Prevalence of drug resistance and virulence features in Salmonella spp. isolated from foods associated or not with salmonellosis in Brazil.

Salmonella is the most common etiological agent of cases and outbreaks of foodborne diarrheal illnesses. The emergence and spread of Salmonella spp., which has become multi-drug resistant and potentially more pathogenic, have increased the concern with this pathogen. In this study, 237 Salmonella spp., associated or not with foodborne salmonellosis in Brazil, belonging mainly to serotype Enteritidis, were tested for antimicrobial susceptibility and the presence of the virulence genes spvC, invA, sefA and pefA. Of the isolates, 46.8% were sensitive to all antimicrobials and 51.9% were resistant to at least one antimicrobial agent. Resistance to more than one antimicrobial agent was observed in 10.5% of the strains. The highest rates of resistance were observed for streptomycin (35.9%) and nalidixic acid (16.9%). No strain was resistant to cefoxitin, cephalothin, cefotaxime, amikacin, ciprofloxacin and imipenem. The invA gene was detected in all strains. Genes spvC and pefA were found in 48.1% and 44.3% of strains, respectively. The gene sefA was detected in 31.6% of the strains and only among S. Enteritidis. Resistance and virulence determinants were detected in Salmonella strains belonging to several serotypes. The high rates of antibiotic-resistance in strains isolated from poultry products demonstrate the potential risk associated with the consumption of these products and the need to ensure good food hygiene practices from farm to table to reduce the spread of pathogens relevant to public health.

INTRODUCTION

Worldwide, Salmonella is the most common etiological agent of foodborne diarrheal illnesses[18,26,45,46]. There is an increasing concern with this pathogen due to the emergence and spread of antibiotic-resistant and potentially more pathogenic strains[37,39]. The increase in resistant strains can be attributed to the inappropriate use of antimicrobials as therapeutic or prophylactic agents in human and veterinary medicine, as well as the use of growth promoters in animal production[54]. Studies suggest that the use of antibiotics in food animal production has contributed to the selection of resistant lineages transferable to humans via the food chain[17]. Antimicrobial-resistant Salmonella spp. have been isolated from different foods of animal origin around the world[5,13,36,38,55,57].

The ability of Salmonella to cause disease can be attributed to an array of virulence genes located in the chromosome or in large virulence-associated plasmids[22]. Gene invA in the Salmonella Pathogenicity Island[33] codes for the production of proteins from the type III secretion system, related to the invasion of Salmonella into eukaryotic host cells[49]. Many Salmonella serotypes harbor virulence plasmids of varying sizes. The Salmonella plasmid virulence (spv) operon, which consists of five genes (spvRABCD), is important for the intracellular survival and replication of Salmonella, and contributes to the systemic phase of the illness[22]. There are many types of fimbriae that mediate Salmonella intestinal adhesion including type 1 fimbriae (Fim), long polar fimbriae (Lpf), aggregative fimbriae (Agf), plasmid-encoded fimbriae (Pef) and Salmonella Enteritidis fimbriae (Sef). The fim and agf operons are highly conserved among isolates of Salmonella, whereas the sef and pef operons are only found in certain serotypes[2].

The objective of this study was to evaluate the prevalence of antimicrobial resistance and to detect the virulence -related genes among Salmonella spp. isolated from foods associated or not with foodborne salmonellosis.

MATERIALS AND METHODS

The study included 237 randomly selected Salmonella spp. strains isolated from food sources between 1983 and 2007, belonging to the culture collection of Food Microbiology Laboratory of the Adolfo Lutz Institute, São Paulo, Brazil. The isolates belonged to 51 serotypes and most of them were S. Enteritidis (32%). Among the selected strains, 42 were isolated from foods associated with foodborne salmonellosis and 195 from food samples not associated with outbreaks or sporadic cases (Table 1). Among the strains from foods associated with salmonellosis, 14.3% were isolated from beef, 11.9% from poultry, 23.8% from confectionery, 21.4% from eggs and mayonnaise, 9.5% from water, 7.1% from vegetables, 2.4% from dairy products and 9.5% from other products. Regarding the strains obtained from foods not associated with salmonellosis, 27.2% were isolated from poultry, 29.7% from beef, 18% from pork, 11.3% from seafoods, 5.6% from condiments, 2% from milk and derivatives, 1.5% from eggs and mayonnaise, 1.5% from vegetables and 2.6% from other products.

