From July to August 2023, fecal samples were collected from diarrheic calves at 10 large-scale dairy or beef cattle farms located in Shizuishan, Wuzhong, Zhongwei, and Guyuan cities in Ningxia Hui Autonomous Region. A total of 100 fecal samples were collected, with 10 samples obtained from each farm (Table 1). Each sample was labeled using a unique patient ID that incorporated the initials of the sampling city in Pinyin.

Fecal samples were 10-fold diluted in peptone buffer, and 0.2 mL of the diluted suspension was inoculated into LB broth and incubated at 37°C until the logarithmic growth phase. The cultures were then transferred onto E. coli chromogenic medium and Salmonella chromogenic medium and incubated at 37°C for 16–18 h. A single purified colony was selected using an inoculation loop for Gram staining and microscopic observation. Bacterial genomic DNA was extracted (24), and PCR amplification was performed using specific primers. The amplification system contained 3.5 μL of Taq PCR Master Mix (Dining, China), 0.5 μL each of forward and reverse primers, 0.5 μL of DNA template, and 5 μL of dd H₂O. PCR products were analyzed by 1.5% agarose gel electrophoresis (Dining, China) and visualized using a gel imaging system. Additionally, PCR was used to detect virulence genes of E. coli and Salmonella. Primers were designed based on sequences from the NCBI database. The bacterial primer sequences and annealing temperatures are shown in Table 2, and virulence gene primers are listed in Table 3. All media were purchased from Qingdao Hope Biotechnology Co., Qingdao, China. The E. coli strain ATCC^® 25922 and Salmonella strain ATCC^® 12002 preserved in our laboratory was used as a control.

Antibiotic susceptibility was determined according to the standards of the Clinical and Laboratory Standards Institute (CLSI, 2023) using the broth microdilution method. A total of nine classes of commonly used veterinary antibiotics were tested (Shanghai Xingbai Biotechnology Co., Ltd., China). The test bacterial strains and reference strains were prepared as bacterial suspensions with a final concentration of 0.5 × 10⁵ CFU/mL. Using an 8-channel pipette, 100 μL of each working solution was added to the wells of the susceptibility testing plate. Negative and positive controls were included. After incubation at 37°C for 16–18 h, results were recorded by observing the wells. The lowest concentration of the antibiotic that showed no visible bacterial growth (clear well) was recorded as the minimum inhibitory concentration (MIC).

The checkerboard method was used to evaluate the combined antibacterial activity of the Chinese herbal formula and Microcin J25 against E. coli and Salmonella isolated from diarrheic calves (25). Briefly, four herbal extracts—Populus tomentosa male flower, Portulaca oleracea, Sanguisorba officinalis, and Euphorbia humifusa (Sichuan Hengrui Tongda Biotechnology Co., Ltd., China)—were mixed in equal proportions to prepare the Chinese herbal formula. All four herbal extracts were prepared according to the water extraction and alcohol precipitation method specified in the Chinese Pharmacopoeia (2025 edition), with a standardized concentration of 1 g extract equivalent to 10 g crude drug. The formula and Microcin J25 (Zhongnong Yingtai Biotechnology Co., Ltd., China) were then diluted to seven gradient concentrations ranging from 1/16 MIC to 2 MIC. In a 96-well microtiter plate, each column contained a fixed concentration of drug A, and each row contained a fixed concentration of drug B. The first row and column served as monotherapy controls, and the well at the intersection (origin) served as the positive control. The initial bacterial concentration in each well was adjusted to approximately 0.5 × 10⁵ CFU/mL. E. coli ATCC^® 25,922 and Salmonella ATCC^® 12,002 were used as standard reference strains. After incubation at 37°C for 16–18 h, the results were recorded. The experiment was repeated in triplicate. FICI was calculating by given formula: FICI was calculating by given formula: FICI was calculated as the sum of the MIC of Chinese Herbal Formula in combination divided by the MIC of Chinese Herbal Formula alone, and the MIC of Microcin J25 in combination divided by the MIC of Microcin J25 alone. FICI≤0.5 means “synergy”; 0.5 < FICI≤0.75 means “partial synergy”; 0.76 < FICI≤1 denotes “additive”; 1 < FICI≤4 denotes “indifferent”; while “antagonistic” in cases which the FIC index > 4 (26).

