Gut microbiome dynamics associated with lifestyle shifts
Wild giant pandas exhibited significantly higher gut bacterial community richness and diversity than did wild-training I/II pandas, consistent with previous studies of primates (Clayton et al 2016) and bears (Borbón-García et al 2017). The increased richness and diversity could be due to the exposure to more diverse microbial meta-communities via habitats of wild giant pandas compared to those of wild-training I pandas (Burns et al 2016, Schmidt et al 2019, Wu et al 2017).
Proteobacteria and Firmicutes dominated the bacterial communities among all hosts with different lifestyles, which is consistent with previous studies (Zhang et al 2018, Zhu et al 2011). However, Proteobacteria was the most dominant phylum in the gut of wild-training I giant pandas, while Firmicutes was the most abundant phylum in wild-training II and reintroduced pandas. Proteobacteria have also been observed as the dominant phylum in the gut communities of wild-training I giant pandas (Wei et al 2015), while Firmicutes have been observed as the dominant phylum in the guts of wild pandas (Zhu et al 2011). Similarly, Firmicutes were the dominant phylum in wild deer mice and musk deer (Li et al 2017, Schmidt et al 2019) that primarily ingest insoluble fibers that are degraded by cellulose and hemicellulose-digesting enzymes including cellulase, beta-glucosidase, and xylan 1,4-b-xylosidase (Costa et al 2012, Zhu et al 2011). Non-captive giant pandas require maximal energy from their diets due to environmental stressors that wild-training I giant pandas do not experience (Schmidt et al 2019). Consequently, we also hypothesized that environmental stressors/threats (including pathogens, intraspecific competition, and interspecific competition, among others) may be important factors that drive the composition of panda intestinal bacterial communities during lifestyle shifts. In addition to the above, Bacteroidetes, Verrucomicrobia, and Actinobacteria abundances were higher in the wild giant panda communities compared to those of the reintroduced pandas (Non-parametric factorial Kruskal-Wallis sum-rank test, LDA>4). Bacteroidetes have been positively associated with the digestion of carbohydrates and proteins, and may help facilitate the development of gut immune systems (Ley et al 2006, Li et al 2017). In addition, Actinobacteria have been positively associated with fat digestion (Wu et al 2011). Thus, these results suggest that wild giant pandas had more complex environmental habitats and food choices that likely affected their GM.
Verrucomicrobia have been notably not previously observed in the guts of giant pandas although they were relatively abundant in the gut communities of wild pandas in our study. Verrucomicrobia have been observed in the gut communities of primates and termites in addition to various other environments (Dedysh et al 2006, He et al 2010, Lee et al 2009, Manjula et al 2016, Su et al 2016). The more microbial species that a host comes into contact with, the more likely it is that those species will persist in the host’s gut microbiome (Schmidt et al 2019, Smits et al 2017, Chave 2010). These observations support our suggestion that wild giant pandas were more adapted to natural environments and interacted with more diverse microbial species than did wild-training I giant pandas.
At the genus level, Escherichia were significantly enriched in the wild-training I panda gut bacterial communities (Non-parametric factorial Kruskal-Wallis sum-rank test, LDA>4).Escherichia have also been observed as the major bacterial taxa in the guts of captive giant pandas (Xue et al 2015), humans (Lagier et al 2012) and pigs (Niu et al 2015) gut. Conversely, Streptococcus andLeuconostoc were more abundant in the wild-training II giant panda gut communities, while Clostridium, Leuconostoc, andTuricibacter were more enriched in the reintroduced panda communities (Non-parametric factorial Kruskal-Wallis sum-rank test, LDA>4). These genera may be involved in the more complete digestion of bamboo in non-captive giant pandas, which could then help the pandas gain adequate energy from limited nutritional sources (Oyeleke and Okusanmi 2008, Zhu et al 2011).