Table 1

Number of isolates and serovars of Salmonella spp. included in the study

Serovar	Number of isolates
	Associated with foodborne salmonellosis	Not associated with foodborne salmonellosis
Enteritidis	35	40
Typhimurium	1	14
Agona	1	9
Infantis	1	8
Brandenburg	1	7
Saintpaul	1	2
I 6,7:r:-	1	3
Sandiego	1	3
Panama	--	11
Anatum	--	12
Bredeney	--	7
Heidelberg	--	6
Ohio, Oranienburg and Schwarzengrund	--	5 each
Madelia, Mbandaka, I 4, 5, 12: i: - and London,	--	4 each
Hadar and Derby	--	3 each
Senftenberg, Cerro, IV 43:Z4, Z24:-, Glostrup, Give, I 4, 12: i:- and Rubislaw	--	2 each
Muenster, I 4, 5, 12: b: -, Javiana, 3, 10: Z: -, Emek, Muenchen, Saphra, Levingstone, I 6, 8: E, H: -, Lexington, Brandenlsing, 1,4, 5, 12, Berta, Poona, I 3, 10: -:1,6, Abaetetuba, I 13, 23: Z: -, Tenessee, I 6,8:Z10:-, Newport, Paratyphi B and Pomona	--	1 each
Total	42	195

Determination of the antimicrobial susceptibility profile: The strains were tested for antibiotic resistance by the plate disk diffusion method, according to the Clinical and Laboratory Standards Institute[8]. The following disks (Oxoid, Basingstoke, UK) were included in the test: ampicillin (10 µg), cefoxitin (30 µg), cephalothin (30 µg), cefotaxime (30 µg), imipenem (10 µg), chloramphenicol (30 µg), amikacin (30 µg), gentamicin (10 µg), kanamycin (30 µg), streptomycin (10 µg), nalidixic acid (30 µg), ciprofloxacin (5 µg), tetracycline (30 µg), trimethoprim-sulphamethoxazole (1.25/23.75 µg) and sulphonamides (300 µg). Escherichia coli ATCC 25922 was used for quality control. Results were interpreted as recommended by CLSI (2012b)[9].

Investigation of virulence genes: The virulence genes pefA, invA, sefA and spvC were investigated as described by CORTEZ et al. (2006)[10] with some modifications. DNA was extracted from the cultures using the Wizard® Genomic DNA Purification System kit (Promega) following the manufacturer's instructions. The pefA gene was investigated by singleplex PCR with the primers pefA-1 (5′- TTCCATTATTGCACTGGGTG-3′) and pefA-2 (5′-AAGCCACTGCGAAAGATGCC-3′)[25]. The genes invA, sefA and spvC were amplified by multiplex PCR, with the following primers: invA-1 (5′-TTGTTACGGCTATTTTGACCA-3′) and invA-2 (5′-CTGACTGCTACCTTGCTGATG-3′)[47]; sefA-3 (5′- GCAGCGGTTACTATTGCAGC-3′) and sefA-4 (5′-TGTGACAGGGACATTTAGCG-3′)[56]; spvC-1 (5′-CGGAAATACCATCTACAAATA-3′) and spvC-2 (5′-CCCAAACCCATACTTACTCTG-3′)[47].

Singleplex PCR was performed in a reaction volume of 25 µL containing PCR reaction buffer (50 mM KCl, 10 mM Tris-HCl, 2.5 mM MgCl2, pH 8.3), 200 µM dNTP, 3 mM MgCl2, 0.1 µM pefA primers, 2.5 U Taq DNA polymerase, and 2 µL DNA template. Amplification was carried out in an automatic thermocycler (GeneAmp PCR System 2400, Perkin Elmer) with the following cycles: 94°C for two minutes, followed by 35 cycles of 94°C for 30 seconds, 50°C for 45 seconds and 72°C for one minute, and a final elongation at 72°C for seven minutes. The samples were stored at 4°C until electrophoresis. The same conditions were used for the multiplex PCR, except that primers were used in the following concentrations: 0.2 µM invA, 0.5 µM sefA and 0.5 µM spvC; in addition, annealing temperature was 55°C rather than 50°C. Electrophoresis of amplified products was carried out using 1.5% agarose gel in TAE running buffer (40 mM Tris-acetate, 1 mM EDTA, pH 8.00) containing 0.5 µg/mL of ethidium bromide. The amplified DNA fragments were visualized under UV light. S. Enteritidis ATCC 13076 and S. Typhimurium ATCC 14028 were used as positive controls.

RESULTS AND DISCUSSION

Antimicrobial susceptibility tests indicated that 51.9% of the strains were resistant to at least one antimicrobial agent and that 10.5% were resistant to more than one antimicrobial. Resistant strains were more common in foods not associated with foodborne salmonellosis (55.9%) than in those related to salmonellosis (33.3%). The resistant Salmonella spp. were grouped in 14 different antimicrobial resistance profiles (Table 2). Resistance was observed in 35 serotypes, mainly in S. Infantis (88.9%), S. Typhimurium (80%), S. Enteritidis (58.7%), S. Anatum (50%) and S. Agona (40%).

