Incidence and Characterization of Salmonella Isolates From Raw Meat Products Sold at Small Markets in Hubei Province, China.

Salmonella is a leading cause of foodborne disease and is often associated with the consumption of foods of animal origin. In this study, sixty-six Salmonella isolates were obtained from 631 raw meat samples purchased at small retail suppliers in Hubei Province, China. The most prevalent Salmonella serotypes were Thompson (18.2%) and Agona (13.6%). Frequent antimicrobial resistance was observed for the sulfonamides (43.9%), tetracycline (43.9%), and the β-lactams amoxicillin and ampicillin (36.4% for each). Interestingly, a high incidence of resistance to cephazolin was observed in strains of the most common serotype, S. Thompson. Class I integrons were found in 27.3% (18/66) of the isolates and five of these integrons contained different gene cassettes (aacA4C-arr-3-dfr2, dfrA12-aadA21, aadA2, dfrA12-aadA2, dfr17-aadA5). Additional antimicrobial resistance genes, including bla TEM-1, bla CTX-M-65, bla CTX-M-15, qnrB, and qnrS, were also identified among these Salmonella isolates. Results of replicon typing and conjugation experiments revealed that an integron with qnrB and bla CTX-M-15 genes was present on incH12 mobile plasmid in S. Thompson strain. Multilocus sequence typing (MLST) analysis revealed 32 sequence types, indicating that these isolates were phenotypically and genetically diverse, among which ST26 (18.2%) and ST541 (12.1%) were the predominant sequence types. The integrons, along with multiple antimicrobial resistance genes on mobile plasmids, are likely contributors to the dissemination of multidrug resistance in Salmonella.

Introduction

Salmonella is an important source of foodborne disease, resulting in morbidity and mortality worldwide (Park et al., 2015). In China, 70–80% of bacterial food poisoning cases are due to Salmonella that originates in poultry, eggs, beef and pork (Yang et al., 2013). Direct and indirect spread of Salmonella between animals and humans are threats to human health. Although more than 2610 serotypes of Salmonella have been identified, the majority of human Salmonella infections are caused by strains of only a few serotypes (Mancin et al., 2015). Therefore, serotype determination is an important aspect of epidemiological surveillance and disease assessment (Hung et al., 2017). Changes in the prevalence of specific serotypes can result from the movements of people, animals, and foods (Hendriksen et al., 2009). At present, serotypes Enteritidis and Typhimurium have been most frequently implicated in salmonellosis outbreaks from foods in Taiwan, Greece, Qatar and South Africa (Chang et al., 2016; Papadopoulos et al., 2016; Hung et al., 2017; Smith et al., 2017).

Though not as facile or rapid as serotyping, multilocus sequence typing (MLST) is a reliable and highly discriminatory molecular typing method. In this technique, the phylogenetic relationship between bacterial strains is determined based upon nucleotide sequence variations within several housekeeping and other conserved genes (Chang et al., 2016). MLST has many advantages compared with other genotyping techniques (Stepan et al., 2011). It is consistent and reproducible, and combines nucleotide sequencing and bioinformatics with a population genetics approach to produce accurate, reliable and highly portable results (Smith et al., 2017). Furthermore, it has been suggested that MLST is superior to serotyping for Salmonella surveillance and outbreak tracking (Achtman et al., 2012).

The widespread use of antibiotics poses problems for antimicrobial resistance, which leads to an increase in treatment costs and even to therapy failure (Hur et al., 2012). Multiple drug resistance (MDR) among Salmonella is prevalent. This applies not only to traditional antimicrobials such as sulfamethoxazole and chloramphenicol, but also against clinically important antimicrobials like the third generation β-lactams and fluoroquinolones (Agada et al., 2014). The selective pressure caused by the application of antimicrobials in poultry production and veterinary practice for growth promotion and prophylaxis has resulted in an increase in antibiotic resistance and an increase in the presence of genes conferring antimicrobial resistance to Salmonella (Zishiri et al., 2016). The spread of antimicrobial resistance in Salmonella is primarily due to integrons (Rajaei et al., 2014). Integrons are mobile DNA elements containing multiple antimicrobial resistant genes, which can move from one bacterium to another. This increases multi-drug resistance in bacterial populations (Asgharpour et al., 2014). Among various integrons, the class I integron is frequently observed among antimicrobial resistant Salmonella and plays a major role in the dissemination of resistance genes (Wright, 2010).

To the best of our knowledge, there is limited information regarding the incidence of Salmonella in retail raw meat products in Central China. Therefore, the objective of this study was to determine the prevalence, serotype distribution, genetic diversity, antimicrobial resistance, and the presence of class I integrons of Salmonella from retail raw meat products in Hubei Province, China. This data can be used for quantitative risk assessments of Salmonella isolates in retail meat and to help establish a series of prevention and control measures against foodborne pathogenic bacteria.

