Toxin gene profile and antibiotic resistance of Staphylococcus aureus isolated from clinical and food samples in Egypt Rehab

The antimicrobial sensitivity, toxin profiles, and agr genotyping of 70 Staphylococcus aureus isolates were determined. The evaluation of 10 antimicrobials showed that 88.5, 52.8, 40, and 25.7% of isolates were resistant to ampicillin, tetracycline, cefoxitin, and oxacillin, respectively. All isolates were sensitive to gentamicin. The prevalence of staphylococcal enterotoxin (SE) genes (sea, seb, sec, sed, see, seg, sei, and seh) and the toxic shock syndrome (tsst-1) gene were estimated by polymerase chain reaction (PCR); 95.7% of S. aureus carried the SE genes. The predominant gene was sed (75.7%), followed by sea and tsst-1 (58.5%), then see and sei (51.4%). The tsst-1 gene was found at a significantly higher rate among food isolates than clinical isolates (P= 0.003). The agr types were identified by multiplex PCR; agr II was more prevalent (58.5%) than agr I (25.7%) and agr III (20%).


INTRODUCTION
The Gram-positive bacterium Staphylococcus aureus is an opportunistic pathogen that is highly dangerous to human health.It can be identified in the anterior nares and skin of human (Kluytmans et al., 1997).This organism has the ability to cause infection of almost every tissue and organ system in human body, doing so by exporting a large array of virulence factors (surface proteins, enzymes and toxins) to the cell surface and extracellular environment of the human host (Vandecasteele et al., 2009).Both high virulence and rapid development of antimicrobial resistance contribute to the pathogenicity of S. aureus (Arvidson and Tegmark, 2001).Although pathogenesis of S. aureus is a multifactorial process that depends on multiple virulence factors' expression, some disease symptoms are particular to specific exotoxins including toxic shock syndrome toxin (tsst-1), enterotoxins (SEs) and exfoliative toxins (ETs) (Dinges et al., 2000).
S. aureus represents a major cause of food poisoning due to SEs produced by some strains.S. aureus produces large array of toxins (23 serologically different toxins) including staphylococcal enterotoxins (SEs) and staphylococcal enterotoxin-like proteins (SEls) (Hu and Nakane, 2014).SEs and SEls can be subdivided into classical (SEA to SEE) and new (SEG to SElX) types.SEs are considered the major cause of food poisoning, because they are resistant to heat and to proteolytic enzymes like trypsin, pepsin and renin; so, pass both cooking process and move through the gastrointestinal tract without losing their activity.They cause systemic effects such as high fever, abdominal cramps, vomiting, and diarrhea (Fernandez et al., 2006).These symptoms appear within few hours (1 to 6 h) of ingestion of food contaminated with very small quantities of SEs (20 ng to 1 μg/1 g of food) (Le Loir et al., 2003;Di Giannatale et al., 2011).Usually, infected individual recovers within one to two days, but some cases require hospitalization.
Food handlers are the primary source for contamination of foods with S. aureus.The foods that are commonly mentioned in staphylococcal food poisoning cases differ greatly from one country to another due to different eating and consumption habits.Contaminated raw meat is considered a major source of such type of food poisoning.The toxic dose of enterotoxigenic strains of S. aureus can be reached easily in ground meat before its consumption (Zargar et al., 2014).
The tsst-1 produced by some strain of the S. aureus is also a member of the superantigen family.It is responsible for toxic shock syndrome in all menstrual cases as well as 50% of non-menstrual ones.This toxin can affect immunity of infected individual causing it to diminish (Podbielska et al., 2011).
The objectives of this study were (i) to study the resistance profile of the isolates, (ii) to investigate the distribution of genes encoding SEs and tsst-1 toxin by polymerase chain reaction (PCR), and (iii) genotyping of isolates.

Bacterial strains
During the period of January 2013 to January 2014, 370 samples were collected from clinical (70) and food (300) sources.Clinical samples (diabetic foot infections, wounds and abscesses) were collected from patients in Mansoura University hospitals, while food samples (75 ready to eat meat, 50 fresh meat, 75 mince beef and 100 beef burger) were collected from supermarkets and butchers shops distributed in Mansoura, Egypt.All samples were processed properly and evenly spread onto dried surfaces of Mannitol salt agar and Baird-Parker agar supplemented with egg-yolk tellurite emulsion, then incubated at 37°C for 24 to 48 h (Khalifa et al., 2014).Colonies exhibiting characteristic morphology of S. aureus were randomly selected and subjected to Gram stain, tests for catalase, coagulase enzymes and finally genotypic confirmation through PCR detection of S. aureus specific nuc gene (Brakstad et al., 1992).The experimental protocol conducted in the study was approved by the Ethical Committee of Faculty of Pharmacy, Mansoura University, Egypt, with code (2015-55).
All discs were purchased from Oxoid (U.K).After 18 to 24 h incubation at 37°C, the zone of inhibition was measured and isolates were recorded as sensitive or resistant according to the interpretation criteria of CLSI (2014).

