Bacterial isolation and identification
During our routine surveillance of bacillary dysentery, a total of 136 Shigella isolates were collected in this study, including 54 S. flexneri (1.63%, 54/3321), 44 S. sonnei (1.32%, 44/3321), and 38 S. dysenteriae (1.14%, 38/3321), and no S. boydii was found. All 38 S. dysenteriae were only found on six farms in Gansu Province.
Detailed information on the study isolates is listed in Table S3. Among them, 20 isolates were isolated from 3 beef cattle farms, and 18 isolates were isolated from 3 dairy farms. There are 14 and 10 S. dysenteriae isolates from the same farms of Zhangye and Jinchang, respectively.
All 38 S. dysenteriae isolates were type 1 according to the results of the serotype reactions. In addition, based on the typical biochemical characteristics of Shigella spp., analysis of biochemical reactions indicated the presence of 3 biotypes among these isolates (Table 1, Table S3). Among these BTs, BT2 (ability to ferment glucose, arabinose, and melibiose) was the predominant biotype, accounting for 86.84% (33/38). Furthermore, BT2 was widely isolated from each locus, with the exception of Baiyin.
MLST-based genotype analysis
Thirty-eight S. dysenteriae isolates belonged to 4 STs, ST57, ST191, ST228 and ST229. Among them, ST57 and ST191 were previously reported, while ST228 and ST229 were unique. The allele number for each locus and the designation of the ST are listed in Fig. 1 and shared on the EcMLST website. The most common STs identified were ST228 (n = 13) and ST229 (n = 15), accounting for 73.68% (28/38). ST228 and ST229 were major ST types for dairy cows and beef cattle, respectively. In addition, these two ST types were different by 3 loci: arcA, clpX, mutS. The other ST type, ST57 includes 2 isolates from beef cattle and 5 isolates from dairy cows. ST191 was only isolated from three steers in Linxia (Fig. 1, Fig. 2).
PFGE-based genotype analysis
The genotypes and genetic relatedness of the 38 isolates were further determined by using PFGE. PFGE patterns of these S. dysenteriae isolates were heterogeneous; however, multiple PFGE patterns were present among these strains (Fig. 3). With approximately 80% similarity, XbaI-digested S. dysenteriae type 1 could be divided into 28 distinctive PFGE patterns (PT) and belonged to two major groups: A (A1-A4) and B (B1-B3), about 66% similarity. Interestingly, most strains in different geographical locations can be clustered individually, suggesting high genetic diversity among S. dysenteriae 1.
Prevalence of virulence genes
The frequencies of some major Shigella virulence genes in this study. A total of five virulence genes were detected in those isolates involving ipaH, ipaBCD, ial, sen, and stx. The most frequently observed virulence genes are ipaH (100%), ipaBCD (92.11%), stx (73.68%), and ial (57.89%). The Shigella enterotoxin genes sen (28.95%) are occasionally present in Wuwei and Jinchang isolates. None of the study strains possessed the set1A or set1B gene.
Regarding the differences in virulence gene distributions, the 38 S. dysenteriae isolates fell into 5 gene profile types (VT) (Table 2). Among these VTs, VT IV (n = 17) and VT V (n = 11) were the most common, accounting for 44.34% and 28.95%, respectively. One interesting finding was the presence of the same and/or similar VTs in the same locus. In addition, 92.11% of isolates carried two or more virulence genes. Three Linxia isolates (7.89%) belonged to VT I, which was only positive for ipaH.
Antimicrobial resistance profiles
The antimicrobial resistance profiles of the 20 antimicrobials for 38 S. dysenteriae are shown in Table 3. All S. dysenteriae isolates were uniformly multidrug resistant to at least 3 types of antimicrobial agents. Among them, resistance to E was the most common (36,94.74%), followed by AMP (35, 92.11%), KZ (33, 86.84%), CRO (32, 84.21%), CTX (32, 84.21%), TE (31, 81.58%), CN (31, 81.58%), ENR (26, 68.42%), LEV (25, 65.79%), CIP (25, 65.79%), NOR (15, 39.47%), OFX (15, 39.47%), C (15, 39.47%), and S (15, 39.47%). Fortunately, all 38 isolates were sensitive to AMC, FOX, FEP, MEM, IPM and AK. However, the resistance rate of fluoroquinolone antibiotics has reached 39.47–68.42%, which is not optimistic.
Moreover, most of the isolates (29/38, 76.32%) were resistant to fluoroquinolone antibiotics, there were 4 isolates resistant to other fluoroquinolones that is not CIP. These resistant isolates were divided into 5 antibiotic-resistance profile patterns (Table 4). All of the fluoroquinolone-resistant isolates were multidrug resistant (MDR), and 41.8% (12/29) of fluoroquinolone-resistant isolates were resistant to each class of antibiotics. In addition, the resistance profiles of strains isolated from the same farms were very similar.
ARG analysis for fluoroquinolone-resistant S. dysenteriae isolates
To determine the molecular characterization in fluoroquinolone-resistant S. dysenteriae, both SNPs in QRDR of gyrA/B and parC/E genes and PMQR genes were analyzed. In 29 fluoroquinolone-resistant isolates, there was no strain that displayed mutations in the gyrB and parE genes, although some mutations in gyrA and parC were identified in each resistant isolate (Table 4). All resistant isolates in the present study carried common mutations in gyrA codon 83 (S→L) and parC codon 83 (S→L). Furthermore, 12 Zhangye isolates showed an additional 87 (D→N) and 80 (S→I) mutations in gyrA and parC, respectively.
The PMQR genes qnr, aac(6’)-Ib-cr, and qepA occur worldwide and are increasingly detected in clinical isolates of Enterobacteriaceae. 20,21 In our study, aac(6’)-Ib-cr was the most common PMQR gene and harbored by each isolate; however, all isolates were negative for qepA, except SD001. Only seven (7/29, 24.14%) isolates harbored the qnr gene, and no isolate harbored qnr, aac(6’)-Ib-cr, and qepA simultaneously. Interestingly, the qnrB (n = 2) gene and qnrS (n = 5) were detected in Jinchang and Zhangye, respectively.