Rapid differentiation of Staphylococcus aureus isolates harbouring egc loci with pseudogenes {psi}ent1 and {psi}ent2 and the selu or seluv gene using PCR-RFLP

The staphylococcal enterotoxins (SEs) belong to the superantigen family of exotoxins and are the only superantigens to possess emetic activity (Le Loir et al., 2003; Proft & Fraser, 2003). The International Nomenclature Committee for Staphylococcal Superantigens proposed that only staphylococcal superantigens that caused vomiting after oral administration in a primate model should be designated SEs, whereas other related toxins that either lacked emetic properties or had not been tested in a primate model should be designated staphylococcal enterotoxin-like toxins (SEls) (Lina et al., 2004). Fifteen years ago only five SEs were recognized, namely, SEA, SEB, SEC1/SEC2/SEC3, SED and SEE. Since then, through gene sequencing and partial and complete genome sequencing, 15 additional SEs and SEls have been described, namely, SEGSEI, SElJSElR, SElU, SElU2 and SElV, and three further variants of SEC (SEC-bovine, SEC-ovine and SEC-caprine) (reviewed by Le Loir et al., 2003; Smyth et al., 2004) (Thomas et al., 2006).

Following the identification of enterotoxins SEG and SEI (Munson et al., 1998), the genes encoding them were demonstrated to be part of a chromosomal operon, named the enterotoxin gene cluster (egc), comprising five genes, now designated selo, selm, sei, seln and seg in transcriptional order, and two pseudogenes, ψent1 and ψent2, between the sei and seln genes (Jarraud et al., 2001a, b; Monday & Bohach, 2001). It was proposed that the egc cluster arose by gene duplication and variation from an ancestral gene through unequal crossing-over generated by recombination involving misalignment between non-allelic regions (Jarraud et al., 2001a). Subsequently, Letertre et al. (2003) have demonstrated that some egc clusters possess a novel gene between the sei and seln genes, designated selu, arising from a 15 bp insertion (5'-CTCTAAAATTGATGG-3') in the ψent1 pseudogene sequence. Moreover, while the selu genes from three strains exhibit 99 % nucleotide sequence identity, one designated a variant (selu_v) yields a gene product with only 9596 % amino acid sequence identity to the other three SElU proteins (Letertre et al., 2003).

Letertre et al. (2003) developed three sets of primers to allow distinction between strains possessing pseudogenes and the selu or selu_v gene in the egc locus. However, verification of PCR findings other than by DNA sequencing has not been described because of the very high sequence identity of the seiseln intergenic region apart from the 15 bp insert. The development and application of a simple procedure for distinction of egc loci of Staphylococcus aureus bearing pseudogenes and the selu or selu_v gene is described herein. The method uses PCR-RFLP with either digestion by restriction endonucleases HindIII and HphI in tandem or endonuclease BccI digestion alone of PCR amplimers from the stop codon of the sei gene through the 5' first quarter of the seln gene obtained using the PSE1 and PSE4 primers of Letertre et al. (2003). The differential susceptibility or resistance of pseudogene- and selu- or selu_v-bearing amplimers allows distinction of these two types of egc loci.

Strains. The fully genome-sequenced S. aureus strains RF122, Mu50 (ATCC 700699), MRSA 252, MW2, COL and NCTC 8325-4 were used as control strains of known egc locus status (Table 1). Ten isolates from chickens, goats and cows, all of which were known to possess the egc locus, were also examined (Smyth et al., 2005). Eight bovine isolates of RAPD type 7 that were known to possess or lack the sec and tst genes (Fitzgerald et al., 2000), but which had not been included in the study of Smyth et al. (2005) and had not been further tested for the presence of SE and SEl genes, were also used (Table 1).

