Heterogeneity of Molecular Characteristics among Staphylococcus argenteus Clinical Isolates (ST2250, ST2793, ST1223, and ST2198) in Northern Taiwan

Staphylococcus argenteus is an emerging pathogen that is recognized as non-pigmented Staphylococcus aureus. However, the molecular characteristics of S. argenteus and its virulence factors have not been well studied. The present study analyzed 96 isolates of S. argenteus recovered from blood. Identification of S. argenteus was based on results of MALDI-TOF MS and lacking crtM gene. All 96 isolates were methicillin-susceptible. Multilocus sequence typing (MLST) revealed four sequence types: ST2250 (n = 72), ST2793 (n = 12), ST1223 (n = 10), and ST2198 (n = 2). All 72 ST2250 isolates harbored CRISPR loci with polymorphism of direct repeats and spacers, but no other STs carried CRISPR loci. To date, ST2793 isolates have rarely been reported in other countries. Collagen-binding adhesin gene (cna) and staphylococcal enterotoxin type C (sec) were detected in 12 (100%) and 8 (67%) ST2793 isolates, respectively. ST1223 has been reported as food poisoning pathogens, and enterotoxin gene clusters (egc) were detected in all 10 isolates, while seb gene was detected in three isolates. Two ST2198 isolates carried bone sialoprotein-binding protein gene (bbp), belonging to agr type IV. Our focus on the heterogeneity of molecular characterization in four ST types of S. argenteus revealed that S. argenteus had been isolated as early as 2000. Each ST type of S. argenteus harbors particular genetic markers that may contribute to their virulence.


Introduction
Staphylococcus argenteus is an emerging pathogen that has been recognized as non-pigmented Staphylococcus aureus displaying white colonies on chocolate agar plates owing to lack of the crtOPQMN gene operon required for staphyloxanthin pigment production. [1,2]. To date, S. argenteus clinical isolates have been reported in many countries, including Belgium, France, Thailand, Japan, China, Bacterial isolates were collected from the National Taiwan University Hospital (NTUH) in 2000,2005, and 2010-2012 in prior studies [8,9]. In brief, NTUH is a 2500-bed teaching hospital providing both primary and tertiary care in northern Taiwan. For all the preserved S. argenteus (previously misidentified as S. aureus) isolates used in this study, species identification was performed by colony morphology, Gram staining results, a positive slide or tube coagulase test, and using the Vitek 2 identification system (bioMerieux, Marcy l'Etoile, France). Antimicrobial susceptibility was determined using the standard disk diffusion method and Vitek 2 identification system. In total, 96 methicillin-susceptible S. argenteus (MSSAg) bloodstream isolates were further confirmed by MALDI-TOF MS, multilocus sequence typing (MLST), and detection of the crtM gene. MLST was conducted as previously described [19], and STs were assigned using the S. aureus MLST database (http://saureus.mlst.net). The previously described amplification PCR primers were crtMupF (5 -TTAGGAAGTGCATATACTTCAC-3 ) and crtMdownR , and conditions were established to amplify a partial crtM gene involved in staphyloxanthin production [20]. The 1660-bp amplicon was generated in isolates containing crtM. The existence of the crt operon was detected by PCR with primers crtOp-F (5 -CCATGAAAAGCACCATATTT-3 ) and crtOp-R (5 -GTTAACAGCAACGGTTCTGT-3 ), whose targets on the crt operon are upstream and downstream, respectively. Amplification was performed using the following conditions: 7 min at 94 • C, followed by 30 cycles at 94 • C for 30 s, 45 • C for 30 s, and 68 • C for 6 min, and ending with a final extension time of 10 min at 68 • C. A 6.4 kb amplicon was obtained in pigmented isolates, while a 977 bp amplicon was generated from non-pigmented isolates. For matrix-assisted laser desorption/ionization time-of-flight mass spectrometer (MALDI-TOF MS, Bruker Daltonik GmbH, Bremen, Germany) identification, all S. argenteus isolates were prepared and analyzed as previously described [9].

