Antimicrobial Resistance and Genomic Characterization of Salmonella Serovars Typhimurium and 4,[5],12:i:- in Huzhou, China

Introduction

Salmonella is a diverse genus comprising over 2600 serovars that can cause a range of infections. It is estimated to cause more than 300,000 deaths annually, mostly in developing countries.^1,2 Based on different pathogenic and disease manifestations, the serotypes of Salmonella can be classified into typhoidal and non-typhoidal (NTS). NTS have a broad host range and in humans, most often cause self-limiting gastroenteritis that is typically acquired through the consumption of contaminated food or water.^3,4 The incidence of NTS was estimated at 627 cases per 100,000 persons in China.⁵ Of all NTS, Salmonella serovar Typhimurium (S. Typhimurium, 1,4,[5],12:i:1,2) is reported among the most common serovar linked with human diseases worldwide.^6,7 S. Typhimurium infections consistently exhibit the highest incidence rate among all salmonellosis cases in China, maintaining its position as the predominant serotype.⁸

Salmonella 4,[5],12:i:- is considered a monophasic variant of S. Typhimurium, lacking the ability to express the second-phase flagellar antigen,⁹ and has become a major global cause of NTS diseases in human during the past two decades.^10–12 After identification in the mid-1990s,¹³ Salmonella 4,[5],12:i:- has been reported in various countries at different times,^9,14 and some reports suggest that it is the global pandemic clone.^15,16 According to European studies, Salmonella 4,[5],12:i:- has become one of the top five serovars among human clinical isolates.¹⁷ In 2015, the detection rate of Salmonella 4,[5],12:i:- surpassed that of other serotypes, establishing it as the predominant serotype among human Salmonella isolates in Guangdong Province, China.¹⁸ This trend was also observed in Huzhou.

The prevalence of antibiotic-resistant strains, which pose serious risks to the public’s health, is a serious challenge. Because of the indiscriminate use of antibiotics for both clinical purposes and food production, antimicrobial resistance (AMR) has become increasingly common.¹⁹ S. Typhimurium and Salmonella 4,[5],12:i:- have been identified as exhibiting a high prevalence of multidrug resistance (MDR) among various Salmonella serovars, raising considerable public health issues worldwide.^20,21 Multidrug-resistant strains show resistance to a wide range of clinical antimicrobial agents, such as third-generation cephalosporins and colistin,²² making clinical treatment difficult and necessitating the study of drug resistance in these two Salmonella serovars.²³

Because whole genome sequencing (WGS) has reduced in price, it is replacing prior conventional technologies for investigating disease outbreaks and public health surveillance.^24–26 WGS could generate vast amounts of genetic data rapidly for species identification, determining virulence and resistance characteristics, and phylogenetic analyses.^27,28 It has been instrumental in improving identification of public health pathogens.²⁹

To provide improved knowledge regarding S. Typhimurium and Salmonella 4,[5],12:i:- infections, we carried out this retrospective analysis to assess the AMR and genomic characterization of 90 S. Typhimurium and Salmonella 4,[5],12:i:- isolates recovered from hospitalized patients and food in Huzhou between 2021 and 2023. This study aimed to explore AMR, distribution characteristics of resistance genes and virulence genes, and the evolutionary relationships of isolates. Our findings could provide valuable insights for the prevention and control of MDR in S. Typhimurium and Salmonella 4,[5],12:i:-.

Materials and Methods

Isolate Collection, Salmonella Identification and Serotyping

Between 2021 and 2023, a total of 80 S. Typhimurium and S. 4,[5],12:i:- isolates were collected from six sentinel hospitals of the Chinese Pathogen Identification Net of Huzhou, as well as 10 strains isolated from retail food in wet markets (raw animal meat, raw poultry meat and pre-made dishes). Sample collection from patients were conducted year-round, whereas food sampling was restricted to the period from March to October annually.

Human origin strains were isolated from hospitals, and food strains were isolated by our laboratory according to the National Food Safety Standard (GB 4789.4–2016) rules.³⁰ After suspicious colonies were isolated by Columbia Blood Agar Plates (Chromagar, France), colonies were identified using matrix-assisted laser desorption ionization time-of-flight (MALDITOF) mass spectrometry (bioMérieux, France). The strains were stored in glycerol broth at –70°C.

