See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/315774827

Host species shapes the co-occurrence patterns

rather than diversity of stomach bacterial
communities in pikas

Article in Applied Microbiology and Biotechnology April 2017

DOI: 10.1007/s00253-017-8254-0

CITATIONS READ

0 1

5 authors, including:

Huan Li Bo Tu
Chengdu insititute of Biology, China Chengdu Institute of Biology, CAS, Chengdu,
11 PUBLICATIONS 18 CITATIONS 2 PUBLICATIONS 0 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Pika gut microbiomenews links

http://www.cib.ac.cn/xwdt/kydt/201702/t20170207_4741801.html View project

All content following this page was uploaded by Huan Li on 06 April 2017.

The user has requested enhancement of the downloaded file.

Appl Microbiol Biotechnol
DOI 10.1007/s00253-017-8254-0
ENVIRONMENTAL BIOTECHNOLOGY
Host species shapes the co-occurrence patterns rather than
diversity of stomach bacterial communities in pikas
Huan Li 1,2 & Tongtong Li 1 & Bo Tu 1,2 & Yongping Kou 1 & Xiangzhen Li 1
Received: 4 December 2016 / Revised: 10 March 2017 / Accepted: 16 March 2017
# Springer-Verlag Berlin Heidelberg 2017
Abstract The mammalian stomach acts as an important bar-
rier against ingested pathogens into the entire gastrointestinal
tract, thereby playing a key role in host health. However, little
is known regarding to the stomach microbial compositions in
wild mammals and the factors that may influence the commu-
nity compositions. Using high-throughput sequencing of the
16S rRNA gene, we characterized the stomach bacterial com-
munity compositions, diversity, and interactions in two com-
mon pika (Ochotona sp.) species in China, including Plateau
pikas (Ochotona curzoniae) and Daurian pikas (Ochotona
daurica) living in the Qinghai-Tibet Plateau and the Inner
Mongolia Grassland, respectively. The bacterial communities
can be divided into two distinct phylogenetic clusters. The
most dominant bacteria in cluster I were unclassified bacteria.
Cluster II was more diverse, predominantly consisting of
Bacteroidetes, followed by unclassified bacteria, Firmicutes
and Proteobacteria. Three dominant genera (Prevotella,
Oscillospira, and Ruminococcus) in pika stomachs were sig-
nificantly enriched in cluster II. In addition, seasons, host spe-
cies, and sampling sites as well as body weight and sex had no
significant impacts on the composition and diversity of pika
Huan Li and Tongtong Li contributed equally to this work.
Electronic supplementary material The online version of this article
(doi:10.1007/s00253-017-8254-0) contains supplementary material,
which is available to authorized users.
* Xiangzhen Li
lixz@cib.ac.cn
1
Key Laboratory of Environmental and Applied Microbiology, CAS;
Environmental Microbiology Key Laboratory of Sichuan Province,
Chengdu Institute of Biology, Chinese Academy of Sciences,
Chengdu 610041, Peoples Republic of China
2
University of Chinese Academy of Sciences, Beijing 10049, Peoples
Republic of China
stomach communities. Interestingly, Plateau pikas harbored a
more complex bacterial network than Daurian pikas, and these
two pika species showed different co-occurrence patterns.
These results suggested that the pika stomach harbors a di-
verse but relatively stable and unique bacterial community,
which is independent on host (host species, body weight,
and sex) and measured environmental factors (sampling sites
and seasons). Interestingly, host species shapes the microbial
interactions rather than diversity of stomach bacterial commu-
nities in pikas, reflecting specific niche adaptation of stomach
bacterial communities through species interactions.
Keywords Bacterial community . Pikas . Stomach .
Diversity . Co-occurrence patterns
Introduction
The mammalian stomach plays a critical role in the break-
down of food, the denaturation of certain protein, and the
prevention against pathogen infection (Adbi 1976;
Tennant et al. 2008). Thus, some potential disease associ-
ated with ingested food can be prevented. The composi-
tion of the stomach microbial community has been inves-
tigated in various mammalian species, including human
(Bik et al. 2006), horse (Perkins et al. 2012), and woodrat
(Kohl et al. 2014). These reports showed that phyla
F i r m i c u t e s , P ro t e o b a c t e r i a , B a c t e ro i d e t e s , a n d
Actinobacteria are dominant in the stomach of mammals.
In addition, laboratory work has demonstrated that the
stomach communities may be influenced by multiply fac-
tors, including disease, immune status, antibiotic medica-
tion, and pH (Aviles-Jimenez et al. 2014; von Rosenvinge
et al. 2013). Wild mammals are exposed to more diverse
microbes associated with complex natural environments.