Table 1. Strains of S. aureus

All isolates were confirmed as S. aureus on the basis of sugar fermentation tests (arabinose, maltose, mannitol, sucrose, trehalose, xylose), the VogesProskauer test, deoxyribonuclease production, and a tube coagulase test using standard methods, and by the Staphaurex rapid latex test kit (Murex Diagnostics), which detects protein A and clumping factor (fibrinogen-binding protein), used according to the manufacturer's instructions. Primers for the femA gene encoding the S. aureus FemA peptidyltransferase (Schneider et al., 2004) were used for additional confirmation of species identity with animal isolates using the thermal cycling conditions described by Mehrotra et al. (2000). Strains and isolates were grown in Tryptic Soya broth (Oxoid) or on Tryptic Soya agar (Oxoid) at 37 °C overnight.

In silico analyses of seiseln intergenic regions of egc loci. FASTA files of DNA sequences of egc loci, seiseg intergenic regions, and selu and selu_v genes were retrieved from the National Centre for Biotechnology Information (NCBI) nucleotide database (Table 2). For strains A900322, FRI572, RF122, 382F, Mu50, N315 and MRSA 252, the DNA sequences of the forward primer PSE1 and the reverse primer PSE4 of Letertre et al. (2003) were used to identify the seiseln intergenic regions bearing pseudogene ψent1 and ψent2 and the selu or selu_v sequences. The ORF Finder (NCBI) was used to identify the pseudogenes ψent1 and ψent2 and the selu or selu_v genes (). Alignments performed using the CLC Free Workbench software version 3.0.2 (CLC bio; ) on these PSE1PSE4 sequences allowed identification of the PSE2 primer sequence of Letertre et al. (2003) in the selu and selu_v genes and of the 15 bp insertion that converts the pseudogene sequences to the selu or selu_v gene. Unless otherwise indicated, the nucleotide numbering system used herein refers to the first base of the PSE1 forward primer (T) as nt 1.

Table 2. DNA sequences of egc loci, seiseg intergenic regions, and selu and seluv genes

The database sequences for the selu genes of strains 352E, 383F and FRI137 were incomplete with respect to the PSE1PSE4 region. Using the BLAST program, the available sequences for strains 352E (AY205305), 383E (AY205307) and FRI137 (AY205306) gave best matches of 1040/1052 (98.9 %), 1041/1052 (99.0 %) and 1024/1024 (100 %), respectively, with that of strain RF122, in line with their own inter-identity values (Letertre et al., 2003). Accordingly, the necessary sequence additions to the 5' and/or 3' ends to obtain complete in silico PSE1PSE4 regions for these three strains were based on the complete PSE1PSE4 sequence of strain RF122. All of the complete or completed PSE1PSE4 sequences were analysed using RestrictionMapper version 3 () to identify restriction enzyme cleavage sites.

DNA preparation. Extraction of genomic DNA from S. aureus was performed as previously described, with some minor modifications (Fitzgerald et al., 1997). The procedure involves lysostaphin lysis, proteinase K treatment, addition of EDTA, sarkosyl and cetyltrimethylammonium bromide (CTAB), extraction with chloroform : isoamyl alcohol and phenol/chloroform : isoamyl alcohol, and precipitation with ethanol.

PCR. All primers were synthesized by MWG Biotech. For amplication of the seiseln intergenic region bearing the ψent1 and ψent2 pseudogenes and the selu or selu_v gene, the forward/reverse primer pairs PSE1/PSE4, PSE2/PSE4 and PSE2/PSE6 described by Letertre et al. (2003) were used. The primers PSE1 and PSE4 are designed to amplify the egc locus region encoding either the pseudogenes ψent1 and ψent2 or the selu or selu_v gene from the stop codon of the sei gene through the first 172 nt of the 5' end of the seln gene. The 27-mer forward primer PSE2 includes the specific 15 nt insertion of the selu or selu_v gene. PCR assays with the primer pairs PSE2/PSE4 and PSE2/PSE6 yield 790 and 142 bp amplimers, respectively, when the selu or selu_v gene is present (Letertre et al., 2003).