Pulsed-Field Gel Electrophoresis and Spa Typing
The genetic associations of 96 S. argenteus isolates were determined by pulsed-field gel electrophoresis (PFGE). The DNA in gel plugs was digested with SmaI (New England BioLabs, Ipswich, MA, USA) and then separated in a CHEF-DR III apparatus (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The plugs were applied to wells of 0.8% (w/v) agarose gels (Bio-Rad Laboratories, Inc., Hercules, CA, USA). PFGE was carried out at 200 V and 12 • C for 20 h, with a pulse angle of 120 • and pulse times ranging from 5 to 60 s. The pulsotypes were analyzed using BioNumerics software version 4.0. (Applied Maths, Sint-MartensLatem, Belgium) In addition, the spa typing was determined by PCR and sequencing, as previously described [21]. The X-region of the staphylococcal protein A gene (spa gene) was amplified with primer pair spa-1095F and spa-1517R [21]. The resulting spa types were assigned using the Ridom Spaserver website (http://www.spaserver.ridom.de) [22].

Sequencing of CRISPR/cas Loci of S. argenteus ST2250
Specific primers (Table S4) based on S. argenteus type strain MSHR1132 were designed to amplify the CRISPR/cas Loci (clustered regularly interspaced short palindromic repeats (CRISPR)-CRISPR associated proteins (Cas)). CRISPR/cas Loci of selected S. argenteus ST2250 isolates were determined by PCR and direct sequencing. Direct repeats (DRs, leader end and inner repeat), degenerated DRs (trailer end repeat), and spacers of the CRISPR region of all 72 S. argenteus ST2250 isolates were sequenced to see the polymorphism.

Phylogenetic Analysis
DNA sequences were aligned using the GeneWorks software (IntelliGenetics, Mountain View, CA, USA). The phylogenetic relationships between the species were analyzed using the neighbor-joining method of phylogenetic tree construction, as shown in the MEGA (molecular evolutionary genetic analysis) analytical package [30]. For neighbor-joining analysis, distances between the sequences were calculated using Kimura's two-parameter model. Levels of similarity were determined between the species. Bootstrap values were obtained for 500 randomly generated trees.

Identification of S. argenteus by MALDI-TOF MS, MLST Typing and Lack of CrtM
All 96 isolates were recovered from blood cultures and first identified as S. aureus by Vitek2 but displayed white colonies (non-pigmented) and lacked the crtM gene, pigment production-associated genes, and whole crt operon (crtOPQMN). All 96 isolates were identified as S. argenteus by MALDI-TOF MS using a previously established database [9]. MLST typing identified four sequence types with ST2250 (n = 72), ST2793 (n = 12), ST1223 (n = 10), and ST2198 (n = 2) ( Table 1). S. argenteus ST2250 lineage accounted for 75% (72/96) of isolates. The second most frequent ST was ST2793 (12.5%), and the third one was ST1223.   Table 2 for new spa type.

Antimicrobial Susceptibility and Biochemical Characteristics
All 96 S. argenteus isolates were susceptible to methicillin, oxacillin, clindamycin, trimethoprim/sulfamethoxazole, and teicoplanin. Resistance rates for erythromycin (2 isolates), fusidic acid (1 isolate), and gentamicin (2 isolates) were low. Comparison of biochemical activities of the 96 S. argenteus isolates tested by VITEK 2 indicated similar results to those of S. aureus and a previous study [2], except for urease, N-acetyl-d-glucosamine, and d-ribose (Table S1). Of the 96 S. argenteus isolates, 62 (65%) were positive for urease, while the overall positivity rate of clinical S. aureus (which may include S. argenteus) was approximately 2% in the VITEK 2 database and 8.4% in the National Taiwan University Hospital (NTUH) database. For N-acetyl-d-glucosamine, 94 (97.5%) isolates were negative, but only 59.8% of S. aureus (including S. argenteus) were negative based on the results of VITEK 2 in the NTUH. ST1223, which displayed a more than 20% difference from the average positive rate in six biochemical test items, may have exhibited its distinct biochemical profiles more than ST2250 and ST2793 isolates.

