Multiomics Reveals Alterations in the Gut Microbiome, Host Proteins, and Host Metabolites Correlating with SADS-CoV Pathogenicity and the Immune Response in Piglets

doi:10.21203/rs.3.rs-4839968/v1

Download PDF

Article

Multiomics Reveals Alterations in the Gut Microbiome, Host Proteins, and Host Metabolites Correlating with SADS-CoV Pathogenicity and the Immune Response in Piglets

https://doi.org/10.21203/rs.3.rs-4839968/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

SADS-CoV, a coronavirus, is known to induce swine acute diarrhea syndrome. To explore the differences and commonalities in the pathogenesis mechanisms between highly pathogenic and low-pathogenic strains of SADS-CoV, we conducted an integrated analysis comprising proteomics, metabolomics, and short-chain fatty acid (SCFA) analysis, along with 16S rRNA sequencing of intestinal mucosa and fecal samples from piglets infected with SADS-CoV P7 (highly pathogenic)or SADS-CoV P83༈low-pathogenic༉. Additionally, we examined molecular events linked to potential pathogenicity and host immune responses subsequent to correlational analysis of diverse omics data.

In the SADS-CoV P7-infected group, the abundance of unidentified members of the family Enterobacteriaceae was markedly greater than in either the control group or the SADS-CoV P83-infected group in the ileum mucosa and feces. The concentration of SCFAs was significantly lower in SADS-CoV P7-infected pigs than in SADS-CoV P83-infected pigs, and SCFA levels were negatively correlated with the abundance of Enterobacteriaceae and the abundance of the species Escherichia coli in the ileum mucosa. Compared to those in the SADS-CoV P83 group, the differentially expressed proteins in the SADS-CoV P7 group were predominantly linked to extracellular matrix (ECM)-receptor interactions and focal adhesion pathways. Following SADS-CoV P7 infection, there was an increase in both the adhesion force and the number of Escherichia coli O157 adherent to IPEC-J2 cells. Moreover, SADS-CoV P7 can modulate the adhesion of E. coli O157 to IPEC-J2 cells by regulating the expression of the ECM-related protein integrin alpha5 (ITGA5), suggesting that ITGA5 plays a pivotal role in the invasion of E. coli O157 into intestinal epithelial cells during SADS-CoV infection.

A correlation exists among the multiomics profiles of the small intestinal mucosa and feces of piglets following infection with various generations of SADS-CoV. Understanding this correlation can help us better prevent the virus from harming piglets.

Biological sciences/Microbiology

Biological sciences/Microbiology/Bacteria

Biological sciences/Microbiology/Communities

Biological sciences/Microbiology/Virology

Multiomics

Microorganisms

Coronavirus

ECM receptor

Escherichia coli

In recent years, numerous emerging and re-emerging porcine enteroviruses have recurrently surfaced in Chinese pig farms, resulting in substantial and irrevocable economic losses. Enterovirus infection precipitates an imbalance in the intestinal flora of afflicted pigs, notably elevating the proportion of conditional pathogens. However, controlling and preventing the onset of enterovirus-related diseases solely through vaccines or medications poses considerable challenges.

Swine acute diarrhea syndrome coronavirus (SADS-CoV), which originated from Rhinolophus spp., was first identified in August 2016 [1]. In January 2017, SADS-CoV triggered a severe outbreak of acute diarrhea in piglets, with high mortality, on a large-scale Guangdong pig farm, with outbreaks reoccurring in Fujian in 2018 and Guangdong in 2019 [2, 3]. The small intestine serves as the primary target organ for SADS-CoV, leading to swine acute diarrhea syndrome, with acute diarrhea being the predominant clinical symptom among piglets [4]. Although SADS-CoV does not affect humans directly, research has indicated that it can infect mammalian species and human cell lines and can replicate within mouse dendritic cells, suggesting potential cross-species transmission [5]. Given its well-defined origin and safe, established cellular and animal infection models, SADS-CoV has emerged as a valuable model pathogen. Investigating the mechanisms underlying acute diarrhea and piglet mortality is highly important for the future prevention and control of SADS-CoV and analogous coronavirus-related diseases.

The intestinal tract constitutes an initial barrier against pathogens in animals. Among the host's intestinal microorganisms, bacteria and viruses play pivotal roles, and their interaction is crucial for maintaining the dynamic equilibrium and health of the host's intestinal environment. Research underscores the significant role of the intestinal flora and associated metabolites in modulating the host immune system, with a flora imbalance contributing to various inflammatory diseases [6]. Clinically, piglets infected with enteric diarrhea virus frequently experience secondary Escherichia coli infections, exacerbating diarrhea severity and mortality rates [7]. Notably, influenza virus neuraminidase can activate the host transforming growth factor-beta (TGF-β) signaling pathway, upregulating fibronectin and integrin alpha 5, thereby promoting Klebsiella pneumoniae adhesion and inducing secondary pneumonia [8]. Porcine transmissible gastroenteritis virus (TGEV), another coronavirus, positively regulates TGF-β, inducing epithelial–mesenchymal transition (EMT) and promoting E. coli K88 adhesion [9]. Our study revealed that secondary E. coli infection following SADS-CoV infection is a significant contributor to neonatal piglet mortality, yet the underlying molecular mechanisms remain elusive.

Recognizing the crucial role of the intestinal flora in host defense against viral infections, it is essential to investigate the inhibitory effects of the intestinal flora on porcine enterovirus infection and pathogenesis and to elucidate the molecular mechanisms of the intestinal flora in regulating viral infection. Our research team serially passaged SADS-CoV/CN/GDWT/2017 in Vero cells. We successfully obtained SADS-CoV-CN/GDWT/2017-P7 (SADS-CoV P7) and SADS-CoV-CN/GDWT/2017-P83 (SADS-CoV P83) and identified that these two strains exhibit different pathogenicities. Specifically, SADS-CoV P7 is highly pathogenic, whereas SADS-CoV P83 shows low pathogenicity [10]. This study revealed significant alterations in the intestinal flora and metabolites of piglets after infection with SADS-CoV P7 and SADS-CoV P83, highlighting the complex interplay and regulatory relationships among SADS-CoV, the porcine intestinal flora, and the host. These pioneering findings offer a fresh perspective for analyzing secondary bacterial infections post-SADS-CoV infection while also suggesting novel strategies for SADS-CoV prevention and control. Moreover, the research results hold potential significance for other coronavirus infections.

Virus strains

The passaged strains SADS-CoV-CN/GDWT/2017-P7 (SADS-CoV P7) and SADS-CoV-CN/GDWT/2017-P83 (SADS-CoV P83), derived from the SADS-CoV-CN/GDWT/2017 strain, were isolated and propagated in our laboratory[1]. According to the results of previous research conducted in our laboratory, P7 is highly pathogenic, while P83 is lowly pathogenic[10]. Vero cells were incubated in serum-free DMEM supplemented with 10 µg/ml trypsin (Gibco) at 37°C in a 5% CO2 atmosphere and monitored daily for cytopathic effects (CPEs). Once CPE was observed in 90% of the cells, cell harvesting was performed by subjecting the cells to three freeze‒thaw cycles, followed by centrifugation at 8000×g for ten minutes. The resulting supernatant was collected and stored at -80°C as a viral stock for subsequent passaging. This procedure was replicated for 75 passages in Vero cells, with purification conducted every ten passages starting from the eighth passage. Viral titers were determined by 50% tissue culture infectious dose (TCID50) assays at intervals of every five or ten passages.

Piglet challenge experiments

The piglets were sourced from Huanong Wen’s Co., Ltd., farm. For the experiments, we specifically selected piglets weighing approximately two kilograms that exhibited characteristics such as alertness, vigor, a strong appetite, and responsiveness to external stimuli. After ear tagging, the piglets were separated into groups A, B, and C for later infection. To prevent any potential cross-contamination, each group was housed separately in designated animal rooms. Prior to commencing the challenge test, anal swabs were collected for baseline health assessment. RT-PCR analyses were conducted to confirm the absence of specific viral infections, including SADS-CoV, porcine deltacoronavirus (PDCoV), TGEV, porcine epidemic diarrhea virus (PEDV), and porcine rotavirus (RV), among other diarrheal pathogens. At 5 days of age, the piglets were transferred to the animal facility and subjected to a challenge test following environmental acclimatization. Group A (consisting of 9 piglets) received an oral dose of 12 mL containing 1.0×10^6.3 TCID50/mL of the virulent parental strain SADS-CoV-CN/GDWT/2017-P7 (SADS-CoV P7). Group B (9 piglets) was similarly administered the avirulent strain SADS-CoV-CN/GDWT/2017-P83 (SADS-CoV P83), which was derived from the parental strain, while control Group C (8 piglets) received an equivalent volume of DMEM.