The stomach is at the front of the digestive system; thus,
this region may have more chances to be invaded by for-
eign pathogenic microbes compared with other hindgut
regions (e.g., small intestine, cecum or colon). It has been
suggested that gut microbiota was influenced by various
host (e.g., host species, body weight, or sex) and environ-
mental (e.g., sampling sites or seasons) factors in wild
mammals (Amato 2013; Amato et al. 2015; Maurice
et al. 2015). However, it remains unclear whether these
factors are able to shape the stomach microbiota in wild
mammals.
Pikas are small, herbivorous mammals that belong to
the order Lagomorpha. They play key roles in grassland
degradation, soil carbon cycle, and ecosystem diversity
(Li et al. 2013; Smith and Foggin 1999). Most pika spe-
cies are mainly distributed in Asia (Yang et al. 2008). In
China, there are two common pika (Ochotona sp.) spe-
cies, including Daurian pikas (O. daurica) living in the
Inner Mongolia Grassland with 1000 m above sea level
(m ASL) and Plateau pikas (Ochotona curzoniae)
inhabiting the Qinghai-Tibet Plateau with more than
3000 m ASL. The habitats of Plateau pikas are generally
alpine meadows that are dominated by herbaceous
perennids (Liu et al. 2008), while Daurian pikas mainly
live in the typical and desert steppe (Pei et al. 2008),
which are dominanted by grasses, shrubs, and forbs.
Different living environments have lead to the divergence
of their hosts genes (Yu et al. 1998) and diets (Fan et al.
1995), which may shape the microbial composition of
gastrointestinal tract. Our previous study has demonstrat-
ed that their cecal microbiota has a significant structural
segregation (Li et al. 2016b). However, it remains
unkown whether the stomach microbiota structure of
Plateau and Daurian pikas has a signicant divergence.
In addition, due to the seasonal changes of diet composi-
tions in Plateau pikas (Liu et al. 2009), the stomach microbiota
may vary in response to seasonal changes. Such variations
may finally influence host nutrition and health. However, the
knowledge on the season shifts of pika stomach microbial
communities remains unclear.
Here, we used the high-throughput sequencing of the
16S ribosomal RNA (rRNA) gene to characterize the
stomach bacterial community compositions, diversity
and interactions of pikas, and host (e.g., host species,
body weight and sex) and environmental (e.g., sampling
sites and seasons) factors that may influence the bacte-
rial communties. Our data suggested that the diversity
and structure of stomach bacterial communities in pikas
were independent on these host and environmental fac-
tors, while Plateau pikas and Daurian pikas had differ-
ent co-occurrence patterns.These results have important
implications for understanding host-microbe interactions
and stomach function in wild mammals.
Appl Microbiol Biotechnol
Materials and methods
Sample collection
Samples were collected in summer between July and August,
2014 with the exception of the group GLW (Sampling from
Guoluo in winter, 2014). In the Qinghai-Tibet Plateau, wild
Plateau pikas were captured from three sampling sites, includ-
ing Xiaderi (XD), Wangjiaxiang (WJ), and Guoluo (GL). The
groups GL and GLW were from the same sampling site
(Guoluo) but different seasons (summer/winter). In the Inner
Mongolia Grassland, wild Daurian pikas were collected from
two sites, including Xiwuqi (XW) and Baiyinxile (BY). The
habitat type of the Plateau pikas is alpine meadow, which is
dominated by the plant species Kobresia humilis. In contrast,
Daurian pikas occupy the typical steppe, which is dominated
by the plant Stipa spp. and Leymus chinensis. Each pika was
euthanized and then dissected in a sterile box after capture.
The method for stomach sampling in this study was similar to
that reported by Kohl et al. (2014). The whole content from
the stomach for each individual was collected. Totally, 61 wild
pika samples were collected, including 36 Plateau pikas and
25 Daurian pikas. The stomach contents were immediately
placed into 50-ml sterile tubes and frozen provisionally at
20 C in a portable freezer. All the samples were transferred
to the laboratory in Chengdu Institute of Biology within 24 h,
and stored at 40 C for DNA extraction. The sample infor-
mation is given in Supplementary Table S1.
This study did not involve any endangered or protected
wild animal species. All animal experiments and protocols
were strictly in conformity with the Institution of Animal
Care and the Ethics Committee of Chengdu Institute of
Biology, Chinese Academy of Sciences.
DNA extraction, PCR amplification, and 16S rRNA
sequencing
The detailed procedures of DNA extraction, PCR amplifica-
tion, and gel extraction were performed as described previous-
ly (Li et al. 2016a). For each sample, PCR amplification was
performed in duplicate using the universal primer pairs
515F/909R, which can amplify the V4-V5 region of bacterial
and archaeal 16S rRNA (Tamaki et al. 2011). The 5-end of the
515F primer included 12-bp unique barcodes, which were
used to split each sample. Finally, the purified PCR products
were pooled in equal concentrations and sequenced using an
Illumina MiSeq sequencer at the Environmental Genomic
Platform of the Chengdu Institute of Biology, CAS.
Bioinformatics analysis
The raw reads were processed and analyzed in QIIME
Pipeline-version 1.7.0 (http://qiime.org/scripts/index.html).
Appl Microbiol Biotechnol
Sequences of each sample were isolated according to their
unique barcodes. Sequence analysis followed the methods of
Li et al. (2016c). Briefly, after removing low-quality se-
quences, chloroplasts, and chimeras, all the remaining se-
quences were then clustered into operational taxonomic units
(OTUs) at a 97% sequence identity using CD-HIT. Eukaryota
and Archaea lineages only occupied a small minority of total
reads (<0.01%) and were thus removed in the current study.
Those sequences not classifying to bacteria and singletons
were also filtered out. Representative sequences for each
OTU were picked using the script Bpick_rep_set.py^ in the
QIIME platform. Finally, the representative sequences were
aligned against the Greengenes 13_8 reference database
(DeSantis et al. 2006) using PyNAST, and were classified
using the RDP classifier (Wang et al. 2007).
In order to compare all samples with different reads,
each sample was normalized to 4698 sequences. Thus,
286,578 total sequences were used for the following anal-
ysis. To evaluate alpha diversity indices, we calculated the
phylogenetic diversity and Shannon diversity of stomach
bacterial communities. To assess beta diversity, the
weighted (for community structure) UniFrac distance ma-
trices (Hamady et al. 2010) were calculated. Differences
in overall bacterial community structure among groups
were visualized using the PCoA plots of weighted
UniFrac dissimilarity matrices. To compare the similarity
of stomach communities across individuals, the Jackknife
hierarchical clustering of weighted UniFrac scores was
conducted. Finally, the phylogenetic tree was visualized
using the Figtree v1.4.2 (Rambaut 2009).
Statistical analysis
In order to compare the differences of stomach bacterial
communities between host species, seasons, or sampling
sites, analysis of similarity (ANOSIM) was used to test
the significant differences of microbial community struc-
ture using the procedure Banosim^ in the R Bvegan^
package (Warton et al. 2012). The model also com-
prised other variables, including body weight and sex.
To assess the optimal number of clusters across all sam-
ples, the Calinski-Harabasz (CH) and the silhouette
score (Moeller et al. 2012) were calculated for the
Jackknife clusters using the procedure BcascadeKM^ in
the R vegan package (Oksanen et al. 2013). The maxi-
mum value of CH usually indicates the correct numbers
of clusters. Taxonomic composition for each cluster was
evaluated at the phylum and genus level. One-way anal-
ysis of variance (ANOVA) and Turkeys post hoc tests
were used to test the differences in relative abundances
of these taxonomic units and alpha diversity measures
among different groups.
Network analysis and identification of putative keystone
taxa in pika stomachs
In order to understand the co-occurrence interactions of
stomach bacterial communities between Daurian and
Plateau pikas, phylogenetic molecular ecological net-
works (pMENs) (http://129.15.40.240/mena/) were
constructed based on the OTUs from the 16S rRNA
gene sequences using Random Matrix Theory (RMT)-
based methods (Deng et al. 2012). Because all groups
were derived from summer except the group GLW from
winter, thus the samples from GLW were excluded for
network analysis. The protocols of network construction
were described previously (Li et al. 2016b). Briefly, only
those OTUs that were present in more than half of all
samples in each pika species were used for network con-
struction. In order to compare the topological features of
bacterial networks between Daurian and Plateau pikas, the
two pMENs were generated with the same threshold (0.
72). For each network, total nodes, total links, average
degree (avgK), average clustering coefficient (avgCC),
density (D), centralization of betweenness (CB), within-
module connectivity (Zi), and among-module connectivity
(Pi) were performed using the window Bglobal network
properties^ and Bmodule separation and modularity
calculation^ in the above website. Finally, visualization
of these ecological networks was made with Cytoscape
3.0.2. (Saito et al. 2012).
A module generally consists of many nodes (or
OTUs), which are tightly connected with each other in
one group, while these nodes have only few connections
outside the group. The values of Zi and Pi can deter-
mine the features and connectivity of each node, and
thus are often used to identify the topological role of
these nodes (Deng et al. 2012). These nodes can be
classified into four categories (Zhou et al. 2011), includ-
ing peripherals (OTUs connected in modules with few
outside connections, Zi <2.5 and Pi <0.62), connectors
(OTUs that connect modules, Pi >0.62), network hubs
(highly connected OTUs within entire network, Zi >2.5
and Pi >0.62), and module hubs (highly connected
OTUs within modules, Zi >2.5). Module hubs, network
hubs, and connectors were considered as putative key-
stone taxa in a microbial community (Shi et al. 2016).
The keystone taxa of stomach communities in Daurian
pikas and Plateau pikas were identified based on the
above rules.
Nucleotide sequence accession numbers
The original 16S rRNA data were available at the European
Nucleotide Archive by accession no. PRJEB18037 (http://
www.ebi.ac.uk/ena/data/view/PRJEB18037).

