The giant panda, a bamboo specialist, harbors hundreds of bacterial operational taxonomic units (OTUs) in its gut as an adult, mostly from Firmicutes or Proteobacteria, and dominated by Escherichia, Shigella, and Streptococcus (Xue et al., 2015). However, despite their high fiber diet, cellulolytic species typically present in herbivore guts, e.g., members of Ruminococcaceae or Bacteroides (Zhang et al., 2018), are conspicuously absent or present at extremely low abundances in the gut of giant panda. Instead, Bacillus (Zhou et al., 2015) and Clostridia, which have limited cellulolytic activity and thus cannot fully utilize bamboo (Zhu et al., 2011), are widely observed (Dierenfeld et al., 1982). Furthermore, the gut microbiota compositions of adult giant pandas exhibited extensive seasonal variation (Xue et al., 2015; Wu et al., 2017). This variation may indicate disturbance, or it may be a normal feature of the healthy adult gut microbiome of giant pandas. Disorder of gut microbiota structure can be triggered by an abnormal increase in internal enteropathogens or invasion by exogenous pathogens (Zhao et al., 2017a; Zhang et al., 2017). In contrast, biofilm (Hall-Stoodley et al., 2004) or surfactin (Zhou et al., 2018) secreted by internal microbes of giant panda can stabilize the gut microbial structure.

Study of gut microbiome transitions in the giant panda cub during the nursing stage can be of high utility to decipher the mechanism shaping the adult gut microbiome. Regarding giant panda cub, the nursing stage is approximately 18 months (Xue et al., 2015). For the first 2 months after birth, cubs rely on breast milk feeding; they obtain essential nutrients, such as oligosaccharides, microorganisms and exosomal microRNAs (Ma et al., 2017), from maternal milk, as distinct from subsequent formula milk (Zhang et al., 2015); this prepares them for rapid postnatal development. Cessation of exclusive breast feeding drives the maturation of cub gut microbiota into adult-like microbiota (Backhed et al., 2015). During the first 6 months after birth, giant panda cubs display a stable and gender-independent growth rate with age; rapid growth begin at 7 ~ 20 days (Zhang et al., 2015), and the inheritable maximum growth rate occurs at 75 ~ 120 days (Che et al., 2015). A recent study of the gut microbiomes of captive-born giant pandas (S1 group: <2-month-old cubs with breast milk diet; S2 group: 3 ~ 12-month-old cubs with formula milk diet; S3 group: >6-month-old juveniles with bamboos stem or leaf diet; S4: >6-month-old juveniles with bamboo shoot diet) indicated substantial increases in microbial diversity and gene abundance with age: 51 16S-rDNA OTUs (from Clostridiaceae, Veillonellaceae, Peptostreptococcaceae, Lachnospiraceae) and 26 metagenome linking groups (MLGs) (from Microbacterium, Clostridium, Citrobacter, Delftia, Enterobacter, Klebsiella, Raoultella, Serratia, Enterococcus, Lactobacillus, and Staphylococcus) showed significant differences between the S1 and S3/S4 groups and among S1, S2, S3, and S4, respectively, none of which belong to potentially cellulolytic species. Furthermore, abundant genes related to hemicellulose degradation were detected (Guo et al., 2018; Zhang et al., 2018). In total, the abundance of 81 KEGG orthologous groups (KOs), which are involved in flagellar assembly, ABC transporters, a two-component system, the secretion system, transcription factors, phenylalanine metabolism, purine metabolism, pentose and glucuronate interconversion, pyruvate metabolism, methane metabolism and glycerolipid metabolism, changed significantly with age. Specifically, certain genes related to phenylalanine metabolism, fatty acid biosynthesis, purine metabolism, glutathione metabolism, antibiotic resistance and streptomycin biosynthesis were abundant in groups S3 and S4 (Zhang et al., 2018).

