Safety Evaluation of the Coagulase-Negative Staphylococci Microbiota of Salami: Superantigenic Toxin Production and Antimicrobial Resistance

The risks of contracting staphylococci food poisoning by the consumption of improperly manufactured salami and the possibility of this food being reservoirs for antibiotic resistance were evaluated. Nineteen coagulase-negative staphylococci (CNS) strains were found in commercial and artisanal salami. The species in commercial salami were S. saprophyticus, S. sciuri, S. xylosus, and S. carnosus. Artisanal salami showed S. succinus, S. epidermidis, and S. hominis but no S. carnosus. Phylogenetic analyses grouped the strains into three major staphylococcal species groups, comprised of 4 refined clusters with similarities superior to 90%. Fifteen strains harbored multiple enterotoxin genes, with high incidence of seb/sec and sea, 57% and 50%, respectively, intermediate incidence of sed/seh/selm and sei/seln/tst-H, 33% and 27%, correspondingly, and low incidence of see/selj/selo and seg, of respectively 13% and 1%. Real time RT-PCR and enzyme-linked-immunosorbent assays confirmed the enterotoxigenicity of the strains, which expressed and produced enterotoxins in vitro. The CNS strains showed multiresistance to several antimicrobials of therapeutic importance in both human and veterinarian medicine, such as β-lactams, vancomycin, and linezolid. The effective control of undue staphylococci in fermented meat products should be adopted to prevent or limit the risk of food poisoning and the spread of antimicrobial-resistant strains.


Introduction
Staphylococcal food poisoning is an illness caused by the ingestion of contaminated food containing enterotoxins produced by bacteria belonging to this genus. Enterotoxins that exhibit superantigenic activities are heat stable proteins and may not be destroyed even during cooking conditions.
In Brazil, according to data from the Ministério da Saúde (Ministry of Health), staphylococcal poisoning is the second most common foodborne disease, ranking only after outbreaks involving Salmonella spp. [1]. Staphylococcus classified as coagulase-positive are considered potential food enterotoxin-producing strains [1], although, recently, the enterotoxigenic potential of coagulase-negative staphylococci (CNS) species in food poisoning has also been recognized [2].
Initially, enterotoxin SEs family members were divided into five serological types (sea through see) based on their antigenicity [3,4]. In recent years, however, newly described types of SEs-SEG, SEH, SEI, SElJ, SElK, SElL, SElM, SElN, SElO, SElP, SElQ, SElR, and SElU-with amino acid sequences similar to the classical SEs, were discovered. These newly described enterotoxins are designated as SE or SElike (SEl), according to their emetic properties displayed in a primate model following oral administration [5].

BioMed Research International
The toxic shock syndrome toxin-1 (TSST-1) is also a member of the SE-related toxin family and has the ability to stimulate large populations of T cells containing a particular V element in their T-cell receptors (TCR). Like other superantigenic toxins, it bypasses normal antigen presentation by binding to class II major histocompatibility complex molecules on antigen-presenting cells and to specific variable regions on the beta-chain of the T-cell antigen receptor. Through this interaction, a massive proliferation of T cells at orders of magnitude above antigen-specific activation occurs, resulting in a massive cytokine release that is believed to be responsible for the most severe features of TSST [6].
Salami is a kind of dry sausage obtained by the microbial fermentation of raw pork meat, using Staphylococcus starter cultures as technological accessories to ferment the product and give it its organoleptic characteristics. In Brazil, Italian type salami, similar to salami produced in Southern Europe, is avidly consumed, with a production trade of 13.093 tons between 2000 and 2014 [8].
S. xylosus, S. equorum, and S. carnosus are part of the starter culture microbiota that participate in the reactions required for creating the flavor and aroma during the maturing period of fermented meat production [9]. In addition, other species, such as S. epidermidis, S. pasteuri, S. sciuri, and S. succinus, may also occasionally be present in meaningful amounts [10].
However, even the combination of physical and chemical barriers cannot always guarantee the stability and microbial safety of starter cultures. Contamination of salami fermentation starter culture microbiota by pathogenic coagulasenegative staphylococci (CNS) strains is perhaps the most harmful factor in the production of cured meat products, since these pathogens are able to produce heat-stable enterotoxins with superantigenic activities in food matrices [11,12].
Staphylococci species are not usually identified at the species level by routine laboratory testing and commercial kits, since phenotypic discrimination cannot reliably identify these species due to the variable expression of some phenotypic traits [13]. For this purpose, molecular techniques, including nucleotide sequencing within the 16S rDNA, hsp60, tuf, sodA, and rpoB genes, have been successfully used to identify Staphylococcus species [14].
Depending on the conditions, some species of coagulasenegative staphylococci can present health risks, since they have shown resistance to several antibiotics of therapeutic importance, such as -lactams [15].
The aim of the present study was to identify the members of the CNS microbiota from salami. We sampled the salami marketed in Brazil, comparing the CNS microbiota in salami produced by industrial companies and in artisanal salami manufactured by small producers. The CNS strains were identified by sequencing of a 16S rDNA region and the phylogenic relationships between the observed species were established. The presence of multiple genes encoding the classical and newly described se/sel and tstH1 toxins in the CNS genomes was investigated. The risk of food poisoning was assessed by evaluating the ability of the CNS strains in transcribing and expressing the classical and newly described enterotoxins in vitro by using real time RT-PCR and enzymelinked immunosorbent assay (ELISA). The resistance of the isolated strains to antimicrobial agents of therapeutic importance in staphylococci infections was also evaluated.