Genotyping by Pulsed-Field Gel Electrophoresis (PFGE) and Spa Types
PFGE analysis revealed four major pulsotypes ( Figure 1). Isolates of the same ST always clustered together. The spa types are listed in Tables 1 and 2. Most spa types are t5078 and t6675 among ST2250 and t5142 among ST1223. The other spa types have not been reported, and non-typeable (N.T.) (new spa types) in S. argenteus. t5078 was the most frequent spa type (55 isolates and 57%), followed by t6675 (5 isolates and 5%). However, spa was not typeable in 18 isolates, including in all of 12 ST2793 isolates that contain eight isolates with one identical spa repeat (t19483).  Table 2.

Phylogenetic Trees of Coa, DnaJ and GroESL and Agr Type
Selected isolates of each ST of S. argenteus were examined for the D1 region sequence of the staphylocoagulase (SC) gene and the (coa) gene. Phylogenetic trees showed that each ST type displayed a distinct type ( Figure 2). SC genes of S. argenteus ST2250, ST2793, ST1223, and ST2198 were close to genotype type XI, type II, type VI, and type V but were assigned to type XId, type XVI, type Figure 1. Pulsed-field gel electrophoresis (PFGE) dendrogram of 96 S. argenteus clinical isolates. PFGE cluster was assigned to isolates having 80% or greater similarity from the dendrogram. For unknown spa type (NT): see Table 2.

Phylogenetic Trees of Coa, DnaJ and GroESL and Agr Type
Selected isolates of each ST of S. argenteus were examined for the D1 region sequence of the staphylocoagulase (SC) gene and the (coa) gene. Phylogenetic trees showed that each ST type displayed a distinct type ( Figure 2). SC genes of S. argenteus ST2250, ST2793, ST1223, and ST2198 were close to genotype type XI, type II, type VI, and type V but were assigned to type XId, type XVI, type XV, and type XIV according to the criteria, respectively [31]. In particular, ST1223 has been reported as serotype VI, and the phylogenetic tree of the coa gene showed that ST1223 is also close to genotype VI [32]. XV, and type XIV according to the criteria, respectively [31]. In particular, ST1223 has been reported as serotype VI, and the phylogenetic tree of the coa gene showed that ST1223 is also close to genotype VI [32].  To further characterize S. argenteus, analyses of partial dnaJ and groESL sequences were performed in selected isolates for each ST. Based on the dnaJ and the groEL sequences, in this study, the S. argenteus clinical isolates and the S. argenteus MSHR1132 were clustered together and separated from S. aureus, forming a stable clade with bootstrap values of 95% and 100% in the phylogenetic trees, respectively ( Figure 3). The BLAST results indicated that dnaJ and groEL showed the highest similarity (>98%) to that of non-pigmented S. argenteus MSHR1132 and approximately 90% similarity to S. aureus N315 (Tables S2 and S3). In addition, the spacers between groES and groEL were analyzed. S. argenteus displayed 71 nucleotides of spacer length, while S. aureus and S. argenteus ST1223 had 75 nucleotides of spacer length (Table 3).

