Introduction

The giant panda (Ailuropoda melanoleuca), an endemic, rare wild animal in China, is threatened by habitat destruction and listed as endangered in the World Conservation Union’s Red List of Threatened Animals (Li et al. 2015; Zhao et al. 2013). With the implementation of a series of conservation strategies (Hu et al. 2011; Tang and Zhang 2001), the population of the species in the wild has increased from 1000 in the 1970s to 1600 in 2013, and about 376 in captivity (breeding centers and zoos) (Huang et al. 2015). As a highly specialized herbivore, the giant panda spends an average of 25 % of their daily time to consume up to 6 % of body weight in dry matter per day, which consists of 99 % bamboo (Mainka and Zhang 1994). In contrast to other herbivores, the giant panda retains a typical carnivorous digestive system (Li et al. 1984). The short and simple gastrointestinal tract (GIT) indicates an extremely rapid transit of digesta and is regarded as an inefficient digester (<20 % dry matter digestibility) (Dierenfeld et al. 1982). Previous studies have suggested that the giant panda genome lacks the coding enzyme that is essential for cellulose digestion (Li et al. 2010), whereas enzymes related to lignin degradation have been discovered in the fecal microbiome of the giant panda through metagenomic library screening (Fang et al. 2012; Zhu et al. 2011). Therefore, intestine microbes play a vital role in the digestive strategy of the giant panda.

As a path to an efficient conservation scheme of giant pandas, a comprehensive investigation of the ecology of the complex intestinal microbiota is crucial. It is known that the intestine microbial composition of cubs significantly distinct from that of adults and juveniles, with a shift from dominance by Lactobacillus to genera within Enterobacteriaceae (Hirayama et al. 1989; Xue et al. 2015) that attribute to a change in diet from protein-rich to fibrous. Seasonal foraging behaviors for processing and consuming bamboo (Hansen et al. 2010; Tarou et al. 2005; Nie et al. 2015) also affect the composition and diversity of the intestinal microbiota (Williams et al. 2013; Xue et al. 2015), especially changes in abundance of Lactobacillus, Enterococcus and Streptococcus. The differences between the captive and the wild pandas’ diet contribute to the different intestinal microbiota, and wild animals have a higher ratio of Firmicutes/Proteobacteria than captive animals (Zhu et al. 2011). Although the associations between the microbiome and diet shift have been extensively studied, investigation into structural changes and compositional evolution from the subadults to geriatrics has just begun recently. Comparisons have been conducted and some considerable differences were detected between geriatric and adult giant pandas (Tun et al. 2014), but rarely involved subadults. The differences of intestinal bacteria composition from three age groups (subadults, adults, and geriatrics) have not been extensively examined.

Denaturing gradient gel electrophoresis (DGGE) is an effective tool to visually compare and analyze the dominant bacterial composition of different habitats, including soil (Siles et al. 2015), water (Danishta 2013), food (Alonso et al. 2015), animals gastrointestinal tract (Duytschaever et al. 2013). Meanwhile, real-time PCR is a powerful advancement of the basic PCR technique with highly sensitive and specific, and it allows for the rapid determination the level of a particular bacterial species/strain in samples through the design of specific primer sets (Furet et al. 2009). In this current study, we aimed to make a comparative analysis of intestinal bacterial communities in giant pandas under three different age groups (subadults, adults, and geriatrics) using a combination of PCR-denaturing gradient gel electrophoresis (DGGE) fingerprinting of the 16S rDNA V3 regions and real-time PCR.

Materials and methods

Animals

In November 2014, twelve fecal samples were obtained from twelve captive giant pandas, including four subadults (subjects A1–A4), four adults (subjects B1–B4), and four geriatrics (subjects C1–C4). Subject B4 and C3 were housed in Chengdu Zoo, whereas the other giant pandas were housed in the Chengdu Research Base of Giant Panda Breeding. Table 1 provides details of the twelve animals characterized in this study. Tables 2 and 3 show the detailed diet composition of giant panda. For adult panda B4, diagnosed with enteric disease during our sampling period, honey water was given as a supplementary food in addition to the normal diet.