Serotypes of Salmonella isolates were determined by slide agglutination for somatic antigen O and flagellar antigens Husing commercially available antiserum (Denka Seiken, Japan) according to the Kauffmann–White–LeMinor Scheme,¹ with normal saline used as a negative control.

Antimicrobial Susceptibility Testing

The broth micro-dilution method was used to determine the minimum inhibitory concentrations (MICs) of 17 antimicrobial drugs for 90 isolates and breakpoints were interpreted according to Clinical and Laboratory Standards Institute (CLSI) protocols.³¹ The following antimicrobials (Thermo, USA) were tested: ampicillin (AMP), ampicillin/sulbactam (AMS), ceftazidime (CAZ), cefotaxime (CTX), ertapenem (ETP), meropenem (MEM), nalidixic acid (NAL), ciprofloxacin (CIP), azithromycin (AZM), amikacin (AMI), streptomycin (STR), tetracycline (TET), tigecycline (TIG), chloramphenicol (CHL), trimethoprim-sulfamethoxazole (SXT), and colistin (CT). Escherichia coli ATCC 25922 was used as the control strain for susceptibility testing.

Whole-Genome Sequencing

Salmonella DNA was extracted from overnight cultures using QIAamp DNA Mini Kits (Qiagen, Germany), following the manufacturer’s instructions. DNA concentration was tested by the Qubit 4 (Thermo, USA). Libraries were constructed using a Metagenomic DNA Library Kit (Matridx Biotechnology, China) and sequenced by the Next Seq 550 high Output Reagent Cartridge (Illumina, USA).

Genomic Analysis

Quality control analysis of raw reads was performed with Fast QC (0.11.9) and fastp (0.23.2) was used to remove low-quality data. Reads were assembled using SPAdes (3.15.4). The serotype of isolates from the Kauffmann–White scheme was confirmed by SeqSero2 (1.1.1). Based on the results of WGS analysis, the CARD (https://card.mcmaster.ca/) and VFDB (https://www.mgc.ac.cn/VFs/) databases were used to predict the resistance genes and virulence genes by the parameter identity of >99.0% and coverage of >99.0%. The core single-nucleotide polymorphism (SNP) for 90 isolates was determined by Snippy (4.4.5) using the reference sequence S. Typhimurium LT2 (GenBank BioSample ID SAMN02604315). FastTree (2.1.11) was used for sequence alignment and homology analysis, and Gubbins was used for reassembly. The phylogenetic tree and the heatmap of genes were visualized using tvBOT.³² The genomes of the 90 isolates have been deposited in the National Center for Biotechnology Information (NCBI) database under the Bioproject number PRJNA1195539.

Statistical Analysis

The data were analyzed using SPSS (Statistical Package for the Social Sciences) 23.0 (SPSS, Chicago, USA). Differences in Salmonella infection rates by sex and age were analyzed using the chi-square test. P-values below 0.05 were considered significant.

Results

Description of S. Typhimurium and Salmonella 4,[5],12:i:-

A total of 30 S. Typhimurium and 60 monophasic variant Salmonella 4,[5],12:i:-isolates were recovered between 2021 and 2023, including 80 from clinical samples of patients with diarrhea and 10 from food samples (Table S1). Of the 80 patients infected with Salmonella, 32 (40.0%) were aged between 0 and 5 years (Figure 1a), and the detection rate in the older age group was >60 was 15.0%. No significant association in the incidence of salmonellosis was noted between sex and age group (p >0.05).

Figure 1 Epidemiological information regarding the 90 isolates in this study: (a) number of isolates in each age group; (b) composition ratios of serotypes in each age group. (c) composition ratios of ST types in each age group; and (d) composition ratios of sex in each age group.

The distribution of serotypes in different age groups was also different, and the main serotypes in patients under 60 years of age were S.1,4,[5],12:i:-, while S. Typhimurium were predominant in those over 60 years of age (Figure 1b).

Multilocus sequence typing (MLST) analysis generated three STs from 90 isolates. ST34 was the most frequent sequence type (67.8%, 61/90), followed by ST19 (30.0%, 27/90) and ST1544 (2.2%, 2/90). The proportion of the ST types also differed between the two serotypes, the isolates of Salmonella 4,[5],12:i:- were all ST34. The distribution of ST types varied across different age groups: ST34 was the predominant ST type in patients under 60 years of age, while ST19 was the primary ST type in those over 60 (Figure 1c).