Studies of the gut microbiome of giant panda cub are quite rare, and no study has yet explored gut microbiome transitions of giant panda cub during the growth process. Answers to questions relating to the giant panda nursing period, for instance the relevance of inheritance of the gut microbiome from parents, transitions in gene functions, and significantly changed microbial pathways, remain to be clarified. A detailed investigation of the gut microbiome of giant panda cub during the growth period after birth is urgently needed. In this study, we concentrated on the gut microbiome of newborn giant panda cubs in the nursing stage (samples were grouped by cub age, C0: 0 ~ 1.5-month-old cub fecal samples; C1: 1.5 ~ 6-month-old cub fecal samples; C2: 6 ~ 9-month-old cub fecal samples) and parents (F: father fecal samples; M: mother fecal samples), to demonstrate their gut microbiome transitions in the 1st year after birth.

Although increasing numbers of studies of the significant gut microbes or pathways in giant panda have been reported in recent years (Xue et al., 2015; Wu et al., 2017; Guo et al., 2018; Yang et al., 2018; Zhang et al., 2018), few focused on giant panda cubs (Xue et al., 2015; Zhang et al., 2018) or established linkages between the significant gut microbes and pathways in cubs. Moreover, the two representative studies of cub gut microbiome omitted the significant microbiome changes in 6 ~ 9 months old of age during the nursing stage: one (Xue et al., 2015) focused on the gut microbiome of cubs (<6 months old), juvenile (2 ~ 5 years old) and adults (6 ~ 22 years old); the other one (Zhang et al., 2018) focused on cubs < 2 months old, cubs 3 ~ 12 months old, and juveniles (mostly >1 year old with bamboo diet). In addition, the samples allocated to each group were collected from different individuals. Individual differences could affect the gut microbiome of giant pandas, even in the same season (Xue et al., 2015). Thus, a consecutive and close study minimizing the factors except diet to trace gut microbiome transitions within the same giant panda cub after birth is urgently needed. Our study was based on the two newborn giant panda cubs and their parents living in Macao, China, to explore significant gut microbes and pathways, and their relevance in the growth of cubs and adults. We enriched microbes on a 0.22 μm membrane, then extracted their DNAs; obtained DNAs were deep sequenced with an average of 14.1 Gb clean data for each sample [~5.25 Gb each in Zhang et al. (2018), ~6 Gb each in Wu et al. (2017), and ~9.6 Gb in Yang et al. (2018)], and yielded a non-redundant gene set with 1,040,648 microbial (or bacterial and viral) genes [681,167 unique genes in Zhang et al. (2018)].

We detected an increase of gut bacterial richness with age, as per previous studies (Xue et al., 2015; Zhang et al., 2018). However, compared to 0 ~ 6-month-old cubs, the diversity decreased in the gut microbiota of 6 ~ 9-month-old cubs. One possible reason for this is that the dramatic expansion (Supplementary Figure S1 and Figures 1A,C) of gut microbiota after birth [particularly on cessation of breast milk (Backhed et al., 2015)] gave rise to increasing unevenness (Figure 1D) of gut microbiota. We detected dominant bacterial genera rarely reported in cubs or adults (Xue et al., 2015; Wu et al., 2017; Yang et al., 2018; Zhang et al., 2018), such as Erwinia, Providencia, Edwardsiella, Pectobacterium, and Serratia. Worthy of mention is Pectobacterium, a group of bacteria that produce pectolytic enzymes to hydrolyze pectins of bamboos. Meanwhile, the reported abundant genera Leuconostoc and Stenotrophomonas (Yang et al., 2018) were in low abundance in adults and all investigated samples, respectively. In addition, rapid enrichment of Pediococcus, Lactobacillus, and Clostridium in 3 ~ 12-month-old cubs (Zhang et al., 2018), and reductions of Klebsiella (Xue et al., 2015), Campylobacter, Pediococcus, and Lactobacillus (Zhang et al., 2018) in adults, were also observed in our study. We found that Lactobacillus, Pediococcus, and Clostridium started to become enriched from the 1st month after birth, and Campylobacter reached its highest abundance in 6 ~ 9-month-old cubs. These changes in Pediococcus, Lactobacillus, and Campylobacter made Lactobacillaceae and Campylobacteraceae the most significant changed microbial communities in the gut microbiota of 6 ~ 9-month-old cubs compared to both 0 ~ 1.5-month-old cubs and adults (Supplementary Figure S3). Besides, Lactobacillus and Campylobacter played significant metabolic roles in 6 ~ 9-month-old cubs: Lactobacillus largely promoted the ‘amino acid metabolism’ in 6 ~ 9-month-old cubs, especially ‘D-Arginine and D-ornithine metabolism.’ It also contributed to ‘secondary bile acid biosynthesis’ in 0 ~ 1.5-month-old cubs, and the low abundance of this pathway is a potential cause of the high content of undigested lipids seen during the DNA extractions. Fortunately, this pathway was later enriched by Lactococcus in adults (whereas Lactobacillus was largely decreased); Campylobacter significantly enriched ‘purine metabolism and pyrimidine metabolism,’ ‘carbon metabolism,’ ‘RNA polymerase,’ and the NOD-like receptor signaling pathway in 6 ~ 9-month-old cubs, which facilitated cell cycle and strengthened the environmental interplays of gut microbiota. Here, I stress that Lactobacillus and Campylobacter also enriched in gut microbiota of 1.5 ~ 6-month-old cubs, so they were excluded from the list of microbes that could discriminate the gut microbiota architecture among 0 ~ 1.5-month-old cubs (C0), 1.5 ~ 6-month-old cubs (C1), 6 ~ 9-month-old cubs (C2) and parents. However, Campylobacteraceae was reported to be scarce in >4-month-old captive pandas (Zhang et al., 2018). This can be understood by taking into consideration the different living conditions of (e.g., diet; region; captive, or wild) (Wu et al., 2017; Yang et al., 2018), and individual differences among (Xue et al., 2015), giant pandas.

