MOLECULAR TYPING OF FOODBORNE COAGULASE-POSITIVE STAPHYLOCOCCUS ISOLATES IDENTIFIED BY MALDI-TOF MS

The aim of the study was the identiﬁ cation and characterisation of coagulase-positive Staphylococcus bacteria obtained from food matrices by mass spectrometry and molecular methods. A total of 46 coagulase-positive Staphylococcus isolates were collected from diﬀ erent foodstuﬀ s. The Staphylococcus isolates were identiﬁ ed by MALDI-TOF MS and conﬁ rmed by the presence and sequence analysis of the Staphylococcus protein A gene. Staphylococcal enterotoxin genes were also investigated by multiplex PCR. Based on the identiﬁ cation of strains by the MALDI-TOF MS technique and spa -typing, all strains were identiﬁ ed as Staphylococcus aureus . Based on their MS peak proﬁ les, the isolates matched the spectra of three S. aureus reference strains in the Bruker MALDI Biotyper database, with identiﬁ cation scores higher than 1.999 in the case of all 46 (100%) isolates. The isolates showed great genetic variability. Twenty spa types were identiﬁ ed, from which most lineages are capable of colonizing humans. Fifty percent of the strains harboured at least one of four enterotoxin genes ( seg , seh , sei , and ser ), but none of the classical enterotoxin genes could be detected.

In the European Union, Staphylococcus aureus (S. aureus) is an important pathogenic agent of food intoxications (D et al., 2017). S. aureus strains have several virulence factors such as lipases, thermonuclease, hyaluronidase, and haemolysins; however, the major virulence factors are the heat-stable enterotoxins that cause staphylococcal food poisoning (SFP) (O et al., 2010). The onset of SFP symptoms is very rapid in contrast to other gastrointestinal infections, generally a short time after ingestion of the contaminated food (M et al., 2015). In outbreaks, the symptoms typically include diarrhoea (89%), vomiting (87%), and abdominal cramps (72%); however, fever (9%) is rarely reported. Usually, pork (41% of pork dishes were ham) or poultry dishes (41% freshly prepared poultry dishes) are implicated (55%) in S. aureus outbreaks (B et al., 2013). Coagulase-positive Staphylococcus species have several defence mechanisms like immunoglobulin binding proteins -including staphylococcal protein A (SpA) -that assist the evasion of the host immune system (F , 2005). Protein A is produced by numerous species of Staphylococcus, and is a highly potent virulence as well as immune evasion factor (B et al., 2018). The 40-60 kDa cell-wall protein is secreted during the exponential growth phase and is encoded by the spa gene (B & S , 1972). The synthesised protein is able to bind to the Fc and F (ab') 2 regions of immunoglobulins (F , 1970) and thus may interfere with the immunising ability of vaccines that depend on the production of opsonophagocytic antibodies, because these inhibit the functions of B-cells (D & S , 2012).

HORVÁTH et al.: MOLECULAR TYPING OF STAPHYLOCOCCUS ISOLATES
The spa-typing of S. aureus isolates has two major aims: to analyse genetic microvariation in outbreak investigations and to analyse genetic macrovariation in population-and phylogenetic-based assaying (H et al., 2009). A variable repeat region (spa) in the gene allows a reliable and rapid method to discriminate S. aureus isolates in outbreaks from those presumed to be epidemiologically unrelated (K et al., 2004). Furthermore, the epidemiological defi nition makes it possible to compare the outbreaks occurring in clinical cases and animal husbandry. Given that spa-typing involves the sequencing of only one gene, it provides signifi cant advantages in terms of time-to-result, standardisation, ease of use, and reproducibility as compared to other molecular biology techniques (A et al., 2007). Matrix-assisted laser desorption ionisation time-of-fl ight mass spectrometry (MALDI-TOF MS) allows the identifi cation of microorganisms isolated from food or clinical cases by a low-cost, rapid, easy, high-throughput, and effi cient identifi cation technique (C et al., 2012). Using standardised and developed procedures, MALDI-TOF MS devices have revolutionised the identifi cation at the species level of most Gram-positive and Gramnegative bacteria (B et al., 2015). The present study aims to discriminate and perform molecular characterisation of foodborne coagulase-positive staphylococci typed molecular techniques after identifi cation by MALDI-TOF MS.