Materials and Methods

Sample Collection and Salmonella Isolation and Serotyping

We collected 631 retail raw meat samples, including 263 chicken meat, 102 duck meat, 80 pork and 186 fish products from 20 small markets, in Wuhan City of Hubei Province, China, from 2014 to 2017. All samples were transported to our laboratory on ice within 6 h of collection. Salmonella isolation and identification methods were carried out according to the standard procedures of National Food Safety Standard Food Microbiological Examination: Salmonella (GB 4789.4-2016). Biochemical identification was done using an HBI microbial biochemical identification kit (Hope Biotechnology, Qingdao, China).

Serotyping of Salmonella isolates was done according to the Kauffmann-White scheme using an O and H antigen slide agglutination kit (Tianrun Biopharmaceutical, Ningbo, China) (Zhao et al., 2017).

Antimicrobial Susceptibility Testing

Antimicrobial susceptibility testing was performed using the Kirby-Bauer disk diffusion method as described by the Clinical and Laboratory Standards Institute (Clinical and Laboratory Standards Institute [CLSI], 2016). Antimicrobial disks (Oxoid Ltd., Basingstoke, United Kingdom) used for testing contained 10 μg amoxicillin, 10 μg ampicillin, 30 μg cefotaxime, 30 μg cefuroxime sodium, 30 μg cephazolin, 30 μg chloramphenicol, 5 μg ciprofloxacin, 30 μg nalidixic acid, 10 μg norfloxacin, 10 μg streptomycin, 10 μg gentamicin, 300 μg nitrofurantoin, 300 μg sulfonamide and 30 μg tetracycline. Escherichia coli strain ATCC 25922 was used as a reference strain. The standard break points were interpreted according to the guidelines of the Clinical and Laboratory Standards Institute (Clinical and Laboratory Standards Institute [CLSI], 2016).

Antimicrobial Resistance Genes and Class I Integron Identification

Bacterial DNA was extracted from Salmonella isolates cultured overnight in Luria–Bertani broth using the MiniBEST Bacteria Genomic DNA Extraction Kit Ver.3.0 (Takara, Beijing, China), according to the manufacturer’s instructions. Eighteen genes known to be associated with resistance to the antibiotics as well as Class I integrons were tested by PCR using the oligonucleotide primers listed in Table 1. The PCRs were carried out using a MyCyler Thermocycler (BioRad, Hercules, CA, United States) in a 25 μL mixture that contained 0.5 mM of each primer, 1 × PCR buffer (Takara), 250 μM of dNTP, 0.5 U of Taq DNA polymerase (Takara), 1.5 mM MgCl2, and 0.5 μL of template DNA. The cycling conditions were 94°C for 5 min, followed by 30 cycles of 94°C for 30 s, 54–61°C for 30 s and 72°C for 1 min, and a final extension of 72°C for 10 min. PCR amplification products were separated on agarose gels and the DNA was purified from agarose plugs using the Takara Agarose Gel DNA Purification Kit (Takara). BGI Tech Solutions (Wuhan, China) performed DNA sequencing, and this data was aligned and analyzed using the Basic Local Alignment Search Tool (BLAST).

TABLE 1

Primer sequences used for gene screening.

Resistance			Amplicon	Annealing
gene type	Gene	Primer Sequence (5′ → 3′)	size (bp)	Temperature (°C)	References
Chloramphenicol	catA1	F: GCAAGATGTGGCGTGTTACGGTGAA	258	56	This study
		R: TCATTAAGCATTCTGCCGACATGGA
	floR	F: AACCCGCCCTCTGGATCAAGTCAA	549	56	Ghoddusi et al., 2015
		R: CAAATCACGGGCCACGCTGTATC
	cmlA	F: ATTTTAGTTGGCGGTACTCCCT	307	61	This study
		R: CCCAGAAAGACTTCAGCCGAT
Tetracycline	tetA	F: GCTACATCCTGCTTGCCTTC	210	56	This study
		R: CATAGATCGCCGTGAAGAGG
	tetB	F: TTGGTTAGGGGCAAGTTTTG	659	56	This study
		R: GTAATGGGCCAATAACACCG
	tetC	F: AAAGGCTATCAAGCTGTTTC	311	56	This study
		R: CCAGTAATGTATTCAAGGCTAA
Sulfonamide	sul1	F: CTTCGATGAGAGCCGGCGGC	437	55	Aarestrup et al., 2003
		R: GCAAGGCGGAAACCCGCGCC
	sul2	F: GCGCTCAAGGCAGATGGCATT	285	55	Aarestrup et al., 2003
		R:GCGTTTGATACCGGCACCCGT
	sul3	F: AGATGTGATTGATTTGGGAGC	443	55	Zhang et al., 2009
		R: TAGTTGTTTCTGGATTAGAGCCT
Quinolone	qnrA	F: AGAGGATTTCTCACGCCAGG	580	60	Cattoir et al., 2007
		R: TGCCAGGCACAGATCTTGAC
	qnrB	F: GGMATHGAAATTCGCCACTG	264	56	Cattoir et al., 2007
		R: TTTGCYGYYCGCCAGTCGAA
	qnrS	F: GCAAGTTCATTGAACAGGGT	428	57	Cattoir et al., 2007
		R: TCTAAACCGTCGAGTTCGGCG
	qnrD	F: ACAGGAATAGCTTGGAAGGGTG	329	58	This study
		R: AAGATCGGAGCCACGAAACA
β-lactamase	blaTEM	F: GAGTACTCACCATCACAGAAAAGC	489	58	Lu et al., 2010
		R: GACTTCCCGTCGTGTAGATAAC
	blaPSE	F: AATGGCAATCAGCGCTTCCC	598	55	Shahada et al., 2006
		R: GGGGCTTGATGCTCACTACA
	blaCTX–M	F: CGATGTGCAGCACCAGTAA	584	58	This study
		R: AGTGACCAGAATCAGCGG
	blaCMY–2	F: GACAGCCTCTTTCTCCACA	1015	57	This study
		R: TGGAACGAAGGCTACGTA
	blaSHV	F: TTCGCCTGTGTATTATCTCC	807	60	This study
		R: TTTGCTGATTTCGCTCGG
Class I integron	qacEΔ1-F	F: ATCGCAATAGTTGGCGAAGT	800	54	Zhou et al., 2014
	sul1-B	R: GCAAGGCGGAAACCCGCGCC
	intI1	F: TGTCCACTGGGTTCGTGCCT	707	56	This study
		R: GCTTCGTGATGCCTGCTTGTT
	5′-CS	F: GGCATCCAAGCAGCAAGC	Random	54.5	Meng et al., 2011
	3′-CS	R: AAGCAGACTTGACCTGAT

