A total of 68 S. Kentucky ST198 isolates were identified from 1,574 cultured Salmonella isolates from diarrhea patients, food, and environmental samples across eleven regional Center for Disease Control and Prevention diagnostic laboratories located in Henan, China, between 2018 and 2022. Serovars of confirmed Salmonella isolates were determined by slide agglutination using commercial antisera (Statens Serum Institute, Denmark), following the Kauffmann–White scheme.

Antimicrobial susceptibility testing of 68 Salmonella Kentucky ST198 isolates was performed against 15 agents from 11 distinct antimicrobial classes using the agar dilution method (Clinical Laboratory Standards Institute (CLSI), 2022), with Escherichia coli ATCC™ 25922 as the quality control strain. The agents and classes include: penicillin (ampicillin, AMP); β-lactam/β-lactamase inhibitor combinations (ampicillin-sulbactam, AMS); cephalosporins (cefotaxime, CTX; ceftazidime, CAZ; cefazolin, CFZ; and cefoxitin, CFX); folate pathway inhibitors (trimethoprim-sulfamethoxazole, SXT); carbapenems (imipenem, IPM); aminoglycosides (gentamicin, GEN); tetracyclines (tetracycline, TET); macrolides (azithromycin, AZM); polymyxins (colistin, CT); phenicols (chloramphenicol, CHL); and quinolones (ciprofloxacin, CIP; and nalidixic acid, NAL). MDR phenotype was defined as resistance to three or more of these antimicrobial classes (Magiorakos et al., 2012).

Genomic DNA from Salmonella was extracted using the QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany). All DNA samples were sequenced on the Illumina NovaSeq 6,000 platform (Illumina, USA), generating 150-bp paired-end reads from libraries with an average insert size of 350-bp and sequencing depth of more than 100×. Raw reads were filtered to remove low-quality sequences and subsequently assembled de novo using SPAdes v3.8.2 (Bankevich et al., 2012). Five representative isolates were further sequenced by long-read sequencing using the MinION platform from Oxford Technologies (ONT, Oxford, UK). Complete genome sequences were obtained through a hybrid de novo assembly strategy with Unicycler (Wick et al., 2017). Draft and complete genomes were annotated using Prokka (Seemann, 2014). AMR genes and chromosomal point mutations were identified using ResFinder4.1 with thresholds of 90% identity and ≥80% minimum length coverage (Bortolaia et al., 2020). The replicon types of plasmids and insertion sequences were detected using PlasmidFinder database and ISFinder database under the same thresholds (Carattoli et al., 2014; Kichenaradja et al., 2010). Core genome multilocus sequence typing (cgMLST) profiles (cgSTs) of the 68 isolates were analyzed based on 3,002 alleles using the cgMLSTFinder database (Alikhan et al., 2018).

Core genomes of all assemblies were calculated using Roary v3.13.0 (Page et al., 2015). Snippy v4.6¹ were used to clean indels in the core genomes, while Gubbins v3.3.1 (Croucher et al., 2015) were used to exclude recombination region. Core genome single nucleotide polymorphisms (cgSNPs) were extracted using SNP-sites v2.5.1 (Page et al., 2016). SNP distance matrices for all isolates were generated using snp-dist v.0.8.2.² Non-repetitive core SNPs were used to construct phylogenetic analysis by the maximum likelihood method with FastTree v2.1.11 (Price et al., 2009) and visualized using ChiPlot (Xie et al., 2023). In addition, 168 published genomes of S. Kentucky ST198 isolates from other regions were downloaded from the National Center for Biotechnology Information (NCBI) and incorporated into the phylogenetic tree with S. Kentucky ST198 isolates of Henan using cgSNPs.

From 2018 to 2022, a total of 1,574 Salmonella isolates were collected from diarrhea patients, food, and environmental samples from 11 cities in Henan, through the laboratory-based foodborne disease and food safety surveillance system in China. Among them, 102 serovars were detected (Supplementary Table S1), with S. Kentucky ranking the fourth most common serovar (Table 1).