Table 2

Antimicrobial resistance among Salmonella spp. strains isolated from foods associated and not associated with foodborne salmonellosis

Resistance to	Associated with foodborne salmonellosis	Not associated with foodborne salmonellosis
STR	2 (4.8%)	57 (29.2%)
NAL	7 (16.7%)	28 (14.3%)
TET	--	3 (1.5%)
AMP	--	1 (0.5%)
GEN+STR	1 (2.4%)	5 (2.6%)
GEN+KAN	--	1 (0.5%)
TET+STR	--	6 (3.1%)
SSS+STR	--	2 (1.0%)
GEN+KAN+STR	1 (2.4%)	--
GEN+NAL+STR	2 (4.8%)	1 (0.5%)
TET+SSS+STR	--	2 (1.0%)
TET+KAN+STR	--	1 (0.5%)
TET+NAL+CHL+STR	--	2 (1.0%)
AMP+STX+SSS+STR	1 (2.4%)	--
Number of resistant strains	14 (33.3%)	109 (55.9%)

AMP, ampicillin; FOX, cefoxitin; CEF, cephalothin; CTX, cefotaxime; CHL, chloramphenicol; TET, tetracycline; AK, amikacin; GEN, gentamicin; STX, trimethoprim-sulphamethoxazole; CIP, ciprofloxacin; NAL, nalidixic acid; IPM, imipenem; KAN, kanamycin; SSS, sulphonamides; STR, streptomycin.

All Salmonella spp. were susceptible to cefoxitin, cephalothin, cefotaxime, amikacin, ciprofloxacin and imipenem. As for other antibiotics, resistance to streptomycin prevailed in strains associated with salmonellosis (16.7%) in comparison with those not associated with salmonellosis (39%). Resistance to streptomycin was observed in strains belonging to 33 different serotypes. Similar rates were reported in Spain[50], the United States[40] and Brazil[29]. According to PEIRANO et al. (2006)[43], resistance to streptomycin might be related to co-selection, since the therapeutic use of this antibiotic in human and veterinary medicine has declined in recent years.

Among nalidixic acid resistant strains, 21.5% were associated with salmonellosis and 15.9% of them were not associated with salmonellosis. Resistance to nalidixic acid was detected in serotypes Enteritidis, Typhimurium and Infantis, and only in strains isolated from poultry products. Fluoroquinolones are extensively used as therapeutic drugs in veterinary medicine and the use of these compounds can select for mutant Salmonella strains resistant to nalidixic acid or with reduced susceptibility to fluoroquinolones[20,21,30,48]. Resistant strains to nalidixic acid were detected mainly from the beginning of 2000. Before that, only one strain isolated in 1993 was resistant to this antibiotic (Table 5). In Europe, the occurrence of resistance to nalidixic acid in S. Enteritidis strains of human origin almost tripled between 2000 and 2004, rising from 10% to 26%[35]. In Brazil, OLIVEIRA et al. (2005)[41] evaluated 31 strains of S. Enteritidis isolated from food involved in outbreaks in the state of Rio Grande do Sul from 1995 to 1997 and verified that only 3.2% of the strains were resistant to nalidixic acid. However, another study carried out by OLIVEIRA et al. (2006)[42] with 79 S. Enteritidis strains isolated from outbreaks in this same region between 2001 and 2002 found high resistance rates (21.5%), which are similar to this study's findings for strains associated with food poisoning.

Table 5

Frequency (%) of resistance to the 15 antimicrobials tested by year of isolation for Salmonella spp. strains