Figure 2.
Phylogenetic tree based on D1 region of coa gene. The phylogenetic tree was generated using the unrooted neighbor-joining method in the MEGA6 package. The numbers at the nodes are confidence levels, expressed as percentages of occurrence in 1000 bootstrapped resamplings. The scale bar indicates the evolutionary distance between sequences, as determined by measuring the lengths of the horizontal lines connecting two organisms. Each sequence type (ST) type of S. argenteus belonged to different clades. The results showed identical partial coa sequences at least two isolates for each ST. GenBank accession numbers are given in parentheses with one isolate for each ST in this study.
To further characterize S. argenteus, analyses of partial dnaJ and groESL sequences were performed in selected isolates for each ST. Based on the dnaJ and the groEL sequences, in this study, the S. argenteus clinical isolates and the S. argenteus MSHR1132 were clustered together and separated from S. aureus, forming a stable clade with bootstrap values of 95% and 100% in the phylogenetic trees, respectively ( Figure 3). The BLAST results indicated that dnaJ and groEL showed the highest similarity (>98%) to that of non-pigmented S. argenteus MSHR1132 and approximately 90% similarity to S. aureus N315 (Tables S2 and S3). In addition, the spacers between groES and groEL were analyzed. S. argenteus displayed 71 nucleotides of spacer length, while S. aureus and S. argenteus ST1223 had 75 nucleotides of spacer length (Table 3).  groEL (B). The phylogenetic tree was generated using the unrooted neighbor-joining method in the MEGA6 package. The numbers at the nodes are confidence levels, expressed as percentages of occurrence in 1000 bootstrapped resamplings. The scale bar indicates the evolutionary distance between sequences, as determined by measuring the lengths of the horizontal lines connecting two organisms. The results showed identical partial dnaJ or groEL sequences at least two isolates for each ST. GenBank accession numbers are given in parentheses with one isolate for each ST in this study.  groEL (B). The phylogenetic tree was generated using the unrooted neighbor-joining method in the MEGA6 package. The numbers at the nodes are confidence levels, expressed as percentages of occurrence in 1000 bootstrapped resamplings. The scale bar indicates the evolutionary distance between sequences, as determined by measuring the lengths of the horizontal lines connecting two organisms. The results showed identical partial dnaJ or groEL sequences at least two isolates for each ST. GenBank accession numbers are given in parentheses with one isolate for each ST in this study. Table 3. Nucleotide sequences and length of spacers between groES and groEL among species. "-" indicated the same nucleotide with MSHR1132 and "." Indicated blank. The results showed identical spacer sequences at least two isolates for each ST. GenBank accession numbers are given in parentheses with one isolate for each ST in this study.

Strain
Spacer Length (bp) Spacer Sequence ( Gene agrD produces a ribosomal propeptide of which the middle section encodes the seven to nine residue autoinducing peptide (AIP) used as a quorum sensing (QS) signal molecule. The agrD alignment is shown in Figure 4. S. argenteus harbors S. aureus arg types I, III, and IV. The most dominant agr are type I (ST1850 and ST2250) and type III (ST1223 and ST2793). S. argenteus ST2198 and S. schweitzeri ST2022 are known to be of animal origin, and these two isolates belong to type IV. Although S. argenteus harbors identical AIP to that of S. aureus, S. argenteus displays its unique agrD amino acid sequences, which are different from those of S. aureus or S. schweitzeri.
Gene agrD produces a ribosomal propeptide of which the middle section encodes the seven to nine residue autoinducing peptide (AIP) used as a quorum sensing (QS) signal molecule. The agrD alignment is shown in Figure 4. S. argenteus harbors S. aureus arg types I, III, and IV. The most dominant agr are type I (ST1850 and ST2250) and type III (ST1223 and ST2793). S. argenteus ST2198 and S. schweitzeri ST2022 are known to be of animal origin, and these two isolates belong to type IV. Although S. argenteus harbors identical AIP to that of S. aureus, S. argenteus displays its unique agrD amino acid sequences, which are different from those of S. aureus or S. schweitzeri.

Examination of Virulence Factors
The results of virulence factors are shown in Table 4. Each ST type displayed distinct characteristics. The hlb, hlg, lukPVL, eta, etb, edin, sea, sed, see, seh, and tst genes were not detected in all S. argenteus isolates, but the hemolysin genes (hla and hld) were distributed universally. CRISPR/cas, collagen-binding adhesin gene (cna), enterotoxin gene cluster (egc), or sialoprotein-binding protein gene (bbp) were found exclusively in S. argenteus ST2250, ST2793, ST1223, or ST2198 lineage, respectively. Immune evasion cluster (IEC) genes were detected in ~90% isolates, and most of them harbored staphylococcal complement inhibitor gene (scn) and staphylokinase (sak). In addition, the gene encoding staphylococcal enterotoxin type C (sec) was found in 8 out of 12 (67%) ST2793 isolates, and that of staphylococcal enterotoxin type B (seb) was found in three S1223 isolates (3/10, 30%) and one ST2198 (1/2, 50%). The expression of seb or sec in these isolates was further tested by the rapid latex agglutination test. All nine isolates carrying sec were exclusively reactive with enterotoxin type C-specific antibodies, but no positive reactions were seen in the isolates carrying seb.