From 3 to 5 days postchallenge, during the peak incidence period, three piglets from each group were randomly selected for blood sampling and euthanasia. Subsequently, the intestines were inspected and sampled following dissection of the abdominal cavities.

16S rRNA gene sequencing

Genomic DNA was extracted from the samples using the CTAB/SDS method. The concentration and purity of the DNA were assessed using 1% agarose gels. Based on the concentration, the DNA was diluted to 1 ng/µL using sterile water.

Specific regions of the 16S rRNA and 18S rRNA genes, and the internal transcribed spacer regions (ITS) (including 16S V4, 16S V3, 16S V3-V4, 16S V4-V5, 18S V4, 18S V9, ITS1, ITS2, and Arc V4) were amplified using specific primers with barcodes. PCRs were carried out in 30 µL volumes containing 15 µL of Phusion® High-Fidelity PCR Master Mix (New England Biolabs), 0.2 µM forward and reverse primers, and approximately 10 ng of template DNA. The thermal cycling protocol included initial denaturation at 98°C for 1 min, followed by 30 cycles of denaturation at 98°C for 10 s, annealing at 50°C for 30 s, and elongation at 72°C for 30 s, with a final extension at 72°C for 5 min.

For PCR product mixing and purification, equal volumes of 1× loading buffer (containing SYBR Green) were mixed with the PCR products, and electrophoresis was performed on a 2% agarose gel for visualization. The PCR products were then mixed in equidensity ratios and purified using the GeneJETTM Gel Extraction Kit (Thermo Scientific).

Sequencing libraries were prepared using the Ion Plus Fragment Library Kit 48 reactions (Thermo Scientific) according to the manufacturer's instructions. Library quality was assessed using a Qubit® 2.0 Fluorometer (Thermo Scientific). Finally, the libraries were sequenced on an Ion S5TM XL platform, generating 400 bp/600 bp single-end reads.

Protein detection and tandem mass tag (TMT) labeling

The tissue samples were individually ground to a fine powder in a mortar with liquid nitrogen. The powder was then mixed with lysis buffer containing 50 mM Tris-HCl (pH = 8), 8 M urea, and 0.2% SDS. The resulting homogenate was subjected to ultrasonication on ice for 5 minutes, followed by centrifugation at 12,000×g for 15 minutes at 4°C. The supernatant was carefully transferred to a clean tube, and the protein concentration was determined using the Bradford assay. Next, 2 mM DTT was added to the sample, which was then incubated at 56°C for 1 hour. Subsequently, iodoacetic acid was added to the sample, and the mixture was incubated for 1 h in the dark at room temperature. The sample extract was then mixed with four volumes of cold acetone, thoroughly vortexed, and stored at -20°C for 2 hours to overnight. Following centrifugation, the pellet was collected and washed twice with cold acetone. Finally, the pellet was resuspended in dissolution buffer containing 0.1 M triethylammonium bicarbonate (TEAB, pH = 8.5) and 8 M urea. The protein concentration was once again determined using the Bradford assay. After confirming the quality of the protein using the Bradford protein quantification kit, the TMT labeling procedure was initiated.

Liquid chromatography‒mass spectrometry (LC‒MS) analysis

The TMT-labeled peptide mixture was fractionated using a C18 column (Waters BEH C18, 4.6×250 mm, 5 µm particle size) on a RigoL L3000 HPLC system operating at a flow rate of 1 mL/min. The column temperature was maintained at 50°C. A binary solvent system consisting of mobile phase A (2% acetonitrile, adjusted to pH = 10.0 using ammonium hydroxide) and mobile phase B (98% acetonitrile, adjusted to pH = 10.0 using ammonium hydroxide) was used for gradient elution. The gradient program was set as follows: 3% B for 5 min, 3–8% B for 0.1 min, 8–18% B for 11.9 min, 18–32% B for 11 min, 32–45% B for 7 min, 45–80% B for 3 min, 80% B for 5 min, 80 − 5% B for 0.1 min, and equilibration at 5% B for 6.9 min. Tryptic peptides were monitored at 214 nm, and the eluate was collected every minute, resulting in 15 fractions. The collected fractions were dried under vacuum and reconstituted in 0.1% formic acid (FA) in water for subsequent analyses.

Shotgun proteomics analyses were conducted using an EASY-nLCTM 1200 UHPLC system (Thermo Fisher) coupled to an Orbitrap Q Exactive HF-X mass spectrometer (Thermo Fisher) operating in data-dependent acquisition (DDA) mode. Samples corresponding to 2 µg of total peptides reconstituted in 0.1% FA were injected onto an Acclaim PepMap100 C18 Nano-Trap column (2 cm×100 µm, 5 µm particle size) for peptide trapping, followed by separation on a Reprosil-Pur 120 C18-AQ analytical column (15 cm×150 µm, 1.9 µm particle size). A 90-minute linear gradient was used for the TMT6-plex, or a 120-minute gradient was used for the TMT10-plex, with elution starting from 5–100% of mobile phase B (0.1% FA in 80% Acetonitrile) in mobile phase A (0.1% FA in H₂O) at a flow rate of 600 nL/min.

The mass spectrometer was operated in positive polarity mode, with a spray voltage of 2.3 kV and a capillary temperature of 320°C. Full MS scans were acquired from 350 to 1500 m/z at a resolution of 60,000 (at 200 m/z), with an AGC target value of 3×10⁶ and a maximum ion injection time of 20 ms. For DDA, up to 40 of the most abundant precursor ions were selected for higher energy collisional dissociation fragmentation analysis. Fragmentation scans were acquired at a resolution of 15,000 for TMT6-plex or 45,000 for TMT10-plex, with an AGC target value of 1×10⁵, a maximum ion injection time of 45 ms, a normalized collision energy of 32%, and a dynamic exclusion duration of 60 s.

Gas chromatography (GC)‒MS targeted detection

Fifty milligrams of tissue samples was removed from a 2 mL eppendorf tube and extracted with 0.5 mL of distilled water, followed by vortexing for 10 seconds. The mixture was homogenized for 4 minutes at 35 Hz using a ball mill and then subjected to ultrasonic treatment for 5 minutes while incubating on ice. The solution was centrifuged at 5000 rpm and 4°C for 20 minutes. Then, 0.3 mL of the supernatant was transferred to a new 2 mL eppendorf tube, 0.5 mL of distilled water was added, and the mixture was vortexed for 10 seconds. The mixture was homogenized again for 4 minutes at 35 Hz using a ball mill, followed by ultrasonic treatment for 5 minutes on ice. The solution was centrifuged at 5000 rpm and 4°C for 20 minutes. Then, 0.5 mL of the supernatant was transferred to another new 2 mL tube. This mixture was combined with the supernatant from the previous step, resulting in a total of 0.8 mL of supernatant. Next, 0.1 mL of 50% H₂SO⁴ and 0.8 mL of 2-methylvaleric acid (a 25 mg/L stock solution in methyl tert-butyl ether) was added as the internal standard, followed by vortex mixing for 10 seconds and oscillation for 10 minutes. The solution was centrifuged at 12000 rpm and 4°C for 15 minutes. The solution was allowed to stand at -20°C for 30 minutes. Then, the supernatant was transferred to a new 2 mL glass vial for GC‒MS analysis.

GC‒MS analysis was conducted on an Agilent 7890B gas chromatograph interfaced with an Agilent 5977B mass spectrometer. The analysis utilized an HP-FFAP capillary column. A 1 µL aliquot of the sample was injected in splitless mode (5:1). Helium served as the carrier gas, with a purge flow rate of 3 mL min^− 1 at the front inlet and a column flow rate of 1 mL min^− 1. The initial temperature was set at 80°C for 1 min, subsequently increased to 150°C at a rate of 5°C min^− 1, and then maintained at 230°C for an additional 12 min at a rate of 40°C min^− 1. The temperatures for the injection, transfer line, quadrupole, and ion source were set to 240°C, 240°C, 230°C, and 150°C, respectively. The electron impact mode was employed with an energy of -70 eV. The mass spectrometer acquired data in Scan/SIM mode with a mass-to-charge ratio (m/z) range of 33–200, following a solvent delay of 5 minutes.

