomes in these two kogiid hosts was comprised of 115 OTUs that accounted for most of the sequence reads recovered from K. breviceps (95.6%) and K. sima (94.0%, Table S4).
The 25 most common OTUs in kogiid microbiomes were examined in further detail to compare their relative abundances between kogiid hosts and presence in the microbiomes of other toothed and baleen whales ( Table 2). All 25 OTUs represented core members of both kogiid species microbiomes, yet exhibited different relative abundances in each host. Together, these 25 OTUs contributed to 66.3% of the variation observed between the microbiomes of K. breviceps and K. sima, with 14 OTUs (56%) exhibiting significantly different abundances between kogiid hosts ( Table 2). Eighteen (72.0%) of these OTUs were unique to kogiid microbiomes (i.e., not detected in the other toothed and baleen hosts included herein) and 15 OTUs (60%) exhibited greater than 3% sequence divergence from any previously characterized sequence (GenBank database), likely representing novel bacterial species. Combined with the unresolved taxonomic status of some dominant OTUs, these results indicate that several novel taxa comprise kogiid microbiomes and are most closely related to symbionts identified in gut communities of other mammalian hosts ( Table 2). For example, the most abundant OTU in both kogiid hosts (OTU1) was unique to kogiid microbiomes, unclassified below the phylum level (Bacteroidetes) and exhibited over 9% sequence divergence from the most closely related bacterium, a symbiont characterized from elephant (Elephas maximus) feces ( Table 2).
The few symbiont OTUs shared between gut communities of kogiid and other cetacean hosts were generally in very low abundance in one or both hosts, with notable exceptions. OTU2 (Firmicutes, Peptostreptococcaceae) was a dominant member of the microbiomes in both kogiid hosts (10.2% in K. breviceps, 6.7% in K. sima) as well as bottlenose dolphins (18.5%). Mycobacterium arupense (Actinobacteria, OTU4) and Oscillospira sp. (Firmicutes, OTU8) were common in K. breviceps (3.8 and 2.6%, respectively) and K. sima (1.3 and 1.0%), with the former symbiont a dominant member of beluga whale microbiomes (65.3%) and the latter symbiont a dominant member of humpback whale microbiomes (14.5%).

Discussion
Characterization of gut microbiomes in pygmy (K. breviceps) and dwarf (K. sima) sperm whales revealed diverse and host-specific microbial communities inhabiting kogiid whales. Gut microbiomes in K. breviceps and K. sima exhibited similar levels of richness and evenness, along with high overlap in symbiont membership, yet

Richness (S)
Diversity (1/D)  Figure 2. Diversity metrics of kogiid gut microbiomes compared with previous data from other toothed (gray bars) and baleen (black bars) whales 25 . Different letters above bars denote significantly different means among host species (P < 0.05). No significant differences in evenness occurred between host species (n.s. = not significant). Error bars represent ±1 standard deviation.
were differentiated by the relative abundance of shared symbiont taxa. As a result, distinct microbial community structure was detected between kogiid species, as well as, among kogiid microbial communities and previously characterized gut microbiomes of other toothed and baleen whales. Notably, these latter differences were driven by differences in community membership and several dominant kogiid-associated OTUs exhibited high levels of differentiation from any described bacteria, indicating the presence of unique symbiont taxa in these hosts. Kogiid gut microbiomes exhibited unique structure and composition and a high-degree of host-specificity, thus highlighting the importance of host identity in structuring cetacean microbiomes. The diversity of gut microbial communities in kogiid whales was unexpectedly high, as previous investigations of toothed whale gut microbiomes revealed low levels of phylum-and OTU-level diversity. Symbiont OTU richness in K. breviceps (432 ± 7) and K. sima (416 ± 18) was nearly five times higher than the gut microbiomes of bottlenose dolphins (T. truncatus 52 ± 8) and beluga whales (D. leucas 84 ± 18) included in comparative analyses  herein. In contrast to the kogiid colonic content sampled from stranded wild individuals, fecal samples from bottlenose dolphins and beluga whales were collected following defecation by captive individuals 25 . However, similarly low diversity communities have been reported in other toothed whale species, including rectal swabs of wild (36 ± 2) and captive (41 ± 8) bottlenose dolphins 20 , striped dolphins (Stenella coeruleoalba, 75 OTUs) 23 and colonic contents of the Yangtze finless porpoise (Neophocaena phocaenoides asiaeorientalis, 30 OTUs) 24 , indicating that different sampling methods and host status (wild vs. captive) alone do not account for the observed differences in microbiome richness between kogiids and other toothed whales. As such, kogiid whales appear to represent a more fertile microbial habitat