Analysis of the intestinal microbial community altered during rotavirus infection in suckling mice

RV SA11 strain (provided by Dr Kobayashi, Osaka University, Japan) and fetal African green monkey kidney cells (MA104 cells; Cell Resource Center, IBMS, and CAMS/PUMC, Beijing, China) were used in this study. MA104 cells were grown in Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher Biochemical, Beijing, China) supplemented with 5% calf serum (FBS; GIBCO, Paisley, UK) and 1% penicillin–streptomycin (Sigma-Aldrich, St. Louis, MO, USA) in 5% CO2 at 37 °C in a humidified incubator. RV strain SA11 was propagated in MA104 cells as in previous studies⁴². The virus titer was counted by fluorescent focus assay (FFA) in MA104 cells resulting of 10⁶FFU/mL, and viral load was quantified by QPCR [22].

Twenty specific-pathogen-free female Kunming (KM) mice were provided by the Laboratory Animal Center at Jinzhou Medical University (Liaoning Province, China). All experimental groups were housed in the same specific-pathogen-free room. The living environment, feeding conditions and microbial conditions of animals in research were always consistent which was maintained on a 12-h light/dark cycle at 22 ± 2 °C with 40–70% humidity.

The female mice were paired with male mice upon delivery. The males were removed the next day, and the pregnant females were monitored daily and allowed to deliver at term. The day of birth was recorded as day 1 of life. Litter statistics and the ratio of males/females in each cage were not calculated. One lactating female mouse and her pups were maintained together in individually ventilated autoclaved cages (IVC). Suckling mice were divided into two groups: RV-infected and uninfected groups. Each group contained 10 litters of mice. There were 7–8 mice in each litter. RV-infected groups of suckling mice were orally administered 50 μL of 10⁶ PFU/mL RV strain SA-11 [12], and the uninfected group were orally administered 50 μL of phosphate-buffered saline (PBS) as a control.

All of the suckling mice were euthanized four days (the time point at which the most severe diarrhea symptoms presented) post-RV infection. Mixed intestinal contents of the colon and rectum were collected using autoclaved tweezers and stored in sterile tubes at − 80 °C. All animal experiments were performed in accordance with the Jinzhou Medical University guidelines. All of the animal experiments in this study were approved by the Animal Welfare and Ethical Review Board at Jinzhou Medical University (approval ID: 2019014). All animal infections and infectious work were performed in biosafety level 2 facilities.

Five sucking mice were randomly selected from each RV group and NC group for sample collection and pathological analysis. The duodenum was harvested from the abdominal cavity immediately after the suckling mice were euthanized. For the histopathological investigation, the duodenum was fixed in a solution of 4% paraformaldehyde (0.01 M PBS, pH 7.4). For paraffin section preparation, the duodenum was dehydrated with an increasing ethanol gradient, cleared with xylene, and then embedded in wax. Three consecutive paraffin sections (5 μm thick) were used for hematoxylin and eosin (H&E) staining. For each slice, fields were randomly selected.

The RV-infected and uninfected control mice groups were euthanize. Each group contained 10 litters of mice, and there were 7–8 mice in each litter. Genomic DNA was extracted from mixed intestinal contents using the QIAamp DNA Stool Mini kit (Qiagen Inc., Valencia, CA, USA) according to the manufacturer protocols. The V3–V4 region of the bacterial 16S ribosomal RNA genes (342F and 806R) was amplified by PCR using primers 341F 5′-CCTACGGGRSGCAGCAG)-3′ and 806R 5′-GGACTACVVGGGTATCTAATC-3′. The PCR protocol used in this study was as follows: 95 °C for 3 min, followed by 30 cycles at 98 °C for 20 s, 58 °C for 15 s, 72 °C for 20 s, and a final extension at 72 °C for 5 min. PCR reactions were performed in a 30-μL mixture containing 15 μL of 2× KAPA Library Amplification ReadyMix (Roche-KAPA, Shanghai, China), 1 μL of each primer (10 μM), 50 ng of template DNA, and ddH₂O. Amplicons were extracted from 2% agarose gels and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) according to the manufacturer’s instructions. The resulting library was analyzed with a Thermo NanoDrop 2000 spectrophotometer (ThermoFisher, Shanghai, China) and 2% agarose gel electrophoresis. Once the library passed the quality inspection, it was quantified by Qubit and mixed according to the data requirements of the quantification and normalization of individual PCR products.