Of these S. enterica strains, a total of 76 S. Kentucky isolates (4.8%, 76/1574) were identified, including 68 ST198 (89.5%, 68/76) and 8 ST314 (10.5%, 8/76) isolates. The ST198 strains were isolated from diverse sources, including patients (n = 18), poultry products (n = 38), beef (n = 1), pork (n = 1), and environment (n = 10) (Supplementary Table S2). The poultry products isolates (n = 38) were obtained from broiler chicken in retail market (n = 25), chicken farm (n = 1), and slaughterhouse (n = 12). The environmental isolates (n = 10) were obtained from chicken farm (n = 2) and slaughterhouses (n = 8).

Antimicrobial susceptibility testing of the 68 S. Kentucky ST198 isolates against a panel of 15 antimicrobial compounds revealed resistance to 14 compounds (Supplementary Table S3). All the isolates were MDR strains, showed resistance to three or more classes of antimicrobials. No resistance phenotype was detected against imipenem. Resistance to nalidixic acid (100%) and ciprofloxacin (100%) was the most common, followed by ampicillin (98.5%), tetracycline (98.5%), cefotaxime (97.1%), cefazolin (97.1%), gentamicin (91.2%), chloramphenicol (83.8%), ceftazidime (82.4%), sulfamethoxazole-trimethoprim (77.9%), and azithromycin (70.6%). Resistance to colistin (2.9%) and cefoxitin (8.8%) was rare.

To investigate the genetic relationship of Chinese isolates, we executed a phylogenetic analysis based on cgSNPs. The Chinese isolates grouped into two groups, namely ST198-1 (9 isolates) and ST198-2 (59 isolates) with distinct QRDR mutations in gyrA gene (Figure 1). The clade ST198-1 isolates originated from human (n = 6), chicken in market (n = 2), and slaughterhouse (n = 1) (Figure 1, Supplementary Table S2). The isolated sources of clade ST198-2 strains were human (n = 12), samples in retail market (chicken, beef and pork, n = 25) and samples in farm or slaughterhouses (broiler chicken n = 12 and environment n = 10). Within subclade ST198-2, pairwise cgSNP differences among strains isolated from chickens and the environmental samples across different stages (incubation, breeding, or slaughtering) within the same slaughterhouse ranged from 0 to 43 (Supplementary Table S7). Furthermore, cgSNP differences between strains isolated from diarrhea patients and poultry products in local supermarkets ranged from 10 to 18. Most strains of clade ST198-2 were resistant to azithromycin (81.4%) and trimethoprim (88.1%), whereas all ST198-1 strains were sensitive, except for one trimethoprim resistant strain (Supplementary Table S2).

The cgMLST was analyzed based on the whole genome sequences to investigate molecular characteristics of S. Kentucky ST198 isolates. Among the S. Kentucky ST198 isolates were grouped into five cgSTs (Figure 1) and among ST198-1 isolates, three cgSTs were detected, with cgST236434 predominating across four cities (Kaifeng, Pingdingshan, Nanyang, and Zhoukou) of Henan (Supplementary Table S2). Among ST198-2 isolates, two cgSTs detected were [cgST230926 98.3% (58/59)] and [cgST296405 1.7% (1/59)]. The predominant cgST296405 were distributed across 10 cities (Hebi, Xuchang, Zhoukou, Nanyang, Kaifeng, Luoyang, Pingdingshan, Puyang, and Jiyuan) in Henan.

To investigate the genetic relatedness of ST198 isolates globally, we constructed a phylogenetic tree combining our isolates with 168 ST198 genomes obtained from the NCBI database. Consistent with previous reports, global ST198 isolates were grouped into multiple clusters (Figure 2). The phylogenetic relationship showed that the global ST198 isolates clustered separately with distinct QRDR mutations in gyrA gene. Additionally, the clade ST198-1 of Henan clustered with strains from Europe (n = 15), other provinces in China (n = 7), and Africa (n = 1) (Figure 2, Supplementary Table S4). The Henan ST198-2 clade clustered with strains from Europe (n = 47), Africa (n = 12), Asia (n = 11), other provinces in China (n = 11), and Oceania (n = 8). Overall, Chinese isolates were mostly related to the European isolates, with ST198-2 isolates showing stronger genetic relationship with strains from Europe than ST198-1.

