Characterisation of Australian MRSA Strains ST75- and ST883-MRSA-IV and Analysis of Their Accessory Gene Regulator Locus

Background Community-acquired methicillin-resistant Staphylococcus aureus have become a major problem in Australia. These strains have now been isolated throughout Australia including remote Indigenous communities that have had minimal exposure to healthcare facilities. Some of these strains, belonging to sequence types ST75 and ST883, have previously been reported to harbour highly divergent alleles of the housekeeping genes used in multilocus sequence typing. Methodology/Principal Findings ST75-MRSA-IV and ST883-MRSA-IV isolates were characterised in detail. Morphological features as well as 16S sequences were identical to other S. aureus strains. Although a partial rnpB gene sequence was not identical to previously known S. aureus sequences, it was found to be more closely related to S. aureus than to other staphylococci. Isolates also were screened using diagnostic DNA microarrays. These isolates yielded hybridisation results atypical for S. aureus. Primer directed amplification assays failed to detect species markers (femA, katA, sbi, spa). However, arbitrarily primed amplification indicated the presence of unknown alleles of these genes. Isolates could not be assigned to capsule types 1, 5 or 8. The allelic group of the accessory gene regulator (agr) locus was not determinable. Sequencing of a region of agrB, agrC and agrD (approximately 2,100 bp) revealed a divergent sequence. However, this sequence is more related to S. aureus agr alleles I and IV than to agr sequences from other Staphylococcus species. The predicted auto-inducing peptide (AIP) sequence of ST75 was identical to that of agr group I, while the predicted AIP sequence of ST883 was identical to agr group IV. Conclusions/Significance The genetic properties of ST75/ST883-MRSA may be due to a series of evolutionary events in ancient insulated S. aureus strains including a convergent evolution leading to agr group I- or IV-like AIP sequences and a recent acquisition of SCCmec IV elements.


Introduction
A variety of community-acquired methicillin-resistant S. aureus (ca-MRSA) strains with divergent genetic backgrounds have emerged globally.In Australia some of these strains have been isolated from Indigenous people living in remote communities who have had minimal previous exposure to healthcare facilities.One of these strains, ST75-MRSA-IV, is a PVL-negative ca-MRSA frequently isolated from skin and soft tissue infections in people living in the northern regions of Australia [1].Recently it has been demonstrated that the multilocus sequence typing (MLST) genes in ST75 differ from other S. aureus.Using array hybridisation, the accessory gene regulator (agr) genes in ST75-MRSA-IV and in a related ST883-MRSA-IV strain could not be detected [2].
Although no phenotypic differences to other S. aureus were observed, it has been proposed to classify these strains as a new subspecies of S. aureus [3].
As a toxin signatory pathway, the agr locus genes encode AgrA, AgrB, AgrC, and AgrD that constitute a virulence regulating quorum-sensing system [4][5][6].This system responds to the extracellular concentration of a secreted auto-inducing peptide (AIP) derived from AgrD.AgrB is a putative processing enzyme, and AgrC serves as receptor for AIP.Quorum sensing via AIP regulates the expression of RNAIII derived from the hld gene.RNAIII up-regulates the production of secreted toxins, but downregulates genes encoding surface proteins.The central segment of the agr locus (the C-terminal two-thirds of AgrB, AgrD, and the Nterminal half of AgrC) shows a striking interstrain variation.
Known alleles cluster into four distinct groups [7].The division of S. aureus strains into these agr groups is based on their experimentally proven ability to inhibit or stimulate expression of virulence-related genes.The agent mediating this modulation of virulence gene expression is AIP.Its sequence is different in each agr group.AIP stimulates expression of virulence genes in its own agr group, but represses these genes in strains of a different agr group, resulting in bacterial interference [8].It has previously been proposed that the four different alleles of the agr cluster reflect an ancient division in S. aureus evolution [7].However, strains belonging to related clonal complexes based on MLST have been shown to have different agr groups.Furthermore, highly divergent clonal complexes have been shown to harbour identical agr alleles [2].Consequently, the concept of a straightforward division of S. aureus into four agr groups attaining some kind of subspecies status is not valid.Proposals for a phylogenetic tree of S. aureus, made to explain the discrepancies between overall genetic background and agr group affiliations, assume multiple steps of diversification and recombination [9;10].
As ST75-MRSA-IV may be a new subspecies of S aureus, the aim of this study was to characterise several of these isolates using phenotypic and genotypic methods.For genotypic characterisation, hybridisation patterns on DNA arrays were analysed.Genes encoding the 16S rRNA and the RNA subunit of RNase P (rnpB) genes were sequenced, as these genes have previously been used to infer taxonomy and phylogeny (see, e.g., [11][12][13][14]).As agr regulation is a vital pathway in S. aureus, the failure to detect known alleles of agr locus genes in ST75-MRSA-IV and ST883-MRSA-IV suggests the presence of new alleles of these genes.Consequently, a characterisation of the agr locus of these isolates was performed in this study.