PCR was performed in a final volume of 25 µl. The reaction mix contained 22.5 pmol each primer. Each reaction contained 15.2 µl H₂O, 1 µl each primer (forward/reverse combinations PSE1/PSE4, PSE2/PSE4, PSE2/PSE6), 1 µl dNTP mix, 0.3 µl Taq polymerase (5 U µl¹), 2.5 µl 10x Thermophilic DNA polymerase buffer (Mg²⁺ free), 3 µl MgCl₂ (15 mM) and 1 µl template DNA. All reagents were supplied by Promega. Thermal cycling conditions were: 94 °C for 4 min, 7 cycles of (94 °C for 30 s, 56 °C for 30 s, 72 °C for 4 min), 21 cycles of (94 °C for 30 s, 56 °C for 35 s, 72 °C for 4 min) and a final elongation step of 72 °C for 7 min. DNA preparations from strains A900322, Mu50, RF122, MRSA 252, MW2 and NCTC 8325-4 were used as controls.

To control for the presence of sufficient template DNA, the test samples were examined by PCR with a primer set that anneals to staphylococcal 16S rRNA genes, generating a 228 bp amplimer (Monday & Bohach, 1999), using the thermal cycling conditions of Smyth et al. (2005). All PCR products were analysed on 2 % (w/v) agarose gels alongside a 100 bp ladder (NEB), stained with ethidium bromide and visualized under UV light.

Restriction enzyme digestion of PSE1PSE4 amplimers. Using the above PCR protocol, forward/reverse primer pair PSE1/PSE4 was used to amplify the pseudogene- or selu/selu_v-containing region. PCR product (5 µl) was run on a 1 % agarose gel to confirm that the reaction had been successful. Restriction endonuclease digestion was performed in a final volume of 30 µl. For endonuclease HindIII cleavage, this contained 17 µl H₂O, 6 µl PCR product, 3 µl 10x Buffer E (60 mM Tris/HCl, pH 7.5, 1 M NaCl, 60 mM MgCl₂, 10 mM DTT), 3 µl 10x BSA (100 µg ml¹) and 1 µl HindIII restriction endonuclease (Promega; 10 U µl¹). The digestion mixture was incubated overnight at 37 °C. Endonuclease HphI cleavage was carried out as above, except that the digestion mixture contained 3 µl 10x NE Buffer 4 (50 mM potassium acetate, 20 mM Tris-acetate, 10 mM magnesium acetate, 1 mM DTT, pH 7.9) and 1 µl Hph1 restriction endonuclease (NEB; 5 U µl¹). For endonuclease BccI, the digestion mixture contained 3 µl 10x NE Buffer 1 (10 mM Bistris propane-HCl, 10 mM MgCl₂, 1 mM DTT, pH 7.0) and 1 µl BccI restriction endonuclease (NEB; 10 U µl¹). All digests were analysed on 2.5 % (w/v) agarose gels alongside a 100 bp ladder (NEB).

Sequencing of PSE1PSE4 PCR product. PCR was performed for the PSE1PSE4 amplimer of strain A900322. The PCR product was purified using the High Pure PCR Product Purification kit (Roche) according to the manufacturer's instructions. The PCR amplimer was sequenced by GATC Biotech. The derived sequence (accession no. DQ993159) was aligned with the complete egc locus sequence of strain A900322 (AF285760) and with the seiseg intergenic region of strain FRI572 (AF156894) using the CLC Free Workbench software version 3.0.2. The BLAST program () was used to interrogate the NCBI database for matches for the derived sequence.

In silico analyses of PSE1PSE4, seiseln intergenic regions of egc loci
Alignment of the PSE1PSE4 region sequences of the four ψent1⁺ ψent2⁺ strains A900322 (1066 nt), and Mu50, N315 and FRI572 (each 1135 nt), revealed a 69 nt deletion in the ψent1ψent2 sequence of strain A900322 between nt G 3462 and nt G 3463 of the original egc cluster sequence AF285760 compared with those of the three other ψent1⁺ ψent2⁺ strains in which this 69 nt sequence is identical (5'-ACCGAGCATGATGGAAATCAAATAGATAAAAATAATTCAACTGATAACTCTCATAATATCTTAATTAAA-3') (Fig. 1). This 69 nt sequence is also 100 % conserved in the sequences of two of the selu⁺ strains (383F, 352E) and the two strains (382F and MRSA252) examined. In the other two selu⁺ strains RF122 and FRI137 there is an inversion of TC to CT at nt 37 and nt 38 of the 69 nt sequence (see TC bold type in 69 nt sequence above). The missing 69 nt in the strain A900322 sequence occur in the 23 nt ψent1 and ψent2 reading-frame overlap region that was originally described by Jarraud et al. (2001a) to harbour the start codon of pseudogene ψent2 and the stop codon of pseudogene ψent1, namely, 5' 3448-ATG[ψent2]TATGGCGGTGTG[Δ69 nt]GTTTATGA [ψent1]-3470 3' (bold type signifies features in the sequence mentioned above). In strain A900322, the ψent1ψent2 pseudogene sequence starts at nt 154 and ends at nt 856 of the PSE1PSE4 amplimer sequence, whereas in strains Mu50, N315 and FRI572, the ψent1ψent2 sequence starts at nt 154 and ends at nt 925.

