Introduction

Enteric fever, including typhoid and paratyphoid fever, caused by Salmonella enterica serovar Typhi (S. Typhi) and Paratyphi (S. Paratyphi) A, B and C, is a global health problem. Typhoid and paratyphoid fever are transmitted primarily by the fecal-oral route and result in a variety of symptoms, including gradual onset of sustained fever, chills, nausea, rash, anorexia, abdominal pain and hepatosplenomegaly1. Despite the improvement in personal hygiene provision of clean water and sanitation systems, the global burden of typhoid and paratyphoid fever remains considerable. Each year, an estimated 11.9–20.6 million cases of typhoid fever occur in developing countries and cause approximately 129,000–223,000 deaths, with the majority occurring in South Asia2.

Antimicrobial therapy is the mainstay of treatment for enteric fever at present. Without receiving appropriate therapy, enteric fever has a case-fatality rate of 10–30%, whilst the percentage will fall to 1–4% when it is managed properly3. However, multi-drug resistant (MDR) Salmonella strains emerged and spread globally in the 1970s and 1980s due to the frequent use of traditional antimicrobials such as chloramphenicol, ampicillin and co-trimoxazole4,5, and since then fluoroquinolones have been regarded as the preferred drugs for the clinical treatment of typhoid fever6. Alarmingly, after the widespread and indiscriminate use of quinolones, cases of Salmonella with resistance to nalidixic acid and decreased susceptibility to fluoroquinolones have been increasingly reported in several countries7,8,9. The emergence of resistance to quinolones in Salmonella making the treatment of enteric fever more difficult or resulting in treatment failure.

Acquired quinolone resistance were attributable to mutations in quinolone resistance-determining region (QRDR) of DNA gyrase and topoisomerase IV, whose subunits are encoded respectively by gyrA, gyrB, parC, and parE genes10,11,12. In gram-negative bacteria such as Salmonella, the primary target of quinolones is gyrA sub-unit of DNA gyrase13,14. Meanwhile, plasmid-mediated resistance (PMQR) determinants qnr genes and aac (6)-Ib-cr4 have also been associated with resistance to quinolone15.

Recently, more studies have been performed to analyze the antimicrobial susceptibility and molecular characterization of S. Typhi and S. Paratyphi, but few were performed in Jiangsu Province. The objective of the present study was to evaluate the prevalence of S. Typhi and S. Paratyphi resistance to quinolones and characterize the underlying mechanism in Jiangsu, China between 2013 and 2017.

Materials and Methods

Specimen collection and isolate identification

This was a retrospective study of archived isolates (from stool samples) biobanked at Jiangsu Provincial Center for Disease Prevention and Control, China, between 2013 and 2017. API 20E test strips (bioMerieuxVitek, Marcy-l′Etoile, France) were used to confirm the identity of the isolates. All the isolates were then serotyped by slide agglutination with commercial antiserum (Tianrun Bio-Pharmaceutical Co., Ltd., China) according to the Kauffmann-White scheme (WHO, 2011).

Ethical considerations

The study protocol was approved by the ethics committee of the Chinese Centre for Disease Control and Prevention (CCDC) and all experiments were performed in accordance with relevant guidelines and regulations. No informed consent was obtained from the patients because the study was retrospective. Informed consent has been waived off by the ethics committee of the Jiangsu Provincial Center for Disease Control and Prevention.

Antimicrobial susceptibility testing

Antimicrobial susceptibility testing was performed using Kirby-Bauer disc diffusion system on Mueller-Hinton agar in compliance with the recommendations of the Clinical and Laboratory Standards Institute (CLSI) guidelines (2016) by using the following agents: ampicillin, chloramphenicol, trimethoprim-sulfamethoxazole, nalidixic acid, ofloxacin, ciprofloxacin, levofloxacin and ceftriaxone. E. coli ATCC 25922 was used as the reference strain.

PCR amplification and DNA sequencing

The polymerase chain reaction (PCR) assays of QRDR of gyrA, gyrB, parC and parE genes were performed on the nalidixic acid-resistant (NAR) or intermediate resistant isolates. All S. Typhi and S. Paratyphi isolates were analyzed using PCR assays for the presence of PMQR determinants of qnrA, qnrB, qnrC, qnrD, qnrS, qepA and aac(6)-Ib-cr. The PCR amplifications were performed using the primers shown in Table 1. Purified PCR products were sequenced by the Sanger Biotech. Sequence data were then analyzed by Bioeditor and analysed by comparison with sequences obtained from NCBI GenBank.

Table 1 Primers for PCR detection of antimicrobial resistance determinants.