compared to other toothed whale species, possibly related to the unique colonic sac that enlarges the large intestine of kogiid whales (e.g. 50 L of feces were extracted from a 295 cm individual K. breviceps) 35 . The presence of this colonic feature, and its potential to sequester feces for extended periods of time, are also hypothesized to contribute to the detectability of the neurotoxin domoic acid in kogiids 36 . Otherwise, kogiid gastrointestinal tracts exhibit similar morphology to those of most other odontocetes (including Physeter macrocephalus) and mysticetes, with a multi-chambered stomach consisting of single fore-, main and pyloric chambers 37 . Interestingly, the low diversity of gut microbiomes in non-kogiid toothed whales contrasts with other body sites of these animals, where high diversity microbial communities have been reported. For example, the oral microbiome of bottlenose dolphins displayed twice the phylum-level and four times the OTU-level diversity compared to their gut microbiomes 20 . Whether the observed differences in microbiome diversity between kogiid and other toothed whales translate into differences in community functionality and stability (e.g. via functional redundancy) are important questions for understanding host-symbiont interactions and will require additional metagenomic and cultivation-based studies of whale microbiomes. Kogiid whale gut microbiomes were dominated by symbiont taxa affiliated with the phyla Firmicutes and Bacteroidetes, two common lineages in mammalian gut communities. Members of the phylum Firmicutes were rare in some non-kogiid toothed whale hosts, for example, representing <2% of gut communities in beluga whales. In bottlenose dolphins, high variability was observed in the relative abundance of Firmicutes symbionts (<1 to 55%), though recent work indicates that Firmicutes are indeed dominant members of the bottlenose dolphin gut microbiome 20,38 . In contrast, Bacteroidetes were greatly reduced (<2%) in gut microbiomes of both non-kogiid toothed whales investigated herein, consistent with previous studies 20, 38 , while comprising a large fraction of symbiont communities in K. breviceps (31%) and K. sima (13%). Bacteroidetes are common members of herbivorous 14  these symbionts are thought to play a key role in the digestive health of mammals and their absence in many toothed whale species may reflect alternative nutrient acquisition strategies. Importantly, Bacteroidetes have been reported as common (even dominant) in other body sites of toothed whale hosts, including the oral, upper gastric and respiratory microbiomes of bottlenose dolphins 20 , thus their absence in the lower gut may indicate environmental selection against these taxa in this particular microhabitat of these hosts. The structural determinants of gut microbiome composition in marine mammals include age, diet and phylogenetic position 17,20,25 , similar to studies of terrestrial counterparts 42,43 . In this study, the similar foraging niches and dietary overlap between K. breviceps and K. sima 27 allowed for the investigation of cetacean microbiomes in a comparative context controlling for broad differences in diet and isolating host-specific factors. Our results suggest that host identity plays an important role in structuring cetacean microbiomes, even at fine-scale taxonomic levels, as distinct microbiome structure was observed between closely-related kogiid hosts. In contrast, host sex and carcass condition showed no significant effects on microbiome structure in kogiid whales. Evidence to date indicates that sex is not an important structuring factor for marine mammal microbiomes, as past studies have similarly reported no significant effect of sex on gut microbiomes of bottlenose dolphins 20 , leopard seals (Hydrurga leptonyx) 17 , manatees (Trichechus manatus) 16 and dugongs (Dugong dugon) 14 . An exception to this trend occurs in elephant seals (Mirounga leonina) which, in contrast to the aforementioned species, exhibit pronounced sexual size dimorphism hypothesized to drive observed differences in gut microbiomes between males and females 17 . Future studies of gut microbial communities in cetacean species that display drastic sexual size dimorphism (e.g. the true sperm whale, Physeter macrocephalus) will yield further insight into the impacts of sex-specific variations in body size and foraging behavior on gut microbiome structure. Importantly, no effect of carcass condition was recovered herein. These results are consistent with recent studies of postmortem mammal microbiomes (e.g. murine models) 44 showing that endogenous gut bacteria continue to dominate gastrointestinal communities until advanced stages of decomposition, specifically intestinal rupture and the exposure of the abdominal cavity to oxygen. The postmortem stability of gut-associated communities indicates that stranded whales in early stages of decomposition can be used to provide accurate representations of cetacean gut microbiomes in wild ranging animals.