Metabolite extraction and identification

One hundred microliters of the liquid sample to be tested (0.1 mg of tissue fluid ground in liquid nitrogen) was transferred to an Eppendorf tube. Then, 400 µL of an 80% methanol-water solution (four times the volume of methanol) was added, the mixture was vortexed, and the mixture was incubated at -20°C for 60 minutes, followed by centrifugation at 14,000×g and 4°C for 20 minutes. An aliquot of the supernatant was transferred to a 1.5 mL centrifuge tube and then vacuum freeze-dried. The residue was dissolved in 100 µL of reconstitution agent, vortexed, and centrifuged at 14,000×g and 4°C for 15 minutes. The supernatant was collected and injected for LC‒MS analysis.

Equal amounts of supernatant from each processed sample were mixed well to prepare a quality control sample. The blank sample consisted of the blank matrix of the experimental sample, and the sample pretreatment process was the same as that used for the experimental sample.

The offline data files of LC-MS were imported into CD search software, where simple screening of parameters such as retention time and mass-to-charge ratio was performed. Then, the peaks of different samples were aligned based on a retention time deviation of 0.2 minutes and a mass deviation of 5 ppm to enhance identification accuracy.

Next, peak extraction was carried out using the following settings: a mass deviation of 5 ppm, a signal intensity deviation of 30%, a signal-to-noise ratio of 3, a minimum signal intensity of 100,000, and summed ions. Simultaneously, the peak area was quantified. Target ions were integrated for the prediction of molecular formulas, which were then compared with the mzCloud database. Background ions were removed using blank samples, and the quantification results were normalized using quality control samples. Finally, the identification and quantification results for the data were obtained.

E. coli O157 culture

A single colony of E. coli O157 from a laboratory culture and another colony from the livestock ecology research laboratory were isolated and transferred to separate 1.5 mL Eppendorf tubes containing LB media. The tubes were incubated on a constant temperature shaker at 37°C for 6 hours.

Quantitative method for assessing the adhesion of E. coli O157 to IPEC-J2 cells

The system for quantitative PCR (qPCR) used a volume of 20 µL that included 1 µL of bacteria at different concentrations diluted in a 10-fold gradient, 0.5 µL of each forward and reverse primer, 10 µL of enzyme, and 8 µL of ddH₂O. The PCR program was as follows: initial denaturation at 95°C for 3 minutes, 40 cycles of denaturation at 95°C for 10 seconds, annealing at 56°C for 30 seconds, and extension at 72°C for 30 seconds, followed by melting curve analysis starting at 65°C. Finally, an absolute quantification standard curve was established based on E. coli O157 gene copy number and Ct values. The primer sequences can be found in Table 1.

SADS-CoV infects IPEC-J2 cells

The SADS-CoV virus suspension was inoculated onto a monolayer of IPEC-J2 cells, followed by the addition of F-12 culture medium containing 8 µg/mL trypsin. The cells were incubated in a 37°C incubator with 5% CO2 for 1 hour. Subsequently, the culture medium was supplemented with an appropriate volume of F-12 culture medium containing 8 µg/mL trypsin to maintain the culture.

SADS-CoV and E. coli o157 coinfect IPEC-J2 cells

IPEC-J2 cells were seeded in a 24-well plate one day in advance. When the cells reached 80% confluency, the culture medium was discarded, and the cells were washed twice with PBS. SADS-CoV was added at an MOI of 1. After SADS-CoV had acted for five hours, E. coli was added at an MOI of 1 for co-culture. Samples were collected after one hour and seven hours of co-culture. The coinfection steps are shown in Fig. 1.

Expression and interference of ITGA5

A eukaryotic expression plasmid encoding ITGA5 was transfected into IPEC-J2 cells at 60%-70% confluency using the transfection reagent Lipofectamine 3000. Following transfection, the cells were incubated for 6 hours before they were switched to complete F-12 culture medium. At 48 hours posttransfection, the cells were inoculated with E. coli at a multiplicity of infection (MOI) of 1.

The experimental groups included those transfected with the ITGA5 eukaryotic expression plasmid and subsequently inoculated with bacteria at an MOI of 1, groups inoculated solely with bacteria at an MOI of 1, and a control group. SiRNA experiments were also conducted following the aforementioned steps.

Detection of the transcription levels of ECM-related proteins

Following the protocol provided by TAKARA's RNAiso reagent, RNA extraction was carried out for the IPEC-J2 cells, followed by reverse transcription. The primer sequences can be found in Table 1.

Table 1

Primers used for realtime fluorescence quantitative PCR
Primers	Sequence (5′–3′)
ITGB1-F	CATCCAGATGACATTGAA
ITGB1-R	GTCCATAAGGTAGTAGAGA
GAPDH-F	GGGTGTGAACCACGAGAAATA
GAPDH-R	GTCATGAGCCCTTCCACAAT
qITGB5-F	TGTCCTACATTACCAGAG
qITGB5-R	CACAGATATTGTCTTCTCC
qTN-F	GGAGACGATGGAATATAATATC
qTN-R	GCTGTCAGGAGTTATATTG
qE. coli-F	CATGCCGCGTGTATGAAGAA
qE. coli-R	CGGGTAACGTCAATGAGCAAA

SADS-CoV infection leads to changes in the structure of the microbiota

High-throughput sequencing of the 16S rRNA gene was conducted to explore the differences in the microbiota composition of the ileum mucosa and feces between healthy and SADS-CoV-infected piglets. On average, 76,676 sequences per sample were generated (range: 53,073–96,405) for the 16S rRNA gene V4 region, with an average effective tag length of 253 bp.

The dominant bacterial composition in the ileum mucosa and feces of piglets was altered by SADS-CoV infection across various taxonomic levels (Supplementary Fig. 1). Furthermore, the Shannon and Simpson indices were greater in piglets in the control group and SADS-CoV P83 group than in those in the SADS-CoV P7group (Supplementary Figs. 2A and 2B).

Beta diversity among the groups was assessed via principal coordinate analysis (PCoA) based on weighted UniFrac distances. The resulting scatterplot revealed six distinct clusters representing gut bacterial communities (Supplementary Fig. 2C). Notably, samples from the SADS-CoV P7-infected group and the control group tended to have different clusters, whereas samples from the SADS-CoV P83-infected group and the control group displayed substantial overlap, suggesting a degree of similarity in microbial composition between these two groups.

To further elucidate the similarity among samples, a clustering tree using the unweighted pair-group method with arithmetic mean (UPGMA) based on weighted UniFrac distances was constructed (Supplementary Fig. 2D). The clustering tree effectively differentiated the SADS-CoV P7-infected samples from the control group samples in both the ileum mucosa and feces of piglets.

To delve deeper into the differences in microbiota distribution, we conducted MetaStat analyses focusing on the top 35 genera across the three groups. As illustrated in Fig. 2A, following SADS-CoV P7 infection, there was a notable increase in the abundance of pathogenic bacteria associated with diseases that cause diarrhea. In particular, the proportion of unidentified Enterobacteriaceae in the ileum mucosa and feces of the SADS-CoV P7-infected group was markedly high (35.48% and 12.50%, respectively), whereas the proportion was substantially lower in the SADS-CoV P83-infected group and the control group, both in the ileum mucosa (1.77% and 1.58%, respectively) and feces (3.43% and 2.93%, respectively). Furthermore, the proportions of Fusobacterium and Megasphaera in the feces of the SADS-CoV P7-infected group were significantly greater than those in the feces of the SADS-CoV P83-infected group (19.10% and 1.35% vs. 2.90% and 0.01%, respectively) and the control group (1.95% and 0.16%, respectively). Additionally, the proportion of Campylobacter in the feces of the SADS-CoV P7-infected group (1.63%) was significantly greater than that in the feces of the control group (0.02%).

Conversely, there was a decrease in the abundance of bacteria beneficial for intestinal health in the SADS-CoV P7-infected group. The proportions of Faecalibacterium, unidentified Ruminococcaceae, Actinobacillus, and Streptococcus in the ileum mucosa of the SADS-CoV P7-infected group were significantly lower than those in the ileum mucosa of the SADS-CoV P83-infected group and the control group. Similarly, the proportions of Turicibacter and Alloprevotella in the feces of the SADS-CoV P7-infected group were significantly lower than those in the feces of the SADS-CoV P83-infected group and the control group. Moreover, the proportions of Faecalibacterium and unidentified Ruminococcaceae in the feces of both the SADS-CoV P7-infected and SADS-CoV P83-infected groups were significantly lower than those in the control group. Furthermore, the proportion of unidentified Clostridiales in the ileum mucosa of the SADS-CoV P7-infected group was significantly lower than that in the ileum mucosa of the control group. Similarly, the proportion of Terrisporobacter in both the ileum mucosa and feces of the SADS-CoV P7-infected group was significantly lower than that in the control group. Notably, Comamonas could no longer be detected in either the ileum mucosa or feces of the SADS-CoV P7-infected group, whereas in the control group, there was a low but identifiable proportion (0.78% and 0.0004%, respectively).