Isolation of Bacterial Strains.
Six samples of distinct brands of salami, 03 from the meat industry and 03 from small artisanal producers, were collected in the municipality of Rio de Janeiro, Brazil. Twenty-five grams of salami was added to 225 mL of 0.1% peptone water. The suspensions were transferred to homogenizer bags (Interscience, Saint Nom, France) and coupled to a Stomacher @ 400 circulator (Seward, Worthing, West Sussex, UK) at 260 rpm for 1 min. The suspensions were serial-diluted from 10 −6 to 10 0 and 100 L of each dilution was transferred onto 20 L of Baird-Parker agar containing egg yolk tellurite emulsion (BPAþ RPF, bioMerieux, France). Eighty presumptive coagulase-negative staphylococci colonies were tested by Gram-staining, catalase, coagulase, and thermostable DNAse activities according to Bergey's Manual of Systematic Bacteriology. Sixty presumptive CNS strains were stored at −80 ∘ C in tryptone soy agar (TSA, Franklin Lakes, New Jersey, USA) plus 45% v/v glycerol.

DNA Preparation.
The strains were cultured aerobically overnight in 10 mL Brain Heart Infusion broth (BD BBL, Le Pont de Claix, France) at 37 ∘ C for 24 h. The suggestive CNS colonies were adjusted to 10 6 UFC/mL in a spectrophotometer and harvested by centrifugation at 5,700 × g for 1 min. The cell pellet was used for DNA extraction using the DNeasy blood and tissue kit (Qiagen, Dusseldorf, Germany), following the manufacturer's instructions. Genomic DNA was quantified using the Qubit fluorometer (Invitrogen, Grand Island, New York, USA) and Qubit assay kits.  (Table 1), 0.5 U Taq DNA polymerase (Invitrogen, Grand Island, New York, USA), and 100 ng of DNA templates. Uniplex-and duplex-PCR assays were performed under the following conditions: 94 ∘ C for 5 min followed by 35 cycles of 94 ∘ C for 2 min, 53 ∘ C for 2 min, and 72 ∘ C for 1 min for extension, ending with a final extension at 72 ∘ C after 7 min [16], with modifications in the annealing temperature, using a thermal cycler (MyCycler, Bio-Rad,

Enterotoxin Expression
Assays. The observed strains were cultured aerobically overnight in 10 mL Brain Heart Infusion Broth (BD BBL, Le Pont de Claix, France) at 37 ∘ C for 72 h. Bacteria supernatants were collected by centrifugation at 4,000 ×g for 10 min and used for the detection of sea, seb, sec, sed, and see by an ELISA assay using a commercial detection kit (RIDASCREEN SET A, B, C, D, E Art. number R4101, R-Biopharm AG, Germany). The assay was performed according to the manufacturer's recommendation and as described elsewhere [16]. The mean lower limit of detection of the assay was 0.25 ng mL −1 . The threshold is defined as the average OD of two negative controls plus 0.15, a constant established by the kit. Samples containing SEs showed absorbance values equal to or greater than the threshold value. All experiments were performed in duplicate.