Results
Differences of stomach bacterial community compositions
between host species or seasons
Across all samples, the pika stomachs were dominated
by unclassified bacteria (49.8%), Bacteroidetes (26.5%),
Firmicutes (19.2%), and Proteobacteria (2.6%). Other
rare phyla (mean relative abundance <1%) included
A c t i n o b a c t e r i a , S p i ro c h a e t e s , C y a n o b a c t e r i a ,
Acidobacteria, and Tenericutes (Fig. 1). At genus level,
only the mean relative abundances of Prevotella and
Oscillospira were more than 1% of all sequences.
Other important genera included Ruminococcus,
Lactobacillus, Treponema, YRC22, Anoxybacillus,
Streptococcus, Sphaerochaeta, CF231, Allobaculum,
and so on. Although Helicobacter spp. have been re-
ported to be present in human (Linz et al. 2007) and
horse (Contreras et al. 2007) stomachs, none of the
286,578 sequences matched this specific genus, indicat-
ing the absence of Helicobacter spp. in our samples.
We compared the differences of stomach community
compositions between Daurian and Plateau pikas (the
group GLW was excluded). At phylum level, only un-
classified bacteria taxa and Bacteroidetes showed signif-
icant differences between the two pika species (one-way
ANOVA, P < 0.05), while other phyla were not signif-
icantly different (Table 1). At genus level, most of the
genera with mean relative abundances >0.1% were rel-
atively stable between host species with the exception
of Streptococcus and CF231.
In addition, we explored the differences of stomach
community between summer (GL) and winter (GLW) in
Plateau pikas. A majority of phyla were not significant-
ly different between seasons. At genus level, only
Treponema, Streptococcus, and Allobaculum showed sig-
nificant differences (one-way ANOVA, P < 0.05).
Fig. 1 The compositions of
bacterial communities at phylum
level in the stomachs of Plateau
pikas and Daurian pikas. Only
those phyla with mean relative
abundance >1% were shown
Appl Microbiol Biotechnol
The alpha diversity of stomach bacterial communities
The alpha diversity measures (phylogenetic diversity and
Shannon diversity) were calculated in each group. The phylo-
genetic diversity and Shannon diversity showed no significant
differences among all groups with the exception of the group
WJ (Fig. 2). In other words, when the group WJ was removed,
host species and sampling sites as well as seasons had no
significant impacts in shaping the alpha diversity of pika
stomach bacterial communities (one-way ANOVA, P > 0.05).
The beta diversity of stomach bacterial communities
Principal coordinate analysis (PCoA) was used to reveal the
structural segregation of pika stomach bacterial communities
between host species, sampling sites, or seasons (Fig. 3).
Seasons had no significant influences in shaping the stomach
community structure of Plateau pikas (ANOSIM r = 0.039,
P > 0.05). Host species had little impacts in determining the
stomach bacterial communities (ANOSIM r = 0.07,
P = 0.033). In addition, the community structure of neither
Plateau pikas (ANOSIM r = 0.112, P > 0.05) nor Daurian
pikas (ANOSIM r = 0.08, P > 0.05) was driven by sampling
sites. In addition to the above factors, we found that sex
(ANOSIM r = 0.03, P > 0.05) and body weight (ANOSIM
r = 0.07, P > 0.05) of pikas also showed no significant effects
in shaping the stomach bacterial communities.
The compositions and diversity of two phylogenetic
clusters in pika stomachs
In order to evaluate the optimal cluster across all samples, the
Calinski-Harabasz (CH) and silhouette scores were calculated
based on the weighted UniFrac distance (Supplementary
Table S2). Because the CH value was the highest when the
stomach communities formed two clusters, the stomach bac-
terial communities across all the pika samples can be divided
Appl Microbiol Biotechnol

Table 1 The comparison of main phyla and genera (mean relative abundance >0.1%) from pika stomachs between host species, seasons or clusters

Taxon (%) Daurian pikas Plateau pikas P value Summer Winter P value Cluster I Cluster II P value