Interestingly, the proportions of unidentified Prevotellaceae, Alloprevotella, Parabacteroides, Faecalibacterium, and unidentified Ruminococcaceae in the ileum mucosa, as well as Turicibacter in the feces of the SADS-CoV P83-infected group, were significantly greater than those in the SADS-CoV P7-infected group and the control group. Additionally, the proportions of unidentified Lachnospiraceae, Romboutsia, Collinsella, and Subdoligranulum in the ileum mucosa of the SADS-CoV P83-infected group were significantly greater than those in the ileum mucosa of the SADS-CoV P7-infected group (Fig. 2B). Detailed interspecies differences at the genus level are shown in the Supplementary Fig. 3.

Using LEfSe analysis, we evaluated the significant differences in microbial abundance in the ileum mucosa (Fig. 3A and B) and feces (Fig. 3C and D) of suckling piglets from the phylum level to the species level and plotted cladograms representing the structure of the host-microbiota axis. As shown in Fig. 3A and 3B, the cladogram indicated significant shifts in each of the three groups. Among the three groups, 24 phylotypes from phylum to species were identified as high-dimensional biomarkers. These microbes mainly belonged to six different phyla, namely, Firmicutes, Bacteroidetes, and Proteobacteria. The abundance of E. coli, which belongs to the phylum Proteobacteria, was significantly greater in the RSMA group. The abundance of the phyla Firmicutes and Proteobacteria in the RSMB group substantially increased due to an increase in the abundance of the order unidentified Clostridiales; family Ruminococcaceae, belonging to the phylum Firmicutes; and unidentified genera of the family Prevotellaceae, belonging to the phylum Proteobacteria.

The RSMC group had distinct mucosal microbial compositions mostly belonging to the phylum Firmicutes, represented by the genus Mitsuokella, which belongs to the class Negativicutes; the family Streptococcaceae; and the species Lactobacillus agilis, the latter two of which belong to the order Lactobacillales. As shown in Fig. 3C and 3D, the cladogram indicated significant shifts in each of the three groups. Among the three groups, 25 phylotypes from phylum to species were identified as high-dimensional biomarkers. These microbes mainly belonged to six different phyla, namely, Firmicutes, Bacteroidetes, Proteobacteria, Fusobacteria, Tenericutes, and unidentified bacteria. The abundance of the species E. coli; the genus Fusobacterium, belonging to the phylum Fusobacteria; the order Campylobacterales, belonging to the phylum “unidentified Bacteria”; and the genus Peptostreptococcus, belonging to the phylum Firmicutes, was significantly greater in the RSMA group. The abundance of the phyla Firmicutes and Proteobacteria in the RSMB group substantially increased due to an increase in the abundance of the order Clostridiales, which belongs to the phylum Firmicutes, and the classes Deltaproteobacteria and Gammaproteobacteria, which belong to the phylum Proteobacteria. The RSMC group had distinct fecal microbial compositions, mostly belonging to the phylum Bacteroidetes and phylum Firmicutes, represented by the family Muribaculaceae and genus Alloprevotella belonging to the phylum Bacteroidetes, and the species Lactobacillus agilis, species Eubacterium coprostanoligenes, and genus Faecalibacterium belonging to the Firmicutes.

Microorganisms in the ileum mucosa and feces were both selected as research objects. As shown in Fig. 3 (E)-(G), the microbial structures of the three groups were obviously different. In the control group, Lactobacillus gasseri and the genus Romboutsia, belonging to the Firmicutes, the genera Stenotrophomonas and Vibrio, and the order Rhizobiales, belonging to the Proteobacteria, were mainly distributed in the ileum mucosa; the species Eubacterium coprostanoligenes and the genus Faecalibacterium, both belonging to the phylum Firmicutes; the genus unidentified Prevotella, and the genus Alloprevotella and the family Muribaculaceae, belonging to the phylum Bacteroidetes; and the class Mollicutes, belonging to the phylum Tenericutes, were mainly distributed in the feces. In the SADS-CoV P83-infected group, the family Rhizobiaceae and genera Actinobacillus and Stenotrophomonas, all belonging to the phylum Proteobacteria; the species Prevotella copri (phylum Bacteroidetes), and the genus Romboutsia (phylum Firmicutes) were mainly distributed in the ileum mucosa; the genus Ruminococcaceae (phylum Firmicutes) was mainly distributed in the feces. In the SADS-CoV P7-infected group, the species E. coli and the family Rhizobiaceae, which belong to the phylum Proteobacteria, were mainly distributed in the ileum mucosa; the class Negativicutes, the order Selenomonadales, and the family unidentified_Lachnospiraceae, all belong to the phylum Firmicutes; the species Fusobacterium mortiferum (phylum Fusobacteria); and the order Campylobacterales (“unidentified Bacteria phylum”) were mainly distributed in the feces.

Microbial functions were predicted using Tax4Fun based on the relative abundance of microbes. Overall, 35 functional groups were identified in the microbiota of the ileum mucosa and feces. “Carbohydrate metabolism”, “membrane transport”, “replication and repair”, “translation” and “amino acid metabolism” were the top five functional annotations in the three groups, followed by “energy metabolism”, “nucleotide metabolism”, “glycan biosynthesis and metabolism”, “metabolism of cofactors and vitamins”, and “signal transduction” (Fig. 4A). After functional prediction of each group, cluster analysis was carried out using a heatmap. As shown in Fig. 4B, the mucosa and fecal microbial functional clusters in the SADS-CoV P7-infected group were quite different from those in the SADS-CoV P83-infected group and the control group. We found that the mucosa and fecal microbial functional clusters were more similar between the SADS-CoV P83-infected group and the control group, although their regions of relative abundance differed.

In summary, highly pathogenic SADS-CoV P7 increased the proportion of harmful bacteria and decreased the proportion of beneficial bacteria in the ileum mucosa and feces of piglets, while low-pathogenicity SADS-CoV P83 altered the microbial structure without obvious clinical manifestations. Notably, our results showed that the abundance of the opportunistic pathogen E. coli was significantly greater in the RSMA group than in the other groups.

Abnormal metabolism of short-chain fatty acids is accompanied by microbial dysbiosis in SADS-CoV-infected piglets

The short-chain fatty acids (SCFAs) in the contents of the large intestine were examined. As shown in Fig. 5A, the concentrations of SCFAs, including acetate, propionate, butyrate, isobutyrate, valerate, isovalerate, and hexanoic acid, in SADS-CoV P7-infected pigs were significantly lower than those in SADS-CoV 83-infected pigs (P < 0.01). The concentrations of acetate, propionate, butyrate, isobutyrate, valerate, and isovalerate were also significantly lower in SADS-CoV P7-infected pigs than in control group pigs (P < 0.01), whereas in the SADS-CoV P83-infected pigs, the concentrations of acetate, propionate, butyrate, and valerate were significantly lower than those in the control group pigs (P < 0.05), indicating some differences in SCFA content between the two groups of infected pigs.

Interestingly, the levels of acetate, propionate, butyrate, isobutyrate, valerate, and isovalerate were negatively correlated with the levels of Enterobacteriaceae and Escherichia coli in the ileum mucosa but were positively correlated with those of the family Clostridiales, genus Clostridiales, and family Peptostreptococcaceae in the ileum mucosa. Acetate, propionate, butyrate, isobutyrate, valerate, and isovalerate levels were negatively correlated with the levels of Fusobacteria, order Fusobacteriales, class Fusobacteriia, family Fusobacteriaceae, genus Fusobacterium, phylum unidentified_Bacteria, class unidentified_Bacteria, order Campylobacterales, and genus Peptostreptococcus in the feces but were positively correlated with the levels of the family Muribaculaceae, class Mollicutes, phylum Tenericutes, genus Alloprevotella, and family Ruminococcaceae in the feces (Fig. 5B).

Global overview of metabolism in the fecal metabolome

By utilizing variable importance in projection (VIP) scores derived from the first principal component of partial least squares-discriminant analysis (PLS-DA) model, in conjunction with T-test P values, we identified differentially abundant metabolites in the fecal metabolome, employing the criteria of a VIP > 2.0, a fold change (FC) exceeding 2.0 or less than 0.5, and P values below 0.05. This methodology revealed 157 differentially abundant metabolites between the SADS-CoV P7 infection group and the control group, comprising 76 cationic and 81 anionic metabolites. For the SADS-CoV P83 infection group versus the control group, 249 differentially abundant metabolites were detected, with 172 in cation mode and 77 in anion mode. Moreover, a comparison of the SADS-CoV P7 and P83 infection groups revealed 148 differentially expressed metabolites, which were evenly distributed, with 74 in both the cation and anion modes.