Real Time RT-PCR Assays.
Total RNA was extracted by using the QIAGEN RiboPure Bacteria kit (Life Technologies, Carlsbad, California, USA) following the manufacturer's instructions and quantified using the Qubit fluorometer (Invitrogen, Grand Island, New York, USA) and Qubit assay kits. The cDNA synthesis was performed by using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, California, USA) and the ABI PRISM 7500 Fast RT-PCR system (Applied Biosystems, California, USA). Samples were plated in triplicate in 96-well plates as follows: 12 L of the SYBR Green PCR Master Mix; 1 L of primer mix (sea, seb, sec, sed, and see); and 4.5 L of the cDNA ultrapure water in each well. Amplification was performed under the following conditions: 95 ∘ C for 15 min, 40 cycles at 95 ∘ C for 15 s, 54 ∘ C for 30s, and 72 ∘ C for 30 s. The dissociation curve was performed at 95 ∘ C for 15 sec, 54 ∘ C for 30 sec, and 95 ∘ C for 15 sec. CT means, the standard deviations, and the cDNA semiquantification were calculated using the Graph-Pad Prism 5 software package. Calibration curves based on five points were constructed in triplicate corresponding to serial dilutions (1, 1 : 10, 1 : 100, 1 : 1000, and 1 : 10000) from 100 ng of a DNA template stock solution.
2.6. 16S rDNA Sequencing. Amplification of the V5 region of 16S rDNA fragment was performed using 50 ng of DNA templates from the 65 strains found in the salami samples. PCR was performed under the following conditions: 95 ∘ C for 10 min, followed by 30 cycles at 95 ∘ C for 30 s, 60 ∘ C for 30 s and 72 ∘ C for 45 s, and a final extension at 72 ∘ C for 10 min. PCR products were purified using the PCR DNA Purification Kit (Applied Biosystems, California, USA) and sequenced using 20 ng purified DNA and 13 L of primer sets in a final volume of 20 L. After amplification, products were purified according to the protocol of the BigDye Terminator Purification X Kit (Applied Biosystems, California, USA) and sequenced on a 3130 sequencer Genetic Analyzer (Applied Biosystems, California, USA). The sequences were compared to the 16S rDNA gene sequences of Staphylococcus species available at the GenBank database (http://www.ncbi.nlm.nih.gov/Genbank/index.html). Multiple sequence alignments were performed using Clustal W (Kyoto University, Bioinformatics Center; http://www .genome.jp/tools/clustalw/).

2.7.
Sequencing of Enterotoxin PCR Products. PCR products were purified using the PCR DNA Purification Kit (Applied Biosystems, California, USA) and sequenced using 10 ng of purified DNA and 3.2 pmoles of each primer set in a final volume of 20 L. After amplification in the same conditions as the PCR step (Section 2.3.3), products were purified according to the BigDye Terminator Purification X Kit protocol (Applied Biosystems, California, USA) and sequenced on a 3130 sequencer Genetic Analyzer (Applied Biosystems, California, USA). The sequences were compared to Staphylococcus aureus and Staphylococcus pasteuri gene sequences available at the GenBank database (http://www.ncbi.nlm.nih.gov/Genbank/index.html). Multiple sequence alignments were performed using Clustal W (Kyoto University, Bioinformatics Center; http:// www.genome.jp/tools/clustalw/).

Phylogenetic Analyses.
Phylogenetic relationships between the CNS strains were performed by sequence alignments using the Clustal X 2.0 software package [24]. The phylogenetic trees were constructed using the software Mega 6.0 and UPGMA methods [25].

Minimal Inhibitory Concentration (MIC) Determinations.
The MICs of vancomycin, linezolid, methicillin, and ampicillin were determined by the macrodilution broth method based on CLSI recommendations, using in-houseprepared panels [26]. Antibiotic concentrations of 0.03, 0.06, 0.125, 0.25, 0.5, 1.0, and 2.0 mg mL −1 were tested. One mL of broth was transferred to the tubes and 100 L of the bacteria suspension was adjusted to 10 6 CFU/mL in saline 0.85% according to a 0.5 McFarland scale and transferred to tubes containing 1 mL of each antimicrobial. Strains were grown in Mueller-Hinton broth (BD BBL Franklin Lakes, New Jersey, USA) and the MIC was estimated as the lowest antibiotic concentration that inhibits visible growth after 24 h [26].