Phyla
Unclassified bacteria 58.16 4.32 43.65 4.06 0.018 47.02 6.77 46.16 7.56 0.994 69.11 2.05 29.86 1.94 <0.001
Bacteroidetes 18.9 2.82 32.61 3.44 0.004 31.13 5.95 26.62 4.47 0.662 13.38 1.61 40.05 2.52 <0.001
Firmicutes 18.29 2.22 20.37 2.28 0.521 17.38 3.60 16.87 2.60 0.935 13.79 0.98 24.86 2.46 <0.001
Proteobacteria 2.57 0.44 1.72 0.28 0.094 2.39 0.53 8.18 4.83 0.069 1.92 0.24 3.3 0.93 0.16
Actinobacteria 0.55 0.08 0.44 0.05 0.22 0.66 0.08 0.29 0.06 0.018 0.46 0.05 0.49 0.07 0.771
Spirochaetes 0.48 0.12 0.37 0.09 0.464 0.45 0.19 0.94 0.20 0.167 0.41 0.11 0.51 0.09 0.472
Cyanobacteria 0.39 0.06 0.45 0.09 0.603 0.63 0.20 0.43 0.1 0.558 0.42 0.09 0.42 0.05 0.958
Acidobacteria 0.26 0.16 0.07 0.01 0.201 0.06 0.01 0.02 0.004 0.113 0.21 0.13 0.08 0.01 0.328
Tenericutes 0.13 0.02 0.13 0.02 0.99 0.13 0.02 0.14 0.02 0.829 0.11 0.02 0.15 0.02 0.222
Genera
Prevotella 1.92 0.33 3.22 0.57 0.073 2.37 0.57 1.15 0.25 0.213 1.05 0.18 4.04 0.53 <0.001
Oscillospira 1.69 0.42 2.21 0.40 0.377 1.49 0.49 1.7 0.29 0.803 0.86 0.15 3.09 0.43 <0.001
Ruminococcus 0.66 0.11 1.06 0.17 0.069 0.91 0.21 0.89 0.21 0.966 0.4 0.04 1.39 0.16 <0.001
Lactobacillus 0.50 0.07 0.40 0.05 0.25 0.31 0.09 0.51 0.07 0.199 0.39 0.04 0.51 0.1 0.102
YRC22 0.22 0.04 0.41 0.10 0.131 0.24 0.06 0.11 0.04 0.251 0.1 0.02 0.52 0.1 <0.001
Treponema 0.3 0.08 0.24 0.04 0.464 0.18 0.05 0.60 0.11 0.001 0.2 0.03 0.39 0.07 0.025
Anoxybacillus 0.05 0.01 0.35 0.34 0.426 0.84 0.81 0.27 0.15 0.679 0.03 0.01 0.41 0.35 0.289
Streptococcus 0.22 0.04 0.13 0.01 0.017 0.17 0.03 0.03 0.02 0.009 0.15 0.02 0.18 0.04 0.553
Sphaerochaeta 0.14 0.08 0.09 0.06 0.62 0.22 0.15 0.31 0.09 0.706 0.18 0.08 0.08 0.04 0.291
CF231 0.05 0.01 0.17 0.03 0.002 0.11 0.03 0.14 0.06 0.558 0.04 0.01 0.2 0.03 <0.001
Allobaculum 0.06 0.03 0.03 0.02 0.393 0.06 0.04 0.94 0.06 1.89E09 0.09 0.04 0.14 0.06 0.412

Bold values indicate significant differences at P < 0.05

into two distinct phylogenetic clusters (Fig. 4). ANOSIM
analysis confirmed that these two clusters had significantly
different bacterial communities (r = 0.937, P < 0.001). For
Daurian pikas, 17 individuals belonged to cluster I, while
eight animals were affiliated with cluster II. In contrast,
Plateau pikas consisted of 14 individuals from cluster I and
22 individuals from cluster II. Most individuals in Plateau
pikas from winter were also in cluster II. The most dominant
bacteria in cluster I were unclassified bacteria, followed by
Firmicutes, Bacteroidetes, Proteobacteria, and some other
rare phyla. However, the predominant bacteria in cluster II
were Bacteroidetes, followed by unclassified bacteria,
Firmicutes, Proteobacteria, and some other bacteria. The rel-
ative abundances of unclassified bacteria, Bacteroidetes and
Firmicutes showed significant differences between the two
clusters (one-way ANOVA, all P < 0.001). At genus level,

Fig. 2 The comparison of alpha diversity measurements (phylogenetic diversity and Shannon diversity) of stomach bacterial communities across
groups. One-way ANOVA was used to compare the differences. Significant differences were indicated by different letters (P < 0.05)

Fig. 3 Principal coordinate plots of the weighted UniFrac dissimilarity metrics comparing the bacterial communities between host species (a) and
seasons (b)
three most dominant genera in both cluster I and cluster II
were Prevotella, Oscillospira, and Ruminococcus. However,
these three genera as well as YRC22, Treponema, and CF231
were significantly different in relative abundances between
cluster I and cluster II (one-way ANOVA, all P < 0.05).
The alpha diversity values (phylogenetic diversity and
Shannon diversity) of cluster I and cluster II were compared
(Fig. 5). The individuals in cluster II showed higher phyloge-
netic diversity and Shannon diversity than those in cluster I
(one-way ANOVA, both P < 0.001).
Network analysis of bacterial communities and putative
keystone taxa
To understand the microbial interactions of pika stomach
bacterial communities, the bacterial network of each pika
Fig. 4 Principal coordinate plots (a) and phylogenetic clustering (b) of the weighted UniFrac dissimilarity metrics comparing the bacterial communities
between two clusters
Appl Microbiol Biotechnol
species was constructed (Fig. 6). Both the networks of
Plateau and Daurian pikas consisted of 125 bacterial
OTUs. Using the same threshold (0.72), a total of 180
and 197 links were identified for Daurian and Plateau
pikas, respectively. Notably, for Daurian pikas, the per-
centage of negative correlations (126 out of 180) was
the dominant, while positive correlations (133 out of
197) were the predominant in Plateau pikas.
Interestingly, we found that the values of total links, av-
erage degree (avgK), and density (D) were higher in
Plateau pikas than those in Daurian pikas, indicating a
relatively simpler network in Daurian pikas. A lower av-
erage clustering coefficient in Daurian pikas indicated that
the bacterial network was mainly composed of relatively
isolated nodes or loose node groups. However, Daurian
pikas had a higher level of centralization of betweenness
Appl Microbiol Biotechnol

Fig. 5 The comparison of alpha diversity measurements (phylogenetic diversity and Shannon diversity) of stomach bacterial communities between two
phylogenetic clusters. One-way ANOVA was used to compare the differences. Significant differences were indicated by P <0.001

than Plateau pikas, implying that Plateau pikas harbored a
lower frequency of centralized nodes.
The putative keystone species in bacterial networks
were identified in each pika species based on the Zi
and Pi values (Table 2). Six key module hubs and three
connectors were identified in Daurian pikas. Among
these OTUs, module hubs were affiliated with
Clostridiales, Ruminococcaceae, Firmicutes, and
unclassified bacteria, while the connectors showed the
highest similarities with Alistipes massiliensis,
Prevotella, and S24-7. In contrast, only three module
hubs were observed in Plateau pikas. These OTUs
belonged to Bacteroidales, S24-7, and unclassified bac-
teria. Notably, these keystone species in networks
showed relatively low abundances in the stomach bac-
terial communities.