This investigation successfully identified differentially expressed metabolites across three comparisons: SADS-CoV P7 versus control, SADS-CoV P83 versus control, and SADS-CoV P7 versus SADS-CoV P83. Particularly, in the comparison between the SADS-CoV P7 group and the control group, cation mode analysis revealed notable upregulation of metabolites such as alfentanil, 2-O-sulfo-L-idopyranuronic acid, and creatinine, whereas ethamivan, toloxatone, and mepivacaine exhibited significant downregulation. Anion mode analysis revealed downregulation of metabolites such as 2,6-diamino-7-hydroxynonanedioic acid and 3a,7a-dihydroxycholanoic acid, with only a few, such as paracetamol sulfate, showing significant upregulation. In the comparison between SADS-CoV P83 and the control group, upregulated cationic metabolites included iscotrizinol and rebamipide, while significant downregulation was observed for metabolites such as ±-lauroylcarnitine. In the anionic mode, notably downregulated compounds included 2-[(dimethylamino)] and hexadecanedioic acid mono-L-carnitine ester, with azelaic acid among the few that were significantly upregulated. Among the SADS-CoV P7 and P83 groups, compared to the control group, the upregulated cationic metabolites included decanoylcarnitine, while the downregulated cationic metabolites included D-panthenol and mepivacaine. Anionic analysis revealed downregulation of prostaglandin A1 ethyl ester and upregulation of metabolites such as linalyl butyrate (Supplementary Fig. 4A).

Hierarchical clustering analysis of the primary differentially expressed metabolites across treatment groups revealed entirely divergent metabolic expression trends between the SADS-CoV P7 and SADS-CoV P83 groups, the SADS-CoV P7 and control groups, and the SADS-CoV 83 and control groups, underscoring the distinct metabolic impacts of each SADS-CoV infection scenario (Supplementary Fig. 4B).

Through enrichment analysis utilizing the KEGG pathway database, we investigated the metabolic pathways associated with differentially expressed metabolites among the SADS-CoV P83 and SADS-CoV P7 infection groups and the control group. As illustrated in Supplemental Fig. 4C, in cation mode, the metabolites that were differentially abundant between the SADS-CoV P7 group and the SADS-CoV P83 group were primarily associated with fatty acid biosynthesis and arginine and proline metabolism pathways; in the anionic mode, the differentially abundant metabolites were mainly linked to benzoxazinoid biosynthesis and the biosynthesis of type II polyketide products. Furthermore, in the comparison with the control group, the differentially abundant metabolites in the SADS-CoV P7 group were predominantly associated with valine, leucine, and isoleucine degradation; beta-alanine metabolism; and the synthesis and degradation of ketone bodies. In cation mode, the metabolites that were differentially abundant between the SADS-CoV P7 group and the control group were primarily related to the drug metabolism-cytochrome P450 and glucosinolate biosynthesis pathways. Additionally, in the cation mode, in the comparison with the control group, the differentially abundant metabolites in the SADS-CoV P83 group were mainly linked to caprolactam degradation and glycine, serine, and threonine metabolism pathways. In the anionic mode, the metabolites that were differentially abundant between the SADS-CoV P83 group and the control group were primarily associated with dioxin degradation and circadian entrainment pathways.

Analysis of the correlations between the microbiota and metabolites

We separately analyzed the NGS (Next-generation Sequencing) or mass spectrometry data of proteomics, metabolomics, short-chain fatty acid analysis, and 16S rRNA sequencing through basic bioinformatics analysis, expression/abundance detection, and differential expression analysis between samples. In this context, "A" represents the SADS-CoV P7 group, "B" represents the SADS-CoV P83 group, and "C" denotes the control group. The outcomes were further analyzed based on the intersections of A-vs-C and A-vs-B, as well as B-vs-C and A-vs-C. Consequently, we identified molecular events related to pathogenicity (P set) and immune responses (I-V set) in each of these groups. We considered the intersection of A-vs-C and A-vs-B to be the P set, and the intersection of B-vs-C and A-vs-C to be the I-V set.

Hierarchical all-against-all association (HAllA) analysis was performed to explore the correlation between gut bacteria and metabolites. Metabolite information is shown in Supplement Table 1. In the P set, in the feces, in the anion mode, indole-3-carbidol; sedanolide; lysoPC (18:4(6Z,9Z,12Z,15Z)); 1-stearoyl-2-arachidonoyl-sn-glycero-3-phosphoethanolamine; and 8,9a-dihydroxy-3,3b,5',5b-tetramethylicosahydro-3H-spiro[benzo[6, 7]-as-indaceno[2,3-b]furan-2,2'-pyran]-10-carbaldehyde were significantly negatively correlated with the genus unidentified_Fusobacteriaceae. Unoprostone, canthiumine, premithramycinone, and 3-methoxy-4-hydroxyhippuric acid were significantly positively correlated with the genus Alloprevotella. In the cation mode, norverapamil, toloxatone, and ethamivan were significantly negatively correlated with the genus Fusobacterium, and ethamivan, toloxatone, norverapamil, and 1-phenyl-1,3-octadecanedione were significantly positively correlated with the genus Alloprevotella (Fig. 6A). In the ileum mucosa, anionic 2,6-diamino-7-hydroxynonanedioic acid and trans-geranic acid were significantly positively correlated with the phylum Firmicutes (Fig. 6B). In cation mode, 2-O-sulfo-L-idopyranuronic acid was significantly negatively correlated with the phylum Firmicutes. R1128B, guaifenesin, and ethamivan were significantly negatively correlated with the genus unidentified_Enterobacteriaceae. Ethamans, n-acetylsphingosine, and toloxatone were positively correlated with the phylum Firmicutes (Fig. 6B). In I-V, in the feces, in the anion mode, unoprostone and methyl3-acetoxy-11-methoxyurs-12-en-28-oate were significantly negatively correlated with the genus Fusobacterium, and unoprostone and 2-(dimethylamino)-5,6-dimethylpyrimidin-4-ol were significantly positively correlated with the genus Alloprevotella. In the cation mode, (Z)-N-[(2R)-2-amino-1-hydroxy-3-{[(1E)-N-hydroxy-7-(methylsulfanyl)heptanimidoyl]sulfanyl}propylidene]glycine, 2-oxohex-4-enoic acid, and N-desmethyltramadol were significantly negatively correlated with the genus Fusobacterium. TDIQ, 2E-crotamiton, and prilocaine were significantly positively correlated with the genus Alloprevotella (Fig. 6C). In the ileum mucosa, in anion mode, 3-amino-5-methylhexanoic acid and ellipticine were significantly negatively correlated with the genus unidentified_Enterobacteriaceae, and 4-methylcatechol, trepibutone, and ellipticine were significantly positively correlated with the phylum Firmicutes. In the cation mode, (+)-castanospermine and (Z)-N-[(2R)-2-amino-1-hydroxy-3-{[(1E)-N-hydroxy-7-(methylsulfanyl)heptanimidoyl]sulfanyl}propylidene]glycine were positively correlated with the phylum Firmicutes, and 2-oxohex-4-enoic acid, (Z)-N-[(2R)-2-amino-1-hydroxy-3-{[(1E)-N-hydroxy-7-(methylsulfanyl)heptanimidoyl]sulfanyl}propylidene]glycine, and N-desmethyltramadol were significantly positively correlated with the family Prevotellaceae (Fig. 6D).

Global overview of proteomics

To identify differentially expressed proteins under specific conditions. Within the SADS-CoV P7 infection group, when compared to the control group (A vs. C), we identified 244 differentially expressed proteins, 137 of which were markedly upregulated and 107 of which were significantly downregulated. Notably, the most substantially downregulated proteins in the A vs. C comparison included the 39S ribosomal protein L53 (mitochondrial) and the S100 calcium-binding protein A9 isoform X1.

Analysis of the SADS-CoV P83 infection group versus the control group (B vs. C) revealed 51 differentially expressed proteins, among which 8 were significantly upregulated and 43 were significantly downregulated. The most notably upregulated and downregulated proteins were membrane-associated phosphatidylinositol transfer protein 1 and putative acyl-coenzyme A thioesterase 6, respectively.