Isolation and Identification of Coagulase-Negative Strains from
Salami. Sixty-five presumable coagulase-negative staphylococci microorganisms from salami were isolated by colony morphology, coagulase slide test, subsequent tube test, and biochemical tests. The sequencing of the V5 region of the 16S rDNA fragment of the strains was discriminative enough to differentiate the Staphylococcus isolated from salami at the subspecies level, with the exception of 1 strain, identified up to the Staphylococcus spp. genus.
S. carnosus and S. xylosus are commonly used as commercial starter cultures for sausage manufacturing [26], but S. succinus has also been observed in dry fermented sausages and its use as a starter culture has been already proposed [27].
The diversity of CNS microbiota found in the samples analyzed in the present study could be related to the origin of the salami. The microbiota from the artisanal salami showed greater biodiversity when compared to the commercial salami. In commercial salami, S. saprophyticus was the predominant species, but S. xylosus and S. carnosus, also observed in commercial salami, are commonly isolated from starter cultures [26]. The predominance of S. saprophyticus followed by S. xylosus has also been reported in salami from South Italy [28], similar to Belgian sausages, where S. saprophyticus was the most frequently detected species [29].
S. carnosus was not observed in artisanal salami, but S. succinus, S. epidermidis, and S. hominis were detected.
During the last decades S. epidermidis and S. saprophyticus have been described as emerging pathogens [34]. S. saprophyticus is considered a frequent contaminant of fermented sausages and raw meats and has been isolated from rectal swabs of cattle carcasses and pigs. In humans, the main reservoir of S. saprophyticus is the gastrointestinal tract [35]. S. saprophyticus and S. epidermidis can be opportunistic pathogens, isolated from the human urinary tract, and the presence of these species in food should be taken into account concerning possible contamination of the starter inoculum and/or improvements in the salami manufacturing process [13].
Other CNS species observed in the present study are mainly associated with ordinary food contaminants, with S. epidermidis and S. hominis being the dominant species in human skin and occasionally isolated from the skin of domestic animals [36].
The subclusters of those clusters showed mismatches of species belonging to the four species groups. The first subcluster grouped S. epidermidis HF088211.1, S. saprophyticus HQ699510.1, and S. carnosus NR116434.1 and the second subcluster grouped S. hominis JX519988.1 and the S. saprophyticus EU430992.1, all of them originally isolated from animals and human beings. The third subcluster includes species previously isolated from the marine environment, such as S. xylosus KF198080.1, while S. carnosus KJ862002.1 is utilized as a probiotic organism in foods. The fourth cluster grouped the S. sciuri JX966436.1 and S. saprophyticus HF937252.1 strains, which are homologous to species found in soil.
The CNS strains clustered into groups near the bottom of the phylogenetic tree are mostly strains found in artisanal salami, whereas the species at the top of the tree are mainly CNS strains found in commercial salami ( Figure 1).
The close similarities between the S. saprophyticus AB697517. 1