Fig. 6 Comparison of stomach bacterial networks between Plateau pika clustering coefficient (avgCC) signifies that the bacterial network is
and Daurian pika. Highly positive correlation is indicated by red color mainly composed of relatively isolated nodes. A low level of
and highly negative correlation by gray color. Lower average degree centralization of betweenness is a result of low frequency of centralized
(avgK) and density (D) means a simpler network. A lower average node (color figure online)
 Appl Microbiol Biotechnol

Table 2 The taxonomic

identification of putative keystone Pika OTUID Taxonomy Average Abundance Nodes type Zi Pi
species in the networks of pika species relative rankingb
bacterial communities abundance
(%)a

Daurian OTU8937 Clostridiales (Fir) 0.024 360th Module 3.9 0

pikas hubs
OTU3315 Ruminococcaceae 0.033 243th Module 3.38 0.26
(Fir) hubs
OTU58710 Firmicutes 0.017 536th Module 3.24 0.35
hubs
OTU806 Unclassified 0.026 333th Module 2.79 0.32
bacteria hubs
OTU2984 Unclassified 0.027 315th Module 2.68 0.47
bacteria hubs
OTU1646 Unclassified 0.035 228th Module 2.59 0
bacteria hubs
OTU2322 Alistipes 0.042 178th Connectors 0.71 0.67
massiliensis
(Bac)
OTU1306 Prevotella (Bac) 0.038 203th Connectors 0.92 0.67
OTU35 S24-7 (Bac) 0.039 195th Connectors 0.17 0.625
Plateau OTU836 Bacteroidales 0.58 24th Module 3.64 0.28
pikas (Bac) hubs
OTU222 S24-7 (Bac) 0.55 26th Module 3.54 0.15
hubs
OTU5277 Unclassified 0.018 604th Module 3.08 0.35
bacteria hubs

Fir Firmicutes, Bac Bacteroidetes

a
The average relative abundance of the OTU in each pika species
b
The abundance ranking of the OTU in each pika species according to the order from high to low