As previously described, the growth curve assay was performed with slight modifications (27). The growth-inhibitory effects of the Chinese herbal formula and Microcin J25, alone or in combination, against E. coli and Salmonella were evaluated by growth curve analysis. The experiment was divided into four groups: control (no treatment), Microcin J25 alone, Chinese herbal formula alone, and a combination treatment in which both Microcin J25 and the Chinese herbal formula were administered together. The concentrations used were 0.25 × MIC for single treatments and 0.25 × MIC each for the combination group. Each group was inoculated with standardized bacterial suspension (final concentration: ~0.5 × 10⁵ CFU/mL) and incubated at 37°C. Aliquots of 100 μL from each group were sampled every 2 h for 18 h, and bacterial growth was monitored by measuring the optical density at OD₆₀₀ using a microplate reader. Each experiment was performed in triplicate, and the growth curves were plotted to compare the inhibitory effects of different treatments over time.

The relative expression levels of virulence genes in isolated strains were determined by real-time quantitative fluorescence PCR (RT-qPCR). RNA extraction was performed using the triazole method, with RNA concentration determined by spectrophotometry and integrity assessed via agarose gel electrophoresis. Subsequently, cDNA synthesis was carried out using an Integrated First-strand cDNA Synthesis kit, followed by RT-qPCR analysis using 2 × Fast HS SYBR QPCR Mixture. The amplification protocol was as follows: 95.0°C for 30 s; 95.0°C for 5 s, 60.0°C for 34 s, for a total of 40 cycles; followed by 95.0°C for 15 s, 60.0°C for 1 min, and 95.0°C for 15 s. The resulting qPCR data were analyzed for relative changes in gene expression levels based on the 2^−∆∆Ct method. Primers listed in Table 3 were utilized for this study.

Eighty healthy Holstein calves aged between 7 and 14 days were randomly divided into eight groups, with ten calves per group. The groups included: a negative control group (fed with standard pasteurized milk, referred to as NC); low-, medium-, and high-dose Chinese herbal formula groups (administered at doses of 0.05 g, 0.1 g, and 0.2 g per kilogram of body weight, referred to as CHFL, CHFM, and CHFH, respectively); a Microcin J25 group (supplemented at a concentration of 0.8 g per liter of milk, referred to as MCJ); and three combination groups receiving the same doses of the Chinese herbal formula along with 0.8 g per liter of Microcin J25 (referred to as COML, COMM, and COMH, respectively). The trial lasted for 7 days. Except for the control group, all groups were supplemented with the respective agents via milk feeding twice daily. Fecal scores were defined as follows: 0 indicated firm, well-formed feces; 1, slightly loose feces; 2, unformed loose feces; and 3, watery diarrhea (28). At the end of the trial, blood and rectal swab samples were collected for further analysis.

Twenty diarrheic Holstein calves aged between seven and 14 days were randomly divided into four treatment groups, with five calves in each group. The groups included: an enrofloxacin group (ENR), a Microcin J25 group (MCJ), a Chinese herbal formula group (CHF), and a combination group (COM). The trial lasted for 7 days. The treatment regimens were as follows: calves in the enrofloxacin group received intramuscular injections of enrofloxacin at a dose of 0.1 milliliters per kilogram of body weight once daily; the Microcin J25 group received 0.8 g of Microcin J25 per liter of milk daily; the CHF group was administered the herbal formula at 0.1 g per kilogram of body weight daily; and the combination group received both the herbal formula (0.1 g per kilogram of body weight) and Microcin J25 (0.8 g per liter of milk) daily. Fecal scores were recorded daily. Clinical resolution was defined as maintaining a fecal score below 2 for ≥2 consecutive days. At the end of the trial, blood and rectal swab samples were collected for analysis.

After thawing the collected calf anal swabs, microbial genomic DNA was extracted using a DNA extraction kit (Meiji Biotechnology Co., Ltd., Guangzhou, China) and detected by agarose gel electrophoresis. The V3-V4 hypervariable regions of bacterial 16S rDNA were amplified using specific primers: 341F (5′-CCTAYGGGRBGCASCAG-3′) and 806R (5′-GGACTACNNGGG TATCTAAT-3′). PCR products were purified with magnetic beads, pooled based on concentration, and target bands were excised for library construction. The constructed libraries were quantified using Qubit® and RT-qPCR. Qualified libraries were sequenced on the Illumina MiSeq platform (PE250 mode, Novaseq 6,000 system). Sample data were demultiplexed based on barcode and primer sequences. After removing barcodes and primers, paired-end reads were merged using FLASH to generate raw Tags. Reverse primer sequences were trimmed with Cutadapt to minimize interference in downstream analyses. Raw Tags were rigorously filtered with fastp to obtain high-quality Tags, which were then aligned against a reference database for chimera removal, yielding final effective OTUs. The validated OTUs were subjected to clustering analysis. Subsequent analyses included OTUs abundance profiling, alpha diversity (within-sample diversity), beta diversity (between-sample diversity), and taxonomic composition.