Genotypic and Phenotypic Characterization of CNS
Strains. Twelve distinct combinations of staphylococcal enterotoxins genes were found in the 15 CNS strains, comprising SEs A-E, G-J, and also the enterotoxin-like toxins (SElL) K-R and U (Table 2). Fifteen strains (79%) carried at least one gene encoding enterotoxins in their genomes (Figure 2). The seb and sec genes were the most predominant, harbored by 57% of the strains, followed by sea, carried by 50%, whereas the sed/seh/selm genes showed intermediate incidence harbored by 33% of the strains, while sei/seln and tsH1 were found in 27% of strains. Finally, see, selj, and selo showed low incidence (13%) and seg was carried by only 1% of strains ( Table 2).
The relationships between superantigenic toxin genotypes and toxin gene-encoding mobile genetic elements in CNS strains were evaluated. Distinct combinations of SaPI and plasmids or plasmids and genes of egc operon enterotoxins were found in the CNS strains obtained from salami.
Four strains presented only the classical enterotoxin genes, namely, S. saprophyticus KJ004623.1 and HF937252.1, S. xylosus CP008724.1, and S. carnosus NR116434.1. Another strain, S. saprophyticus AB697717.1, presented only the newly described enterotoxins seh, sei, and selm and, finally, 10 strains presented a combination of classical and newly described enterotoxin genes in their genomes, with at least one of each enterotoxin type (Table 2).
Previous studies have shown that sea is the most common toxin associated with Staphylococcus food poisoning, followed by sed and see, with SEH and SEI being considered as playing only a minor role [39]. Among the 15 enterotoxigenic strains found in salami, 73% of them harbored at least the sea gene or combinations of these 05 genes. Strains S. carnosus KJ862002.1, found in commercial salami, and S. epidermidis KF600589.1, found in artisanal salami, carry 04 of these genes. The sea and seb enterotoxin genes are known to occupy the same locus on the chromosome, which may explain why these enterotoxins are commonly found together in food poisoning outbreaks [36]. The combination of sea and seb genes was found in 05 strains: S. saprophyticus EU430992.  shown to carry the selm, seln, and selo genes and the selm and seln sequences were detected in the genome of a single strain, S. saprophyticus subsp. bovis KJ699151.1. S. epidermidis KF600589.1 showed a combination of seg and sei genes and 03 strains harbored a single gene, sei or selm or seln, perhaps due to the high degree of genetic polymorphism in the chromosomal assembly [41]. Indrawattana et al., 2013, detected the seg-sei-selm-seln-selo of the highly prevalent egc locus was 26.1% in contrast to this study where 13% CNS presents the combination of selm, seln, and selo genes [42]. The CNS strains clustered into groups near the bottom of the phylogenetic tree carried the classical enterotoxin genes, while the species at the top of the tree showed high diversity among the enterotoxin genes, combining the classical and the newly described genes in their genomes (Figure 1).
To assess the risk of staphylococcal food poisoning, the ability of the identified strains in harboring the se and sel genes and in expressing and producing enterotoxins in vitro was evaluated. The mRNA for each enterotoxin gene was evaluated by real time RT-PCR assays and enterotoxin content was estimated by a sandwich enzyme immunoassay for the combined detection of Staphylococcus enterotoxins (SET) A, B, C, D, and E.
The fifteen enterotoxigenic CNS strains were able to express the classical enterotoxins SEA, seb, sec, sed, and see, in concentrations ranging from 0.3 ng mL −1 to 1.4 ng mL −1 , as assessed by the in vitro assays ( Table 2) Although the sandwich enzyme immunoassay is considered the most sensitive method to detect sea-see enterotoxins, able to detect 0.125 ng mL −1 , differences in the specificity and sensitivity of the assays for the detection of staphylococcal enterotoxins from foods are expected [43]. A single strain, Staphylococcus spp. KF135445.1, which harbors both the sea and seb genes, was unable to produce sea-see enterotoxins.
The mRNA levels evaluated by the real time RT-PCR assays for enterotoxins were detected in 12 of 15 strains (86%) ( expressed mRNA for multiple se and/or sel genes. There is a differential transcription among these genes, where the most frequent among the classical ones were seb and see/sea, transcribed by 04 and 02 strains, respectively, and the most frequent among the newly described genes were sei and seh/seln/selo, transcribed by 03 and 02 strains, respectively.
No mRNA for the sec gene was detected, although S. saprophyticus AB697717.1 was able to produce the enterotoxin in vitro, as detected by the ELISA assays.

Enterotoxin Gene Homologies between Salami CNS and CPS Strains.
The nucleotide sequencing of six enterotoxin genes-two of them, sec and see, encoding classical enterotoxins-and four of them, encoding the newly described genes, seg, seh, selm, and seln from CNS strains, was compared to the enterotoxin genes from S. aureus or S. pasteuri, two coagulase-positive staphylococci strains (Figure 3). The homology between the CNS and CPS enterotoxin genes varied from 65% to 98%. The homology of sec, seg, seh, selm, and seln between CNS from salami and S. aureus was 98%, 60%, 98%, 65%, and 70%, respectively. The homology between see from CNS found in salami and in S. pasteuri was 98%.
Although further studies should be performed, it seems that the sequences encoding enterotoxins can be conserved among coagulase-negative and coagulase-positive staphylococci, as shown in the phylogenetic analysis of the 06 enterotoxin genes (Figure 4).