Conjugation Experiments

To determine whether resistance determinants were present in transferable genetic elements in Salmonella isolates, conjugation experiments were carried out using the broth mating method as previously described (Olsen et al., 2004). The Salmonella isolates positive for blaCTX–M and/or integrons served as donors, and E. coli C600 was used as a recipient. Transconjugants were selected on Mueller–Hinton agar (Beijing Land Bridge Technology Co., Ltd.), containing 750 μg/ml rifampin and 100 μg/ml ampicillin. Susceptibility patterns of the transconjugants were determined by the methods described above. Co-transfer of resistance determinants was identified by amplifying the relevant genes from the transconjugants by PCR and determining the nucleotide sequencing of the PCR products. Plasmid DNA from transconjugants and donor strains was extracted utilizing a Tiangen Plasmid Purification Mini Kit (Tiangen Biotech, China) according to the manufacturer’s instructions. Replicon typing of plasmids was performed by a PCR-based method as previously described (Carattoli et al., 2005).

MLST

Seven housekeeping genes (aroC, dnaN, hemD, hisD, purE, sucA, and thrA) were used for MLST as described online[1]. All PCR products were purified from agarose gels and sequenced by BGI Tech Solutions as described above. Alleles and STs were assigned according to the MLST scheme.

Results and Discussion

Salmonella Serotype Prevalence Among the Animal Products

Out of the 631 meat samples, 66 (10.5%) tested positive for Salmonella, and chicken had the highest prevalence (14.8%; 39/263), followed by duck meat (11.8%; 12/102), pork (8.8%; 7/80), and fish (4.3%; 8/186).

A total of 19 serotypes were identified among 66 Salmonella isolates and one isolate from chicken was un-typeable (Table 2). The 19 serotypes included 11 from chicken, and 8, 5, and 6 from duck meat, pork and fish, respectively. S. Thompson was the serotype most frequently identified (18.2%; 12/66), followed by S. Agona (13.6%; 9/66), S. Takoradi (12.1%; 8/66), and S. Typhimurium (10.6%; 7/66). S. Thompson was not isolated from pork but it was the most common serotype isolated from both duck meat and fish with 33.3% (4/12) and 37.5% (3/8), respectively. Although S. Thompson was the most prevalent serotype overall and chicken was the most likely to be contaminated with Salmonella, S. Thompson was not the most common serotype isolated from chicken. Both S. Agona and S. Takoradi were isolated from chicken at a higher rate (20.5%; 8/39 for each of the two serotypes) than S. Thompson (12.8%; 5/39) (Table 2).

TABLE 2

Distribution of Salmonella serotypes in retail meat.