Results

Distribution of S. Typhi and S. Paratyphi isolates

Between January 2013 to December 2017, altogether 239 Salmonella isolates were collected from Jiangsu, China; 106 (48.6%) were male and 112 (51.4%) were female (21 cases unknown). The age of the patients ranged from 1 month to 88 years. Among 239 Salmonella strains, there are 164 strains of S. Typhi, 44 strains of S. Paratyphi A, 30 strains of S. Paratyphi B and 1 strain of S. Paratyphi C. The demographic characteristics of S. Typhi and S. Paratyphi isolates are shown in Fig. 1.

Figure 1
figure 1

Distribution of Salmonella enterica serovar Typhi and Salmonella enterica serovar Paratyphi (A–C) isolates in patients by (A) number (B) gender (C) age and(D) months (2013–2017) in Jiangsu, China.

During 2013–2015, enteric fever was caused mostly by S. Typhi (n = 83; 80.6%) and only 20 (19.4%) cases of S. Paratyphi (A and B) infection were identified. A double increase in the number of S. Paratyphi cases were observed with a total of 55 (40.4%) cases between 2016 and 2017, during which the number of S. Typhi cases remained relatively stable (Fig. 1).

Antimicrobial susceptibility testing

As shown in Table 2, except for the 50% resistance rate to nalidixic acid, the other 10 antibacterial agents showed a relatively low resistance rate to S. Typhi and S. Paratyphi (both < 20%). The resistance to quinolones varied by subspecies. The dominating serotype resistant to nalidixic acid was S. Typhi (n = 94; 73.4%). Although only eleven Salmonella isolates showed resistance to ciprofloxacin (including 4 S. Typhi, 6 S. Paratyphi A and 1 S. Paratyphi A), intermediately resistant to ciprofloxacin were up to 66 (27.6%). The proportion of Salmonella isolates with susceptibility to ciprofloxacin was 100% in 2013 and down to 34.4% in 2015.

Table 2 Antimicrobial sensitivity patterns of S. Typhi and S. Paratyphi A, B and C isolates in Jiangsu Province, China, between 2013 and 2017.

MDR was observed in 21 isolates of the Salmonella, in which 12, 8, 2 and 1 strains were found in S. paratyphoid B, S. Typhi, S. Paratyphi A and S. Paratyphi C respectively. Interestingly, both S. Typhi and S. paratyphoid A isolates were highly susceptible to gentamicin, ceftazidime, cefotaxime and amoxicillin-clavulanic acid. On the contrary, S. Paratyphi B was severely resistant to 9 antimicrobials except for gentamicin and ciprofloxacin (> 20%). In addition, a case of serotype Paratyphi C was detected in 2016 which was susceptible to nalidixic acid but intermediate to ciprofloxacin (Table 2).

Identification of quinolone resistance-encoding genes in Salmonella Typhi

Number of different cases of major mutations among Salmonella are shown in Fig. 2. The mutations of QRDR were examined in 94 nalidixic acid–resistant S. Typhi strains. There were 69 (73.4%) isolates containing one or more mutations encoding at least. The most common target mutation identified was gyrA: E133G, accounting for 88.4%, followed by mutation at codon 83 in the gyrA gene (S83F/S83Y, 65.2%). 22 S. Typhi isolates harbored gyrA mutation at amino acid position 87, leading to aspartic acid replaced by asparagine or glycine. In the gyrB, mutation was only identified in S426G which was first reported in S. Typhi. Additionally, parC gene mutations were detected in 4 S. Typhi isolates including E84K (n = 1) and novel mutation D79G (n = 3). Four other novel mutations were found in parE in 7 isolates of S. Typhi (including E543K, V560G, I444S, Y434S). Particularly, all the mutations of gyrB, parC and parE were detected with the concomitant presence of E133G in gyrA. The correlation of quinolones and nucleotide changes within the QRDR is shown in Table 3.

Figure 2
figure 2

Number of different cases of major mutations among Salmonella enterica serovar Typhi and Salmonella enterica serovar Paratyphi isolates.

Table 3 Combinations of quinolone phenotypes and genotypes identified in Salmonella Typhi and Salmonella Paratyphi A and B in Jiangsu, China between 2013 and 2017.

In addition, qnrB4 (n = 1), qnrS1 (n = 1) and aac (6)-Ib-cr4 (n = 2), which were PMQR determinants, were detected in four S. Typhi isolates. Two aac (6)-Ib-cr4 positive isolates coexisted with the gyrA: D87N mutation, one of which was MDR and resistant to ciprofloxacin.