The PCR products were quantified using Qubit (Invitrogen, Carlsbad, CA, USA), multiplexed at an even concentration, and subjected to 400–450 bp pair-end sequencing. The adaptor was added to the products, and the samples were sequenced on an Illumina MiSeq platform (Illumina, Inc., San Diego, CA, USA). The reads were assembled and used for subsequent 16S analysis. The assembled reads were filtered to acquire clean reads. Reads with an average quality value below 20 and number N that was more than 3 was removed, then the sequences spanning the entire V3–V4 amplicon were matched using PANDAseq [23]. Merged sequences with 97% nucleotide sequence identity (97% identity) were binned into operational taxonomic units (OTUs) using UPARSE [24]. Based on the RDP classifier, a representative sequence of each OTU was assigned to a taxonomic level in the RDP database using 0.8 as the minimum confidence threshold [25].

Alpha diversity indices, including the number of OTUs, observed species diversity, and Shannon and Simpson indices, were calculated by normalizing the number of clean reads in all samples to 47,504 sequences using mother software [26]. Rarefaction curves were analyzed with mothur and plotted using R. A representative sequence was chosen from each OTU by selecting the sequence that had the largest number of hits in the OTUs. The non-parametric Wilcoxon test (Wilcox. Test in R) was performed for each index of alpha diversity. Rank sum test was used to screen the alpha diversity indices with significant differences under different conditions.

Beta diversity was calculated for the normalized OTU table using UniFrac distance matrices [27, 28] in order to determine the amount of bacterial diversity shared between the two groups. Principal coordinate analysis (PCoA) of bacterial communities was performed using weighted UniFrac distances based on the presence and absence of OTUs, and the plot was generated using PERMANOVA, which was also performed using weighted UniFrac distance to test for differences in bacterial community composition in the samples from the two groups.

Linear discriminant analysis effect size (LEfSe) analysis was used to determine the features most likely to explain differences between the RV-infected group and healthy control group. Different features with an LDA score were identified [29]. The p value was set at p < 0.05, and the threshold on the logarithmic linear discriminant analysis (LDA) score was 2 [30]. In addition, heatmap analysis was performed to compare the significant differences at the genus level in the RV-infected and uninfected groups. PICRUSt (version 1.0.0, http://​picrust.​github.​com/​) [31] was performed to predict the microbial community function of the two groups. Finally, a statistical analysis of the taxonomic and functional profiles (STAMP, version 2.1.3, http://​kiwi.​cs.​dal.​ca/​Software/​STAMP) [32] was used for further exploration in level 2 of the KEGG analysis using Reconstruction of Unobserved States (PICRUSt) analysis.

All RV-infected suckling mice in the RV group had watery feces after infected 4 days. These clinical sign were not shown in the suckling mice in the normal control (NC) group. Quantification of the viral titer via fluorescent focus assay (FFA) showed values ranging from 3 × 10⁵ to 5.2 × 10⁷ PFU/mL. No virus was detected in the NC group (Fig. 1a). Five sucking mice were randomly selected from each RV and NC group for sample collection and pathological analysis. About 90% of the samples showed above reported phenomenon. The pathological characteristics of the duodenum showed vacuolar degeneration of the intestinal epithelial cells and congestion and edema of the stroma distributing within the villi of the small intestine in the RV group (Fig. 1b). There was no pathological damage in the NC group.

After runs on an Illumina MiSeq PE250 platform and quality filtering as described in the methods, a total of 1,182,607 merged sequences produced from all samples result in an average yield of 59,130 ± 2571 sequences/sample. According to alpha diversity analysis, which takes into account both sequencing saturation and sample integrity, 47,504 reads were randomly selected for each sample. The sample sequence diversity and richness were assessed on account of the operational taxonomic unit (OTU) counts in every sample as shown in Table 1. A Venn diagram showed that there were 58 OTU sequences in the RV groups and 64 OTU sequences in the NC group, with 58 OTUs shared (Fig. 2a). The number of OTUs in the NC group was significantly higher than that in the RV group, manifesting a significant difference in the number of OTUs between the two groups (Wilcox, p = 0.034) (Fig. 2b). However, There were no significant difference in Shannon and Simpson indices between the RV and NC groups within the experiment (Wilcox, p = 0.91 and p = 0.39, respectively) (Fig. 2c–d).