Serotype	Number (%) of isolates
	Chicken	Duck	Pork	Fish	Total
Thompson	5 (12.8)	4 (33.3)		3 (37.5)	12 (18.2)
Agona	8 (20.5)			1 (12.5)	9 (13.6)
Takoradi	8 (20.5)				8 (12.1)
Typhimurium	3 (7.7)	2 (16.7)	2 (28.6)		7 (10.6)
Enteritidis	4 (10.3)	1 (8.3)		1 (12.5)	6 (9.1)
Kentucky	5 (12.8)				5 (7.6)
Derby	1 (2.6)		2 (28.6)		3 (4.5)
Lindenburg	1 (2.6)	1 (8.3)		1 (12.5)	3 (4.5)
London	1 (2.6)		1 (14.3)		2 (3.0)
Anatum	1 (2.6)				1 (1.5)
Indiana	1 (2.6)				1 (1.5)
Galiema				1 (12.5)	1 (1.5)
Madelia				1 (12.5)	1 (1.5)
Paratyphi A			1 (14.3)		1 (1.5)
Senftenberg			1 (14.3)		1 (1.5)
Brunei		1 (8.3)			1 (1.5)
Brijbhumi		1 (8.3)			1 (1.5)
Muenchen		1 (8.3)			1 (1.5)
Rissen		1 (8.3)			1 (1.5)
Untypable	1 (2.6)				1 (1.5)
Total	39 (59.1)	12 (18.2)	7 (10.6)	8 (12.1)	66 (100)

Salmonella is a major cause of foodborne illness worldwide. In this study, Salmonella were recovered from retail raw meat products in Hubei Provence, China. The overall prevalence of Salmonella in retail meat products sampled in this study (10.5%) agreed with that reported in Henan Province, China (9.7%) (Yu et al., 2014) but was higher than that reported from Xinjiang, China (6.8%) (Yin et al., 2016). However, a simple conclusion could not be made from direct comparisons among studies due to regional differences, sampling seasons and the types of products sampled. In this study, the higher contamination rate found may be related to fewer biosecurity and hygiene measures inside these small markets.

We also identified 19 serotypes among the 66 isolates. The most common serotype in retail raw meat products was S. Thompson. While this result was consistent to two previous studies conducted in Iran (Soltan-Dallal et al., 2010; Sodagari et al., 2015), it differed from other reports in which the dominant serotype was S. Hadar in Xinjiang, China (Yin et al., 2016), S. Enteritidis in Henan Province, China (Yang et al., 2013) and S. Typhimurium in Brazil (Ristori et al., 2017). Other serotypes recovered in this study included S. Agona, S. Typhimurium, S. Enteritidis, S. Kentucky, and S. Derby, which have all been previously found in food products (Ma̧ka et al., 2014; Shah et al., 2017; Wang et al., 2017). The differences in dominant serotypes may be due to geographical differences, sample types or seasons and serotype pathogenicity.

Antimicrobial Susceptibility

Among 66 Salmonella isolates, a high percentage of the isolates were resistant to sulfonamides (43.9%), tetracycline (43.9%), amoxicillin (36.4%), ampicillin (36.4%), chloramphenicol (27.3%), nalidixic acid (21.2%), streptomycin (18.2%), and cephazolin (15.2%) (Table 3). It was found that 17 (25.8%) of the 66 isolates were resistant to 1 antibiotic, 10 isolates (15.2%) were resistant to 2–3 antibiotics, 20 isolates (30.3%) were resistant to 4–6 antibiotics, 3 isolates (4.5%) were resistant to 7–9 antibiotics and a single isolate (1.5%) was resistant to 11 antibiotics (Table 4). Thus, 51 of the 66 Salmonella isolates (77.3%) were resistant to at least one antibiotic. Salmonella isolated from chicken were most likely to be resistant with 92.3% (36/39) of the isolates showing resistance to at least one antibiotic, and a single isolate exhibiting resistance to 11 antibiotics. Salmonella isolates from fish were the most susceptible to the antimicrobials tested with only 2 strains (25%) showing any resistance (Table 4).

TABLE 3

Antimicrobial resistance of Salmonella isolates from retail meat.

Antimicrobial agent	Number of isolates (%)
	Chicken	Duck	Pork	Fish	Total
	(n = 39)	(n = 12)	(n = 7)	(n = 8)	(n = 66)
Sulfonamides	21 (53.8)	4 (33.3)	4 (57.1)		29 (43.9)
Tetracycline	20 (51.3)	3 (25)	6 (85.7)		29 (43.9)
Amoxicillin	16 (41.0)	4 (33.3)	3 (42.9)	1 (12.5)	24 (36.4)
Ampicillin	16 (41.0)	4 (33.3)	3 (42.9)	1 (12.5)	24 (36.4)
Chloramphenicol	15 (38.5)		3 (42.9)		18 (27.3)
Nalidixic acid	9 (23.1)	2 (16.7)	1 (14.3)	2 (25)	14 (21.2)
Streptomycin	10 (25.6)		2 (28.6)		12 (18.2)
Cephazolin	7 (17.9)	2 (16.7)	1 (14.3)		10 (15.2)
Gentamicin	3 (7.7)	1 (8.3)	1 (14.3)		5 (7.6)
Cefuroxime	2 (5.1)				2 (3.0)
Ciprofloxacin	1 (2.6)	1 (8.3)			2 (3.0)
Norfloxacin	1 (2.6)				1 (1.5)
Nitrofurantoin			1 (14.3)		1 (1.5)

TABLE 4

Multidrug resistance strains identified among Salmonella isolates recovered from retail meat.