(12K): Fig. 1. Schematic diagram of PSE1PSE4 PCR products bearing pseudogenes ψent1 and ψent2 and the selu or seluv gene. The hatched area in the pseudogene-bearing PSE1PSE4 PCR product (1135 bp) corresponds to the 69 bp sequence missing in the pseudogene overlap region of strain A900322. The cleavage sites of restriction endonucleases HindIII and BccI for a pseudogene-bearing PCR product, e.g. from strain Mu50, are shown with the sizes of the fragments indicated in bp. One of the BccI sites falls within the 69 bp sequence lacking in the strain A900322 sequence. The filled area in the selu- or seluv-bearing PSE1PSE4 PCR product (1149 or 1150 bp) corresponds to the signature 15 bp insert in the ψent1 pseudogene, which provides the unique cleavage site for restriction endonuclease BccI. The cleavage sites of restriction endonucleases HphI and BccI for a selu- or seluv-bearing PCR product, e.g. from strain RF122 and 382F, respectively, are shown with the sizes of the fragments indicated in bp.

The PSE1PSE4 sequences of three of the selu⁺ (RF122, 352E, FRI137) and the two (382F, MRSA252) strains are 1149 nt long. That of strain 383F (selu⁺) is 1 nt longer due to the insertion of a thymine in its PSE1PSE4 sequence outside the selu ORF between three thymines (T 68T 70). In PSE1PSE4 sequences, the selu or selu_v gene begins at nt 154 (nt 155 in strain 383F) and ends at nt 939 (940 in strain 383F). Analysis of the 1149/1150 nt long selu and selu_v sequences compared with PSE1PSE4 sequences of strains A900322, Mu50, N315 and FRI572 revealed the 15 nt insertion between G 218 and T 219 of the ψent1 sequences (5' 219-CTCTAAAATTGATGG-233 3' of the selu or selu_v genes) [this 15 nt sequence comprises nt 1327 of the PSE2 primer of Letertre et al. (2003)]. All six of the selu- or selu_v-bearing sequences have a single adenine deletion compared with the ψent1 sequences of strains A900322, Mu50, N315 and FRI572 from a run of seven adenines (A 365A 371 of the ψent1 pseudogenes corresponding to nt A 380A 385 of the selu and selu_v genes). This single adenine deletion abolishes the opal/umber stop codon TGA present at the end of pseudogene ψent1.

Each of the 10 PSE1PSE4 sequences was analysed using the RestrictionMapper version 3 program (Table 3). Of the enzymes with one restriction site, endonuclease HindIII cut the PSE1PSE4 regions between A 958 and A 959 in strains Mu50, N315 and FRI572 (Fig. 1), A 889 and A 890 in strain A900322, 33 nt downstream of the ψent2 stop codon TAA and 5 nt upstream of the start codon of the seln gene in the strains with pseudogenes ψent1 and ψent2, but did not cut PSE1PSE4 regions in strains bearing the selu and selu_v genes (Table 3). In contrast, restriction endonuclease HphI cut the PSE1PSE4 region between T 850 and A 851 in strains with selu or selu_v genes (Fig. 1), 86 nt upstream of the stop codon, but did not cut PSE1PSE4 regions in strains bearing the pseudogenes ψent1 and ψent2.