Table 1 Clinical characteristics of the twelve giant pandas
Full size table
Table 2 Diet of giant panda
Full size table
Table 3 Composition and energy levels of concentrate
Full size table

Fecal collection, preparation and DNA extraction

All fecal samples were collected immediately after defecation, placed into an icebox, transported to the laboratory, and pretreated according to previously reported methods (Wei, et al. 2007). Briefly, 100 g of fecal sample was suspended in 500 mL of sterile 0.05 M phosphate-buffered saline (PBS, pH 7.4), and residual bamboo leaves were filtered after vigorous vortexing. The suspension was centrifuged three times at 200×g for 5 min to remove coarse particles. Supernatants were collected and pooled. Bacterial cells were then collected and washed three times through centrifugation at 9000×g for 5 min with 20 mL of sterile PBS. Bacterial precipitation was finally resuspended in 10 mL of sterile PBS, and 1 mL aliquots were stored at −80 °C for further analysis. All samples were subjected to DNA extraction by using the E.Z.N.A.® Stool DNA kit (Omega Biotechnology, USA) according to the protocol for isolation of DNA for pathogen detection. The isolated DNA was eluted in 100 μL of elution buffer and then stored at −20 °C.

PCR amplification for DGGE analysis

The V3 region of the 16S rDNA gene (position 339–539 in Escherichia coli gene) was amplified with the primers HDA1-GC (5′-GC clamp-ACT CCT ACG GGA GGC AGC AGT-3′) and HDA2 (5′-GTA TTA CCG CGG CTG CTG GCA C-3′) (Walter et al. 2001) by using MyCycler™ Thermal Cycler (Bio-Rad Laboratories, USA). Each 25 μL of the reaction mixture contained 12.5 μL of 2× Taq Master Mix (Tiangen Biotechnology, China), 1 μL of each primer (Invitrogen Life Technologies, China), and 1 μL of template DNA of the fecal samples. Amplification program was as follows: 94 °C for 4 min; followed by 30 cycles of 94 °C for 30 s, 58 °C for 30 s, and 72 °C for 2 min; and a final extension of 10 min at 72 °C.

PCR-DGGE analysis

The 16S rDNA V3 PCR products were separated through DGGE using a D-Code TM universal mutation detection system (Bio-Rad Laboratories, USA). Polyacrylamide (8 %) gels formed with 35–65 % linear DNA-denaturing gradients [100 % denaturant consisted of 40 % (vol/vol) formamide and 7.0 M urea] were run with 1× Tris-acetate-EDTA (pH 7.4). 10 μL PCR products were loaded on the DGGE gel. Electrophoresis was performed at 100 V and 60 °C for approximately 16 h. After electrophoresis, the gel was stained with 0.2 % (w/v) AgNO3 (van Orsouw et al. 1997) and scanned using a Bio-Rad GS-800 Calibrated Imaging Densitometer. PCR-DGGE profiles were digitized and normalized with Quantity One analysis software version 4.6.2 (Bio-Rad). A dendrogram was obtained through unweighted pair group mean average with NTSYS-pc software version 2.10 (Applied Biostatistics, Inc.). Shannon–Wiener index (H′), Species evenness index (E) and Species richness index (R) were calculated with the following formulas (Zwielehner et al. 2009):

$${\text{H}}^{\prime } = - \sum {\left( {pi} \right)\,\left( {\ln pi} \right)} ;\quad {\text{E}} = {\text{H}}^{\prime } /\ln \;S\hbox{max} ;\quad R = S$$

where pi is the proportional abundance of the species i and S is the number of bands.

Sequencing of characteristic DGGE bands

Nineteen characteristic bands were excised from gels by observing the changes in the presence/absence or in the variation of intensity of a single band (Ibekwe et al. 2001). The bands diffused into 20 μL of 0.1 % Triton X-100 buffer overnight at 4 °C. 1 μL of eluted DNA of excised bands was amplified with the primers HDA1-GC and HDA2 under similar PCR conditions and purified through DGGE until a single band with a specific mobility was obtained. Same primers omitted from their GC clamp were used to amplify the excised DNA by using PCR. The purified PCR products were ligated into pMD®19-T simple vector (TaKaRa Biotechnology, China) and then transformed into E. coli DH5α cells (TianGen, Beijing, China). Colony PCR was used to validate positive clones which were sequenced (BGI Technologies Co., Ltd) using Sanger by an ABI-3730xl (Applied Biosystems). Searches in GenBank databases were performed using BLASTn to identify closest relatives. All sequences determined in this study were deposited in GenBank (http://www.ncbi.nlm.nih.gov/Genbank) under accession numbers KR363133–KR363144 and KU510542–KU510545.