Sources	Number (%) of isolates resistant to indicated number of antimicrobials
	1	2–3	4–6	7–9	11	Total (> 1)
Chicken (n = 39)	11 (28.2)	8 (20.5)	14 (35.9)	2 (5.1)	1 (2.6)	36 (92.3)
Duck (n = 12)	3 (25)		4 (33.3)			7 (58.3)
Pork (n = 7)	2 (28.6)	1 (14.3)	2 (28.6)	1 (14.3)		6 (85.7)
fish (n = 8)	1 (12.5)	1 (12.5)				2 (25)
Total (n = 66)	17 (25.8)	10 (15.2)	20 (30.3)	3 (4.5)	1 (1.5)	51 (77.3)

The increasing frequency of antimicrobial resistance among Salmonella has become an emerging challenge to public health. Our results indicated that 30.3% of isolates were resistant to 4–6 antimicrobial agents and 4.5% were resistant to 7–9 antimicrobials. This indicated that antimicrobial resistance among Salmonella isolates from retail meat in Hubei Province was lower than that in other areas of China (Zhu et al., 2014; Wang et al., 2017). In our study, the highest rates of antimicrobial resistance were to the sulfonamides, tetracycline, amoxicillin, ampicillin and chloramphenicol. These results were consistent to studies from other regions in China and other countries (Ma̧ka et al., 2014; Yu et al., 2014; Abdi et al., 2017). The reasons for this may lie in the continuous and extensive use of these antimicrobials in food animals for disease treatment and growth promotion.

Cephalosporins and fluoroquinolones have been effectively used for the treatment of invasive and systemic salmonellosis in humans and animals (Bhan et al., 2005). Our study revealed that 15.2 and 3.0% of Salmonella isolates were resistant to first-generation (cephazolin) and second-generation cephalosporins (cefuroxime), respectively, and two isolates showed intermediate resistance to third-generation cephalosporins (cefotaxime) (Table 5). Previous work has described a decreasing susceptibility of Salmonella from retail foods to these antimicrobials (Hindermann et al., 2017; Kim et al., 2017). Interestingly, our results indicated that a high prevalence of cephazolin resistance was observed in isolates of S. Thompson, the predominant serotype. In this study, Salmonella resistance to nalidixic acid (21.2%) was not very common, however, 31.8% of the isolates also showed intermediate resistance to nalidixic acid (Table 5). Only 3.0 and 1.5% of the isolates were resistant to ciprofloxacin and norfloxacin, respectively. These rates were lower than those reported in other studies from China (Pan et al., 2009; Yin et al., 2016).

TABLE 5

Resistance phenotypes and antibiotic resistance genes from Salmonella isolates recovered from retail meat.