The frequency and severity of S. Typhimurium infections varied between males and females. In this study, we observed a higher incidence of S. Typhimurium infections among male patients below 30 and over 60 years of age. Females aged 30–60 years exhibited higher susceptibility (58.3%, 14/24) to S. Typhimurium infection (Figure 1d).

Phenotypic AMR Among S. Typhimurium and Salmonella 4,[5],12:i:-

Antibiotic susceptibility tests were performed for the 90 isolates by 17 antimicrobials (Table 1 and Table S2). The highest recorded resistance rates were observed for tetracycline (85/90, 94.4%), with ampicillin (65/90, 72.2%) and trimethoprim/sulfamethoxazole (63/90, 70.0%). Conversely, the resistance rates for ceftazidime (0/90, 0.0%), amikacin (1/90, 1.1%) and meropenem (2/90, 2.2%) were relatively low. Only two isolates (2/90, 2.2%) demonstrated susceptibility to all 17 antibiotics tested. When analyzing antimicrobial resistance patterns by serovar, Salmonella 4,[5],12:i:- demonstrated significantly elevated resistance rates to tetracycline (83.3%), ampicillin (80.0%), streptomycin (63.3%), chloramphenicol (63.3%), and trimethoprim-sulfamethoxazole (63.3%). For S. Typhimurium, there were some differences: the highest resistance rates were to tetracycline (60/60, 100.0%), trimethoprim/sulfamethoxazole (44/60, 73.3%), and streptomycin (40/60, 70.0%). Source of isolate resulted in different resistance rates: however, the highest resistance rate for both was to tetracycline (patients: 93.8%; food: 100.0%).

Table 1 Antimicrobial Resistant Rate of S. Typhimurium and Salmonella 4,[5],12:i:-Isolates Against 17 Antimicrobials

Eighty-four (93.3%) isolates showed resistance to a minimum of three antimicrobial classes and were chosen as demonstrating MDR. S. Typhimurium and Salmonella 4,[5],12:i:- demonstrated 86.7% and 96.7% MDR, respectively, revealing 47 multidrug-resistant spectra, among which NAL-TET-SXT, seen in 16 isolates, was predominant. The most resistant isolate was detected from a nine-month-old boy in 2023, which was resistant to 16 drugs examined (Table 2).

Table 2 ====Number and Percentage of Multidrug-resistant Profiles in S. Typhimurium and Salmonella 4,[5],12:i:-Isolates

AMR Genes

As shown in Figure 2, we identified 39 different AMR genes of 9 classes in the 90 Salmonella isolates, which included aminoglycoside, ampenicol, beta-lactam, fosfomycin, macrolide, quinolone, rifamycin, sulfonamides, and tetracycline. The aac(6’)-Iaa point mutation, linked to aminoglycoside resistance, was the most common (100.0% of isolates), followed by blaTEM-1 (30/90, 33.3%), tet(B) (24/90, 26.7%), and tetR (24/90, 26.7%). Some AMR genes (such as aac[3]-IId, aadA5, floR, blaCTX-M-14, fosA3, fosB6, mphA, qnrD1, qnrS2, oqxA, oqxB, dfrA17, and tet[K]) were identified in just one isolate. Each strain of S. Typhimurium carries an average of 4.5 AMR genes, while Salmonella 4,[5],12:i:- carries 3.5 AMR genes. Apart from aac(6’)-Iaa, the most prevalent AMR genes in Salmonella 4,[5],12:i:-were blaTEM-1 (19/60), followed by qnrS1 (15/60), tet(B) (15/60), and tetR (15/60). In contrast, the top three AMR genes carried by S. Typhimurium were blaTEM-1 (11/30), tet(B) (9/30) andtetR (9/30).

Figure 2 Annotation heatmap of AMR genes among 90 isolates.

Virulence Factors

Virulence analysis using the Virulence Factor Database (VFDB) identified 184 virulence genes across 90 S. Typhimurium and Salmonella 4,[5],12:i:- isolates. These genes were classified into eight categories based on their roles in pathogenesis. The eight categories included adherence (csgC, bcfA, lpfA, etc.), effector delivery system (tae4, hilC, orgC, etc.), exotoxin (spvB), immune modulation (rck), exotoxin (spvB), nutritional/metabolic factor (mgtB, mgtC, iroB, etc.), stress survival (sodCI), antimicrobial activity/competitive advantage (mig-14 and mig-5), and regulation (fur, phoP, rpoS, etc.). All detected virulence genes categorized by VF class are outlined in Figure 3.