Quantification of microflora by quantitative PCR (qPCR)

All bacterial groups detected and specific primers used are listed in Table 4. qPCR calibration was performed on a Bio-Rad CFX96™ real-time PCR Detection System with CFX Manager Software version 2.0 (Bio-Rad Laboratories, USA). Each reaction was performed in triplicate in 25 μL of the reaction mixture contained 12.5 μL of SYBR® Premix Ex Taq™ II (TaKaRa Biotechnology, China), 1 μL of each primer (Invitrogen Life Technologies, China), 1 μL of fecal sample DNA, and 9.5 μL of sterile deionized water. The PCR program consisted of 95 °C for 1 min, 35 cycles of 95 °C for 15 s, annealing at the optimal temperature (Table 4) for 30 s, 72 °C for 30 s, and finally 95 °C for 15 s. Melting curve analyses were performed by slow heating from 55 to 95 °C (1 °C per cycle of 10 s) to monitor purity of the PCR product.

Table 4 Primers used for real-time PCR analysis
Full size table

PCR products of different primers were purified by using a TIANgel Midi Purification Kit (TianGen, China), and cloned into the pMD19-T vector (TaKaRa, Dalian, China), then transformed into DH5α (TianGen, Beijing, China) for amplification as described Rinttilä et al. (2004). LB agar medium with X-gal (100 μg ml−1; TianGen, Beijing, China), ampicillin (100 μg ml−1; TianGen, Beijing, China) and IPTG (100 μg ml−1; TianGen, Beijing, China) was used to screen the positive clones. Clones containing correct recombinant plasmid were confirmed by PCR amplification and sequenced to further verify the specificity of the primers. Plasmid DNA was extracted by using the E.Z.N.A.™ plasmid mini kit (Omega Bio-Tek), and the concentration was measured using a Nano Drop spectrophotometer (Nano Drop Technologies, Wilmington, DE). The standard curves were generated by tenfold serial dilutions of plasmid DNA. Measurements were performed in triplicate. Plasmid DNA was used in the positive control wells, while a well containing no template DNA was served as a negative control. Copy numbers of the target bacteria were calculated according to the standard curves.

Statistical analysis

All alphanumeric data were expressed as mean ± standard deviation. SPSS 17.0 software (SPSS Inc., Chicago, Illinois, USA) was used for statistical analyses using one-way ANOVA and Duncan’s range test to compare the three age groups. Statistical significance was set at p ≤ 0.05.

Results

Diversity and similarity of bacterial community of fecal samples from the three age groups

A total of twelve fecal samples from three different age giant pandas were analyzed by DGGE fingerprinting to characterize and compare their bacterial community compositions. As shown in Fig. 1, the PCR-DGGE profiles of the different age groups revealed significant differences in the structure and composition of the fecal microflora. An average of 23 and 25 discernable bands were observed in the geriatric and subadult pandas, respectively, whereas an average of 29 bands was observed in adult pandas. Moreover, in comparison to healthy individuals, PCR-DGGE profiles of adult panda B4 and geriatric panda C3 appeared more complex bacterial communities. Diversity index of the fecal microbiota on DGGE profiles was shown in Fig. 2. The species richness index, species evenness index and Shannon–Weiner diversity index of DGGE profiles ranged from 21 to 33, 0.73 to 0.84 and 3.04 to 3.50, respectively. Significant difference was observed between adults and geriatrics (p < 0.05), but not between the subadults and adults (p > 0.05). Dice coefficient of the DGGE profiles was determined by NTSYS-pc (Fig. 3). Banding patterns from different individuals presented a dice similarity coefficient between 40 and 88 %, with a mean of 64 %. However, the similarity between C2 and the other individuals was only 40 %.