Among the 68 isolates, 31 acquired AMR genes were detected, along with a single-point mutation in gyrA and parC (Figure 1). All isolates harbored multiple aminoglycoside resistance genes. Four bla_CTX-M subtypes were identified among the 66 ESBL-positive Salmonella isolates, including bla_CTX-M-14b (n = 9), bla_CTX-M-55 (n = 55), bla_CTX-M-64 (n = 1), and bla_CTX-M-27 (n = 1). The bla_TEM subtypes encoding resistance to ampicillin were also detected among the ST198 isolates, including bla_TEM-1B (n = 15) and bla_TEM-141 (n = 40). For fluoroquinolone resistance, all the isolates carried multiple single-point mutations in gyrA and parC (GyrA S83F, ParC T57S, and ParC S80I). Additionally, different single-point mutations were detected in gyrA including D87N (n = 59) and D87G (n = 9) along with the plasmid-mediated quinolone resistance (PMQR) gene, including qnrS1 (n = 19). Alleles of dihydrofolate reductase (dfr)-encoding genes that endowed resistance to trimethoprim were noticed in 54 isolates, comprising two different variants, dfrA14 (n = 53) and dfrA1 (n = 1). In addition, six types of acquired AMR genes were also present in the majority of the isolates. These included tet(A) (n = 67, encoding resistance to tetracycline), sul1 (n = 67, sulphonamides), floR (n = 58, phenicols), arr-3 (n = 53, rifamycin), Inu(F) (n = 51, lincomycin), and mph(A) (n = 48, azithromycin). Streptomycin resistant genes strA and strB were detected in 11 ST198 isolates. Importantly, the emerging mobile colistin resistance gene mcr-1 was also detected in two isolates.

The distribution of resistance genes clearly differed between ST198-1 and ST198-2 isolates (Figure 1). Acquired AMR genes conferring resistance to fluoroquinolone (qnrS1), third/fourth generation cephalosporins (bla_CTX-M-55, bla_CTX-M-64, and bla_CTX-M-27), ampicillin (bla_TEM-1B, and bla_TEM-141), aminoglycosides (aac(3)-IId, aac(3)-IVa, aadA17, ant(3″)-Ia, aph(4)-Ia, and rmtB), rifamycin (arr-3), trimethoprim (dfrA14), sulphonamides (sul2), lincomycin (ermB and Inu(F)), phenicols (floR), fosfomycin (fosA3), and colistin (mcr-1) were only detected in ST198-2 isolates. In contrast, resistance genes for third/fourth generation cephalosporins (bla_CTX-M-14b), aminoglycosides (aacA5), and trimethoprim (dfrA1) were restricted to ST198-1 isolates (Supplementary Table S6). Genes strA and strB encoding streptomycin resistance were more predominant in ST198-1. For multiple single-point mutations detected in gyrA, different amino acid substitutions were more common in each clade; GyrA D87G was predominant in ST198-1, while GyrA D87N was predominant in ST198-2.