Table 3. In silico analysis of restriction endonuclease cleavage of seiseln intergenic sequences of the egc locus between primers PSE1 and PSE4 containing pseudogenes ψent1 and ψent2 or the selu or seluv gene

The only enzyme found to cut the sequence of the 27-mer PSE2 forward primer (Letertre et al., 2003) was restriction endonuclease BccI. The insertion of the 15 bp sequence creates a unique BccI cleavage site between A 376 and A 377 (nt 17 and nt 18 of the PSE2 primer) of PSE1PSE4 amplimers of selu⁺ and strains (Fig. 1). PSE1PSE4 regions bearing pseudogenes ψent1 and ψent2 are cleaved into three fragments, whereas those bearing selu and selu_v genes are cleaved into four fragments (Table 3). Pseudogene- and selu- or selu_v-bearing strains can be readily distinguished by the presence of a 496 nt BccI fragment and 376/377 nt and 135 nt BccI fragments, respectively (Table 3).

PCR-RFLP analysis of PSE1PSE4 amplimers using restriction endonucleases HindIII and HphI in tandem
Strains Mu50 and A900322 (ψent1⁺ ψent2⁺), RF122 (selu⁺) and MRSA 252 () were examined using the PSE1/PSE4, PSE2/PSE4 and PSE2/PSE6 primer sets of Letertre et al. (2003) (Fig. 2ac). All four strains yielded PCR products of approximately 11351150 bp with the PSE1/PSE4 primers. Strains Mu50 and A900322 did not give amplimers with the PSE2/PSE4 and PSE2/PSE6 primer pairs, confirming their ψent1⁺ ψent2⁺ status. In contrast, strains RF122 and MRSA 252 produced PCR products of approximately 790 and 142 bp with primers PSE2/PSE4 and PSE2/PSE6, respectively (Fig. 2b, c), confirming the presence of selu and selu_v genes, respectively. Strains NCTC 8325-4, MW2 and COL were subjected to PCR as negative controls for the egc locus. None yielded PSE1/PSE4, PSE2/PSE4 or PSE2/PSE6 amplimers.

(71K): Fig. 2. PCR amplification and PCR-RFLP analysis of seiseln intergenic regions of S. aureus harbouring egc loci. PCR amplimers are shown of strains A900322 (lane 1), Mu50 (lane 2), RF122 (lane 3) and MRSA 252 (lane 4), using (a) primers PSE1 and PSE4 showing an amplimer of approximately 1150 bp, (b) primers PSE2 and PSE4 showing an amplimer of approximately 790 bp from strains RF122 and MRSA 252, and (c) primers PSE2 and PSE6 showing an amplimer of approximately 142 bp for strains RF122 and MRSA 252. PCR-RFLP analysis of PSE1PSE4 amplimers digested with restriction endonucleases (d) HindIII, (e) HphI and (f) BccI. PCR amplimers of strains A900322 and Mu50 bearing pseudogenes ψent1 and ψent2 were cut to give bands at approximately 958 and 176 bp by endonuclease HindIII (d) but not by endonuclease HphI (e), whereas PCR amplimers of selu+ and strains RF122 and MRSA 252, respectively, were not cut by endonuclease HindIII (d) but were digested by endonuclease HphI to give bands at approximately 850 and 298 bp. Undigested PSE1PSE4 PCR products of 1135 bp from strains A900322 and Mu50 can be seen in lanes 1 and 2 of (d) and of 1150 bp from strains RF122 and MRSA 252 in lanes 3 and 4 of (e) (Fig. 1a). PCR products of pseudogene-positive strains A900322 and Mu50 were cut by endonuclease BccI to give bands at approximately 496 and 586 bp (f), while those of selu-positive strains RF122 and MRSA 252 were cut to show bands at approximately 586, 376 and 135 bp (white arrow). Lanes M contain 100 bp ladders; 500 bp markers are asterisked.