All experimental data were analyzed and visualized using GraphPad Prism 8. In all comparisons, p < 0.01 or p < 0.05 was considered statistically significant.

Among the 100 fecal samples collected from diarrheic calves, a total of 97 E. coli and 20 Salmonella isolates were recovered (Figure 1A). All 97 E. coli isolates were sensitive to carbapenems. The highest resistance rate was observed against sulfonamides (83.5%), followed by resistance to amphenicols (24.8%), cephalosporins (19.6%), aminoglycosides (12.4%), quinolones (10.3%), tetracyclines (9.3%), penicillins (5.2%), and polymyxins (2.1%) (Figure 1B). All 20 Salmonella isolates were also sensitive to carbapenems and polymyxins, but exhibited the highest resistance to sulfonamides (75.0%), followed by aminoglycosides (50.0%), tetracyclines (25.0%), penicillins (15.0%), quinolones (10.0%), amphenicols (10.0%), and cephalosporins (5.0%) (Figure 1C).

Among the E. coli isolates, the F41 virulence gene had the highest detection rate (39.2%), followed by eaeA (23.7%), STX1 (6.2%), STX2 (2.1%), and STa (1%) (Figure 1D). For Salmonella, the most frequently detected virulence gene was spvC (45%), followed by sopE and invA (both 35%), fimA (30%), and stn (25%) (Figure 1E).

To evaluate the potential synergistic effect of the Chinese herbal formula in combination with Microcin J25, the minimum inhibitory concentrations (MIC) of each agent against 97 previously identified E. coli and 20 Salmonella isolates were determined. Subsequently, checkerboard assays and bacterial growth curve analyses were conducted on three E. coli strains (GY3-5E, GY3-7E, ZW1-5E) and three Salmonella strains (GY3-2S, GY2-4S, ZW1-10S) harboring multiple virulence genes. The MIC of individual agents are shown in Figure 2A. As illustrated in Figure 2B and detailed in Table 4, the combination of the Chinese herbal formula with Microcin J25 exhibited partial synergy against the E. coli reference strain and either partial synergy or additive effects against the remaining E. coli isolates. For Salmonella, synergistic effects were observed against the reference strain, while partial synergy or indifferent interactions were observed for the clinical isolates.

Importantly, as shown in Figure 2C, the combination treatment consistently suppressed the growth of both E. coli and Salmonella. In E. coli, no significant changes were observed in the mRNA expression levels of F41, eaeA, and STX1 (p > 0.05). However, the expression of STX2 was significantly reduced in the Microcin J25, Chinese herbal formula, and combination groups compared with the control group (p < 0.001), with the most pronounced inhibition observed in the combination group. In addition, the expression of STa was significantly reduced in the Microcin J25 group compared with the control (p < 0.05) (Figure 2D). For Salmonella, the expression levels of stn, fimA, and invA were unchanged following treatment (p > 0.05). The expression of sopE was significantly downregulated in the combination group compared to the control (p < 0.05), and spvC expression was markedly decreased in the Microcin J25, Chinese herbal formula, and combination groups (p < 0.0001), with the greatest effect observed in the combination treatment (Figure 2E).

As shown in Table 5, all treatment groups exhibited lower fecal scores compared to the NC group. On day 5, the fecal scores in the MCJ, CHFM, CHFH, COML, COMM, and COMH groups were significantly lower than those in the NC group (p < 0.05). The highest diarrhea incidence was observed in the NC group at 50%, whereas the MCJ group showed a marked reduction to 20%. The diarrhea rates in the CHFL, CHFM, and CHFH groups were also reduced compared to the NC group, with the CHFM group showing the best effect at 20%. A further reduction in diarrhea incidence was observed in the combination groups, particularly in the COMM and COMH groups, which showed the lowest incidence at 10% (Figure 3A).