The distribution of bla_CTX-M subtypes genes differed between isolates in ST198-1 and ST198-2 (Figure 1). The ST198-1 strains carried bla_CTX-M-14b, while ST198-2 strains carried bla_CTX-M-55, bla_CTX-M-27, and bla_CTX-M-64. To further confirm the genetic context of resistance genes, we determined the complete genomes of five representative Chinese isolates from the two clades by long-read sequencing and obtained complete and circular genome sequences. The architecture of the antibiotic resistance complement was complex, with multiple locations in the bacterial genome harboring antibiotic resistance genes. Further analysis showed that bla_CTX-M-14b, bla_CTX-M-55, and bla_CTX-M-64 genes were chromosomally located on different regions in the chromosome whereas bla_CTX-M-27 bounded by IS903B was detected in a p0111 plasmid (denoted as pHA22057-1) (Figure 3). In ST198-1 isolates, genomic alignment showed that bla_CTX-M-14b was located on a 2848-bp translocatable unit mediated by ISECP1, inserted into the coding gene clusters of the type VI secretion system (T6SS) in the chromosome (Figure 3A). The 2848-bp unit was absent in the chromosome among ST198-2 isolates (non-bla_CTX-M-14b). Among ST198-2 isolates, both bla_CTX-M-55 and bla_CTX-M-64 genes were detected in a MDR region (Figure 3B). The MDR region was inserted into bcfH gene at downstream of bcfABCDEFG, disrupting bcfH. This MDR region, absent in the chromosome among ST198-1 isolates, carried multiple resistance genes (qnrS1, bla_TEM-1B, bla_TEM-141, aac(3)-IId, rmtB, arr-3, dfrA14, Inu(F), floR, and fosA3) detected in ST198-2 isolates. The molecular structures and resistance genes of MDR regions were different due to the plasticity of the MDR region whereas the sequence comparison indicated similar backbone structures. The MDR chromosomal regions were bilaterally flanked by IS26 containing multiple insertion sequences that may contribute to the spread of antibiotic resistance genes. These MDR regions harbored 10–14 resistance genes, including bla_CTX-M-55/bla_CTX-M-64, bla_TEM-141, bla_TEM-1B, aph(3′)-Ia, ant(3″)-Ia, aac(3)-IId, fosA, rmtB, lnu(F), mph(A), dfrA14, arr-3, qnrS1, and floR. In this study, bla_CTX-M-27 was the only plasmid-borne bla_CTX-M subtype identified among the ST198 isolates from China. This gene, mediated by IS903B, was carried on a 103,445-bp plasmid belonging to the p0111 type (designated pHA22057-1) (Figure 3C).

Two slightly different SGI1-K variants (type I and II) were detected in S. Kentucky strains in this study (Figure 4). Additionally, type I and type II were detected in clades ST198-1 and ST198-2, respectively. The variant type I of SGI1-K contains several deletions compared to the prototype. The deleted regions include S026, resG, the tnpR gene of Tn5393, a bla_TEM-1b-bearing Tn2, and ΔS044. One IS26 composite-like transposonal module were inserted in downstream of S025 gene, consisting of IS26-[aph(3′)-Ia]-IS26. Compared to the prototype SGI1-K, ISEc78 was inserted into the backbone gene traG, and a large absence of backbone genes were detected in the SGI1-K variant type II. The MDR region of SGI1-K variants type II was bounded by two copies of IS26 with opposite orientation. Regarding the repertoire of antimicrobial resistance genes, the variant type I maintained strAB, tet(A), sul1, aadA7, and aacCA5, whereas the variant type II carried only tet(A), sul1, and aadA7.

Plasmid replicons belonging to 15 different plasmid types were detected in 68 Chinese S. Kentucky isolates. ColRNAI-type (n = 63) plasmid was the most predominant one, followed by Col156 (n = 44) and Col440I (n = 39). Eight different plasmid incompatibility types (IncI1, IncFIB, IncX1, IncR, IncY, and IncI2) were also identified among 29 isolates. In addition to the a 103445-bp p0111 plasmid (Figure 3C) harboring bla_CTX-M-27 gene mentioned above, three plasmid incompatibility types (IncI1, IncY, and IncI2) carrying additional AMR genes were identified among the complete genomes of five representative Chinese isolates. One 103079-bp IncI1 plasmid harbored ermB gene conferring resistance to lincomycin. One 97831-bp IncY plasmid was found to carry a 4,748-bp transposon Tn6330(ISApl1-mcr-1-pap2-ISApl1), whereas another mcr-1 gene was found in a 60793-bp IncI2 plasmid.