Identifi cation of Staphylococcus spp. with MALDI-TOF MS
For the identifi cation, the colonies were grown on Columbia Blood agar. A formic acid suspension preparation protocol was used: a single colony was picked up with a sterile inoculation loop and suspended for 30 s in 40 μl of formic acid in a microtube. Then, 40 μl of acetonitrile was added to the suspension and mixed extensively. Finally, from the suspension, 1 μl was transferred onto the target plate; when dried, it was overlaid with 1 μl α-HCCA (10 mg ml -1 ) matrix solution and left to dry.
Mass spectra were collected by application of a Bruker Microfl ex LT MALDI-TOF mass spectrometer operating in the molecular mass range of 2.0-25 kDa, in positive linear mode. Staphylococcus isolates were identifi ed by MALDI BioTyper 3.1 software. More than 200 shots were required to give adequate spectra with an appropriate signal-to-noise ratio. Before each measurement, calibration was carried out with the Bruker Bacterial Test Standard using lyophilised Escherichia coli. During the measurement, 640 shots were performed with the aim of analysing the mass spectra of Staphylococcus spp.
The 46 isolates were analysed in parallel; the mass spectra obtained were compared with the Bruker MALDI Biotyper database, where mass spectra of six diff erent Staphylococcus aureus isolates were available. The results are reported as numeric scores based on similarity with the reference spectra based on a proprietary algorithm of the Bruker MALDI Biotyper software comparing the presence and symmetry of peaks in the mass spectra of the unknown strain and the database strain entries. The software sets the following threshold score values for identifi cation: scores between 2.300-3.000 are designated as "highly probable species identifi cation", scores between 2.000-2.299 as "secure genus identifi cation", scores of 1.700-1.999 as "probable genus identifi cation", and scores below 1.699 are reported as "nonreliable genus identifi cation". The closest matches are listed in order of score value, the highest indicating the highest similarity in the mass spectra.

Molecular analysisspa-typing
The cheaper and less laborious spa-typing is a single-locus technique, which off ers a subtyping resolution comparable to MLST and PFGE. During spa-typing, the sequence of a polymorphic VNTR (variable-number tandem repeat) sequence was investigated, which is located in the 3' coding region of the S. aureus-specifi c SpA. The spa type of a strain is determined by the repeat succession. There are exceptions; however, the special repeat length for the spa VNTR is usually 24 bp. The method used is based on the second version of the Protocol for PCR Amplifi cation of spa Recommended by the EURL-AR (2012).

Staphylococcal enterotoxin (SE) gene PCR
Genomic DNA was obtained from S. aureus strains using PrepMan Ultra lysis buff er (Thermo Scientifi c, Biocenter Ltd, Szeged, Hungary) according to the manufacturer's protocol. The supernatant was diluted fi ve-fold with TE buff er and used for PCR amplifi cation. To detect the major SE genes (sea, seb, sec, sed, and see), multiplex PCR protocols were performed using the primer sets of the European Union Reference Laboratory for Coagulase-Positive Staphylococci (EURL-CPS) as published elsewhere (B et al., 2014) with a slight modifi cation. Primers detecting the seh and ser genes were run as a third, separate reaction to avoid overlapping of the PCR products of similar size during the gel electrophoresis. Reference strains of S. aureus DSMZ 18586 (SEB), DSMZ 18587 (SEC, SEG, SEH, SEI), DSMZ 18588 (SED, SEG, SEI, SEJ, SER), DSMZ 18589 (SEE), and ATCC 29213 (SEA) were used as positive controls. Amplicons were separated by agarose gel electrophoresis and identifi ed according to their predicted size. Table 1 includes the results of spa-typing and identifi cation of strains by MALDI-TOF MS. Furthermore, the type of food and the best value and best match of identifi cation are also listed, as are the enterotoxin genes harboured by the isolates. Most of the samples containing coagulase-positive staphylococci were collected from meat (70%) including pork (n=20), poultry (n=8), beef (n=3), and venison (n=1) and food made from them. As the total number and type distribution of sampled materials was not available, further conclusion on the isolation frequency of coagulase-positive staphylococci from diff erent food matrices is not within the aims of this study. Based on the presence of the spa gene and on the best match and best score, all strains were identifi ed as S. aureus. Based on the MALDI-TOF MS data, the isolates matched the spectra of three S. aureus reference strains in the Bruker MALDI Biotyper database. The number of the best-matching isolate is indicated in Table 1.