Discussion Some microbes may play important functional roles in the

Our data revealed that the pika stomach harbors a diverse
bacterial community dominated by Bacteroidetes,
Firmicutes, and Proteobacteria, indicating similar overall
compositions of stomach bacterial communities at phylum
level as previously described in other mammalian hosts,
such as humans (Bik et al. 2006), horses (Perkins et al.
2012), and woodrats (Kohl et al. 2014). However, we
found that almost one half of total sequences in pika
stomachs were affiliated with unclassified bacteria. This
finding was similar with previous studies in wild rodents
(Neotoma spp.) showing a high proportion of novel se-
quences from the gastrointestinal tract (Kohl et al. 2011,
2014). Thus, pika stomachs may represent a large potential
to discover novel bacterial species. In addition, although
host (body weight, host species, and sex) and environmen-
tal (seasons and sampling sites) factors had little impact on
the compositions and diversity of stomach bacterial com-
munities, the pika stomach communities can be divided
into two distinct phylogenetic clusters. Plateau pikas and
Daurian pikas had different co-occurrence patterns. These
results laid the basis for understanding stomach function
and host-microbe interactions in wild mammals.
stomach of pikas. For example, some members of the genera
Prevotella, Oscillospira, Ruminococcus, and Treponema were
associated with the fermentation of plant-based food (Bekele
et al. 2011; Dai et al. 2014; Mackie et al. 2003) and considered
as fiber-degrading bacteria. Although the stomach is not main-
ly responsible for food fermentation, these bacterial genera
may flow into the cecum of pikas with digesta, and finally
improve the utilization ratio of indigestible plant polysaccha-
ride. Recent studies have shown that these four bacterial gen-
era were dominant microbes in pikas cecum (Li et al. 2016a
c). The genera Lactobacillus and Streptococcus detected in
our data were also common in the stomach of healthy horses
(Perkins et al. 2012) and humans (Aviles-Jimenez et al. 2014;
Bik et al. 2006), indicating that the bacteria may be widely
distributed in the mammalian stomach. Certain members of
these two genera were able to stimulate host immune re-
sponses (Veckman 2003). Notably, despite the clear interac-
tions between Helicobacter pylori infection and gastric dis-
eases (e.g., inflammation, ulceration, and cancer) in humans
(Dorer et al. 2009; Marshall and Warren 1984), Helicobacter
sp. sequences were not detected across all samples, indicating
that the H. pylori may be not a common inhabitant or patho-
gen in pika stomachs.
Appl Microbiol Biotechnol
Our results showed that the compositions and diversity of
microbial communities in the pika stomachs were almost not
influenced by host (host species, body weight, and sex) and
environmental factors (seasons and sampling sites). In gener-
al, the mammalian stomach is typically considered as a harsh
environment for bacterial growth and development due to low
pH values (Beasley et al. 2015). The acidic environment in
pika stomachs may filter out non-acid-resistant bacteria but
enrich those microbes that genetically better adapted to the
stomach. The low pH values may lead to the relatively stable
and similar bacterial communities in pika stomachs under dif-
ferent hosts or environmental factors. Even so, the stomach
microbiota showed great variations among host individuals. In
addition to the above factors, we infer that the individual mi-
crobial community in stomachs is also likely influenced by
other factors, such as diseases (Aviles-Jimenez et al. 2014),
proton pump inhibitors (Wu et al. 2014), immune status, and
antibiotic medication (von Rosenvinge et al. 2013). Further
research is needed to investigate the relative importance of
these factors in shaping individual microbial communities.
In addition, we found that stomach bacterial communities
in pikas can be divided into two phylogenetic clusters, includ-
ing cluster I and cluster II (we defined it as Bstomach types^).
This pattern was similar to the three Benterotypes^ in human
gut microbiota (Arumugam et al. 2011). Notably, the identifi-
cation of the enterotypes was performed at genus level, but our
identification of the stomach types was calculated at the OTU
level. This method may be more precise when calculating the
phylogenetic clusters of microbial communities. We found
that the host individuals from Plateau pikas and Daurian pikas
were mainly distributed in cluster II and cluster I, respectively,
indicating distinguishable stomach type patterns in these two
pika species. Cluster II was more diverse than cluster I, sug-
gesting that the intraspecific bacterial community diversity in
each pika species was not homogeneous. In humans, diet is
considered as a primary contributor to a hosts enterotype (Wu
et al. 2011). As the Plateau pikas and Daurian pikas had dis-
tinct food profiles (Fan et al. 1995), this may give rise to the
differences of the dominant stomach type. The intraspecific
and interspecific variabilities in the stomach type may provide
the potential for bacterial community to adapt to individual
difference in physiology and genetics through plastic adjust-
ment (species sorting) and species interactions. In addition,
these two pikas are thought to have diverged 6.5 million years
(Yu et al. 1998) from their shared ancestors; thus, different
host genetics may also lead to the discrepancy of stomach
types. For seasonal variations, we found that dominant stom-
ach type in Plateau pikas remained relatively stable although
food profiles were different between seasons. One possible
explanation was that other factors (e.g., host immunity) that
were not investigated in this study might control the stomach
types. Further work should demonstrate which internal and
external factors are able to influence the pika stomach types.
There are complex microbial interactions in a microbial
community, such as competition, mutualism, predation, para-
sitism, amensalism, and commensalism (Faust and Raes
2012). However, these microbial interactions are quite diffi-
cult to be observed in natural systems. Thus, we used the
network analysis to speculate the species interactions accord-
ing to their mutual exclusion (negative) or co-occurrence
(positive) patterns. A positive interaction may result from
the mutualism or commensalism, while a negative relationship
is most likely due to competition, amensalism, predation, and
so on (Faust and Raes 2012). In our constructed networks, the
Plateau pikas were dominated by a number of positive corre-
lations, while Daurian pikas were predominated by a high
amount of negative correlations. We speculate that the plant-
based diet is enough for the stomach bacterial communities in
Plateau pikas in the Qinghai-Tibet Plateau, but is scanty for
the communities in Daurian pikas in the Inner Mongolia
steppe. When sampling, we observed less stomach content
in Daurian pikas than in Plateau pikas. Bacteria may compete
for food resources when food is scarce. This may be the reason
why the bacterial communities of Plateau pikas may harbor
more cooperation interactions than those of Daurian pikas. In
addition, we found that the total links, average degree, and
density of Plateau pikas were relatively higher than those of
Daurian pikas, indicating that the Plateau pikas harbored a
more complex ecological network, mostly also more metabol-
ic redundancy. More complex interactions may provide a
more stable metabolic network for microbial communities.
Further research is needed to demonstrate this hypothesis.
A lower average clustering coefficient signifies that the
bacterial network is mainly composed of relatively isolated
nodes or loose node groups, reflecting lower functional redun-
dancy (Sun et al. 2013). Notably, our data also showed that the
average clustering coefficient in Plateau pikas was relatively
higher than that in Daurian pikas, implying that the former
may harbor higher functional redundancy than the latter. In
addition, a low level of centralization of betweenness, as a
result of low frequency of centralized nodes, was observed
in the bacterial network of Daurian pikas. Centralized nodes
represent keystone species, which may exert an important in-
fluence on the Binformation^ flow among the microbes
(Bissett et al. 2013). These nodes provide important links
among the connections of many other microbes and are gen-
erally regarded as critical control points in the networks
(Bissett et al. 2013; Ruiz-Moreno et al. 2006). Thus, a lower
centralization of betweenness in the Daurian pikas indicated
that the microorganisms had similar status in bacterial net-
works, implying that the loss of one or several keystone spe-
cies may not affect the stability of the whole community. Our
results also confirmed that Daurian pikas harbored more key-
stone species than Plateau pikas. Here, 9 module hubs and
three connectors were observed in the Daurian pikas, while
only three module hubs were detected in the Plateau pikas, but