In our study, the gut microbiota of 0 ~ 1.5-month-old cubs and 1.5 ~ 6-month-old cubs were strongly correlated (Figure 1F), while the maximum growth rate was seen at about 70 ~ 100 days, implying that the fast growth of cubs negatively affected the gut microbiota composition, and that longer interaction with the environment (e.g., diet) is needed for maturity of gut microbiota in the giant panda (Backhed et al., 2015). Here, we do not intend to deny the inheritability of entire gut microbial communities; a small group of microorganisms, especially pioneer species at birth, likely possess high inheritability potential (van Opstal and Bordenstein, 2015). However, the maturity of gut microbiota was largely accelerated after birth, when environmental factors (e.g., diet) intervened (Backhed et al., 2015); for example, when fed with solid food, a shift in gut microbial flora of cubs from Lactobacillus and Bifidobacterium to Enterobactereacea genera was seen (Hirayama et al., 2010). The ‘two-component system,’ ‘Protein export’ and ‘Bacterial secretion system’ pathways strongly relate to environmental interaction (Wu et al., 2017). We detected enrichment of these three pathways in the gut microbiome of 6 ~ 9-month-old cubs and adults, but they were in relatively lower abundance in 1.5 ~ 6-month-old cubs when compared to other stages, suggesting strengthened ability of environmental community with age. Similar abundance profiles were also observed in ‘Pyruvate metabolism,’ ‘Fatty acid metabolism,’ ‘Lysine metabolism,’ ‘Valine, leucine and isoleucine metabolism,’ ‘D-Glutamine and D-glutamate metabolism’ and ‘Alanine, aspartate and glutamate metabolism,’ regardless of their low abundances. These pathways showed higher abundances of genes catalyzing the degradation reactions than biosynthesis in the gut microbiome of adult giant panda (Guo et al., 2018). Moreover, we detected slightly higher abundances of ‘Fatty acid degradation’ than ‘Fatty acid biosynthesis,’ but lower abundance of genes taking part in the degradation of the amino acids, than biosynthesis with age.