Antibiotics	Year (number of isolates)
	1983 (n=1)	1984 (n=2)	1986 (n=13)	1987 (n= 2)	1988 (n=3)	1990 (n=10)	1991 (n=8)	1992 (n=19)	1993 (n=37)	1994 (n=4)	1995 (n=5)	1996 (n=5)	1997 (n=16)	1998 (n=3)	1999 (n=6)	2000 (n=12)	2001 (n=13)	2002 (n=29)	2003 (n=7)	2004 (n=31)	2005 (n=2)	2006 (n=3)	2007 (n=6)
AMP	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	3.4	0	0	0	0	16.7
FOX	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
CEF	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
CTX	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
CHL	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	6.7	0	0	0	0	0
TET	0	0	0	33.3	0	10	0	5.3	8.3	0	0	0	6.25	0	0	0	0	20.7	0	0	0	0	0
AK	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
GEN	0	0	0	0	0	0	0	0	0	0	0	0	25	0	0	8.3	15.4	13.8	0	0	0	0	0
STX	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	16.7
CIP	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
NAL	0	0	0	0	0	0	0	0	2.8	0	0	0	0	0	0	25	0	20.7	57.1	74.2	50	33.3	16.7
IPM	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
KAN	0	0	0	0	0	0	0	5.3	0	0	0	0	0	0	0	8.3	7.7	0	0	0	0	0	0
SSS	0	0	0	0	0	0	25	0	2.8	0	0	0	6.25	0	0	0	0	0	0	0	0	0	16.7
STR	0	100	38.5	66.6	0	50	50	57.9	41.7	0	0	20	37.5	33.3	16.7	33.3	23.1	41.4	14.3	12.9	0	66.7	66.7

AMP, ampicillin; FOX, cefoxitin; CEF, cephalothin; CTX, cefotaxime; CHL, chloramphenicol; TET, tetracycline; AK, amikacin; GEN, gentamicin; STX, trimethoprim-sulphamethoxazole; CIP, ciprofloxacin; NAL, nalidixic acid; IPM, imipenem; KAN, kanamycin; SSS, sulphonamides; STR, streptomycin.

Concerning resistance to tetracycline, gentamicin, sulphonamides, kanamycin, ampicilin, chloramfenicol and trimethoprim-sulphamethoxazole, the prevalence of resistant Salmonella spp. was low - below 6%. Although the use of tetracycline as a growth promoter has been banned in Brazil since 1998, this antimicrobial agent remains one of the most frequently used therapeutic drugs in animal production[41] thus contributing to the high levels of resistance to this antimicrobial[32]. In this survey, resistance to this antibiotic was low (5.9%) whereby eight strains were isolated in years prior to the prohibition and six were isolated in 2002 (Table 5) and found in strains belonging to serotypes Typhimurium, Bradenburg, Hadar, I 4, 5, 12: i:-, Derby, Bredeney and I 6,8: Z10:-. The low resistance observed for the sulphonamides is surprising, since other Brazilian studies have reported a high frequency of resistance to this antibiotic in Salmonella [3,23,36].

Regarding S. Enteritidis, the serotype with the highest prevalence in humans in Brazil[19], 31.4% (11/35) of the strains associated with foodborne salmonellosis and 82.5% (33/40) of the strains not involved in salmonellosis were resistant to one or more antimicrobial agents. The higher prevalence of resistance among strains not involved in salmonellosis cases or outbreaks may be related to food sources. Among the strains not associated with foodborne salmonellosis, 81.8% were isolated from poultry and related products, whereas among the strains involved in salmonellosis, only 21.4% were obtained from these products. In Brazil, resistant S. Enteritidis has been frequently isolated from foods of animal origin, mainly poultry-related products[1,16,36,41,52]. Concerning nalidixic acid, a total of 70% of strains not associated with salmonellosis were resistant to this antibiotic, while 25.7% of the strains involved in salmonellosis presented this phenotype, certainly due to the higher number of strains not associated with salmonellosis isolated from poultry and poultry-related products. On the other hand, similar resistance to streptomycin and gentamycin was observed in strains associated or not with salmonellosis: 11.4% and 15% for streptomycin, respectively, and 11.4% and 12.5% for gentamycin, respectively.

The virulence gene invA was detected in all strains, which is consistent with the findings from previous reports[7,28,40]. InvA is conserved among Salmonella serotypes and is a useful marker for molecular detection of this pathogen by PCR[15,31].

The genes spvC, sefA and pefA were observed in 48.1%, 31.6% and 44.3% of the strains, respectively. Studies have shown a heterogeneous distribution of these virulence factors in Salmonella isolated from different origins[4,53,58].

The spvC gene has an important role in systemic infection of Salmonella [24,34] and in the present survey this gene was found in 80.9% (34/42) and 41% (80/195) of the strains associated and not associated with foodborne salmonellosis, respectively (Table 3). The spv operon is located in virulence plasmids and limited to specific serotypes (i.e. Enteritidis, Typhimurium, Dublin, Cholerae-suis, Gallinarum, Pullorum and Abortus-ovis)[44]. Among the strains associated with salmonellosis, the serotypes that commonly have plasmids (S. Enteritidis and S. Typhimurium) corresponded to 86% of the analyzed strains, whereas among those not involved in salmonellosis, these two serotypes corresponded to only 27.7% of the strains.