Phenotypic characterisation
Morphologically on Columbia blood agar, isolates formed yellow-greyish colonies with beta-haemolysis.The clumping factor test was positive.Standard biochemistry was in accordance with the Vitek 2 S. aureus profile (Supplemental File S1).The urease test was negative for ST75-MRSA-IV, but positive for ST883-MRSA-IV.ST75-MRSA-IV and ST883-MRSA-IV isolates were also identified as S. aureus using MALDI-TOF.

Array hybridisation
Detailed hybridisation results are provided in Supplemental File S2.
ST75-MRSA-IV and ST883-MRSA-IV provided hybridisation patterns distinct from other S. aureus strains.Some of the characteristic S. aureus markers including coa, spa, katA, femA or sbi did not yield hybridisation signals in both amplification protocols, or were detectable only when using the sequenceindependent random-primer amplification (Supplemental File S2).This can be attributed to the present of distinct sequences which are sufficiently divergent to prevent their detection.Indeed, a previously published ST75 coa sequence (GenBank AB436988, strain JCSC1469, [15]) shows mismatches affecting primer and probe binding sites.The spa gene is known to be present in both strains, as a spa types have previously been reported ( [16 16607 was positive in the femA, sbi and katA (in frame) PCRs, but did not amplify in the katA (out frame) PCR.
Other species-specific genes such as gapA, nuc, setC, hld, and some of the genes encoding microbial surface components recognising adhesive matrix molecules (MSCRAMM), such as bbp, clfA/B, eno, fnbA/B and sdrC/D and the biofilm-associated gene icaA were detected.Other MSCRAMM genes, cna, ebpS, fib, map and vwb, as well as biofilm-associated genes icaC/D and bap were not detected.The sasG gene was detected in ST75, but not in ST883.Isolates did not yield signals with probes for capH, capI, capJ and capK genes corresponding to the capsule types 1, 5 or 8, in either the primerdirected, or the random amplification protocols.
The affiliation to agr groups could not be determined as no signals were obtained with primers and probes derived from all previously described S. aureus agr locus sequences.Other regulatory genes, sarA, saeR, saeS, vraS, were detectable, while vraR yielded weak (in ST75-MRSA-IV) or no signals (in ST883-MRSA-IV).
ST75-MRSA-IV carried the egc enterotoxin gene cluster locus including seg, sei, sem, seo and seu.Another gene from this cluster, sen, was only detectable by random amplification.The ST75-MRSA-IV isolates differed in the presence of the enterotoxin B gene seb.Eleven isolates yielded either no signals, or in the random amplification protocol only, while seven isolates were positive in the primer-directed amplification.Two isolates from the former group failed to produce enterotoxin B, while two isolates from the latter group were positive for toxin production (Supplemental File S3).Enterotoxin genes were not detected in ST883-MRSA-IV, and this strain did not produce enterotoxin B. This isolate also differed from ST75 with regard to some other markers such as ssl/ set and hysA alleles (Supplemental File S2).The gamma haemolysin locus (hlgA, lukF/S-hlg), PVL genes (lukF/S-PV) and the leukocidin homologues lukD/E and ''lukX/Y'' (NC_002951; SACOL2004 and 2005) were not detected.
Approximately 2,045 base pairs of agrB, agrD and agrC of the ST75-MRSA-IV isolate 03-17848 were sequenced (GenBank FJ154839).Sequences revealed differences to known agr S. aureus sequences.Although they were not completely identical, agrB, agrD and agrC sequences from the ST883-MRSA-IV isolate 06-16607 (GenBank HQ260328) resembled ST75-MRSA-IV.BLAST yielded an identity of 92% with query coverage of 100%.Both sequences are more closely related to S. aureus agr alleles than to other staphylococcal agr sequences (Figure 2).Among the S. aureus agr alleles, they are more closely related to agr group I and IV than to agr groups II or III.A protein sequence alignment (Supplemental File S4) revealed that the putative AIP sequence of ST75 was identical to agr group I, and that the predicted AIP sequence of ST883 was identical to agr group IV.