Examination of Virulence Factors
The results of virulence factors are shown in Table 4. Each ST type displayed distinct characteristics. The hlb, hlg, luk PVL , eta, etb, edin, sea, sed, see, seh, and tst genes were not detected in all S. argenteus isolates, but the hemolysin genes (hla and hld) were distributed universally. CRISPR/cas, collagen-binding adhesin gene (cna), enterotoxin gene cluster (egc), or sialoprotein-binding protein gene (bbp) were found exclusively in S. argenteus ST2250, ST2793, ST1223, or ST2198 lineage, respectively. Immune evasion cluster (IEC) genes were detected in~90% isolates, and most of them harbored staphylococcal complement inhibitor gene (scn) and staphylokinase (sak). In addition, the gene encoding staphylococcal enterotoxin type C (sec) was found in 8 out of 12 (67%) ST2793 isolates, and that of staphylococcal enterotoxin type B (seb) was found in three S1223 isolates (3/10, 30%) and one ST2198 (1/2, 50%). The expression of seb or sec in these isolates was further tested by the rapid latex agglutination test. All nine isolates carrying sec were exclusively reactive with enterotoxin type C-specific antibodies, but no positive reactions were seen in the isolates carrying seb.   (17) and serotype was according to previous described (27). c IEC (immune evasion cluster) type B contains sak+, scn+, and chp (60 isolates) and type E contains sak+ and scn+ (one isolate).

Polymorphism of CRISPR/Cas System in ST2250 Isolates
Since S. argenteus MSHR1132 harbored the CRISPR/cas system, we tried to determine whether S. argenteus isolates have the same one. The whole CRISPR/cas system of selected S. argenteus ST2250 isolates was determined by PCR and direct sequencing. S. argenteus NTUH_9546-1 and NTUH_4415 (ST2250) and MSHR1132 (ST1850) shared the same type III CRISPR/cas system, and the cas loci was cas1, cas2, csm1(cas10), csm2, csm3, csm4, csm5, csm6, and cas6 in order. (Figure 5) These isolates did not possess the SCCmec found in MSHR1132 and displayed different direct repeats (DRs, leader end and inner repeat) and spacers. Upstream (CL region) and downstream (CR region) regions of cas loci were sequenced for all 72 ST2250 isolates. A total of four DRs, two degenerated DRs (DDR, trailer end repeat), and 16 different spacers were found in these isolates (Table 5). Half isolates (36 isolates, 50%) displayed the same spacers of CL and CR regions as NTUH_9546-1 isolates (Table 6). for agr and coa typing, coa genotype was based on D1 region and previous described (17) and serotype was according to previous described (27). C IEC (immune evasion cluster) type B contains sak+, scn+, and chp (60 isolates) and type E contains sak+ and scn+ (one isolate).