none of them were the top 20 most abundant species. The
most abundant module microorganism OTU836 was at the
24th position, indicating that the keystone species in the net-
works were not necessarily the most dominant microbes in the
pika stomach communities. However, the topological features
of our networks relied on sequencing depth and a data pro-
cessing method by only retaining OTUs that were present in
more than half of all samples; these data processing methods
may influence the topological features of the bacterial
networks.
In conclusion, this study described the compositions, diver-
sity, and structure of bacterial communities in wild pikas.
Based on 16S rRNA sequencing data, we found that host
and environmental factors almost had no significant impacts
in shaping the stomach bacterial communities, indicating a
relatively stable community in pika stomach. Interestingly,
two stomach types were found in pika stomachs, and high-
altitude Plateau pikas had different dominant stomach types
compared to low-altitude Daurian pikas. Notably, the Plateau
pikas also had more complex microbial interactions than
Daurian pikas, possibly reflecting specific niche adaptation
of stomach bacterial communities through species interac-
tions. The differences of microbial interactions may be related
to food intake in different pika species. These results may help
us understand the assembly of bacterial community in the
stomachs of wild mammals. Future studies should focus on
the relationships between the microbial communities and the
animal diseases in order to make efficient strategies to im-
prove animal health for wildlife conservation.
Acknowledgements The work was supported by Sichuan Province
Science and Technology Project (2017SZ0004, 2017JY0231), Open
Found of Key Laboratory of Environmental and Applied Microbiology
CAS (KLCAS-2016-03), and China Biodiversity Observation Networks
(Sino BON).
Author contributions The experimental design was conceived by H.L
and X.L; experimental work was performed by H.L., T.L., B.T., and Y.K.;
H.L. finished the data analysis and original manuscript; and H.L., T.L.,
and X.L. revised the manuscript.
Compliance with ethical standards Animal ethics approval for the
present project was obtained from the Animal Ethics Committee of
Chengdu Institute of Biology.
Conflict of interest The authors declare that they have no competing
interests.
References
Adbi S (1976) Intestinal phase of protein assimilation in man. Am J Clin
Nutr 29:205215
Amato KR (2013) Co-evolution in context: the importance of studying
gut microbiomes in wild animals. Microb Sci Med 1:1029
Appl Microbiol Biotechnol
Amato KR, Leigh SR, Kent A, Mackie RI, Yeoman CJ, Stumpf RM,
Wilson BA, Nelson KE, White BA, Garber PA (2015) The gut
microbiota appears to compensate for seasonal diet variation in the
wild black howler monkey (Alouatta pigra). Microbial Ecol 69(2):
434443
Arumugam M, Raes J, Pelletier E, Le Paslier D, Yamada T, Mende DR,
Fernandes GR, Tap J, Bruls T, Batto JM, Bertalan M, Borruel N,
Casellas F, Fernandez L, Gautier L, Hansen T, Hattori M, Hayashi T,
Kleerebezem M, Kurokawa K, Leclerc M, Levenez F, Manichanh C,
Nielsen HB, Nielsen T, Pons N, Poulain J, Qin J, Sicheritz-Ponten T,
Tims S, Torrents D, Ugarte E, Zoetendal EG, Wang J, Guarner F,
Pedersen O, de Vos WM, Brunak S, Dore J, Meta HITC, Antolin M,
Artiguenave F, Blottiere HM, Almeida M, Brechot C, Cara C,
Chervaux C, Cultrone A, Delorme C, Denariaz G, Dervyn R,
Foerstner KU, Friss C, van de Guchte M, Guedon E, Haimet F,
Huber W, van Hylckama-Vlieg J, Jamet A, Juste C, Kaci G, Knol
J, Lakhdari O, Layec S, Le Roux K, Maguin E, Merieux A, Melo
Minardi R, MRini C, Muller J, Oozeer R, Parkhill J, Renault P,
Rescigno M, Sanchez N, Sunagawa S, Torrejon A, Turner K,
Vandemeulebrouck G, Varela E, Winogradsky Y, Zeller G,
Weissenbach J, Ehrlich SD, Bork P (2011) Enterotypes of the human
gut microbiome. Nature 473(7346):174180
Aviles-Jimenez F, Vazquez-Jimenez F, Medrano-Guzman R, Mantilla A,
Torres J (2014) Stomach microbiota composition varies between
patients with non-atrophic gastritis and patients with intestinal type
of gastric cancer. Sci Rep 4:4202
Beasley DE, Koltz AM, Lambert JE, Fierer N, Dunn RR (2015) The
evolution of stomach acidity and its relevance to the human
microbiome. PLoS One 10(7):e0134116
Bekele AZ, Koike S, Kobayashi Y (2011) Phylogenetic diversity and
dietary association of rumen Treponema revealed using group-
specific 16S rRNA gene-based analysis. FEMS Microbiol Lett
316(1):5160
Bik EM, Eckburg PB, Gill SR, Nelson KE, Purdom EA, Francois F,
Perez-Perez G, Blaser MJ, Relman DA (2006) Molecular analysis
of the bacterial microbiota in the human stomach. Proc Natl Acad
Sci U S A 103(3):732737
Bissett A, Brown MV, Siciliano SD, Thrall PH (2013) Microbial com-
munity responses to anthropogenically induced environmental
change: towards a systems approach. Ecol Lett 16(Suppl 1):128
139
Contreras M, Morales A, Garcia-Amado MA, De Vera M, Bermudez V,
Gueneau P (2007) Detection of Helicobacter-like DNA in the gas-
tric mucosa of thoroughbred horses. Lett Appl Microbiol 45(5):553
557
Dai X, Tian Y, Li J, Su X, Wang X, Zhao S, Liu L, Luo Y, Liu D, Zheng
H, Wang J, Dong Z, Hu S, Huang L (2014) Metatranscriptomic
analyses of plant cell wall polysaccharide degradation by microor-
ganisms in cow rumen. Appl Environ Microbiol 81:13751386
Deng Y, Jiang Y, Yang Y, He Z, Luo F, Zhou J (2012) Molecular ecolog-
ical network analyses. BMC bioinformatics 13:113
DeSantis T, Hugenholtz P, Larsen N, Rojas M, Brodie E, Keller K, Huber
T, Dalevi D, Hu P, Andersen G (2006) Greengenes, a chimera-
checked 16S rRNA gene database and workbench compatible with
ARB. Appl Environ Microbiol 72:50695072
Dorer MS, Talarico S, Salama NR (2009) Helicobacter pyloris uncon-
ventional role in health and disease. PLoS Pathog 5(10):e1000544
Fan N, Jing Z, Zhang D (1995) Studies on the food resource niches of
plateau pika and Daurian pika. Acta Theriologica Sinica 15:3640
Faust K, Raes J (2012) Microbial interactions: from networks to models.
Nat Rev Microbiol 10(8):538550
Hamady M, Lozupone C, Knight R (2010) Fast UniFrac: facilitating