The taxonomic composition of kogiid gut microbiomes also offers insight into the putative functionality of gut-associated bacteria, as several core symbiont taxa were classified into bacterial lineages with known physiological capabilities. Facultative and obligate anaerobes with fermentative metabolism were common, including members of the families Ruminococcaceae (Faecalibacterium, Oscillospira), Enterobacteriaceae (Citrobacter) and Lachnospiraceae (Buytrivibrio). Microbial fermentation transforms undigested dietary components into end-products readily metabolized by mammal hosts, namely short-chain fatty acids (SCFA). Similar to ruminant mammals, cetacean gastrointestinal tracts include a non-glandular forestomach that likely functions as a fermentation vessel, based on cultivation studies of forestomach bacteria 45 , SCFA profiling of this chamber 46 and metagenomic characterization of carbon metabolism in whales 25 . The taxonomic data herein, combined with past evidence of fermentation in sperm whales (based on bile acid composition) 47 , indicates that fermentative metabolism occurs in the gastrointestinal tract of K. breviceps and K. sima and potentially contributes to energy acquisition during prey digestion. Other core members of kogiid gut microbiomes were affiliated with sulfate reducing bacteria, including Desulfovibrio (Desulfovibrionaceae) and Desulfomonile (Syntrophaceae), which may persist in the anaerobic habitat of whale intestines by deriving energy from the respiration of sulfate present in small volumes of seawater ingested with prey. Finally, several putative pathogens were detected in K. breviceps and K. sima, including Campylobacter fetus, Clostridium perfringens and Mycobacterium arupense. The clinical relevance of these bacteria is unclear, although their role as opportunistic pathogens in other hosts 48 , taxonomic distinction from pathogens reported for toothed whale species 49 and prevalence in diverse cetacean hosts 50 suggest low virulence and potentially a greater threat to immunocompromised individuals.
In summary, this study characterized the structure and taxonomic composition of gut microbiomes of kogiid whales, a critical first step in understanding the role of cetaceans as microbial habitats and the contributions of symbiont communities to host metabolism and health. The selective pressures that contribute to the establishment and maintenance of host-specific microbiomes in kogiid whales may include unique microhabitats within each host, functional contributions of specific microbial guilds to host fitness, or a combination of both phenomena. Future studies targeting how these diverse, host-specific microbiomes form (e.g. comparative analyses across cetacean life stages) and their physiological characteristics (e.g. metagenomic and culture-based investigations) will provide further insight into microbiome development and functionality. Ultimately, a holistic understanding of host-microbe interactions in kogiid whales may yield insights into the impacts of gut dysbiosis on health and aid in health assessments and rehabilitation efforts for these species, which currently exhibit very low success rates due to mortality from gastric and intestinal stasis 35 .  Table 3, Fig. 4) between 2009 and 2014 and stored at −80 °C. All sampled individuals were adults (length range = 220-328 cm, weight range = 184-515 kg) stranded in fresh to moderate carcass condition (Carcass Classification Codes 1-3) 51 and exhibited no signs of direct, human-induced mortality (e.g. boat collisions, fishery interactions). Sampled individuals included both sexes (Table 3).

Ethics
DNA extracts were prepared from 200 mg of fecal material using the Powersoil DNA Extraction kit (MoBio), following Earth Microbiome Project standard protocols (http://press.igsb.anl.gov/earthmicrobiome/protocols-and-standards/dna-extraction-protocol/). DNA extracts were sent to Molecular Research LP (Shallowater, TX) for amplification, library construction and multiplexed sequencing of partial (V4) 16S rRNA gene sequences on an Illumina MiSeq platform. DNA extracts were used as templates for PCR amplifications using the HotStarTaq Plus Master Mix kit (Qiagen) and the universal bacterial/archaeal forward primer 515 f and reverse primer 806r 52 , with a multiplex identifier barcode on the forward primer. Thermocycler conditions consisted of an initial denaturation step at 94 °C for 3 min; 28 cycles of 94 °C for 30 s, 53 °C for 40 s, and 72 °C for 1 min; and a final elongation step at 72 °C for 5 min. Samples were pooled in equimolar concentrations, purified using Agencourt Ampure XP beads (Beckman Coulter) and sequencing on an Illumina MiSeq following the manufacturer's guidelines. For comparative analyses, previously characterized cetacean gut microbiome datasets 25 were downloaded for bottlenose dolphins (Tursiops truncatus), beluga whales (Delphinapterus leucas), humpback whales (Megaptera novaeangliae), right whales (Eubalaena glacialis) and sei whales (Balaenoptera borealis, Table S5). These datasets were constructed using the same extraction kit, primer pair and sequencing platform as the data generated herein. DNA Sequence Processing. Combined datasets from this study and Sanders et al. 25 were processed simultaneously in mothur 53 using a modified version (full details and mothur code in Table S6) of a previously described pipeline 54 . Briefly, sequence reads were quality-filtered, aligned to the SILVA database (release 119, non-redundant, mothur-formatted) and trimmed to the V4 region, screened for sequencing anomalies (e.g. chimeras) and errors, then assigned to taxonomic groups using a naïve Bayesian classifier 55 and Greengenes taxonomy (May 2013 release, mothur-formatted). Following the removal of non-target amplicons (chloroplasts, mitochondria and eukaryotic reads), the remaining high quality sequences were clustered into operational taxonomic units (OTUs) at 97% sequence identity (average neighbor algorithm) and taxonomy assigned to each OTU by majority consensus 56 . Singleton OTUs (occurring once in the global dataset) were removed and datasets were subsampled to lowest read count (n = 24, 389) to avoid artifacts of varied sampling depth on diversity calculations. Raw sequence data were deposited as FASTQ files in the Sequence Read Archive of the National Center for Biotechnology Information (SRA NCBI) under the accession no. SRP097888.