DNA extraction
DNA was extracted by making a suspension of single colony from each isolate in 100 µl DNase/RNase-free water.Then, suspensions were maintained in a boiling water-bath for 10 min to lyse the cells, chilled on ice and centrifuged.Supernatants containing extracts of DNA were transferred to new eppendorf tubes and stored at -20°C for subsequent PCR amplification (Englen and Kelley, 2000).

PCR detection of toxins genes
The primers used for detection of SEs and tsst-1 genes were specifically designed and ordered from Invitrogen (U.K).Primers sequences and expected amplicon sizes are shown in Table 1.The enterotoxin genes (sea, seb, sec, sed, see, seg, seh and sei) and tsst-1 gene were detected using 25 μl reaction: 12.5 μl Dream Taq™ Green PCR Master Mix (2X) (Fermentas, U.K), 1 μl of forward primer (10 μM), 1 μl of reverse primer (10 μM), 3 μl of the DNA extract and 7.5 μl of nuclease free water.The PCR reactions were started with initial denaturation step at 95°C for 5 min, followed by 40 cycles of denaturation at 95°C for 30 s, annealing at 54°C for sea, seb, sec, sed, see, seg, seh and tsst-1 and 56°C for sei for 30 s and extension at 72°C for 1 min and final extension at 72°C for 5 min.Positive control for the PCR-based assays were included with each primer set.For negative control, ddH 2 O was used instead of DNA extract.
Following amplification reactions, the PCR products were analyzed via electrophoresis in a 2% agarose gel (stained with ethidium bromide) at 100 V for 45 min in 1 X Tris-borate-EDTA buffer and illuminated under UV transilluminator and photographed.The Gene Ruler™ 100 bp Plus DNA Ladder (MBI Fermentas, St. Leon-Rot, Germany) was used as a DNA size marker.A reference strain for the seh gene was not available; therefore, the presence of a band at the expected product size was considered a positive result.

agr genotyping
The agr allele types (I to IV) were detected by multiplex PCR as described previously (Gilot et al., 2002;Xie et al., 2011).Positive control strains for agr allele types (I and IV) were included in each run of PCR.Reference strains for agrII and agrIII were not available, so the presence of bands at the expected product size was considered a positive result.
Electrophoresis agarose gels of toxins and agr groups were analyzed visually and scored using a binary code that was analyzed using DendoUPGMA.A dendogram was constructed applying the unweighted pair group method with arithmetic mean Table 1.Primers sequences used for PCR amplification of the tested genes.

Gene Sequence
Amplicon size (bp) Reference

Statistical analysis
Data analysis was performed using GraphPad Instat 3.10, chisquare test was used.The difference is considered significant if the P value ≤0.05.

Isolation of S. aureus
A total of 70 strains were separated and identified as S. aureus [35 strains out of 300 food samples (11.7%) and another 35 strains out of 70 clinical samples (50%)].The genotypic identification was in accordance with the phenotypic characterization results.

Antimicrobial susceptibility test
The resistance percentage of S. aureus isolates to different antimicrobials is shown in Table 2.No significant association was detected between isolates sources and antimicrobials resistance (p >0.05 for each antimicrobial).A multidrug resistance phenotype (resistance to ≥3 classes of antimicrobial agents) was exhibited by 12 (17.1%)isolates; 7 of them were derived from food.Forty percent of the tested isolates were found to be methicillin resistant (MRSA).Methicillin resistance was equally distributed among the food and clinical isolates.Resistance to large number of tested antimicrobials (3 to 5 and 6 to 8 antimicrobials) was mostly associated with  MRSA regardless of the source of isolates (P<0.0001).
There was a wide range of resistance patterns among isolates, as 24 different patterns were observed (Table 3).Nine patterns were found among isolates of both origins (A1, A2, A3, A4, A5, A8, A13, A17 and A19).The most common resistance pattern was A1 that was exhibited by 22.8% of isolates showing resistance to ampicillin only.