Identification of quinolone resistance-encoding genes in Salmonella Paratyphi A, B and C

Out of the 34 strains of Salmonella Paratyphi with resistance to nalidixic acid, high levels of mutations in QRDR were observed in Salmonella Paratyphi A (80%) and B (88.9%). The most common mutations in QRDR were detected in gyrA: S83F in S. paratyphoid A (n = 18) and gyrA: S83Y in S. paratyphoid B (n = 5) (Fig. 2). All eight isolates which carried parC: T57S mutation were S. Paratyphi A accompanied with the presence of S83F in gyrA. Among them, susceptible to ciprofloxacin was found in 7 strains, except one isolate carrying additional mutations in gyrA: D87N and parC: E84K. Only 2 point mutations were observed in parE, I444S (S. Paratyphi A) and S498T (S. Paratyphi B), which were novel. All parC and parE were detected with the concomitant presence of S83F in gyrA.

PMQR genes in S. Paratyphi A and S. Paratyphi C were only found in aac (6)-Ib-cr4 and qnrS1, respectively, and both of them were MDR. Among all 30 S. Paratyphi B isolates, 11 strains contained PMQR genes including qnrS1 (n = 7), aac (6)-Ib-cr4 (n = 4), qnrB4 (n = 2), qnrA1 (n = 2) and qnrD1 (n = 1). None of the isolates were positive for qnrC and qepA. The five strains carrying the aac (6)-Ib-cr4 gene were all MDR and had reduced susceptibility to quinolones. In addition, 5 strains of S. Paratyphi B with a single point mutation in qnrS1 were susceptible to nalidixic acid and ciprofloxacin.

Discussion

In this study, we explored the correlation between the level of resistance and the associated mechanism in S. Typhi and S. Paratyphi isolates obtained from Jiangsu Province of China during 2013–2017. Among the 239 isolates, 164 (68.9%) were S. Typhi and 75 (31.4%) were S. Paratyphi (including 44 S. Paratyphi A, 30 S. Paratyphi B and 1 S. Paratyphi C) showing a predominance of S. Typhi over S. Paratyphi, which is consistent with the previous studies16,17,18,19. However, the prevalence of S. Paratyphi collected between 2016 and 2017 is much higher than that between 2013 and 2015, showing an increasing trend and a change in species of Salmonella in Jiangsu Province of China. Overall, many regions in the world have already seen the increase in the prevalence of S. Paratyphi16,20,21. The reason for this change may be related to the improvements in environmental conditions and the increased use of Vi polysaccharide vaccine in recent years22.

The continued increase in the resistance of Salmonella to quinolones is a global problem, especially in Asian countries including India, Pakistan and Bangladesh23. The S. Typhi and S. Paratyphi isolates in this study showed a high resistance rate (53.8%) against nalidixic acid, similar to previous reports in Jiangsu Province24. Given the emergence of resistance to nalidixic acid for Salmonella, ciprofloxacin has been developed as a new quinolone and proved to be a highly effective treatment alternative. However, Salmonella resistance to ciprofloxacin has also become a troublesome problem due to continued abuse of quinolones in patients with diarrhea in local communities without considering the cause of diarrhea. In our study, all Salmonella were susceptible to ciprofloxacin in 2013, but by 2015, only 34.4% remained susceptible. In these five years, a total of 77 strains have reduced susceptibility to ciprofloxacin, which was associated with clinical failure25.

In the current study, we have analysed the sequence of gyrA, gyrB, parC, and parE genes in all the Salmonella isolates with resistant or intermediate to nalidixic acid. In general, the majority of mutations were found in three highly prevalent codons at 83, 87 and 133 in gyrA. Mutations at codon 83 and 87 in Salmonella isolates associated with reduced susceptibility to quinolones have been described previously26,27,28,29. Nevertheless, what noteworthy is that the most frequent mutation in S. Typhi was observed at codon 133 in gyrA, outside the usual QRDR amino acid mutations between 67 and 106. Of the 70 isolates of S. Typhi containing at least one mutation in QRDR, 57(81.4%) shared a common mutation in gyrA: E133G and only 2 isolates showed a single point mutation in gyrA: E133G. According to previous experiments performed in Salmonella, mutation at codon 133 alone could not necessarily lead to quinolone resistance, unless combination with either a second mutation in the same gene at codon 83 or 8730,31,32.