In a previous study, 51 OTUs from Megasphaera, Veillonella, and Clostridium, and 26 MLGs from Microbacterium, Clostridium, Citrobacter, Delftia, Enterobacter, Klebsiella, Raoultella, Serratia, Enterococcus, Lactobacillus, and Staphylococcus, were identified as key variables for differentiation the microbial structure between <2-month-old cubs with breast milk diet and >1-year-old cubs with bamboo diet, and among <2-month-old cubs with breast milk diet, 3 ~ 12-month-old cubs with formula milk diet and 6-month-old juveniles with bamboo diet, respectively (Zhang et al., 2018). We found seven important genera that can best discriminate the gut microbial profiles of 0 ~ 1.5-months-old cubs, 1.5 ~ 6-month-old cubs, 6 ~ 9-month-old cubs and adults: Shewanella, Helicobacter, Oblitimonas, Haemophilus, Aeromonas, Listeria, and Fusobacterium. Many species of Shewanella have been reported to reduce metals and toxins (Harris et al., 2017); the other six genera Helicobacter (Fischbach and Malfertheiner, 2018), Oblitimonas (Drobish et al., 2016), Haemophilus (Bakaletz and Novotny, 2018), Aeromonas (Cardoso et al., 2018), Listeria (Salama et al., 2018), and Fusobacterium (Guven and Dizdar, 2018) have been implicated in human diseases. Helicobacter was even expressed in higher abundance in 6 ~ 9-month-old cubs than in adults, suggesting that giant panda would be more susceptible to digestive diseases in this stage. Listeria (Listeria ivanovii, L. monocytogenes, L. welshimeri, and L. seeligeri), rarely reported in previous studies (Xue et al., 2015; Wu et al., 2017; Zhang et al., 2018), was the only important genera with low abundance in the gut microbiome of cubs and adults. The Listeria species detected in age groups except 6 ~ 9 months old were non-pathogenic L. welshimeri and L. seeligeri, rather than pathogenic L. ivanovii and L. monocytogenes. The finding that pathogenic L. ivanovii and L. monocytogenes appeared in the gut of 6 ~ 9-month-old cubs reflects the disadvantages of dramatic gut biomass expansion after birth. Therefore, subsequent adjustments of early life gut microbiome along with adult gut metabiome are crucial for giant panda health. At 1 year of age, the gut microbiome of giant panda cub shows many similar functional attributes to that of adults (Zhang et al., 2018); 2-year-old juveniles exhibited similar gut microbiota diversity to adults (Xue et al., 2015). The significances of these (opportunistic) pathogens suggest that they play important ecological role beyond invasion in the gut ecosystem of giant panda in evolution: in our study, we detected that these pathogens were widely involved in ‘cell cycle,’ ‘metabolism,’ and ‘environmental information processing,’ despite little significances; in another metagenomic study (Yang et al., 2018), Serratia, Bacillus, and Pseudomonas were found participating cellulose degradation. However, significant increases of many pathogens enriched infectious pathway abundance (Figure 5) with age. We found that five infectious pathways implicated in invasions of Escherichia, Shigella, or Vibrio were significantly enriched in adults. Legionellosis (contributed mainly by Escherichia coli) was the only infectious pathway that significantly changed with age. Taken together with the enrichment of (opportunistic) pathogens and infectious pathways and large decreases of anti-pathogenic Bifidobacterium and Lactobacillus (Liu et al., 2017), it is intuitive that adult giant panda experiences digestive diseases (e.g., diarrhea) every year.

Escherichia and Klebsiella, but especially Escherichia, contributed most to the metabolic pathways in the age groups in our study. Klebsiella participated in many metabolic pathways in 0 ~ 1.5-month-old cubs, before being replaced by Escherichia. We detected no enrichment of pyruvate or glycerolipid metabolism in 0 ~ 2-month-old cubs compared with adults (Zhang et al., 2018). Instead, 16 pathways in the gut microbiome of 0 ~ 1.5-month-old cubs were significantly enriched in adults, most of them involved in infection, the cell cycle and glycan metabolism. Moreover, enrichment of the purine biosynthesis pathway and genes involved in cellular processes was observed in 6 ~ 9-month-old cubs when compared with 0 ~ 2-month-old cubs. Enrichment was also found in >1-year old cubs (Zhang et al., 2018).