Detection of SE and tsst-1 genes
The PCR amplification of toxin genes showed that one or more toxin genes were carried by 95.7% of the isolates (Figure 1).The sed gene was the most abundant toxin gene among isolates with a rate of 75.7%, while seh gene was the least frequently detected one (11.4%)as shown in Table 4.No significant differences were found between isolates from food and clinical origins except for the tsst-1 gene where it was significantly higher in the food isolates (P=0.003).Through analysis of PCR results, forty nine toxin genes patterns were detected (Table 5).Seven toxicity patterns were exhibited by both food and clinical isolates (T1, T2, T8, T14, T37, T41, and T45).While 19 toxicity patterns were found in food isolates

agr typing
The majority of our isolates (58.5%) belonged to agr type II, followed by agr type I (25.7%) and agr group III (20%).None was positive for agr type IV.
In the 70 S. aureus isolates, the constructed dendrogram showed two large clusters (Figure 2).Cluster I comprises 42 isolates of agr II type (except one isolate of agr III type).It was subdivided into five smaller clusters (A to E).The second large cluster II contained 28 isolates of either agr I, agr III or mixed agr I/III which was subdivided into three smaller clusters (F to H).The number of isolates in each small cluster varied greatly.
Whereas cluster A contained only two isolates, clusters H comprised 17 isolates.Each of clusters E and H had three pairs of isolates that showed 100% identity.Moreover, clusters C, D and H contained 2 mini clusters each and only one mini cluster in cluster E. These mini clusters showed more than 70% similarity.