To test whether isolates bearing either the pseudogenes ψent1 and ψent2 or the selu gene or the selu_v gene could be distinguished using restriction endonucleases HindIII and HphI, the PSE1PSE4 amplimers of strains A900322, Mu50, RF122 and MRSA 252 were incubated overnight with these restriction enzymes. As anticipated from the in silico analysis, amplimers of strains bearing the pseudogenes (A900322 and Mu50) were digested by endonuclease HindIII, but not by endonuclease HphI, whereas amplimers of strains harbouring the selu or selu_v gene (RF122 and MRSA 252) were restricted by endonuclease HphI, but not by endonuclease HindIII, yielding fragments of the expected approximate sizes (Fig. 2d, e).

Eight bovine isolates of RAPD type 7 (Fitzgerald et al., 2000), which were known to possess or lack the sec and tst genes but had not been otherwise screened for SE or SEl genes, were tested for the presence of pseudogenes or the selu/selu_v gene by PCR and PCR-RFLP analysis. In all cases PSE1PSE4, PSE2PSE4 and PSE2PSE6 amplimers characteristic of isolates possessing an egc locus with the selu or selu_v gene were obtained. These isolates were all confirmed to have the selu or selu_v gene on the basis of the HphI susceptibility and HindIII resistance of their PSE1PSE4 amplimers (data not shown).

PCR-RFLP analysis of PSE1PSE4 amplimers using restriction endonuclease BccI
PCR-RFLP analysis of PSE1PSE4 amplimers using endonuclease BccI was carried out for the egc archetypal strain A900322 and the genome-sequenced strains Mu50, RF122 and MRSA 252 (Fig. 2f). The pseudogene-positive strains A900322 and Mu50 yielded a restriction product of approximately 496 bp, whereas strains RF122 and MRSA 252, with selu and selu_v genes, respectively, yielded restriction products of approximately 376 and 135 bp (Fig. 2f, Table 3).

Ten animal isolates known to possess an egc locus (the selo, selm, sei, seln and seg genes) five from cows, two from goats and three from chickens were identified as having pseudogenes and the selu or selu_v gene using PCR with the PSE primer sets and PCR-RFLP with endonucleases HindIII and HphI in tandem. The three chicken isolates and two of the bovine isolates (DS37 and 1007) possessed pseudogenes, while the two goat isolates and three of the bovine isolates had the selu or selu_v gene (Table 1). Their PSE1PSE4 amplimers were then examined by PCR-RFLP using restriction endonuclease BccI in an operator-blinded manner. The egc loci were all correctly identified, those bearing pseudogenes giving a restriction product of approximately 496 bp and those bearing a selu or selu_v gene giving restriction products of approximately 376 and 135 bp, respectively (data not shown).

Sequencing of the PSE1PSE4 amplimer of strain A900322
There was no evidence from the PCR-RFLP studies to suggest that the egc locus of strain A900322 differed from that of the ψent1⁺ ψent2⁺ strain Mu50. Given that the missing 69 nt were present in all other published sequences of selu or selu_v genes and ψent1ψent2 pseudogenes in the NCBI database, resequencing of the pseudogene region of the egc locus of strain A900322 was performed. The 694 nt-long sequence obtained was from nt A 64 to nt G 757 of the PSE1PSE4 amplimer (accession no. DQ993159). This encompasses all of pseudogene ψent1 and the first 288 nt of pseudogene ψent2, including the crucial pseudogene ψent1ψent2 overlap region in which the 69 nt sequence in question would occur, if present.

Using the CLC Free Workbench software version 3.0.2, this 694 nt sequence was aligned with the complete egc locus sequence of strain A900322 (AF285760; 6418 nt) and the seiseg intergenic spacer region of strain FRI572 (AF156894; 2016 nt). With the exception of 1 nt difference (C 342 in AF156894 versus T 189 in DQ993159), the derived sequence showed complete nucleotide concordance with nt A 154G 847 of the strain FRI572 intergenic spacer region of Monday & Bohach (2001). In the case of the complete egc locus sequence of Jarraud et al. (2001a), there was a 100 % match between the derived sequence DQ993159 and the complete egc locus sequence AF285760 of strain A900322 from nt A 2982 to nt G 3606, barring the absence of the 69 nt between nt G 3462 and nt G 3463 of the AF285760 sequence. Using BLAST, the 694 nt DQ993159 sequence gave a 100 % match with ψent1ψent2 regions of the genomic sequences of strains Mu50 and N315, both of which possess the 69 nt sequence.