Fig. 1
figure 1

PCR-DGGE DNA profiles of 16S rDNA V3 region of fecal microbes in different age captive pandas. A1–A4 subadults feces, B1–B4 adults faces, C1–C2 geriatrics feces. Bands that marked in the DGGE gel were identified by cloning and sequencing

Full size image
Fig. 2
figure 2

Diversity index of the fecal microbiota on DGGE profiles. a Shannon-Wiener index (H′), b species evenness index (E), and c species richness index (R) of GIT from three different age captive pandas

Full size image
Fig. 3
figure 3

Dendrograms generated from PCR-DGGE profiles

Full size image

Sequence analysis

Based on the PCR-DGGE profiles, 19 characteristic bands were selected to investigate the predominant intestinal bacteria in captive giant panda (Fig. 1). Bands 5, 9 and 13 could not be re-amplified and therefore no sequences were achieved. Table 5 shows all the identified bands. Almost all sequenced bands were closely related that of Proteobacteria (62.5 %), Firmicutes (12.5 %), and uncultured bacterium (25.0 %). Meanwhile, Acinetobacter johnsonii was appeared in subadults and geriatrics. Clostridium sp. was restricted to adults. Acidovorax sp., Citrobacter freundii, Klebsiella pneumoniae, and Klebsiella variicola were presented in geriatrics.

Table 5 Sequence analysis of bands from PCR-DGGE of fecal samples from 12 captive giant pandas
Full size table

Quantification of bacterial populations in feces

Real-time PCR showed differences in the number of various bacterial species in fecal samples from the three age groups (Fig. 4). Although the dominant bacteria were relatively stable in the three groups, the abundance of Lactobacillus in adults was significantly higher than those in subadults and geriatrics (p < 0.05). Other bacterial groups and species detected in this study showed no significant difference among the three age groups (p > 0.05). However, Bifidobacterium was not detected among all age groups.

Fig. 4
figure 4

Real-time PCR quantitation of bacterial genome copies in fecal samples from different age captive pandas. Duncan’s multiple-range test was used for compare the three subject groups with each other. Bars with different small letter superscripts mean significant difference (p < 0.05)

Full size image

Discussion

Establishment of the intestinal flora in a host is a gradual process, and factors affecting bacterial composition are very complex to be completely understood at present. In this study, we chose to use a culture-independent technique (DGGE) to assess the differences in fecal microbial community of three different age giant pandas. In agreement with previous reports (Tun et al. 2014), we found that adults possessed a significantly higher bacterial diversity than geriatrics (Figs. 1, 2), while the overall structure of the microbiota in subadults was similar to adults. These age-dependent differences in the intestinal microbiota are most likely due to differences in diet. Compared with adults and subadults, geriatric giant panda shown an accelerated decline in masticatory function, smell and taste sensation (Table 1), which may lead to decreased food consumption (Altenhoevel et al. 2012; Nordin et al. 2007), and causing changes in the amount of nutrients available for the host and its intestinal bacteria (Woodmansey 2007). Given the health benefits attributed to intestinal microbiome, their diversity is often used as a biomarker for a well-balanced intestine microbiota (Le Chatelier et al. 2013). Therefore, the significant decline in microbial diversity observed in the present study could have severe repercussions for the geriatric’s health.

Antibiotic therapy has been demonstrated to favor the selection of antibiotic-resistant strains and disrupt colonization resistance, eliciting overgrowth of potentially pathogenic bacteria (Al-Nassir et al. 2008; Engelbrektson et al. 2009). Gastrointestinal diseases are the most common causes of mortality in captive and wild giant pandas (Qiu and Mainka 1993; Zou et al. 1998). Adult panda B4 and geriatric panda C3 both had a history of antibiotic therapy prior to or during the sampling period. In comparison to healthy subjects, PCR-DGGE profiles from these two giant pandas appeared more complex bacterial communities (Fig. 1). Characteristic bands 18 and 19 from the fecal sample of C3 were 100 % identical to Klebsiella (Table 5), which has been described as a pathogen for human and animals (Lee and Kim 2011; Mansour et al. 2014). These findings could suggest that antibiotic use may lead to microbial dysbiosis. In addition, DGGE bands showed that the predominant bacteria shared in three age groups were belonged to Firmicutes and Proteobacteria, which is consistent with the results of previous studies (Tun et al. 2014; Xue et al. 2015).