DISCUSSION
In this study, 70 S. aureus strains were isolated from 370 samples (18.9%).During sampling, high prevalence rate of S. aureus was expected originating from hands of food handlers, because of poor hygiene processes employed, but only 11.7% of food samples were positive for S. aureus.On the other hand, 50% of clinical samples were positive for S. aureus which may indicate the ease of transmission of this microbe between patients.Similar finding were reported previously on prevalence of S. aureus in food (Normanno et al., 2007) and clinical samples (Ahmed et al., 2014).
Ninety five percent of tested isolates showed antimicrobial resistance to at least one antimicrobial; this is similar to previous studies (Normanno et al., 2007;Rhee and Woo, 2010;Aydin et al., 2011).Resistant to ampicillin was 88.5% which is consistent with previous reports that showed high resistance rates of S. aureus (>90%) to ampicillin (Pereira et al., 2009;Daka and Yihdego, 2012).
The prevalence of MRSA has increased globally (Jones et al., 2003).In this study, the rate of MRSA occurrence was found to be 40%; distributed equally between food and clinical isolates.Pereira et al. (2009) found MRSA among food strains in an approximately the same percent (Pereira et al., 2009).Lee do et al. ( 2008) reported a higher occurrence rate of MRSA among Staphylococcus isolated from clinical samples and raw meats (Lee do et al., 2008).
Resistance profile recorded in our study revealed that multidrug resistance to at least 3 antimicrobials was associated with MRSA.These results were in accordance with two previous studies in 2009 (Al-Khulaifi Manal et al., 2009;Galkowska et al., 2009).An explanation is that all MRSA strains have SCCmec (staphylococcal cassette chromosome mec) which act as reservoir for the different Staphylococcal genes, especially those encoding resistance (Al-Khulaifi Manal et al., 2009).Also, 24 resistance patterns were demonstrated by isolates.Ampicillin and ampicillin/tetracyclin resistance was the most recorded pattern exhibited by isolates (22.8 and 15.7%, respectively).
Previous studies reported that microbial isolates from food acquired resistance against most antibiotics (Valsangiacomo et al., 2000;Yucel et al., 2005).This observation was already recorded in our study where strains from food samples showed high resistance levels to antimicrobials similar to clinical ones.This may be explained by the use of antimicrobials in food production and veterinary medicine resulting in increased resistance to antimicrobials used for human therapy (Yucel et al., 2011).
PCR-based detection of enterotoxins sea-see, seg-sei and tsst-1 genes showed that one or more toxin genes were carried by 95.7% of isolates.Similar results were reported previously (Pu et al., 2011;Xie et al., 2011).In our study, sed gene was the most prevalent toxin gene among isolates (75.7%).This finding agrees in part with Normanno et al. (2007) who reported prevalence of sed gene compared to other SEs detected.Several studies reported higher prevalence to other toxin genes such as sea, sec and tsst-1 in the tested isolates (Klotz et al., 2003;Chiang et al., 2008;Pumtang-on et al., 2008).The difference in prevalence of superantigenic toxin genes may be attributed to that enterotoxin genes as well as tsst-1 gene are mostly found on mobile genetic element (Balaban and Rasooly, 2000) which facilitate its transfer among the same or different species of bacteria (Varshney et al., 2009;Malachowa and DeLeo, 2010).So, prevalence of such toxins may depend on the possibility of its transfer between isolates.In accordance with Alibayov et al. (2014) seh gene was the least prevalent gene detected.
The coexistence of seg and sei genes was recorded in only 10% of our isolates.This finding was partly consistent with Mashouf et al. (2015) where seg and sei genes were detected separately.Several authors have reported the presence of seg and sei genes individually or in different combinations with other SEs genes (Cremonesi et al., 2007;Zouharova and Rysanek, 2008;Arcuri et al., 2010).The prevalence of classical enterotoxin genes (sea-see) in the tested isolates was 92.85%, while that of the new enterotoxin genes was 55.7%.This result agrees with the results of the previous studies (Rall et al., 2008;Wu et al., 2010;Mashouf et al., 2015).
The tsst-1 gene was detected in 58.5% of isolates.A similar result was reported in a study of staphylococcal food-poisoning outbreaks in Taiwan (Chiang et al., 2008).No significant differences were found between S. aureus isolates of food and clinical origins except for the tsst-1 gene where it was significantly higher in the food isolates than in the clinical ones (P=0.003).A previous study reported higher prevalence of tsst-1 gene in food isolates than human and animal isolates (Adesiyun et al., 1992).On analysis of toxin gene profile results, 94% of enterotoxigenic isolates were found to harbor more than one toxin gene.This percentage is higher than data reported previously (Becker et al., 2003;Udo et al., 2009).Forty nine toxin patterns were detected in our isolates.This finding was in accordance with Xie et al. (2011) who found 47 toxin patterns.The high diversity of toxin gene profiles of S. aureus can be used in addition to the present genotyping methods to achieve high discrimination between isolates.
The expression of many virulence genes such as exotoxins and capsular polysaccharides type 5 and 8 during S. aureus infections is influenced by agr locus (Luong et al., 2002).This is considered a mechanism for isolating bacterial populations and a basis for species subdividing (Robinson et al., 2005).
During this study, the prevalence of agr specificity groups have been investigated using multiplex PCR which revealed that all analyzed strains harbored agr gene based on the amplicon size differences.The results showed that the agr group II was the most prominent detected in 58.5% of isolates; followed by agr group I (25.7%) and agr group III (20%).This was in accordance with data supplied by Ayepola (2012) who showed that 82% of S. aureus isolated from clinical sources belonged to agr II.
On the other hands, an earlier study showed that agr I was the most prevalent (van Leeuwen et al., 2000).Two food and one clinical isolates showed mixed agr type as they carried both agr I and III.Mixed agr I/IV was reported by previous studies (Goerke et al., 2005;Robinson et al., 2005).Further investigations are required for more characterization of these isolates.
No significant correlation was found between agr type and the isolation sources (P > 0.05) as prevalence of a particular agr group was not found among either food or clinical isolates.Concerning toxin genes, agr I was significantly associated with sei and tsst-1 genes (P= 0.0033 and 0.0005, respectively).Besides that, agr II was correlated with sed and see genes (P =0.0003, 0.0213, respectively).In accordance with Xie et al. (2011), no correlation was found between agr types and toxin gene profile of the S. aureus strains.
Hierarchical clustering of isolates on the basis of agr type, resistance pattern and toxin genes pattern results in two large clusters.These clusters contained isolates from both sources, but the majority of clinical isolates (74.2%) was found in cluster I. Isolates of agr I/III type were grouped in one cluster (H).Although three pairs of isolates showed 100% similarity [(33c, 34c), (3f, 8f) and (19f, 1c)], they were isolated from different locations.This result suggests the need for other typing techniques to analyze the genetic background of these isolates.

Conclusions
High resistance levels were detected in S. aureus isolates either from food or clinical sources in Egypt.This result is an alert of the growing problem of bacterial resistance mostly due to the improper use of antibiotics.Although incidence of S. aureus isolates in food samples was not high, these isolates were highly enterotoxigenic.Moreover, toxin gene profile analysis can be used in genotyping of enterotoxigenic strains as it allows a better discrimination for isolates than agr typing which is limited only to four groups.

Figure 1 .
Figure 1.Number of toxin genes per isolate in clinical and food isolates.

Figure 2 .
Figure 2. Dendrogram of genotypic relationship of S. aureus isolates.Each pattern listes the sample sources, antimicrobial resistance pattern, toxin gene contents and agr types of the isolates.C: Clinical; F: food.

Table 2 .
Resistance percent of S. aureus isolates to the tested antimicrobial agents.

Table 3 .
Antibiotic resistance patterns of the tested S. aureus isolates.

Table 4 .
Prevalence of toxin genes in S. aureus isolates.

Table 5 .
Toxin genes patterns of S. aureus isolates.