Results and discussion
Using the formic acid suspension preparation method, it was possible to obtain identifi cation scores higher than 1.999 in the case of all 46 (100%) isolates by comparison with the Bruker MALDI Biotyper database.The vast majority, 83% of the isolates, gave MS scores ≥2.300, so the species were reliably identifi ed. Identifi cation at the genus level was secure in 17% of isolates, and no isolates had a probable genus level identifi cation with a score in the 1.700-1.999 range. M and U (2017) obtained similar results when investigating foodstuff s: 34 out of 36 coagulase-positive Staphylococcus isolates were identifi ed with a score >2.000, when using another method of direct sample preparation; however, no results of identifi cation with a score over 2.300 were demonstrated.
The strains showed great genetic variability: 20 spa types were identifi ed among the 46 isolates. The distribution of the spa types in food matrices is shown in Table 2. The most common spa type was t084 with an isolation frequency of 19.6% (9/46), belonging to humanrelated clonal complex (CC)15, and was found mostly in dried pasta and meat. Other frequent types were t091 (ST7) and t1491, both with an isolation frequency of 13.0% (6/46). These spa types are also frequently isolated from humans, and t1491 is associated with CC1. This lineage is widely distributed and consists of methicillin-susceptible and -resistant strains with low host specifi city, thus frequently isolated from livestock and related products (C et al., 2015). The connection between the Bruker MALDI Biotyper database closest match reference strain and the spa type cannot be directly inferred: the 20 spa types were distributed among the three database reference strains, and several spa types were found to be assigned to several database entries (t084, t091, t213, t330, t1491).
Among the strains examined, 50% of the isolates contained at least one enterotoxin gene. In two strains (SA-17 and SA-45), two diff erent types of SE were detected (Table 1). Four diff erent enterotoxin genes were identifi ed among the 46 isolates: sei, seg, seh, and selp, of which seh was the most prevalent at 17% (8/46) and which was found mostly in strains from pork meat. In SFP outbreaks, the predominance of SEA is well documented at 70-80% prevalence, independent of the region where the investigation has taken place (A et al., 2010). Though other SEs and SE-like toxins (SEl), like SEH, may also be involved in clinical cases, their role in SFP is still controversial (F et al., 2018). None of the major enterotoxin genes were present in our isolates, and only other SE genes of minor signifi cance occurred in lower numbers, which may refl ect favourable conditions in comparison with other studies on the matter (L et al., 2015; Q & S , 2019).

Conclusions
Based on the results of spa-typing and the best match and best score with the MALDI-TOF MS technique, all strains were identifi ed as S. aureus. The isolates matched the spectra of three S. aureus reference strains in the Bruker MALDI Biotyper database; 83% of the isolates gave scores ≥2.300, so the species were safely identifi ed by mass spectrometry. The MALDI-TOF MS technique is suitable for rapid and cost-eff ective identifi cation of S. aureus among foodborne coagulase-positive Staphylococcus species in routine diagnostics. The correlation between the three MALDI Biotyper database Staphylococcus aureus reference strain entries and the spa types and enterotoxin genes could not be unambiguously established. This can be attributed to the fact that the six Staphylococcus aureus MALDI Biotyper database entries do not represent the genetic diversity within the species, and though it is enough for strain identifi cation, but strain typing is not feasible due to such limitations.
Within the scope of the study, 4 diff erent enterotoxin genes and 20 diff erent spa types were identifi ed among the 46 isolates. Around 50% of the isolates harboured enterotoxin genes (sei, seg, seh, and selp), but these are the SEs of lesser signifi cance. *