Statistical Analyses.
To compare community diversity among cetacean hosts, alpha-diversity indices were calculated for OTU richness (observed richness, S), diversity (inverse Simpson, 1/D), evenness (Simpson, E 1/D ) and dominance (Berger-Parker, d), as implemented in mothur. Analyses of variance (ANOVAs) were used to compare diversity index means across cetacean host species, followed by Tukey's honest significant difference (HSD) tests for multiple pairwise post hoc comparisons.
To compare community structure among cetacean hosts, beta-diversity indices were calculated based on an OTU-dependent metric (Bray-Curtis similarity) and an OTU-independent (i.e. phylogenetic) metric (UniFrac distance) 57 . Two iterations of each beta-diversity metric were performed to differentiate between the effects of taxon abundance (OTU relative abundance Bray-Curtis, weighted UniFrac) and  Table 3. Stranded individuals of K. breviceps and K. sima examined in this study and associated metadata for each sample (n.a. = data not available). * Carcass condition at stranding (left) and at examination (right): 1 = Alive, 2 = Fresh Dead, 3 = Moderate Decomposition.
SCieNtifiC REPORTs | 7: 7205 | DOI:10.1038/s41598-017-07425-z taxon membership (OTU presence-absence Bray-Curtis, unweighted UniFrac) on the structure of cetacean microbiomes. Bray-Curtis similarity values were calculated from OTU tables in PRIMER (version 6.1.11) and visualized in cluster dendrograms. UniFrac distances were calculated in mothur based on an approximate maximum-likelihood phylogeny constructed in FastTree 2.1.5 58 with unique sequence reads (n = 52, 867). Permutational multivariate ANOVAs (PERMANOVA+, version 1.0.1) were used to test for differences in microbiome structure among host species, with significance determined by Monte Carlo asymptotic P-values corrected for multiple pairwise comparisons (Benjamini-Yekutieli false-discovery rate control 59 and an experiment-wise error rate of α = 0.05), and to estimate components of variation ascribed to the factor 'host species' and to residual variation. Permutation multivariate analyses of dispersion (PERMDISP) were conducted to test for homogeneity of multivariate dispersions (i.e. deviations from centroids) among host species, with significance determined by permutational P-values similarly corrected for multiple pairwise comparisons. Additional analyses of kogiid microbiome data were conducted to identify individual symbiont taxa contributing to community-level differences in microbiome structure between K. breviceps and K. sima. At the phylum level, significant differences in the relative abundance of bacterial phyla between kogiid hosts were determined using Student's t tests. At the OTU level, a one-way similarity percentage species contributions (SIMPER) analysis was conducted to investigate the contribution of each symbiont OTU to the observed community dissimilarity between kogiid hosts. In addition, symbiont OTUs that were differentially abundant between kogiid hosts were identified using Metastats (non-parametric t tests) 60 and LefSe (non-parametric Kruskal-Wallis sum-rank tests) 61 .
To test for microbiome differentiation based on sex and carcass condition, separate two-way analyses of similarity (ANOSIM) were conducted for the factors 'sex' and 'carcass condition' , each crossed with the factor 'host species' to control for differences between kogiid species. To exclude potentially transient members of kogiid symbiont communities, core microbiomes were identified for each kogiid host and strictly defined as symbiont OTUs detected in all samples within a host species.