Discussion
Recent studies have displayed the increasing emergence of S. argenteus from many countries, which has been isolated from both humans and animals [33][34][35]. Isolation of S. argenteus emphasizes its clinical significance and the importance of correct identification in clinical laboratories [8,15]. MALDI-TOF MS is a simple and accurate method for phenotypic screening and identification of S. argenteus [8]. Furthermore, in the present study, the biochemical results of VITEK 2 for the 96 S. argenteus isolates showed a higher percentage of urease and were almost negative for N-acetyl-d-glucosamine. The results suggest that these two tests also could be screening phenotypes for S. argenteus. In the present study, all 96 S. argenteus isolates with bacteremia were methicillin-susceptible and comparable to recent reports from Thailand, Japan, and Myanma, while isolates from Europe and Australia were more methicillin-resistant [5,10,[14][15][16].
Notably, the earliest and the only S. argenteus isolate in the year 2000 was ST2793 with novel spa type (t19483). ST2793 lineage is rarely reported in other countries [10][11][12][13][14]36]. To date, ST2793 has been found in Europe and USA but not in Asian counties other than Taiwan [10,13,36]. In addition, a total of 11 new spa types across three ST types were first found in the present study. Some known spa types were identical to those in Japan and Myanmar isolates [14,16] (Table 2). The t5078 clone occupied more than half of S. argenteus ST2250 isolates in Asian countries.
Besides MLST typing and whole genome sequencing, several genetic classification methods have been proposed to distinguish S. argenteus from S. aureus. PCR-based method targeting the NRPS gene, sequencing of the rpoB or the nuc gene, and lack of the crtM gene have been used previously to determine S. argenteus [3,7]. The dnaJ and the groEL genes have been used previously to identify Staphylococcus species [23,37]. Phylogenetic analysis based on the dnaJ and the groEL sequences indicates that these genes may be used as candidate genes for species identification.
Detection of SE-like toxin genes revealed that nine of 12 ST2793 S. argenteus isolates carry sec. A report from Denmark using whole genome sequencing in four ST2793 and 21 non-ST2793 revealed that none of the isolates carried sec [10]. Moreover, the sec partial sequence (245 bp) from ST2793 is identical to that of S. aureus in the Genbank database and almost identical to sec2 or sec3 genes of S. argenteus ST2250 from Japan [14]. ST2250 was described in Thailand as carrying an enterotoxin gene cluster composed of seC-bov (enterotoxin C bovine) and entQ (staphylococcal enterotoxin Q), and these two genes should be the two novel superantigens sel27 and sel26 [14,38]. In addition, some of the ST1223 and the ST2198 isolates in the present study carried enterotoxin type b gene (seb) and others did not, which shows the diversity in Taiwan isolates. However, no isolates carrying seb were found in Hokkaido, Japan, or Myanmar [14,16]. Interestingly, all ST1223 isolates harbored egc clusters as in other studies, but only a small portion carried seb gene, which may cause Staphylococcal food poisoning [32,39].
IEC (immune evasion cluster) was carried by bacteriophage and into hemolysin b gene (hlb) of Staphylococci [29]. Additionally, 60 isolates of ST2250 carried both sak and scn (IEC type E), and only one ST2198 isolate harbored sak, scn, and chp (IEC type B), which is comparable to previous studies [11,40]. Regarding other virulence factors analyzed in this study, collagen binding gene (cna) and spa (protein A) belong to microbial surface components recognizing the adhesive matrix molecules (MSCRAMM) family, and these are associated with bacterial adhesion and pathogenesis [41]. We found that all ST2793 isolates, but not other S. argenteus lineages, harbored the cna gene and carried new and different spa sequences (Table 2). In addition, the ST2793 isolates are methicillin-susceptible, while isolates from Europe carrying SCCmec are methicillin-resistant [10]. This may indicate that the ST2793 isolates in Taiwan reveal genetically diverse clones, which is definitely worth further study.
In the present study, all 72 ST2250 isolates harbored the CRISPR/cas system. Previous studies have mentioned the polymorphism of CRISPR/cas system in S. aureus and S. argenteus [42,43]. However, the present study is the first to study the CRISPR/cas system solely for S. argenteus ST2250 lineage, including four direct repeats with 37 bp (DRs, one upstream and three downstream) and 16 different spacers with 33-39 bp (nine upstream and seven downstream) ( Table 5). S. argenteus ST2250 showed greater diversity of CRISPR/cas than that of most other S. aureus ST types and may account for the prevalence and the success of S. argenteus ST2250 clones.
In summary, this study focused on the molecular characterization of four ST types of S. argenteus isolates in northern Taiwan, finding that S. argenteus was isolated as early as the year 2000. Each ST type of S. argenteus harbors particular genetic markers that may contribute to their virulence Supplementary Materials: Supplementary materials can be found at http://www.mdpi.com/2076-2607/8/8/1157/s1. Table S1: Key biochemical tests used for identification of S. argenteus (N = 96), Table S2: Partial dnaJ (889bp) nucleotide and amino acid sequence similarities, Table S3: Partial groEL (780bp) nucleotide and amino acid sequence similarities, Table S4: Primers list for CRIPSR/cas, Table S5: GenBank accession numbers in the present study.