Antimicrobial Multiresistance of CNS Strains (MRCNS).
Another safety hazard associated with CNS strains besides the ability of producing enterotoxins in food matrices is the antimicrobial resistance to antimicrobial agents commonly used to treat staphylococci infections. The antimicrobial resistance carried by CNS strains from food matrices can be spread to the population by the consumption of an apparently safe food.
Surprisingly, the resistance to chloramphenicol is very similar to that estimated for MRCNS strains isolated from human clinical samples. Indeed, there is no direct correlation between the researched features and the origin of the staphylococci strains, since different virulence factors are widespread, such as antibiotic resistance [44], reinforcing the fact that the ability of food matrices strains to produce SE and SEl and their multiresistance character must be considered when evaluating the safety hazards of food poisoning.
S. epidermidis KF600589.1 and S. hominis JX519988.1 showed the highest MAR indexes (0.93 and 0.80, resp.). These strains found in artisanal salami are homologous to strains isolated from human skin microbiota and should be carefully considered among the potential pathogenic staphylococci found in food from animal origins, which can be caused by contamination by poor hygienic conditions during salami manufacturing.
Depending on the conditions, some species of coagulasenegative staphylococci can present health risks, since they have shown resistance to antibiotics of therapeutic importance, such as beta lactams [15]. However, multiresistant strains like S. carnosus KJ1862002.1 and S. xylosus CP007208.1 and CP008724.1 were intentionally introduced in the food matrix, since they are part of the culture starter used in Italian-salami manufacturing. As previously discussed, food can be a reservoir of multiresistant microorganisms that can spread by the consumption of an apparently safe food. Due to the intensive and indiscriminate use of antibiotics for human and veterinarian therapeutic purposes, multiresistant staphylococci strains are being selected and reproduced in food matrices [45].
The cefoxitin and oxacillin disk diffusion tests are recommended for determining the resistance and susceptibility breakpoint for MRSA surveillance cultures [46]. In this study, there was a good correlation between the disc test   (Table 4).
Previous studies have demonstrated that linezolid is active against Gram-positive bacteria, including methicillinresistant staphylococci [47]. In the present study, 05 strains showed linezolid resistance in the disc diffusion test, but it was impossible to establish MIC values for S. epidermidis KF600589.1, Staphylococcus hominis JX519988.1, and Staphylococcus carnosus KJ862002.1. The remaining strains S. saprophyticus EU430992.1 and S. xylosus CP007208.1 presented an MIC for linezolid of 0.125 mg mL −1 and 0.25 mg mL −1 , respectively.
The MIC of 0.03 mg mL −1 for penicillin is in accordance with the previous values estimated for resistant S. aureus strains found in several food matrices, such as meat, dairy products, and ready-to-eat food [48].
There is still a lack of information on the antimicrobial resistance of staphylococci strains from food matrices, although they are the most worrisome vehicles of dissemination of antibiotic-resistant pathogens [49].
The high resistance found for salami staphylococci strains can be ascribed to their inappropriate use as growth promoters of antimicrobial agents like oxacillin, vancomycin, chloramphenicol, neomycin, and erythromycin, which are commonly used in veterinary medicine to treat infections [50]. Antimicrobial therapy of infections staphylococci is based on results of susceptibility tests in vitro [51]. Methicillinresistant coagulase-negative Staphylococcus spp. found in ready-to-eat products such as meats, fish, and dairy products also offer risks and, although they are not classical food poisoning bacteria, their presence in food offers significant risks to public health due to the possible spread of antibiotic resistance [52].
The standardization of salami quality should include diagnostic methods to screen and quantify the presence of classical and newly described enterotoxins directly in the food matrices as a routine procedure to be conducted by the meat product industry in Brazil.
The safety of salami consumption could also be enhanced by the inclusion of a microbial barrier such as the inclusion of probiotic strains producing natural antibiotics or competitive flora or even the addition of natural bioagents against spoilage or pathogenic microorganisms.