high-throughput phylogenetic analyses of microbial communities
including analysis of pyrosequencing and PhyloChip data. ISME J
4(1):1727
Appl Microbiol Biotechnol
Kohl KD, Miller AW, Marvin JE, Mackie R, Dearing MD (2014)
Herbivorous rodents (Neotoma spp.) harbour abundant and active
foregut microbiota. Environ Microbiol 16(9):28692878
Kohl KD, Weiss RB, Dale C, Dearing MD (2011) Diversity and novelty
of the gut microbial community of an herbivorous rodent (Neotoma
bryanti). Symbiosis 54(1):4754
Li H, Li T, Beasley DE, Hedenec P, Xiao Z, Zhang S, Li J, Lin Q, Li X
(2016a) Diet diversity is associated with beta but not alpha diversity
of pika gut microbiota. Front Microbiol 7:1169
Li H, Li T, Yao M, Li J, Zhang S, Wirth S, Cao W, Lin Q, Li X (2016b)
Pika gut may select for rare but diverse environmental bacteria.
Front Microbiol 7:1269
Li H, Qu J, Li T, Li J, Lin Q, Li X (2016c) Pika population density is
associated with composition and diversity of gut microbiota. Front
Microbiol 7:758
Li N, Li H, Lei G, Zhou Y (2013) Ecological function of plateau pika
(Ochotona curzoniae). Chinese J Wildl 34:238242
Linz B, Balloux F, Moodley Y, Manica A, Liu H, Roumagnac P, Falush
D, Stamer C, Prugnolle F, van der Merwe SW (2007) An African
origin for the intimate association between humans and
Helicobacter pylori. Nature 445(7130):915918
Liu W, Zhang Y, Wang X, Zhao J, Xu Q, Zhou L (2008) Food selection
by plateau pikas in different habitats during plant growing season.
ActaTheriologica Sinica 28:358366
Liu W, Zhang Y, Wang X, Zhao J, Xu Q, Zhou L (2009) Food selection
pattern for plateau pika in winter. Acta Theriologica Sinica 29:1219
Mackie RI, Aminov RI, Hu W, Klieve AV, Ouwerkerk D, Sundset MA,
Kamagata Y (2003) Ecology of uncultivated Oscillospira species in
the rumen of cattle, sheep, and reindeer as assessed by microscopy
and molecular approaches. Appl Environ Microbiol 69(11):6808
6815
Marshall B, Warren J (1984) Unidentified curved bacilli in the stomach of
patients with gastritis and peptic ulceration. Lancet 1:1311
Maurice CF, Cl Knowles S, Ladau J, Pollard KS, Fenton A, Pedersen AB,
Turnbaugh PJ (2015) Marked seasonal variation in the wild mouse
gut microbiota. ISME J 9(11):24232434
Moeller AH, Degnan PH, Pusey AE, Wilson ML, Hahn BH, Ochman H
(2012) Chimpanzees and humans harbour compositionally similar
gut enterotypes. Nat Commun 3:1179
Oksanen J, Blanchet F, Kindt R, Legendre P, Minchin P (2013) Vegan:
community ecology package. R package version 2:08 Available at:
http://CRAN.R-project.org/package=vegan
Pei S, Fu H, Wan C (2008) Changes in soil properties and vegetation
following exclosure and grazing in degraded Alxa desert steppe of
Inner Mongolia, China. Agric Ecosyst Environ 124(12):3339
Perkins GA, den Bakker HC, Burton AJ, Erb HN, McDonough SP,
McDonough PL, Parker J, Rosenthal RL, Wiedmann M, Dowd
SE, Simpson KW (2012) Equine stomachs harbor an abundant
and diverse mucosal microbiota. Appl Environ Microbiol 78(8):
25222532
Rambaut A (2009) FigTree, version 1.4.2. Computer Program Distributed
by the Author. http://tree.bio.ed.ac.uk/software/figtree/
Ruiz-Moreno D, Pascual M, Riolo R (2006) Exploring network space
with genetic algorithms: modularity, resilience and reactivity. In:
View publication stats
Pascual M, Dunne JA (eds) Ecological networks: linking structure
to dynamics in food webs. Oxford University Press, UK, pp 187
208
Saito R, Smoot ME, Ono K, Ruscheinski J, Wang PL, Lotia S, Pico AR,
Bader GD, Ideker T (2012) A travel guide to Cytoscape plugins. Nat
Methods 9:10691076
Shi S, Nuccio EE, Shi ZJ, He Z, Zhou J, Firestone MK (2016) The
interconnected rhizosphere: high network complexity dominates
rhizosphere assemblages. Ecol Lett 19(8):926936
Smith AT, Foggin JM (1999) The plateau pika (Ochotona curzoniae) is a
keystone species for biodiversity on the Tibetan plateau. Anim
Conserv 2:235240
Sun MY, Dafforn KA, Johnston EL, Brown MV (2013) Core sediment
bacteria drive community response to anthropogenic contamination
over multiple environmental gradients. Environ Microbiol 15(9):
25172531
Tamaki H, Wright CL, Li X, Lin Q, Hwang C, Wang S, Thimmapuram J,
Kamagata Y, Liu W (2011) Analysis of 16S rRNA amplicon se-
quencing options on the roche/454 next-generation titanium se-
quencing platform. PLoS One 6:e25263
Tennant SM, Hartland EL, Phumoonna T, Lyras D, Rood JI, Robins-
Browne RM, van Driel IR (2008) Influence of gastric acid on sus-
ceptibility to infection with ingested bacterial pathogens. Infect
Immun 76(2):639645
Veckman V (2003) Lactobacilli and streptococci induce inflammatory
chemokine production in human macrophages that stimulates Th1
cell chemotaxis. J Leukoc Biol 74(3):395402
von Rosenvinge EC, Song Y, White JR, Maddox C, Blanchard T, Fricke
WF (2013) Immune status, antibiotic medication and pH are asso-
ciated with changes in the stomach fluid microbiota. ISME J 7(7):
13541366
Wang Q, Garrity GM, Tiedje JM, Cole JR (2007) Naive Bayesian classi-
fier for rapid assignment of rRNA sequences into the new bacterial
taxonomy. Appl Environ Microbiol 73(16):52615267
Warton DI, Wright ST, Wang Y (2012) Distance-based multivariate anal-
yses confound location and dispersion effects. Methods Ecol Evol
3(1):89101
Wu GD, Chen J, Hoffmann C, Bittinger K, Chen YY, Keilbaugh SA,
Bewtra M, Knights D, Walters WA, Knight R, Sinha R, Gilroy E,
Gupta K, Baldassano R, Nessel L, Li H, Bushman FD, Lewis JD
(2011) Linking long-term dietary patterns with gut microbial
enterotypes. Science 334(6052):105108
Wu WM, Yang YS, Peng LH (2014) Microbiota in the stomach: new
insights. J Dig Dis 15(2):5461
Yang J, Wang ZL, Zhao XQ, Wang de P, Qi de L, Xu BH, Ren YH, Tian
HF (2008) Natural selection and adaptive evolution of leptin in the
ochotona family driven by the cold environmental stress. PLoS One
3(1):e1472
Yu N, Zheng C, Shi L (1998) The correlation between the environmental
changes and the evolution of the two sibling species of pika (genus
Ochotona). Acta Theriologica Sinica 18:127130
Zhou J, Deng Y, Luo F, He Z, Yang Y (2011) Phylogenetic molecular
ecological network of soil microbial communities in response to
elevated CO2. MBio 2(4):e00122-11