Comparing the SADS-CoV P7 infection group with the SADS-CoV P83 infection group (A vs. B), there were 175 differentially expressed proteins, including 141 significantly upregulated and 34 significantly downregulated proteins. The adenylate cyclase type 5 isoform X1 was the most significantly upregulated protein, whereas the pleckstrin homology-like domain family A member 2 was the most significantly downregulated.

Enrichment analysis utilizing the KEGG pathway database was conducted on differentially expressed proteins to elucidate their involvement in specific biological pathways. For proteins differing between the SADS-CoV P7 group and the control group, the analysis highlighted pathways such as drug metabolism-cytochrome P450, extracellular matrix (ECM)-receptor interaction, focal adhesion, and cell adhesion molecules (CAMs) as significantly affected (Fig. 7A). In contrast, proteins that differed between the SADS-CoV P83 group and the control group were predominantly associated with pathways such as viral myocarditis and focal adhesion (Fig. 7B). Moreover, the differentially expressed proteins identified when comparing the SADS-CoV P7 group and the SADS-CoV P83 group were largely involved in pathways related to dilated cardiomyopathy, hypertrophic cardiomyopathy, ECM-receptor interaction, focal adhesion, and Staphylococcus aureus infection, revealing the diverse metabolic and pathophysiological processes influenced by different infection scenarios (Fig. 7C).

Analysis of the correlations between the microbiota and proteins

We analyzed the correlation between proteins and bacteria. Protein information is shown in Supplementary Table 2. In the P set, in feces, the genus Alloprevotella was significantly negatively correlated with transmembrane 4 L6 family member 1 and transgelin isoform X1 and significantly positively correlated with adenylate cyclase type 5 isoform X1 and aldo-keto reductase family 1, member C-like 1. Additionally, the genus unidentified_Enterobacteriaceae was significantly positively correlated with N-acetyllactosaminide alpha-1,3-galactosyltransferase isoform 1 and significantly negatively correlated with carbonic anhydrase 2 and tubulin-specific chaperone cofactor E-like protein isoform X6. The genus Fusobacterium was significantly positively correlated with bcl-2-like protein 15, UPF0184 protein C9orf16 homolog, histone H1.0, and significantly negatively correlated with cytochrome P450 2C42 precursor, cytochrome b reductase 1, and cytochrome P450 3A29 (Fig. 8A). In the ileum mucosa, the genus unidentified_Prevotellaceae was significantly negatively correlated with myristoylated alanine-rich C-kinase substrate and significantly positively correlated with angiotensin-converting enzyme 2 isoform X1, glycerol-3-phosphate dehydrogenase [NAD(+)], and cytoplasmic lactase-phlorizin hydrolase. The genus unidentified_Enterobacteriaceae was significantly positively correlated with C-type lectin domain family 14 member A, tropomyosin alpha-1 chain isoform X3, and somatostatin precursor and significantly negatively correlated with phospholipase D4 isoform X1, phytanoyl-CoA dioxygenase, and peroxisomal isoform X1 (Fig. 8B). In the I-V set, in feces, the genus Alloprevotella was significantly positively correlated with adenylate cyclase type 5 isoform X1 and negatively correlated with protein S100-A2. The genus Fusobacterium was significantly positively correlated with XP_001929591.1 and negatively correlated with adenylate cyclase type 5 isoform X1 and NAD(P)H dehydrogenase [quinone] 1. The genus Peptostreptococcus was significantly negatively correlated with the protein S100-A2 (Fig. 8C). In the ileum mucosa, the genus unidentified_Enterobacteriaceae was significantly positively correlated with somatostatin precursors and negatively correlated with integrin alpha-IIb precursors (Fig. 8D).

Correlation networks

From the results, it can be observed that, in the P set, in the feces, TUBB2B was significantly negatively correlated with the phylum Tenericutes, and the phylum Tenericutes was negatively correlated with Com_804_pos (ethamivan) (Fig. 9A). The class Mollicutes was positively correlated with Com_804_pos and ADCY5 (Fig. 9A). In the ileum mucosa, the species Escherichia coli was negatively correlated with 14(S)-HDHA and PECAM1. The familyn Ruminococcaceae was positively correlated with Com_565_pos (N,N-dimethylaniline) and negatively correlated with PECAM1 (Fig. 9B). In the I-V set, no mutually correlated network among microbiota, metabolites, and proteins were identified (Fig. 9C and 9D).

The promotion of E. coli adhesion by SADS-CoV

Analysis of the 16S rRNA sequencing data revealed that the abundance of E. coli was significantly greater in the SADS-CoV P7 group than in the SADS-CoV P83 group and the control group (Fig. 10A). Moreover, proteomic analysis of intestinal tissues from piglets in the SADS-CoV P7 challenge group, SADS-CoV P83 challenge group, and control group revealed that, compared with those in the control group, the differentially expressed proteins in the SADS-CoV P7 challenge group were associated with ECM-receptor interactions (Fig. 10A). Compared with those in the control group, the differentially expressed proteins in the SADS-CoV 83 challenge group were also associated with the ECM-receptor interaction pathway. Compared with control group, proteins related to the ECM pathway, including tenascin (TN), integrin alpha5 (ITGA5), and integrin beta-1 (ITGB1), were significantly upregulated after SADS-CoV infection (Fig. 10A).

An interaction model between SADS-CoV infection and E. coli adhesion was established using IPEC-J2 cells, and the association between host ECM-related proteins and E. coli adhesion was investigated. Changes were analyzed in the mRNA levels of ITGB1, ITGA5, and TN in IPEC-J2 cells infected with SADS-CoV at an MOI of 1 at 1 hour and 7 hours postinfection. The results showed that the expression levels of ITGB1, ITGA5, and TN were significantly upregulated at both 1 hour and 7 hours after infection (Fig. 10A).

In the experiment on the effect of SADS-CoV on Escherichia coli adhesion, using the Escherichia coli fluorescence quantitative counting method established by our laboratory, the gene copy number can be used as the adhesion count of the bacteria. Simultaneous inoculation of both virus and E. coli serotype O157 resulted in a significant increase in adhesion counts at 1 hour and 7 hours compared to inoculation with E. coli O157 alone (p < 0.05) (Fig. 10B). Using fluorescence quantitative PCR, the transcription levels of ITGB1 and ITGA5 were detected in IPEC-J2 cells that were either separately inoculated with SADS-CoV or E. coli O157 or simultaneously inoculated with SADS-CoV and E. coli O157. At 1 hour postinfection, there was no significant difference in the expression level of ITGB1 between the coinfection group and the group inoculated solely with the virus (virus-only group) (Fig. 10C). However, the expression level of ITGB1 in the coinfection group was significantly upregulated compared to the level in the group inoculated solely with bacteria (bacteria-only group) (p < 0.001); ITGB1 expression was also significantly upregulated in the virus-only group when compared to expression in the bacteria-only group (p < 0.001) (Fig. 10C). At 7 hours postinfection, there was no significant difference in ITGB1 expression between the coinfection group and the virus-only group. However, the expression level of ITGB1 in both the coinfection group and the virus-only group remained significantly upregulated compared to the level in the bacterial-only group (p < 0.001) (Fig. 10C). At 1 hour postinfection, there was no significant difference in the expression level of ITGA5 between the coinfection group and the bacteria-only group. However, compared to the level in virus-only group, the expression level of ITGA5 was significantly downregulated in the coinfection group (p < 0.05), and the expression of ITGA5 was significantly lower in the bacteria-only group compared to that in the virus-only group (p < 0.01). At 7 hours postinfection, the expression level of ITGA5 was a significantly lower in the bacteria-only group compared to that in the coinfection group. However, compared to the virus-only group, there was a significantly upregulated in ITGA5 expression in the coinfection group (p < 0.005), and the expression level of ITGA5 was significantly downregulated in the bacteria-only group compared to the virus-only group (p < 0.001). Additionally, coinfection led to upregulation of ITGA5 expression when compared to the level in the bacteria-only group (Fig. 10C). These results demonstrated that, at the transcriptional level, the expression patterns of ITGB1 and ITGA5 were largely consistent with the observed increase in bacterial adhesion.