Taken together, these findings revealed that the missing 69 nt were present in the resequenced pseudogene ψent1ψent2 overlap region of strain A900322. The 23 nt ψent1 and ψent2 reading-frame overlap region of strain A900322 originally described by Jarraud et al. (2001a) to harbour the start codon of pseudogene ψent2 and the stop codon of pseudogene ψent1 is thus 26 nt long, as it is in the other ψent1⁺ ψent2⁺ strains Mu50, N315 and FRI572.

To date, confirmation of PCR findings in relation to the nature of the egc locus has only been possible by sequencing of PCR products. As demonstrated herein, PCR-RFLP can be utilized for verification of the nature of egc loci determined by PCR when applying the primer sets of Letertre et al. (2003). Chini et al. (2006) have recently used restriction endonuclease HindIII to distinguish MRSA strains bearing the egc1 locus (selo selm sei ψent1 ψent2 seln seg) and the egc2 locus (selo selm sei selu seln seg). Using a clearly defined amplimer, the present findings not only confirm the value of using endonuclease HindIII, but further extend the use of restriction enzymes by inclusion of endonuclease HphI in tandem for reverse susceptibility of the egc1 and egc2 loci. Furthermore, the usefulness of endonuclease BccI is pinpointed by the fact that the 15 bp insert leading to the formation of the selu and selu_v genes from pseudogenes ψent1 and ψent2 creates a unique cleavage site. PCR-RFLP either with endonucleases HindIII and HphI in tandem or with endonuclease BccI alone with the PSE1PSE4 amplimer can also be used as a primary screening test for these egc loci. The described PCR-RFLP procedure is rapid and amenable to the screening of batches of clinical isolates. However, strains with atypical egc loci are known to occur (Blaiotta et al., 2004; Thomas et al., 2006). Both the PCR procedure of Letertre et al. (2003) and the PCR-RFLP procedure described herein may be affected by insertion sequences and by egc gene recombination or deletion events such as those reported by Thomas et al. (2006).

Direct repeats ATTT, AAGG and CATGAT were identified as the insertion sites of the newly described transposase 8-like/retrovirus integrase-like cassette in the sei, seln and seg genes of isolates apparently lacking these genes by PCR analysis (Thomas et al., 2006). Analysis of the seiseln intergenic region reveals that the nucleotide motif ATTT occurs 21 times in strain Mu50 and 22 times in strains RF122 and MRSA 252 (equivalent to 21 % of the total ATTT direct repeats in the egc locus of archetypal strain A900322). Thus, on this basis the seiseln intergenic region would seem to be a potential hot spot for insertion sequences such as the transposase 8-like/retrovirus integrase-like cassette.

The in silico analysis of egc loci based on the available NCBI database sequences annotated for the first time that the seiseln intergenic region of the archetypal egc⁺ strain A900322 (AF285760; Jarraud et al., 2001a) differed from the inter-pseudogenic sequences of three other ψent1⁺ ψent2⁺ strains, one of which (strain FRI572) is the egc prototype strain of Monday & Bohach (2001), by the absence of a 69 nt-long stretch, as well as from the sequences of six selu⁺ or strains. This strongly suggested either that the seiseln intergenic region of strain A900322 was atypical due to a unique deletion or that the sequence lodged in the database was incorrect or faulty. Resequencing (DQ993159) and sequence alignments confirmed the presence of the missing 69 nt in the ψent1ψent2 overlap region of strain A900322.