Table 3

Prevalence of the spvC, sefA and pefA genes among Salmonella spp. strains isolated from foods associated and not associated with foodborne salmonellosis

Serovar	Strains associated with food poisoning				Strains not associated with food poisoning
	n	positive			n	positive
		spvC	sefA	pefA		spvC	sefA	pefA
S. Enteritidis	35	33 (94.3%)	35 (100%)	34 (97.1%)	40	35 (87.5%)	40 (100%)	38 (95%)
S. Typhimurium	1	1 (100%)	0	1 (100%)	14	10 (71.4%)	0	9 (64.3%)
S. Agona	1	0	0	0	9	3 (33.3%)	0	1 (11.1%)
S. Infantis	1	0	0	0	8	3 (37.5%)	0	4 (50%)
S. Brandenburg	1	0	0	0	7	3 (42.8%)	0	3 (42.8%)
S. Saintpaul	1	0	0	0	2	0	0	0
S. enterica I 6,7:r:-	1	0	0	0	4	1 (25%)	0	1 (25%)
S. Sandiego	1	0	0	0	3	0	0	0
S. Panama	0	n.a.	n.a.	n.a.	11	7 (63.6%)	0	0
S. Anatum	0	n.a.	n.a.	n.a.	12	9 (75%)	0	2 (16.7%)
S. Hadar	0	n.a.	n.a.	n.a.	3	2 (66.7%)	0	3 (100%)
S. Heidelberg	0	n.a.	n.a.	n.a.	6	1 (16.7%)	0	1 (16.7%)
S. enterica I 4, 5, 12:i:-	0	n.a.	n.a.	n.a.	4	0	0	2 (50%)
S. Bredeney	0	n.a.	n.a.	n.a.	7	5 (71.4%)	0	5 (71.4%)
S. enterica I 4,12:i:-	0	n.a.	n.a.	n.a.	2	1 (50%)	0	1 (50%)

n.a. = non applicable

The sefA gene was detected in 83.3% (35/42) and 20.5% (40/195) of the strains associated and not associated with salmonellosis, respectively. This gene was only present in S. Enteritidis (Table 3), which is an expected result since the genes of the sef operon (sefABCD) are restricted to the serotypes of serogroup O:9, which include S. Enteritidis, S. Typhi, S. Dublin, S. Berta, S. Gallinarum and others[14]. Due to the specificity of the sef gene, it has been used for the molecular identification of Salmonella Enteritidis[10,11,12].

The pefA gene was present in 83.3% (35/42) and 35.9% (70/195) of the strains associated and not associated with foodborne salmonellosis, respectively (Table 3). Like the spvC gene, pefA is located in virulence plasmids, and limited to certain serotypes such as Enteritidis, Choleraesuis and Typhimurium[2].

All S. Enteritidis isolated from samples associated or not with foodborne salmonellosis contained the invA and sefA genes. The prevalence of invA and sefA genes in strains associated (94.3% and 97.1%, respectively) and not associated with salmonellosis (87.5% and 95%, respectively) was similar. Regarding these four genes, S. Enteritidis strains evaluated in this study were grouped into three profiles (Table 4), with a predominance of strains in the P1 profile (presence of the four virulence genes). These results highlight the pathogenic potential of S. Enteritidis, which is the major serotype associated with human salmonellosis in Brazil[19].

Table 4

Virulence profile of Salmonella Enteritidis strains associated and not associated with foodborne salmonellosis

Profile	Number of isolates		Genes
	Associated with food poisoning	Not associated with food poisoning	spvC	invA	sefA	pefA
P1	33 (94.3%)	35 (87.5%)	+	+	+	+
P2	1 (2.8%)	2 (5%)	-	+	+	+
P3	1 (2.8%)	3 (7.5%)	-	+	+	-

This survey illustrates the occurrence of antibiotic-resistance and a wide distribution of virulence genes in Salmonella spp. regardless of their association with foodborne salmonellosis. Resistance and virulence determinants were detected in strains belonging to several serotypes. The high rates of antibiotic-resistance in strains isolated from poultry demonstrate the potential risk associated with the consumption of these products and the need to ensure good food hygiene practices from farm to table in order to reduce the spread of antibiotic-resistant microorganisms from animals to humans.

The development of antimicrobial resistance in zoonotic bacteria, such as non-typhoidal Salmonella, constitutes a public health risk with many consequences for human health. As bacteria become increasingly more resistant to drugs of clinical importance, there is an increased risk of treatment failure due to limited therapeutic choices[39]. Furthermore, antibiotic-resistance and virulence determinants may be present in the same plasmid, which may be selected by antibiotic pressure[6,27,37] resulting in increased systemic infections and hospitalizations of patients infected with resistant non-typhoid Salmonella [51].