Bamboo diet selects for fiber-digesting microbes (McKenney et al., 2018). In this study, we noted a similar phenomenon in the early life of giant panda under breast or formula feeding. Lactobacillus and Bifidobacterium were enriched during breast or formula feeding, yet sharply decreased in adults. Bifidobacterium significantly promoted the pathway ‘proteasome’ in 6 ~ 9-month-old cubs, aiding proteolytic degradation and maintaining stability of cells (Wang and Maldonado, 2006), which expanded the gut microbiota defense system of giant panda (Hall-Stoodley et al., 2004; Zhou et al., 2018). Besides, Bifidobacterium was reported to take part in cellulose degradation (Yang et al., 2018). The dramatic decreases of Lactobacillus and Bifidobacterium in adult giant panda would likely impair the nutrition conversion and stability of the gut microbiota; Clostridium was regarded to be the only genera relating to giant panda age (Zhang et al., 2018) and was involved in cellulose and hemicellulose degradation (Xue et al., 2015). Clostridium was enriched in the gut microbiota of 6 ~ 9-month-old cubs, when the cubs were under formula feeding, suggesting that it has diverse roles: in a previous study (Yang et al., 2018), the major contributors to cellulase were the bacterial genera Bacillus, Thermoanaerobacter, and Pseudomonas or the fungal genera Aspergillus and Bipolaris, rather than Clostridium or Bacteroidetes genera; in our study, Clostridium significantly increased the abundance of the ‘Legionellosis,’ ‘Homologous recombination,’ ‘Ribosome,’ and ‘Cell cycle - Caulobacter’ pathways in adults.

Regarding the bamboo digestion, low abundances of genes relating to the degradation of cellulose, hemicellulose xylan backbone, pectin and lignin were seen in our samples, as reported in similar studies (Zhu et al., 2011; Guo et al., 2018; McKenney et al., 2018; Zhang et al., 2018). However, a recent study of gut microbiome of captive, semiwild and wild giant pandas (adult) detected 1739 genes homologous to cellulase, β-glucosidase and cellulose 1,4-β-cellobiosidase (Yang et al., 2018), demonstrating a high potential of giant panda gut microbiome with respect to cellulose digestion. Furthermore, another study detected a higher abundance of hemicellulose-degrading genes (e.g., CEs and GH4, 8, 31) in juvenile giant pandas (>1 year old) than in giant panda cubs (0 ~ 12 months old), omnivores and carnivores. In contrast, low abundance of cellulose and hemicellulose enzymes (e.g., GH5, 28, 9, 10, 26) was reported elsewhere (Zhu et al., 2011; Guo et al., 2018). And, an experimental assay of cellulase and xylanase activity in fecal samples demonstrated low-level activity of these enzymes as carnivores (Guo et al., 2018). In our study, we detected low abundance of genes of GH4, 8, 31, 5, 9, 10, 26.

We noticed that 6 ~ 9-month-old cubs possessed a higher abundance of GH families. This may also have been caused by the dramatic explosion of gut biomass after birth (particularly at 6 months after birth), which led to a fast increase of gut microbial richness and abundance (Figure 1A and Supplementary Figure S1). Despite the fact that proteobacteria (e.g., Enterobacteriaceae) started to reduce, while Firmicutes (e.g., Clostridium) was enriched, after 12 months old of age (Zhang et al., 2018), the biomass in the gut of 6 ~ 9-month-old cubs may already be present at high levels (particularly Firmicutes and Enterobacteriaceae). The huge biomass and high richness would contribute to higher abundance of GH families in 6 ~ 9-month-old cubs than in other developmental phases. Besides, we found abundant genes of GH1 and GH39 that accommodate many xylosidases in 6 ~ 9-month-old cubs and adults. This was further supported by Pfam annotation (abundant α-amylase and GH1) and similar study (most abundant GH families: GH13, GH23, GH3, GH1, and GH2) (Yang et al., 2018). This suggests that xylosidase acts positively in bamboo (hemicellulose) digestion. Apart from low abundance of many digestive enzymes, huge spatial barriers between enzymes and substrates also restrict the digestion of bamboos. In the bamboo cell wall, hemicellulose, cellulose and lignin were joined by covalent bonds (hemicellulose and lignin) or hydrogen bonds (intracellulose, cellulose and lignin, hemicellulose and lignin). Thus, massive lignins, which are much harder to degrade in the gut of giant panda, firmly embed celluloses and hemicelluloses, which greatly impedes the approach of carbohydrate enzymes to cellulose or hemicellulose substrates. Fortunately, 12 significantly promoted energy metabolism pathways which were mainly contributed to by Escherichia and abundant pectolytic Pectobacterium, would aid low utilization of bamboo celluloses in the gut microbiome of adult giant panda.