Figure 3 Annotation heatmap of virulence factors among 90 isolates.

Phylogenetic Analysis

To assess genetic variation among these isolates, we conducted whole genome single nucleotide polymorphism (cgSNP) phylogenetic analysis of these 90 isolates. As shown in Figure 4, 90 isolates were classified into six major clades, designated A–F. There was no association between lineages and years, district, origin of isolates; however, the lineages were closely related to serotypes and epidemiology. Salmonella 4,[5],12:i:- was distributed primarily in clade E, which contains 16 strains from one outbreak, while S. Typhimurium isolates were found in the other five clades. The serotype of isolates from clades A to D, and clade F were all 1,4,[5],12:i:1,2. In contrast, 59 isolates from clade E were 1,4,[5],12:i:-, while one isolate was 1,4,[5],12:i:1,2. The ST types of clade A (2 strains), clade C (1 strain), clade D (14 strains) and clade F (2 strains) were ST19. There were two ST types in clade B, with two strains of ST1544 and eight strains of ST19. For clade E, all 60 strains were ST34.

Figure 4 Core genome SNP-based phylogenomic tree of all 90 Salmonella isolates.

Discussion

S. Typhimurium and monophasic Salmonella 4,[5],12:i:- are common causes of foodborne illness. With increasing antibiotic resistance, especially MDR, they have become more and more prevalent in the world, posing a serious public health problem. Salmonella 4,[5],12:i:- has become a global new epidemic clone associated with human food and environmental samples.^13,33 Similar to these studies, S. Typhimurium isolated from food was the top serotype in Huzhou;³⁴ however, the prevalence of S. Typhimurium and Salmonella 4,[5],12:i:- with human infections in Huzhou were unknown.

In our study, the younger age group (aged 0–5 years) was the main population infected by S. Typhimurium and Salmonella 4,[5],12:i:-, which was consistent with previous findings.¹⁰ The underdeveloped immune systems of children under 5 years old render them particularly susceptible to Salmonella infections, which may account for the disproportionately high incidence of severe gastroenteritis in this age group.³⁵ In terms of serotype, monophasic Salmonella 4,[5],12:i:- replaced S. Typhimurium as the most prevalent serotype of Salmonella.²³ ST34 and ST19 were the most frequent sequence type in our study, accounting for 97.8%, which was consistent with several previous studies in China.^36,37 Enhanced surveillance targeting S. Typhimurium sequence type 34 (ST34) was imperative to curb the dissemination of multidrug-resistant strains, given that ST34 isolates exhibit heightened adaptive capabilities.³⁸ The frequency and severity of Salmonella infections vary between men and women, and we observed that males below 30 and over 60 years of age and females 30–60 years of age were more vulnerable. The different incidences of Salmonella according to gender could be attributed to pathogen exposure, eating habits, genetic factors, and so on.³⁹

With the irrational clinical and veterinary use of antibiotics, the problem of resistance is becoming more and more serious.⁴⁰ The 90 isolates from different sources were tested against 17 antimicrobials, and 49 different patterns of resistance were recorded, confirming the wide diversity of resistance profiles in Salmonella strains. Cephalosporins and fluoroquinolones are considered the first treatment for severe salmonellosis;¹⁰ however, our findings indicate potential therapeutic failure: 35.6% of isolates were resistant to cefotaxime, and 56.7% were resistant to nalidixic acid, so that is a serious problem in Huzhou. High levels of resistance were observed to tetracycline, ampicillin, and trimethoprim/sulfamethoxazole, consistent with previous reports,^23,41,42 indicating that these antibiotics may be ineffective in the treatment of Salmonella infections.