The composition of gut bacterial communities in the reintroduced giant panda were closer to those of the wild pandas, although significant differences were observed between the reintroduced and wild panda communities at the phylum and genus levels (Non-parametric factorial Kruskal-Wallis sum-rank test, LDA>4). These observations suggest that the reintroduced giant pandas still maintained differences in their GM relative to wild pandas despite living in the same environment. For example, Clostridium , Leuconostoc ,Turicibacter, and Acinetobacter were more abundant in reintroduced panda communities. Conversely, Pseudomonas ,Sphingobacterium, and Flavobacterium were more abundant in the communities of wild pandas. All of these genera are associated with cellulose, hemicellulose, and lignin degradation (Dahal and Kim 2016, Jiménez et al 2015, Williams et al 2016, Zhu et al 2011), which is an important characteristic for the complete convergence of reintroduced giant pandas to wild pandas. Different habitats can influence the composition of GM through contact with different habitats, foods, and other materials (Borbón-García et al 2017, Li et al 2017, Chave 2010). Thus, these results suggest that additional time is needed for the complete conversion of reintroduced giant panda gut communities to those of wild pandas.
As with the bacterial communities, the richness and diversity of reintroduced and wild giant panda fungal communities were significantly higher than in those of wild-training I pandas (p < 0.05, ANOVA). Thus, captivity could lead to decreased richness and diversity of gut fungal communities of giant pandas. The underlying mechanism behind these differences are likely the same as for the bacterial communities, wherein food, space, and interactions with human keepers are limited for captive giant pandas (Schaller et al 1985), and thereby limit potential interactions with meta-communities relative to wild pandas. Ascomycota and Basidiomycota were the dominant phyla in the communities of wild-training I, wild-training II, and reintroduced pandas, although their relative abundances varied among groups. Ascomycota and Basidiomycota have been observed as dominant in the vaginas of giant pandas (Chen et al 2017), the guts of humans (Christian et al 2013), guts of dogs (Handl et al 2011), bamboo (Zhou et al 2017), soils (Xu et al 2012), and in the near-surface atmosphere (Bowers et al 2013). Indeed, Ascomycota and Basidiomycota are ubiquitous and abundant among most environments. Given that gut microbiome composition is driven by the frequency of contact with microbial species by hosts (Schmidt et al 2019, Smits et al 2017, Wheeler et al 2012), these results suggest that environmental microbiota may be one of the most important lifestyle factors that affect giant panda gut fungal communities.
Considerable variation was observed among the fungal genera associated with lifestyles.Candidawas the dominant genus in the gut communities of captive pandas, which may help in the digestion and absorption of carbohydrates (Christian et al 2013, Iannotti et al 1973). Captive giant pandas are ensured a fixed amount of carbohydrates (e.g. shoots and panda cakes) compared to semi-captive and reintroduced pandas (Schaller et al 1985). In contrast, Cryptococcus was enriched in the gut communities of wild-training II giant pandas, while Mrakiella was abundant in those of the reintroduced pandas.Cryptococcusis common in natural environments and can remain in non-infective states in bodies while later reactivating and spreading to other body areas, causing serious diseases in hosts with weakened immune systems (Hagen et al 2017, Litvintseva and Mitchell 2009).Mrakiella are found in soils and waters, especially in low-temperature environments within various regions (Thomas-Hall et al 2010). These results consequently suggested that diet and the microbial species within specific environments may be important factors that shape the composition of intestinal fungal communities during lifestyle shifts.
A total of 31 fungal genera exhibited significantly different abundances in the gut communities of reintroduced and wild giant pandas (Non-parametric factorial Kruskal-Wallis sum-rank test, LDA>4), albeit with low relative abundances exceptCalycina. Calycina was enriched in the communities of the wild pandas and belongs to the Helotiales order (Zhang and Zhuang 2004) that is associated with root endophytes (Tedersoo et al 2010). These results support that wild giant pandas may have more comprehensive dietary structures or more contact with microbial meta-communities within environments compared to reintroduced pandas, which would then enhance the dietary diversity of reintroduced pandas and contribute to the recovery of natural gut fungal community compositions as seen in wild pandas.