Discussion
Australian ca-MRSA strains characterised as ST75-MRSA-IV and ST883-MRSA-IV have previously been shown to have MLST alleles that are unique compared to other S. aureus genotypes (see http://saureus.mlst.net/).In this study we have shown that the array hybridisation patterns and the agr and rnpB gene sequences for these two strains also differ from other S. aureus strains.Furthermore, species-specific genes such as femA, coa, sbi or spa and capsule genes were either absent or their sequences were sufficiently divergent in the isolates studied to prevent their detection by DNA array hybridisation.Whether these differences justify the recognition of ST75 and ST883 as another species or as distinct subspecies [3] is a matter of debate.Of practical consequence is that nucleic acid based tests in general may not yield a correct result because their performance depends on the underlying sequence information.New variants or alleles of target genes may not be detected if sequences of probe or primer binding sites deviate from known sequences on which the design of the test relied.Thus, our DNA array results may indicate that known genome sequences do not reflect the whole biodiversity of S. aureus, and consequently whole genome sequencing of divergent strains such as ST75-MRSA-IV or ST883-MRSA-IV would be helpful.
While ST75-MRSA-IV are different from other S. aureus, the isolates from this strain are rather homogenous.Variability however was detected in the presence of seb gene and, correspondingly, in the production of enterotoxin B. Only isolates, in which seb was detected by primer-directed amplification were shown to produce enterotoxin B. Complete or partial absence of the seb gene or its replacement by a geno-and phenotypically divergent variant are likely explanations for this observation.The only other variations observed in ST75-MRSA-IV affected possibly mobile resistance markers.Although ST883-MRSA-IV lacks seb as well as the enterotoxin gene cluster egc and differs in some MLST alleles (see http://saureus.mlst.net/),this strain shares, in terms of hybridisation profile and rnpB gene sequences, some features with ST75-MRSA-IV (see above and Supplemental File S2).
The agr sequences of ST75 and ST883 are distinct from previously known agr sequences.Furthermore, an 8% difference between the agr sequences of these two strains was identified.For comparison, a fragment of identical length from agr II sequences of Mu50/N315 (CC5) and RF122 (CC151) differ by only 4%.
Based on DNA sequence, ST75/ST883 agr alleles could be considered as representatives of a novel agr group although they are less uniform than the other agr alleles.However, agr groups are conventionally defined by phenotype rather than by genotype.Although it was beyond the scope of this study to test ST75/ ST883 for bacterial interference with other S. aureus strains or to determine the amino acid sequence of the mature AIP from ST75 and ST883, predicted AIP amino acid sequences derived from gene sequences were analysed.Predicted AIP amino acid sequence differed for ST75 and ST883.ST75 yielded an agr group I AIP amino acid sequence, while the AIP amino acid sequence of ST883 was identical to agr group IV (Supplemental File S4).For this reason, despite the differences demonstrated in the agr gene sequences, these strains cannot be regarded as prototypic for a novel agr group.
The agr alleles I to IV are distributed world-wide.The agr groups I to III are known to include a variety of pandemic MRSA and MSSA strains.The agr allele IV is less diverse, found only in a few S aureus strains including ST50 and ST121 [9].Although no major MRSA strains belong to agr group IV, MSSA of ST121 are known to be common and widespread in several countries [17][18][19][20].In contrast, strains with ST75/ST883-like agr alleles have only been found in remote regions of northern Australia, where ST75-MRSA-IV is frequently isolated [1].Although no data on methicillin-susceptible precursor strains are available, we assume ST75-and ST883-MRSA strains originated in northern Australia, and have evolved from pre-existent, local MSSA strains by acquisition of SCCmec elements.While their genetic background differs from other S. aureus strains, ST75-and ST883-MRSA strains harbour SCCmec IV elements suggesting a recent acquisition of these elements.Thus we assume that ST75/ ST883 may have evolved independently from other S. aureus populations over an extended period of time and in recent years has been transformed to MRSA by the global ''infection-like'' spread of SCCmec elements affecting several staphylococcal species.
The genetic differences in ST75/ST883 may be due to a series of evolutionary events in an ancient, insulated S. aureus strain.S. aureus is a well adapted human parasite, which may have coevolved with humans resulting in a prehistoric radiation of S. aureus into a number of divergent agr alleles.Much later, in the ages of discovery and colonisation, a few S. aureus clonal lineages of European, East Asian or African origin may have been spread worldwide.For example, a recently published analysis of sequence polymorphisms in ST5 (a global sequence type belonging to agr group II) suggests a higher degree of diversity in Europe, and an expansion approximately 1,000 years ago [21].The agr-mediated bacterial interference [8] could have been a driving force for an evolution towards a very small number of competing allelic variants (I to IV) resulting either in extinction of other agr alleles, or in convergent evolution leading to similar AIP and agr gene sequences.Evolution towards a certain uniformity might be a stable strategy for competitors facing the same selective pressure, as economists have noted already long ago [22].Ancient, native S. aureus strains may have been able to survive due to convergent evolution or to recombination events resulting in the acquisition of one of the four globally predominant agr alleles.We observed that predicted AIP amino acid sequences of ST75 and ST883 are identical to AIP sequences of agr group I and IV, respectively, although their genetic background including the rest of their agr sequences is different.This could indicate a convergent evolution towards a limited number of main agr types.S aureus of ST93 (including another unique Australian caMRSA, the Queensland Clone, ST93-MRSA-IV, [23]) may be another example.Its overall genetic background also differs from the currently known sequences resulting in a highly divergent hybridisation profile in microarray experiments [2].Although in this regard it resembles ST75 and ST883, ST93 harbours an inconspicuous agr type III locus [2;24].
In order to prove this concept of S. aureus evolution, historical isolates and isolates from remote human communities and/or native populations in different parts of the world should be collected and scrutinised.This could provide an insight into the evolution and biodiversity of an important human pathogen.No patient data were stored or used for this study; thus formal ethics committee approval and informed consent from the patients were not required.