Thomas et al. (2006) have reported a new seiseln intergenic gene in S. aureus strain A900624 that they designate selu2. The selu2 gene results from a single adenine deletion which converts the 772 nt ψent1ψent2 region into a 771 nt ORF. However, as pointed out herein, this same single adenine deletion is present in all the selu and selu_v gene sequences in the NCBI database. Moreover, the selu2 ORF differs in crucial respects from all published selu and selu_v sequences. The latter are 786 nt long and possess the 15 nt signature sequence 5' 219-CTCTAAAATTGATGG-233 3' between nt G 218 and nt T 219 of pseudogene ψent1. From the description of the selu2 gene, namely, that nt 1 to nt 402 demonstrated 99.3 % (399/402) identity with pseudogene ψent1 of strain Mu50, and that nt 376 to nt 771 had 100 % identity to pseudogene ψent2 of strain Mu50, the selu2 gene is not a selu or selu_v gene sensu stricto but merely a 1 nt frameshift of the ψent1ψent2 sequence that deletes the opal/umber stop codon of pseudogene ψent1.

Furthermore, while the natures of three nucleotide differences in the first 402 nt of the selu2 gene are not available through the NCBI database, use of RestrictionMapper version 3 in silico with the PSE1PSE4 sequence of strain Mu50 with the single adenine removed confirmed that the selu2 gene is cleaved by endonuclease HindIII, but not by endonuclease Hph1, and by endonuclease BccI to yield a 496 bp fragment. It is thus not possible to distinguish selu2-bearing from ψent1ψent2-bearing isolates using the PCR-RFLP procedures described herein. Since the selu2 ORF is distinctly different from all selu genes described to date in the absence of the signature 15 nt descriptor of selu genes and its endonuclease HindIII, HphI and BccI cleavage characteristics, we propose that this gene be redesignated the selw gene. This is coincidentally appropriate, since the single adenine deletion from the ψent1ψent2 sequence removes the sixth-from-last tryptophan residue of the translated ψent1 sequences of strains A900322, Mu50, N315 and FRI572.

Neither the PCR-RFLP analyses herein nor the procedure of Chini et al. (2006) distinguishes selu- from selu_v-bearing isolates. However, this appears to be possible using the PCR-RFLP scheme of Blaiotta et al. (2004, 2006). These authors demonstrated seven different PCR-RFLP groups based on restriction of a 3375 bp selmseg amplimer using endonucleases EcoRI, TaqI, AluI and CfoI. Strain A900322 of the egc1 locus type belonged to REA group 2, strain FRI137 of the egc2 locus type belonged to REA group 2, and strain AB-8802, probably of the egc3 locus type because it possesses the sei_v and seg_v genes that co-exist with the selu_v gene in strain 382F, belonged to REA group 6. However, these PCR-RFLP differences are not based on differences in the seiseln intergenic regions of egc loci, as endonucleases TaqI, AluI and CfoI do not cut the PSE1PSE4 amplimer, and endonuclease EcoRI has the same single cleavage site on the basis of the restriction analyses carried out herein using RestrictionMapper version 3, but rather are dependent on the presence or absence of sequence variation in other genes, e.g. sei_v, seg_v and seln_v.

While there has been a plethora of reports on the occurrence of the five common genes of the egc locus (selo selm sei seln seg) in strains of S. aureus of human and animal origin over the past 5 years, very limited data exist on the relative occurrence of the ψent1 and ψent2 pseudogenes and of the selu and selu_v genes within the egc locus (Letertre et al., 2003; Blaiotta et al., 2004; Bania et al., 2006a, b; Chini et al., 2006; Thomas et al., 2006). The combined PCR and PCR-RFLP methodology described should aid in the determination of the occurrence of egc1 and egc2/egc3 type loci in isolates of different human and animal clinical origin, and in the evaluation of their association, if any, with the disease-causing potential of S. aureus harbouring these different intergenic seiseln egc loci types.

M. M. C. was in receipt of a studentship from the Sarah Purser Medical Research Fund and of a Trinity College postgraduate research studentship. Thanks to Amy Wong and Marie-Laure De Buyser, who provided information on the origins of Food Research Institute (FRI), University of Wisconsin, and Agence Française de Sécurité Sanitaire des Aliments (AFSSA) isolates, to Tim Foster and Sophie Jarraud, who provided control strains of known egc type, and to Davida Smyth and Ross Fitzgerald, who originally characterized the animal isolates used.