Last but by no means least, bacteriophages closely relate to bacterial microbiome in gut ecosystems in early life (Lim et al., 2015). We harvested the bacteriophages inside bacteria on a 0.22 μm membrane, which would miss many bacteriophages with a smaller size. However, we discovered those bacteriophages closely related to bacteria. Consistent with a study of bacteriophage diversity in the gut microbiota of adult giant pandas (Yang et al., 2018), we found that Caudovirales were dominant, indicating that tailed bacteriophages play important roles in regulating the gut bacterial microbiome of giant panda. Further, we detected Caudovirales genera P22virus, Lambdavirus, P1virus, P2virus, and T4virus in high abundance in cubs and adult giant pandas, whereas some abundant genera, namely Epsilon15 virus, Nona33 virus, Rtp virus, G7c virus, and Tl2011 virus (Yang et al., 2018), were not detected in our study. Meanwhile, we observed that Muvirus, Sk1virus, Sfi21dt1virus, and Sfi11virus were enriched after cubs began formula feeding. Members of Muvirus mainly infect Escherichia; Shigella; Sk1virus, Sfi21dt1virus, and Sfi11virus infect Lactococcus or Streptococcus. Besides, the major hosts of bacteriophages were Escherichia, followed by Enterobacteria, Salmonella, Shigella, and Lactococcus in the gut microbiota of adult giant pandas in a previous study (Yang et al., 2018). By contrast, in our study, the major hosts in adults were Enterobacteria, followed by Salmonella, Escherichia, Klebsiella, Lactococcus; in cubs were Enterobacteria, followed by Escherichia, Salmonella, Lactococcus, and Streptococcus. All involved bacteria were dominant in samples. A study on Canine distemper virus, an ssRNA virus of genus Morbillivirus, demonstrated that it induced death in giant panda by reducing the richness of intestinal Escherichia and Clostridium, and increasing the diversity of gut microbes, to finally break the structure of intestinal microbes of giant panda (Zhao et al., 2017b). Thus, a detailed study on the relationship between virome and bacterial microbiome in the gut of giant panda is urgently needed in future.

Gut microbial transitions, inheritability and diet are important issues in policies for protecting giant panda. This study showed gradually increased richness, but reduced inheritance, of gut microbiota during the growth of giant panda cubs after birth, and the vital role of Enterobacteriaceae members in the gut microbiota of giant panda cubs. Cubs at 6 months old enter a significant period, during which a comprehensive and substantial increase of microbiome occurs. In addition, rapid enrichment of diet-related bacteria or phages was observed in the gut microbiota of cubs under breast and formula feeding, yet rapidly decreased in their parents, which indicated diet-stimulated gut microbiome transitions in the gut of giant panda in early life.

All raw reads of samples are available in National Center for Biotechnology Information (NCBI) database, and have been deposited under BioProject ID: PRJNA477424. The SRA accession number is SRP154209. The data reported in this study are also available in the CNGB Nucleotide Sequence Archive (CNSA: https://db.cngb.org/cnsa; accession number CNP0000135).

The sample collection and all the experiments were performed in a manner to minimize risk to the giant pandas and the environment. All experimental protocols of this study were approved by Institute of Chinese Medical Sciences - Animal Ethics Committee (ICMS-AEC) of the University of Macau.

MG and SL designed the research. MG and JwC performed the research. MG, JwC, JM, YW, and GF contributed to analytic tools. MG, JC, YF, QL, LP, and LZ analyzed data. MG and SL wrote the paper. All authors read and approved the final manuscript.

LZ and JC were employed by company Realbio Technology Co., Ltd. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.