In our study, all S. Paratyphi A with parC: T57S and gyrA: S83F were susceptible to ciprofloxacin. The mere combination of these double mutations does not appear to reduce the susceptibility of ciprofloxacin. The data corroborated the previous demonstration that parC: T57S is a spontaneous compensatory mutation that makes Salmonella resistant to nalidixic acid but more sensitive to ciprofloxacin33,34. However, when additional mutation gyrA: D87N and parC: E84K were carried with mutations in parC: T57S and gyrA: S83F, it turned to resistant. Another strain harbored the mutation parC: E84K in S. Typhi also showed resistance. Accumulation of mutations at codon 83 and 87 in gyrA and simultaneous mutation of parC: E84K may be associated with full resistance to ciprofloxacin35,36.

Quinolone resistance in Salmonella is mainly mediated by mutations in gyrA and parC19, with few reported cases of gyrB and parE37. However, 8 isolates were detected containing mutations in gyrB: S426G, and parE: [S498T, E543K, V560G, I444S, Y434S] in the current study. To our knowledge, these mutations were the first reports in Salmonella. Interestingly, almost all Salmonella isolates with novel mutations were simultaneously detected common mutations in gyrA: [S83F and E133G], except one strain that had single E133G in gyrA. Five novel mutations including gyrB: S426G, parC: D79G and parE: [S498T, E543K, V560G and Y434S] showed susceptibility to ciprofloxacin. It showed a very strong correlation between novel mutations and mutations at codon 83 and 133. We speculate that the current mutations at 83,133 may induce the generation of other novel mutations, making the mutation genotype more complicated. However, whether these novel mutations are related to quinolone resistance remains to be confirmed by further experiments. In addition, 5 strains of S. Typhi with the triple mutation combination gyrA: [S83F, E133G] and parE: I444S were found intermediate to ciprofloxacin. This suggests that the novel mutation at codon 444 of parE gene might attenuate the activity of ciprofloxacin. The change at codon 444 could indirectly affect the combination of topoisomerase IV and fluoroquinolone, thus decreasing drug sensitivity38. Overall, the presence of new mutations among S. Typhi and S. Paratyphi isolates increases the complexity of quinolone-resistant genotypes and raises a critical warning to the prevalence of resistance.

Since the PMQR determinants were initially identified in 1998, various PMQR genes have been constantly detected in Salmonella across the world39,40,41. In this study, we found that the prevalence of PMQR genes was 7.1% in a collection of 239 S. Typhi and S. Paratyphi isolates. Overall, PMQR genes were detected with lower frequencies. Among them, the qnrS1 was the most prevalent gene, followed by aac (6)-Ib-cr4 and the qnrB4, qnrA1 and qnrD1, which is consistent with previous reports42,43,44. PMQR genes alone in S. Typhi and S. Paratyphi could not lead to quinolone resistance, unless they co-existed with gyrA QRDR mutations18. In the current study, we found that 2 strains contained PMQR genes and mutations at codons 83 or 87 had reduced susceptibility to ciprofloxacin. Additionally, almost all Salmonella carrying the aac (6 ′)-Ib-cr4 gene described in this report were MDR, indicating that aac (6′)-Ib-cr4 genes could easily lead to MDR phenotype.

Although S. Typhi and S. Paratyphi shared similarities in many aspects such as resistance and mechanisms, there are still differences between them, which is more obvious in S. Paratyphi B45. In the dataset described herein, antibiotic resistance in S. Paratyphi B isolates was more severe. The resistance rate of S. Paratyphi B to the tested antibiotics other than nalidixic acid and ciprofloxacin was 3–30 times that of S. Typhi. This suggests that the dispersed clone of S. Paratyphi B may has unique mechanisms. In the present study, PMQR genes were mainly detected in S. Paratyphi B, accounting for 64.7% and all S. Paratyphi B resistant to nalidixic acids can detect mutations at QRDR or PMQR. As a result, the ratio of MDR in S. Paratyphi B was about 10 times that of S. Paratyphi A and S. Typhi. This means that typhoid fever caused by S. Paratyphi B is more difficult to treat clinically, so typhoid and paratyphoid fever should be considered as different diseases5,46.

In conclusion, this study described antimicrobial resistance and the mechanisms of S. Typhi and S. Paratyphi strains isolated between 2013 and 2017 in Jiangsu Province of China. Because of the increasing prevalence of S. Typhi and S. Paratyphi isolates and indiscriminate use of antimicrobials, the rise in quinolone-resistant Salmonella strains and the spread of quinolone resistance-encoding genes are extremely worrying. Moreover, the emergence of novel mutations in gyrB, parC and parE genes increases the complexity of quinolone-resistant genotypes and poses a threat to public health. To better manage and prevent the spread of antimicrobial resistance, it is necessary to provide clinicians and local governments with accurate epidemiological information concerned. Further research and continuous dynamic monitoring of antibiotic susceptibility would be useful in the treatment and control of this infection.