Effectively treating the MDR of salmonellosis requires costly antibiotics, which is an added burden for developing countries.⁴³ We found that 100.0% and 92.5% of food and patient samples with diarrhea-associated isolates were multidrug resistant in this study, respectively. However, previous studies in Shenzhen showed higher MDR in clinical isolates (81.6%) than in food isolates (66.7%),⁴⁴ which indicated the diversity of multidrug-resistant isolates in different regions. Similarly, high percentages of MDR were also observed in Salmonella isolated from food⁴⁵ and patient⁴¹ samples in various regions of China. The surge in multidrug-resistant Salmonella isolates is recognized as a crucial public health issue,⁴⁶ which may lead to therapeutic failure, especially with the limited number of alternative antibiotics available.

Previous studies have reported some concordance between antimicrobial resistance genes and phenotypes;^47,48 however, they did not consider confounders in bacterial genome-wide associations such as population structure.⁴⁹ In this study, we found a wide distribution of the aminoglycoside-encoding gene aac(6’)-Iaa in all 90 isolates, which were not consistent with resistance phenotypes of aminoglycoside (67.8% were resistant to STR, 1.1% to AMI). aac(6’)-Iaa usually are transcriptionally silent and rarely become transcriptionally active, which means the mere presence of aac(6’)-Iaa can not confer aminoglycoside resistance to Salmonella.⁵⁰ Antimicrobial resistance genes were not usually sufficient to cause changes in resistance phenotypes unless isolates had mutations that could increase their expression.⁵¹ ESBLs (encoded by balCTX-M-14, balCTX-M-65, balTEM-1and balOXA-1) are modified broad-spectrumbeta-lactamases that hydrolyze beta-lactam antibiotics. We found a wide distribution of the ESBL-encoding gene balTEM-1 (33.3%) in these two serotypes. It has been reported that balTEM-1 is mostly linked with ampicillin resistance⁵² (72.2%). Resistance to tetracyclines and chloramphenicol is often associated with efflux pumps by tet and floR, while resistance to sulfonamides is associated with antibiotic target replacement mechanism by sul and dfrA.⁵³

Virulome analysis detected the typical genes implicated in the virulence and pathogenicity mechanisms of Salmonella. Virulence factors (VFs) can overcome host defense mechanisms and cause disease in a host.⁵⁴ The fimbrial genes (eg, bcf, lpf, fim) may influence isolates’ tropism to interact with host epithelial cells, and enhance ability to adhere to different host cell surfaces, aiding in the colonization and infection process.⁵⁵ In fact, 50.4% of VFs (93/184) belong to the effector delivery system, involving the Type VI secretion system (T6SS) and Type III secretion system (T3SS). With the exception offljB, the differences in VFs between S. Typhimurium and Salmonella 4,[5],12:i:- were spvB, rck and mig-5 (see Figure 3). These genes were detected in only some isolates of S. Typhimurium, while the overall presence rates of the other genes were not significantly different between the serovars.⁵⁶ The spv is associated with the survival and proliferation of Salmonella within macrophages.⁵⁷

Whole genome sequencing can reveal the complete DNA make-up of an organism and facilitate the detection of variations both within and between species. Phylogenomic analysis of 90 strains, which were categorized into six main clades (A to F), showed that these strains were closely related and clustered together with serovars. Sixteen strains from outbreaks were clustered in clade E. In addition, the isolate of S2021364 was clustered in clade E while its serotype was 1,4,[5],12:i:1,2; however, the serotype of the remaining 60 strains in clade E were 1,4,[5],12:i:-. Between S2021364 and S2023228(1,4,[5],12:i:-), there were 116 differences in SNP compared to S2023365(1,4,[5],12:i: 1,2), in which there were 468. The classification of S2021364 into clade E was driven by its distinct SNP, rather than the serotype. The clades of a phylogenetic tree did not exhibit particular geographic distributions, origin, or years of isolation.

Conclusion

In conclusion, the data presented here offer an overview of the distribution, antibiotic resistance profile and genome characteristics of S. Typhimurium and its monophasic variant (Salmonella 4,[5],12:i:-) strains isolated in Huzhou City, China. A high prevalence of multidrug-resistant S. Typhimurium and Salmonella 4,[5],12:i:- strains (93.3%) was found and these isolates showed a worrying resistance to TET, AMP and SXT, which are currently used as first-line treatment in human infections. A total of eight virulence factor categories were detected, with effector delivery system being the most abundant. The relationships between resistance phenotypes and resistance genes warrent further investigation. Hence, continuous monitoring of multidrug-resistant isolates using WGS is necessary for public health.