IPEC-J2 cells were subjected to overexpression of ITGA5, followed by bacterial inoculation for adhesion quantification and transcription level measurement. Compared to levels in the control group, ITGA5 overexpression resulted in a significant increase in bacterial adhesion at 1 hour postbacterial inoculation (p < 0.05), along with a highly significant increase in ITGA5 transcriptional levels (p < 0.001) (Fig. 10D). At 7 hours postbacterial inoculation, in the group with ITGA5 overexpression, there was a significant increase in bacterial adhesion compared to in the control group (p < 0.05), accompanied by a highly significant increase in ITGA5 transcriptional levels (p < 0.001) (Fig. 10D). These results indicate that increased expression of ITGA5 leads to increased adhesion of E. coli O157 to cells. At 1 hour after bacterial inoculation, there was no significant difference in bacterial adhesion between the ITGA5-silenced group and the control group, while the transcriptional level of ITGA5 decreased significantly (p < 0.01) (Fig. 10E). At 7 hours postbacterial inoculation, in the ITGA5-silenced group, bacterial adhesion significantly decreased compared to that in the control group (p < 0.05), while there was no significant difference in the transcriptional level of ITGA5 (Fig. 10E). Based on these results, we can conclude that silencing ITGA5 leads to a decrease in bacterial adhesion. Our combined results from the overexpression and silencing of ITGA5 suggest that this gene is a key factor in secondary bacterial infection following SADS-CoV infection in piglets and that increased expression of ITGA5 leads to an increase in E. coli adhesion.

This study employed multiomics approaches to investigate alterations in the gut microbiome, metabolites, SCFAs, and functional proteins in the feces and small intestinal mucosa of piglets infected with two strains of SADS-CoV, namely, SADS-CoV P7 and SADS-CoV P83. SADS-CoV P83 was derived from SADS-CoV P7 through multiple passages. Additionally, this research explored the molecular events associated with potential pathogenic mechanisms and the host immune response to SADS-CoV infection through association analysis of data from the various omics studies. The findings demonstrated that SADS-CoV infection induces changes in the dominant microbial communities in the small intestinal mucosa and feces of piglets across all microbial taxonomic levels. Samples infected with SADS-CoV P7 exhibited distinct spatial separation in microbial composition from the control group, accompanied by significant differences in microbial functions. Conversely, samples infected with SADS-CoV P83 showed considerable overlap with the control group, indicating greater similarity between these two groups in terms of microbial composition and function. This may be related to the virulence differences between the SADS-CoV P83 and SADS-CoV P7 strains, with SADS-CoV P83 exhibiting lower pathogenicity.

In the samples of small intestinal mucosa and feces from the SADS-CoV P7-infected group, the abundance of pathogens associated with disease, including diarrhea, such as Escherichia, Clostridium, and Megamonas, was significantly greater than in the SADS-CoV P83-infected group and the control group. Moreover, the abundance of bacteria beneficial to intestinal health, such as Faecalibacterium, unidentified Ruminococcaceae, Actinobacillus, and Streptococcus, was notably lower in the SADS-CoV P7-infected group than in the SADS-CoV P83-infected and control groups. Existing research suggests that the severity of porcine epidemic diseases largely depends on the occurrence of secondary bacterial infections [7]. As organs directly connected to the external environment, the respiratory and gastrointestinal tracts are the primary targets of pathogenic assaults. Specifically, in cases of gastroenteritis, coinfection with intestinal pathogens plays a crucial role [6]. For instance, infection with TGEV can induce intestinal cells to produce inflammatory cytokines, thereby increasing the susceptibility of the intestine to E. coli infection [11]. According to previous findings of our research group, piglets infected with SADS-CoV P83 exhibited mild diarrhea symptoms and promptly recovered, displaying thickened intestinal walls and normal mesenteries that closely resembled those of mock-infected piglets, based on clinical manifestations [10]. In our present study, we observed a significant increase in the abundance of E. coli in the SADS-CoV P7-infected group compared to both the control group and the SADS-CoV P83-infected group. Unidentified Enterobacteriaceae and the metabolite sedanolide exhibited a significant negative correlation. Previous studies have shown that sedanolide can alleviate dextran sulfate sodium (DSS)-induced colitis in mice by modulating the FXR-SMPD3 pathway in the intestinal tract. Additionally, the SADS-CoV P7-infected group displayed a notable decrease in the concentration of SCFAs in both the small intestinal mucosa and feces compared to concentrations in the control and SADS-CoV P83-infected groups. SCFAs are known to play a critical role in maintaining the gut microbiome balance and are essential for overall health [12]. SCFAs levels can indirectly respond to modulation of intestinal flora and reduce intestinal inflammation by reducing the production of chemokines or adhesion molecules. Moreover, our research revealed a negative correlation between SCFA levels in the small intestinal mucosa and the abundance of the Enterobacteriaceae family. Strains of E. coli, such as O157:H7, which is known for producing Shiga toxin, can inflict damage on intestinal cells, leading to watery or bloody diarrhea [13]. This observation might explain the more pronounced clinical symptoms observed in the SADS-CoV P7-infected group.

Viral infections followed by secondary bacterial infections can considerably increase the complexity and severity of disease. For example, in mice coinfected with the H1N1 virus and Streptococcus pneumoniae, there was a notable exacerbation of disease conditions, resulting in a mortality rate of up to 100% [14]. Additionally, in humans, concurrent infection with viruses and bacteria not only exacerbates the severity of pneumonia but also elevates the incidence and mortality rates among patients with acquired pneumonia, thereby resulting in poorer prognoses [15].

Our proteomic analysis revealed that the differences in protein expression between the SADS-CoV P7-infected group and the control group, as well as between the SADS-CoV P7-infected group and the SADS-CoV 83-infected group, were primarily linked to ECM-receptor interactions and focal adhesion pathways. However, the proteins that differed between the SADS-CoV P83-infected group and the control group did not exhibit significant associations with ECM-receptor interactions or focal adhesion pathways. This suggests that SADS-CoV P7 significantly affects ECM-receptor interactions or focal adhesion pathways, whereas SADS-CoV P83 does not exhibit this effect. In the comparisons among the SADS-CoV P7 group, SADS-CoV P83 group, and control group, notable disparities in the abundance of bacteria adhering to the small intestinal mucosa, particularly E. coli, were observed. Previous studies have suggested that E. coli colonization of the small intestinal mucosa represents the initial stage of its pathogenicity [16, 17], and the ECM, which consists of adhesion molecules such as fibronectin and collagen, acts as a substrate for bacterial attachment [18]. Bacteria can also employ integrins to facilitate processes related to infection, such as cell adhesion and internalization [19]. Our findings revealed that SADS-CoV infection enhances the adhesion strength and number of E. coli to IPEC-J2 cells. A similar increase in susceptibility to bacterial infections was observed in coinfections with other virus-bacteria combinations. For example, susceptibility to E. coli infections in chickens increased following infection with infectious bronchitis virus [20, 21]. Additionally, studies have demonstrated that influenza virus enhances the adhesion and invasion of Streptococcus pneumoniae and E. coli [22]. These findings align with our results, indicating that viral infections can indeed impact the body's susceptibility to bacterial infections. Moreover, in our subsequent investigations, we observed that, compared to the control, SADS-CoV infection significantly upregulated the expression of ITGB1, ITGA5, and TN. In previous studies from our research group, it was shown that ITGA5 expression was significantly higher in the SADS-CoV P7-infected group compared to the SADS-CoV P83-infected group[23]. Notably, SADS-CoV modulate the adhesion of E. coli O157 to IPEC-J2 cells by regulating the expression of ITGA5. This finding suggested that ITGA5 plays a pivotal role in facilitating the invasion of E. coli into intestinal epithelial cells during SADS-CoV infection, providing insights into the more severe clinical symptoms observed in the SADS-CoV P7 group than in the SADS-CoV P83 group. By integrating multiomics data obtained from the small intestinal mucosa and feces of piglets following infection with different generations of SADS-CoV, this study elucidated the intricate relationships among the microbiome, proteins, metabolic products, and SCFAs (Fig. 11). Specifically, we showed that SADS-CoV modulates the adhesion of E. coli O157 to PEC-J2 cells by influencing the expression of ITGA5, a protein associated with the ECM. However, a comprehensive understanding of the underlying mechanisms involved in this process requires further investigation.

In summary, comprehending the molecular mechanisms of coinfections is paramount for fully understanding disease processes. When viral and bacterial coinfections exacerbate disease, leading to unfavorable prognoses and increased mortality rates, regardless of whether the secondary bacterial infections are triggered by the virus infection or by the suppression of probiotic growth by viruses, such scenarios complicate clinical diagnosis and treatment and significantly impact the animal husbandry industry. Our research provides novel insights into the importance of prevention and control of SADS-CoV infection.

Ethics approval and consent to participate

This study was carried out in accordance with the recommendations of the National Standards for Laboratory Animals of the People’s Republic of China (GB149258-2010). The protocol was approved by the Animal Research Committees of South China Agricultural University. Pigs used for the study were handled in accordance with good animal practices required by the Animal Ethics Procedures and Guidelines of the People’s Republic of China.