In order to contrast the differences in the bacterial community composition between the RV and NC groups, the beta diversity in both groups were calculated. We utilized principal coordinate analysis (PCoA) of variance (PERMANOVA) (R vegan package function Adonis) analyses to visualize the differences in the microbial community structures. Figure 3 is a weighted UniFrac PCoA on account of comparison of the microbial community from the colorectal contents of both groups (Adonis test p = 0.001, R² = 0.438). These results showed that the RV and NC groups had a clear separation and distinct differences in the bacterial composition between the RV and NC groups.

According to the results of species annotation, the abundance within samples was analyzed at the phylum, class, order, family, and genus levels. An examination of the forecasted taxonomic profiles at the phylum level for all the samples showed that Proteobacteria (56.40%) was the major phylum within the mixed colorectal contents of the RV group, exceeding both Firmicutes (39.55%) and Fusobacteria (3.58%) (Fig. 4a). An examination of the forecasted taxonomic profiles at the family level for all of the samples showed that Enterobacteriaceae (38.12%) was the main family in the mixed intestinal contents of the RV group, exceeding Lactobacillaceae (35.57%) and Pasteurellaceae (18.44%) (Fig. 4b). Notably, at the genus level, there was a decrease in the abundance of both Lactobacillus (70.57 vs.46.29 percentage) and Helicobacter (1.89 vs. 0%) in the RV group as compared to that in the NC group. Additionally, the abundance of Escherichia/Shigella (7.40 vs. 43.48%) and a decrease in the abundance of Fusobacterium (12.43 vs. 4.74%) in the RV group as contrasted to the NC group (Fig. 4c).

Linear discriminant analysis effect size (LEfSe) analysis was carried out to identify the salient features of the two groups. Firmicutes was the predominant phylum in the NC group, while the enriched phylum in the RV group was Proteobacteria. Notably, at the genus level, Lactobacillus, Paenibacillus, Brevundimonas, Parabacteroides, Bacteroides, and Alloprevotella were the predominant genera in the NC group, while the enriched genera in the RV group included both Enterococcus and Escherichia/Shigella (Fig. 5a). The distinctive features between the two groups at the genus level are listed in Table 2. All species at the genus level were analyzed. The species that failed to pass the p value screening were not shown. There were 10 genera with obvious differences between the RV and NC groups that were identified via a Wilcoxon rank sum test. Significant differences at the genus level between the RV and NC groups were illustrated by a heatmap (Fig. 5b).

In order to gain insight into the differences between the microbiota functions between the RV and NC groups, Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) was used to forecast the potential metagenomes from the community profiles of the normalized 16S rRNA genes. LEfSe was used to analyze the influence of metabolic pathways identified via the Kyoto Encyclopedia of Genes and Genomes (KEGG, http://​www.​genome.​jp/​kegg/​) database and calculate the significant differences between the two groups. The results showed that there were 10 pathways from level 2 that were enriched in the RV group (p < 0.001, White’s non-parametric t test) and 13 pathways that were enriched in the NC group. STAMP (Statistical Analysis of Metagenomic Profiles) analysis was used to predict the differences of the metabolic pathways of KEGG. The results showed that there were 9 pathways enriched in the NC group: (1) carbohydrate metabolism, (2) replication and repair, (3) translation, (4) nucleotide metabolism, (5) lipid metabolism, (6) xenobiotics biodegradation and metabolism, (7) folding, sorting, and degradation, (8) metabolism of terpenoids and polyketides, and 9) cell growth and death. Eight pathways were enriched in the RV group: (1) membrane transport, (2) poorly characterized, (3) cellular processes and signaling, (4) metabolism of cofactors and vitamins, (5) transcription, (6) signal transduction, (7) cell motility, and (8) infectious diseases (Fig. 6). In short, these data indicated that RV infection changes the metabolic functions of the intestinal microbiota in suckling mice. The metabolic functions identified here need further investigation in order to understand the role they in RV infection.