Strain			Sequence		Intermediate
no.	Source	Serotype	Type	Resistance phenotype	phenotype	Integrons/resistance genes
1	Chicken	Not detected	463	AML, AMP, C, NA, TE		Class I, qnrS, sul2, tetA
2	Chicken	S. Agona	13		ST	blaTEM–1, sul1
3	Chicken	S. Agona	13	AML, AMP, ST, S3, TE		blaTEM–1,sul3, tetA
4	Chicken	S. Agona	13	AML, AMP, ST, S3, TE	NA	blaTEM–1, qnrS, sul3, tetA
5	Chicken	S. Agona	13	AML, AMP, ST, S3, TE	NA	blaTEM–1, qnrS,sul3,tetA, floR
6	Chicken	S. Agona	13	C	NA, ST	qnrS, floR
7	Chicken	S. Agona	13	AML, AMP, NA		blaTEM–1, qnrB, sul1, sul2
8	Chicken	S. Agona	1215	C	NA, ST	qnrS, floR
9	Chicken	S. Agona	1328	C	NA,ST	qnrS, floR
10	Chicken	S. Anatum	516	NA		blaTEM–1, blaPSE–1, sul1
11	Chicken	S. Derby	20	S3, TE	NA, ST	blaTEM–1,sul1,sul2,tetA
12	Chicken	S. Enteritidis	11	AML, AMP, NA, ST, S3		Class I, blaTEM–1, sul1
13	Chicken	S. Enteritidis	11	AML, AMP, NA		Class I, qnrB, qnrS, sul1, sul2, tetA
14	Chicken	S. Enteritidis	11	AML, AMP, NA		Class I, blaTEM–1, qnrB, sul1, sul2
15	Chicken	S. Enteritidis	11	NA		blaTEM–1, sul1
16	Chicken	S. Indiana	17	AML, AMP, CXM, KZ, C, CIP, NA, NOR, ST, CN, S3	CTX, TE	blaTEM–1, blaCTX–M–65, sul1, sul2, tetA
17	Chicken	S. Kentucky	314	C, S3, TE	NA, ST	Class I, qnrB, sul2, tetA
18	Chicken	S. Kentucky	314	C, S3, TE	NA, ST	qnrB, sul3, tetA
19	Chicken	S. Kentucky	314	C, S3, TE	NA, ST	Class I, qnrB, sul4, tetA
20	Chicken	S. Kentucky	314		ST
21	Chicken	S. Kentucky	314		ST	floR
22	Chicken	S. Lindenburg	46	S3		blaTEM–1, sul1
23	Chicken	S. London	155	AML, AMP, C, S, CN, S3, TE		Class I, blaTEM–1, qnrB, sul1,floR,
24	Chicken	S. Takoradi	541	TE	NA, ST	qnrS, tetA
25	Chicken	S. Takoradi	541	TE	NA	qnrS, tetA
26	Chicken	S. Takoradi	541	C, ST, S3, TE		sul2, tetA
27	Chicken	S. Takoradi	541	TE	NA	qnrS, tetA
28	Chicken	S. Takoradi	541	C, ST, S3, TE	NA	qnrS, sul2,tetA, floR,
29	Chicken	S. Takoradi	541	C, ST, S3, TE	NA	sul2,tetA, floR
30	Chicken	S. Takoradi	541	C, ST, S3, TE		sul2,tetA,floR
31	Chicken	S. Takoradi	541	C, ST, S3, TE		floR, sul3, tetA
32	Chicken	S. Thompson	26	AML, AMP, KZ, S3, TE	CXM, NA, ST	Class I, blaTEM–1, qnrB, sul1, tetA
33	Chicken	S. Thompson	26	AML, AMP, KZ, S3	CXM, NA	qnrB, tetA
34	Chicken	S. Thompson	26	AML, AMP, CXM, KZ, C, CN, S3, TE	CTX, ST	Class I: dfr17-aadA5, blaTEM–1, blaCTX–M–15, qnrB, sul1
35	Chicken	S. Thompson	26	AML, AMP, KZ, S3	ST	Class I, blaTEM–1, sul1
36	Chicken	S. Thompson	26	AML, AMP, KZ, S3	ST	Class I, sul1
37	Chicken	S. Typhimurium	19	NA		blaTEM–1, sul1
38	Chicken	S. Typhimurium	1682	KZ	AML	blaTEM–1
39	Chicken	S. Typhimurium	1682	AML, AMP, TE	S3	blaTEM–1,sul1
40	Duck	S. Braenderup	22	NA		blaTEM–1
41	Duck	S. Brunei	2256
42	Duck	S. Enteritidis	11	AML, AMP, CIP, CN		Class I, sul1
43	Duck	S. Lindenburg	46	S3		blaTEM–1, sul1
44	Duck	S. Muenchen	82
45	Duck	S. Rissen	469	AML, AMP, S3, TE		sul1, sul3, tetA
46	Duck	S. Thompson	26		ST	blaTEM–1, sul1
47	Duck	S. Thompson	26	AML, AMP, KZ, S3, TE	NA	Class I: dfrA12-aadA2, qnrB, sul1, tetA
48	Duck	S. Thompson	26	AML, AMP, KZ, S3, TE	CXM, NA, ST	Class I, blaTEM–1, qnrB, sul1, tetA
49	Duck	S. Thompson	26
50	Duck	S. Typhimurium	19	NA		blaTEM–1, sul1
51	Duck	S. Typhimurium	99		ST	blaTEM–1, sul1
52	Pork	S. Derby	40	TE	C, NA, S3	sul1, tetA
53	Pork	S. Derby	40	TE	C, NA	sul1, sul2, tetA
54	Pork	S. London	155	AML, AMP, C, ST, S3, TE	NA	Class I: aacA4C- arr-3- dfr2, blaTEM–1, qnrB, sul1, sul2, tetA
55	Pork	S. ParatyphiA	463	KZ, S3, TE		sul1, tetC
56	Pork	S. Senftenberg	14
56	Pork	S. Typhimurium	313	AML, AMP, C, ST, S3, TE		Class I:aadA2, blaTEM–1, qnrB, sul1, sul2
58	Pork	S. Typhimurium	19	AML, AMP, C, NA, CN, F, S3, TE	CIP, NOR, ST	Class I:dfrA12-aadA21, blaTEM–1, sul1, tetB, tetC
59	fish	S. Agona	13
60	fish	S. Enteritidis	11	AML, AMP, NA		Class I, blaTEM–1, qnrB, sul1
61	fish	S. Galiema	501		ST	blaTEM–1, sul1
62	fish	S. Lindenburg	45		S3	blaTEM–1, sul1
63	fish	S. Madelia	226	NA		blaTEM–1, sul1
64	fish	S. Thompson	26
65	fish	S. Thompson	26
66	fish	S. Thompson	26

The Occurrence of Antimicrobial Resistance Genes and Class I Integrons

All 66 Salmonella isolates were screened for the presence of antimicrobial resistance genes and class 1 integrons. Numerous oligonucleotide primer pairs (Table 1) were used to screen the 66 Salmonella isolates by PCR for the presence of genes encoding resistance to chloramphenicol, tetracycline, sulfonamides, quinolones, and β-lactam antibiotics. Overall, the results showed a strong correlation between antimicrobial resistance genotypes and phenotypes (Table 5).