Studies reported that the intestinal microbial composition in captive giant pandas differed significantly from those in wild (Zhu et al. 2011). Despite in captivity for 28 years, geriatric panda C2 harbored a distinct bacterial community relative to the other individuals (only 40 % similarity). This apparent discrepancy may be attributed to early intestinal bacterial colonization. It has been shown that the fecal microbial communities from three giant pandas with different sex, ages, and locations were similar (similarity coefficients ≥60 %) (Wei et al. 2007). In the present study, the dice similarity coefficient among the eleven individuals born in captivity ranged from 65 to 88 %, with a mean of 76.5 %. Such high similarity may be the result of the same diet composition. With PCR-DGGE, we achieved a view of the intestinal bacterial diversity from three age groups at a primary level. However, there was a limit for detecting bands in the DGGE profiles of complex communities (Murray et al. 1996). Some minor bacterial populations (those representing less than 1 % of the target organisms in terms of relative proportion) in samples might not be detected by this method. As a result, the DGGE fingerprinting patterns probably reflected the relative abundance of the dominant bacterial populations, rather than to its total richness (Muyzer and Smalla 1998). Moreover, from the complex DGGE profiles obtained in our study, we could not quantitatively compare the bacterial populations in three different age groups. As such, we performed real-time PCR to quantify Bacteroides–Prevotella–Porphyromonas, Streptococcus, Enterococcus, Lactobacillus group, Clostridium cluster XIVa, Enterobacteriaceae family, Bifidobacterium, and the domain bacteria in the giant panda.

Dietary adaptation is a major driving force for the evolution of intestinal microbiota (Nelson et al. 2013). To better adapt to highly fibrous diet, most herbivores have successfully developed a series of highly diverse cellulolytic obligate anaerobes to enhance nutrient absorption (Bian et al. 2013; Hess et al. 2011; Zeng et al. 2015), such as Ruminococcaceae, Fibrobacteresand, Clostridiales and Bacteroides bacteria. In contrast, our results defining a standard giant panda profile, together with previous reports, showed that Enterobacteriaceae, Enterococcus, Streptococcus and Lactobacillus represent the four dominant groups of the giant panda fecal microbiota (Tun et al. 2014; Wei et al. 2007; Xue et al. 2015). Subdominant groups are Clostridium cluster XIVa and Bacteroides–Prevotella–Porphyromonas (Tun et al. 2014). Clostridium cluster XIVa is a genus of strict anaerobes associated with cellulose-digesting (Zhu et al. 2011), and Bacteroides is well known degrader of polymeric organic matter (Flint et al. 2008). In this study, the numbers of Bacteroides–Prevotella–Porphyromonas and Clostridium cluster XIVa were higher in adults than that of geriatrics. Pyrotag sequencing of fecal samples from four captive giant pandas suggested that no significant variation of enterobacteria composition was observed in the intestines of both geriatric and adult pandas (Tun et al. 2014). At the family and genus level, we also did not observe any significant differences in the abundance of Enterobacteriaceae, Enterococcus, Streptococcus Bacteroides–Prevotella–Porphyromonas and Clostridium cluster XIVa among different age groups (p > 0.05). However, it is proper to notice that differences may exist at strain level. Therefore, a more advanced technique would detect significant differences among individuals.

The genera Bifidobacterium and Lactobacillus are considered to be important in preventing pathogen colonization and maintaining intestinal homeostasis (Turroni et al. 2014). Shifts from protein-rich diet to fibrous diets decreased Lactobacillus populations (Hirayama et al. 1989; Xue et al. 2015), and the absence of Bifidobacterium was discovered in giant panda faecal samples (Tun et al. 2014; Wei et al. 2007; Xue et al. 2015). Not surprisingly, Bifidobacterium was not detected in any panda from this study. Conversely, in the present study, the most marked difference between the adult group and the other two groups was the number of Lactobacillus, with significantly higher 16S rRNA gene copy numbers in the adult group (p < 0.05). Nevertheless, the beneficial role of Lactobacillus in the intestine of the giant panda remains to be determined.

In conclusion, this is the first cross-sectional study quantitatively monitoring the intestinal bacterial community of captive giant pandas from three different age groups (subadults, adults, and geriatrics) and we found that animals with different ages showed different intestinal bacteria composition. Geriatric giant pandas harbored lower microbial diversity and Lactobacillus group than adults. Further studies combining metagenomics must be conducted to explore these declines at the functional level. Moreover, considering that antibiotic use may lead to gastrointestinal dysbiosis, isolation of beneficial Lactobacillus may provide an optimal platform for the development of probiotics specific to the giant panda.