Phenotypic characterisation
Culture morphology was assessed on Columbia blood agar.The clumping factor test was performed using a commercial assay (Pastorex Staph-Plus, Bio-Rad, Munich, Germany).Standard biochemistry was determined by the Vitek 2 Gram Positive Identification card (BioMerieux, Nuertingen, Germany).MALDI-TOF analysis was performed using the MALDI Biotyper (Bruker, Bremen, Germany) according to manufacturer's specifications.

Array procedures
The DNA array (StaphyType system by Alere Technologies GmbH, Jena, Germany) used in this study included probes for 334 different target sequences which corresponds to, depending on the nomenclature used, approximately 180 distinct genes and their different alleles.This included species controls, typing markers (agr, capsule and SCCmec-related genes), genes encoding exotoxins or MSCRAMMs as well as resistance markers.
After overnight incubation on Columbia blood agar, colonies were harvested and lysed using lysostaphin, lysozyme and ribonuclease A followed by digestion with proteinase K.The DNA was purified using the QIAGEN device EZ1 (QIAGEN, Hilden, Germany) according to the manufacturer's tissue lysis protocol.
Two different protocols were used for linear amplification and biotin labelling of amplicons.The first, a target sequencedependent, primer-directed amplification protocol, [2;24;29;30] is a thermally synchronised iterated linear primer elongation method in which a mixture of antisense primers for all targets facilitated a simultaneous copying of all target sequences to single stranded DNA.Within this step, amplicons are labelled by incorporation of biotin-16-dUTP.
The second protocol is a target sequence-independent random amplification procedure [31][32][33][34] resulting in biotin labelled amplicons representing the complete input genome sequence.This method detects target genes for which the binding sites of the primers used in the first protocol were deleted or changed by nucleotide polymorphisms.Thus, this method is more suited to characterise strains for which sequence data are not available yet.Primers were used which consisted of a random sequence and of an invariant part.The latter was used as a primer binding site for a second round of amplification.Amplicons were also labelled by incorporation of biotin-16-dUTP.
Post amplification, the labelled samples were denatured and hybridised to the array.This was followed by several washing steps and the addition of a blocking reagent.Horseradish-peroxidasestreptavidin conjugate was added to the array, followed by incubation and washing.A precipitating dye (Seramun Green, Seramun, Heidesee, Germany) was then added, and the hybridisation was analysed using a designated reading device and software (ArrayMate by Alere Technologies GmbH).

Table 1 .
Primers used for amplification and sequencing of genes in ST75-MRSA-IV and ST883-MRSA-IV.