Serum IgA levels in the COML, COMM, and COMH groups were significantly higher than those in the NC group (p < 0.05), while no significant changes were found in the MCJ, CHFL, CHFM, or CHFH groups (p > 0.05). Serum IgM levels were significantly elevated in the COMM and COMH groups compared to the NC group (p < 0.05), whereas no significant differences were observed in the MCJ, CHFL, CHFM, CHFH, or COML groups (p > 0.05). Serum IgG levels were significantly increased in the CHFM and CHFH groups (p < 0.05), and were markedly elevated in the COML, COMM, and COMH groups compared to both the NC and MCJ groups (p < 0.01) (Figure 3B).

Based on the above experimental results, the CHFM and COMM groups, which showed the best performance, were selected along with the NC and MCJ groups for 16S rDNA sequencing to explore their regulatory effects on the gut microbiota. As shown in Figure 4A, the Venn diagram of OTUs revealed that the CHF group had 268 OTUs, followed by the NC group with 128, the MCJ group with 81, and the COM group with 50. Figure 4B indicates that the sample points of the NC group were farther from those of the CHF and COM groups in the PCoA analysis, suggesting significant differences in microbial structure. The Chao1 and Shannon indices in the COM group were significantly decreased, while those in the CHF group increased significantly (p < 0.05), as shown in Figure 4C. At the phylum level (Figure 4D), Firmicutes was the dominant phylum in the NC, MCJ, and CHF groups, while both Firmicutes and Actinobacteriota dominated in the COM group. UCG-005 was the dominant genus in the NC and MCJ groups, Lactobacillus in the CHF group, and Bifidobacterium in the COM group. LEfSe analysis (Figure 4E) further showed that Lactobacillus mucosae, Methanocorpusculum, and Methanocorpusculaceae were significantly enriched in the CHF group. Meanwhile, Bifidobacterium breve and Collinsella were enriched in the COM group. These findings indicate that CHF improved gut microbiota diversity and stability in calves, increasing the abundance of beneficial bacteria, while COM mainly enhanced the abundance of specific probiotics.

The cure rate was lowest in the CHF group at 20%, followed by the ENR group at 40%, while the COM group exhibited the highest cure rate at 80% (Figure 5A). The COM group showed a significantly shorter average recovery time than the MCJ group (p < 0.05) and a markedly shorter time than the CHF group (p < 0.01), indicating superior therapeutic efficacy (Figure 5B). Moreover, the COM group had significantly lower serum IL-1β and IL-6 levels compared to the other groups (p < 0.05) (Figure 5C).

As shown in the Venn diagram (Figure 6A), the number of OTUs detected in the ENR, CHF, MCJ, and COM groups was 40, 43, 120, and 165, respectively. The COM group exhibited a distinct scatter distribution from the ENR and MCJ groups in PCoA analysis, indicating substantial differences in microbial community structures between these groups (Figure 6B). Compared to the other three groups, the COM group showed significantly higher Chao1 and Shannon indices (p < 0.05), suggesting increased richness and diversity (Figure 6C).

At the phylum level, Firmicutes was the dominant phylum in both the CHF and COM groups, while Firmicutes and Proteobacteria were dominant in the ENR group. At the genus level, Escherichia-Shigella was predominant in the ENR and MCJ groups, while Enterococcus, Streptococcus, and Peptostreptococcus were dominant in the CHF group. In contrast, Collinsella, Lactobacillus, and Blautia were enriched in the COM group (Figure 6D). LEfSe analysis further revealed higher abundance of Escherichia-Shigella and Enterococcus in the ENR group; increased Ruminococcus, Blautia, and Clostridium in the MCJ group; elevated Lactobacillus, Bacillus, and Peptostreptococcus in the CHF group; and increased Lactobacillus, Bifidobacterium, and Clostridium in the COM group (Figure 6E).

These findings indicate that the combination of Chinese herbal formula and Microcin J25 effectively reduced the abundance of harmful bacteria while enriching beneficial microbes in the gut of diarrheic calves, thereby improving the overall microbial structure.

Subsequently, correlation analyses were performed among anti-inflammatory factors, immunoglobulins, and dominant bacterial taxa detected in rectal swabs (Figure 7). The abundance of Euryarchaeota showed a highly significant positive correlation with serum IgM levels (p < 0.01). A significant positive correlation was observed between Lactobacillus abundance and serum IL-6 levels (p < 0.05), whereas Bifidobacterium abundance exhibited a significant negative correlation with serum IL-6 levels (p < 0.05). These findings suggest that the combined treatment of Chinese herbal formula and Microcin J25 may improve anti-inflammatory status and immune function in calves through modulation of gut microbiota composition.