Among 26 β-lactam-resistant isolates in this study, four different β-lactamase genes were identified, including blaTEM–1, blaCTX–M–65, blaCTX–M–15, and blaPSE–1 (Table 5). The blaTEM–1 gene had the highest occurrence (69.2%, 18/26) among β-lactam-resistant isolates; a rate similar to that previously reported for Salmonella isolated from meat and dairy products in Egypt (Ahmed et al., 2014). The extended-spectrum β-lactamase (ESBL) genes, blaCTX–M–65 and blaCTX–M–15, were each observed in a single S. Indiana and S. Thompson isolate, respectively. The blaPSE–1 gene was also identified in only one isolate of S. Anatum. A cephalosporin resistant S. Thompson isolate carried the gene encoding CTX-M-15, which is the most prevalent ESBL worldwide (Chon et al., 2015). In this study, the rarely observed blaCTX–M–65 was present in a ciprofloxacin- and cefuroxime-resistant isolate of S. Indiana. Similarly, it was reported that the blaCTX–M–65 gene was harbored in ciprofloxacin- and cefotaxime-resistant S. Indiana isolates in Henan Province of China (Bai et al., 2016). In addition, the blaCTX–M–65 gene was identified previously in an outbreak of S. Infantis in Ecuador (Cartelle Gestal et al., 2016).

The qnr family of genes encodes a pentapeptide repeat protein that protects DNA gyrase and topoisomerase IV from quinolone and fluoroquinolone inactivation. Plasmids containing qnr can be transferred among bacteria via conjugation and this is a very important emerging public health concern (García-Fernández et al., 2009). Although qnr confers only low-level quinolone resistance, mutations in the quinolone resistance-determining region (QRDR) could result in higher-level resistance to fluoroquinolones, especially to ciprofloxacin. The existence of qnr genes has been reported in Salmonella from other regions in China and in other countries (Yang et al., 2013; Wasyl et al., 2014). In the current study, 62.9% (22/35) of the 14 nalidixic acid-resistant and 21 intermediate resistant Salmonella isolates harbored qnrB and/or qnrS genes (Table 5). The qnrB gene was present in 34.3% (12/35) of the nalidixic acid resistant or intermediate resistant Salmonella isolates while qnrS was present in 31.4% (11/35) of the isolates. Both qnrB and qnrS were found in only a single isolate of S. Enteritidis. The genes qnrA and qnrD were not found among any of the isolates. The blaCTX–M–15 gene was found in combination with qnrB, while the isolate with blaCTX–M–65 carried no quinolone resistance genes.

The tetracycline resistance genes (tetA, tetB, and tetC) and sulfonamide resistance genes (sul1, sul2, and sul3) were identified in 86.7 and 96.9% of the strains resistant or intermediate resistant to the cognate antibiotic. The floR gene was identified in 40% (8/20) of chloramphenicol-resistant isolates (n = 18) and intermediate resistant isolates (n = 2) (Table 5).

An integron is a mobile DNA element, which usually contains one or more antimicrobial resistance genes that can be transferred among bacteria. In this study, Class I integrons (intI1) were found in 18 isolates (27.3%) and 13 of them did not possess antimicrobial resistance gene cassettes. The remaining 5 intI1-positive isolates contained five groups of resistance gene cassettes including aacA4C-arr-3-dfr2, dfrA12-aadA21, aadA2, dfrA12-aadA2, and dfr17-aadA5 (Table 5). Our study also indicated that all of the class I integron-positive isolates were resistant to at least three classes of antimicrobials. This supports the hypothesis of the association between class I integrons and MDR in Salmonella isolates (Abraham et al., 2016; Meng et al., 2016).

Previous investigators identified class I integrons in foodborne Salmonella isolated in China at rates that ranged from 20 to 61% (Zhu et al., 2014; Meng et al., 2016; Li et al., 2017). In this study, the gene cassettes aacA4C-arr-3-dfr2 from S. London exhibited >98.0% nucleotide sequence identity to that of the variable region of Integron In1021 from a clinical E. coli isolated in China (accession no. KR338349), suggesting that some of its components may have originated from other species of Enterobacteriaceae such as Escherichia. The existence of class I integrons containing genes for antimicrobial resistance in Salmonella isolates poses a threat to public health and food safety.