Consent for publication

Not applicable

Availability of data and materials

The raw sequencing data have been deposited in the Science Date Bank (ScienceDB) (the review link: https://www.scidb.cn/en/s/UF7Jvm)

The original contributions presented in the study are included in the article. Further inquiries can be directed to the corresponding authors.

Competing interests

The authors declare that they have no competing interests.

Funding

The authors would like to acknowledge the Guangdong Major Project of Basic and Applied Basic Research (No. 2020B0301030007) and the Natural Science Foundation of Guangdong Province (2020A1515010220, 2022A1515012473)

Authors' contributions

XY. T. Conceptualization, Investigation, Methodology, Formal analysis, Writing - original draft, Project administration. CY. L. Investigation, Methodology, Writing - review & editing. JS. S. Writing - review & editing, Methodology. Y.S. Methodology. QN. L. Investigation, Writing - review & editing. T.L. Formal analysis. JY. M. Methodology, Formal analysis, Project administration, Funding acquisition. All authors reviewed the manuscript.

Acknowledgments

Not applicable

Zhou, P. et al. Fatal swine acute diarrhoea syndrome caused by an HKU2-related coronavirus of bat origin. NATURE. 556, 255 (2018).
Li, K. et al. Complete Genome Sequence of a Novel Swine Acute Diarrhea Syndrome Coronavirus, CH/FJWT/2018, Isolated in Fujian, China, in 2018. MICROBIOL RESOUR ANN. 7, (2018).
Zhou, L. et al. Retrospective detection and phylogenetic analysis of swine acute diarrhoea syndrome coronavirus in pigs in southern China. TRANSBOUND EMERG DIS. 66, 687 (2019).
Xu, Z. et al. Isolation and characterization of a highly pathogenic strain of Porcine enteric alphacoronavirus causing watery diarrhoea and high mortality in newborn piglets. TRANSBOUND EMERG DIS 66, 119 (2019).
Yang, Y. L., Yu, J. Q. & Huang, Y. W. Swine enteric alphacoronavirus (swine acute diarrhea syndrome coronavirus): An update three years after its discovery. VIRUS RES. 285, 198024 (2020).
Zhou, J. & Sun, Y. Effect of COVID-19 protective measures on the epidemiology characteristics of rotavirus, adenovirus, and coinfections among pediatric patients with acute gastroenteritis in Hangzhou, China. MICROBIOL SPECTR. 12, e400723 (2024).
Brockmeier, S. L. et al. Comparison of Asian porcine high fever disease isolates of porcine reproductive and respiratory syndrome virus to United States isolates for their ability to cause disease and secondary bacterial infection in swine. VET MICROBIOL. 203, 6 (2017).
Magira, E. E. & Zakynthinos, S. Legionnaire's Disease and Influenza. INFECT DIS CLIN N AM. 31, 137 (2017).
Xia, L., Dai, L., Yu, Q. & Yang, Q. Persistent Transmissible Gastroenteritis Virus Infection Enhances Enterotoxigenic Escherichia coli K88 Adhesion by Promoting Epithelial-Mesenchymal Transition in Intestinal Epithelial Cells. J VIROL. 91, (2017).
Sun, Y. et al. Attenuation of a virulent swine acute diarrhea syndrome coronavirus strain via cell culture passage. VIROLOGY. 538, 61 (2019).
Xia, L., Dai, L., Zhu, L., Hu, W. & Yang, Q. Proteomic Analysis of IPEC-J2 Cells in Response to Coinfection by Porcine Transmissible Gastroenteritis Virus and Enterotoxigenic Escherichia coli K88. PROTEOM CLIN APPL. 11, (2017).
Li, Y. et al. Review: Effect of Gut Microbiota and Its Metabolite SCFAs on Radiation-Induced Intestinal Injury. FRONT CELL INFECT MI. 11, 577236 (2021).
Ray, R. & Singh, P. Prevalence and Implications of Shiga Toxin-Producing E. coli in Farm and Wild Ruminants. PATHOGENS. 11, (2022).
LeMessurier, K. S. et al. Allergic inflammation alters the lung microbiome and hinders synergistic co-infection with H1N1 influenza virus and Streptococcus pneumoniae in C57BL/6 mice. SCI REP-UK. 9, 19360 (2019).
Lim, Y. K., Kweon, O. J., Kim, H. R., Kim, T. H. & Lee, M. K. Impact of bacterial and viral coinfection in community-acquired pneumonia in adults. DIAGN MICR INFEC DIS. 94, 50 (2019).
Izquierdo, M., Alvestegui, A., Nataro, J. P., Ruiz-Perez, F. & Farfan, M. J. Participation of integrin alpha5beta1 in the fibronectin-mediated adherence of enteroaggregative Escherichia coli to intestinal cells. BIOMED RES INT. 2014, 781246 (2014).
Muenzner, P. et al. Uropathogenic E. coli Exploit CEA to Promote Colonization of the Urogenital Tract Mucosa. PLOS PATHOG. 12, e1005608 (2016).
Bisht, S. et al. Expression of fibronectin-binding protein of L. acidophilus NCFM and in vitro refolding to adhesion capable native-like protein from inclusion bodies. PROTEIN EXPRES PURIF. 145, 7 (2018).
LaFoya, B. et al. Beyond the Matrix: The Many Non-ECM Ligands for Integrins. INT J MOL SCI. 19, (2018).
de Wit, J. J. & Cook, J. K. Factors influencing the outcome of infectious bronchitis vaccination and challenge experiments. AVIAN PATHOL. 43, 485 (2014).
Alber, A. et al. Dose-dependent differential resistance of inbred chicken lines to avian pathogenic Escherichia coli challenge. AVIAN PATHOL. 48, 157-167 (2019).
Narayana, M. A. et al. In vivo and in vitro studies on the roles of neutrophil extracellular traps during secondary pneumococcal pneumonia after primary pulmonary influenza infection. FRONT IMMUNOL. 4, 56 (2013).
Li, Q. et al. Combined Analysis of the Whole Transcriptome of Piglets Infected with SADS-CoV Virulent and Avirulent Strains. MICROORGANISMS. 11, (2023).

No competing interests reported.

Supplementfigure1.png
Supplementalfigure 1. Changes in dominant bacteriain the ileum mucosa and feces of piglets causedby SADS-CoV infection at each microbial taxonomic level
Supplementfigure2.png
Supplementalfigure 2. Alpha and beta diversity (A) and (B) α diversity, including the bacterial diversity indices (Shannon and Simpson indices). (B) PCoA of microbiota from the six groups using weighted UniFrac distances. (C) href="https://www.sciencedirect.com/topics/agricultural-and-biological-sciences/upgma" title="Learn more about UPGMA from ScienceDirect's AI-generated Topic Pages">UPGMACluster analysis at the phylum level comparing the distribution of samples between groups.
Supplementfigure3.png
Supplemental Figure3. Detailed interspecies differences at the genus level. (A) Genus-level inter-group differential bacteria in the feces. (B) Genus-level inter-group differential bacteriain the ileum mucosa.
Supplementfigure4.png
Supplemental Figure4. Mmetabolite profiling analysis. (A) Relative abundance plot of differentially abundant metabolites. (B) Heatmap of clustered differentially abundant metabolites. (C) KEGG pathway enrichment bubble plot of differentially abundant metabolites.
Supplementtable1.xlsx
Supplementtable2.xls

Download PDF

Version 1

posted

You are reading this latest preprint version

Multiomics Reveals Alterations in the Gut Microbiome, Host Proteins, and Host Metabolites Correlating with SADS-CoV Pathogenicity and the Immune Response in Piglets

Status:

Version 1

Abstract

Figures

INTRODUCTION

MATERIALS AND METHODS

Virus strains

Piglet challenge experiments

16S rRNA gene sequencing

Protein detection and tandem mass tag (TMT) labeling

Liquid chromatography‒mass spectrometry (LC‒MS) analysis

Gas chromatography (GC)‒MS targeted detection

Metabolite extraction and identification

SADS-CoV infects IPEC-J2 cells

SADS-CoV and E. coli o157 coinfect IPEC-J2 cells

Expression and interference of ITGA5

RESULTS

SADS-CoV infection leads to changes in the structure of the microbiota

Abnormal metabolism of short-chain fatty acids is accompanied by microbial dysbiosis in SADS-CoV-infected piglets

Global overview of metabolism in the fecal metabolome

Analysis of the correlations between the microbiota and metabolites

Global overview of proteomics

Analysis of the correlations between the microbiota and proteins

Correlation networks

DISCUSSION

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1