Transferability of blaCTX–M Genes and Integrons

Six Salmonella strains harboring the ESBL blaCTX–M gene and integrons were selected as donor strains in conjugation experiments. These donor strains were mated with an E. coli C600 recipient strain. Four transconjugants were obtained and characterized. PCR based replicon typing revealed that plasmids belonging to incompatibility types H12 and N were transferred from Salmonella to the E. coli C600 (Table 6). However, an incA/C plasmid in a S. Indiana isolate was not transferred. Genes present in the transconjugants included class I integrons, blaTEM–1, blaCTX–M–65, blaCTX–M–15, qnrB, sul1, sul2, tetA, and tetB. The co-transfer of blaCTX–M–15 and qnrB was observed with an integron from an S. Thompson isolate to E. coli strain C600 (Table 6). When mated with the S. London donor, the E. coli C600 transconjugants showed an identical antimicrobial resistance phenotype to the donor. For the other Salmonella donors, the E. coli C600 recipients showed similar, but not identical antimicrobial resistances (Table 6).

TABLE 6

Characteristics of Salmonella isolates and E. coli transconjugants.

Isolate				Replicon
No.	Serotype	Resistance profile	Resistance genes	typing
16	S. Indiana	AML, AMP, CXM, KZ, C, CIP, NA, NOR, ST, CN, S3	blaTEM–1, blaCTX–M–65, sul1, sul2, tetA	H12, A/C
16-T		AML, AMP, CXM, KZ, C, CIP, NA, NOR, ST, CN,	blaTEM–1, blaCTX–M–65, tetA	H12
34	S. Thompson	AML, AMP, CXM, KZ, C, CN, S3, TE	Class 1: dfr17-aadA5, blaTEM–1, blaCTX–M–15, qnrB, sul1	H12
34-T		AML, AMP, CXM, KZ, CN	Class 1: dfr17-aadA5, blaTEM–1, blaCTX–M–15, qnrB	H12
54	S. London	AML, AMP, C, ST, S3, TE	Class 1: aacA4C-arr-3-dfr2, blaTEM–1, qnrB, sul1, sul2, tetA	N
54-T		AML, AMP, C, ST, S3, TE	Class 1:aacA4C-arr-3-dfr2, blaTEM–1, qnrB, sul1, tetA	N
58	S. Typhimurium	AML, AMP, C, NA, CN, F, S3, TE	Class 1:dfrA12-aadA21, blaTEM–1, sul1, tetB, tetC	H12
58-T		AML, AMP, C, NA, CN, F, TE	Class 1:dfrA12-aadA21, blaTEM–1, tetB	H12

Of concern, the findings of conjugation tests demonstrated the presence of genes including blaTEM–1, blaCTX–M–15, and qnrB on mobile plasmids. These resistance genes and integron were conjugally transferred from a Salmonella donor strain and conferred a MDR phenotype to a naïve E. coli recipient. Few studies have shown the co-existence of integron with plasmid-mediated quinolone resistance (PMQR) determinants and ESBL genes on transferable plasmids (Lai et al., 2013). It has been suggested that the transmissible elements containing these genes could be selected and transmitted during exposure to any of the antimicrobials (González and Araque, 2013; Jiang et al., 2014).

MLST Analysis

Thirty-two STs were identified among the 66 isolates, and ST26 (n = 12) was the most common followed by ST541 (n = 8), ST13 (n = 7), ST11 (n = 6), and ST314 (n = 5) (Table 5). The STs in this study were correlated with specific serotypes, such as ST11 with S. Enteritidis, ST13 with S. Agona, ST19 with S. Typhimurium, ST314 with S. Kentucky and ST26 with S. Thompson. Though a particular ST was always associated with the same serotype, some serotypes could be represented by multiple STs. For example, the isolates characterized as S. Derby belonged to ST20 and ST40.

The diversity of samples we collected was also reflected in a diversity of the 32 STs that we identified. ST26 was the most common especially among poultry samples and this ST corresponded to S. Thompson that has a low rate of occurrence in China (Ni et al., 2018). ST11 and ST19 were the most frequent sequence types recovered from meat samples and these STs are commonly associated with human salmonellosis outbreaks (Gunel et al., 2015; Ashton et al., 2016). Additionally, our results revealed that the serotypes and STs were highly correlated, which is consistent with previous research (Achtman et al., 2012). We did find sequence type differentiation within serotypes such as with S. Typhimurium and S. Derby. Therefore, MLST analysis provided a greater discriminatory power than serotype determination.

Conclusion

This study identified S. Thompson and S. Agona as the most common serotypes among Salmonella isolates recovered from retail raw meat products in Hubei Province, China. Antimicrobial resistance was prevalent and correlated well with the presence of cognate resistance genes. Salmonella isolates containing class I integrons showed broader antimicrobial resistance than those without them. In addition, conjugation experiments revealed a co-transfer of class I integron with PMQR and ESBL genes. The reasonable use of antibiotics in livestock breeding and management of the food supply chain is of importance to ensure food safety.

Data Availability Statement

Publicly available datasets were analyzed in this study. This data can be found here: http://enterobase.warwick.ac.uk/species/senterica/allele_st_search.

Author Contributions

XS and HW designed the study and revised the manuscript. MZ conducted the experiments and interpreted the results. XL and WH assisted in the completion of the experiments. GP rewrote the discussion and revised the manuscript.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.