SmaI Restriction Site-Based Multiplex PCR for Typing of Hospital- and Community-Acquired Staphylococcus aureus

Abstract

Methicillin-resistant Staphylococcus aureus (MRSA) is an important nosocomial pathogen, and morbidity and mortality rates associated with this pathogen have increased markedly in recent years. MRSA strains are generally resistant to several classes of antibiotics and are therefore difficult and costly to treat. A major issue is to identify the sources of MRSA infections and to monitor their epidemic spread. In this study, we report the development of a typing technique for S. aureus, based on single-nucleotide polymorphism (SNP) variations in and around SmaI-restriction sites (CCCGGG). An assessment of the SmaI restriction site-based multiplex PCR (SmaI-multiplex PCR) typing (SMT) with respect to pulsed-field gel electrophoresis (PFGE) and multilocus sequence typing (MLST) revealed a high level of concordance in the clustering of the test strains. The SmaI-multiplex PCR was found to be more discriminatory than MLST/staphylococcal cassette chromosome mec (SCCmec) typing but less discriminatory than PFGE. SMT can provide real-time information for the investigation of ongoing S. aureus hospital outbreaks. SMT meets the criteria of a practical typing method: it is simple, reproducible, and highly discriminatory and does not require expensive equipment or specialist expertise. Consequently, SmaI-multiplex PCR has the potential to be used in routine clinical microbiology laboratories.

INTRODUCTION

Since its emergence in 1961, methicillin-resistant Staphylococcus aureus (MRSA) has become an important nosocomial pathogen, and morbidity and mortality rates associated with this pathogen have increased markedly in recent years. The National Nosocomial Infections Surveillance System (NNISS) of the Centers for Disease Control and Prevention (CDC) reports that approximately 80,000 patients/year worldwide acquire MRSA infections (1). Consequently, the control of MRSA has been made a high priority among health care professionals (11). Although a consensus has yet to be reached as to the most effective means of control, the majority of studies indicate that aggressive screening is effective in the early detection of colonized and infected patients (8, 10, 12). A major issue is to identify the sources of MRSA infections in a timely manner once an outbreak is suspected and to monitor the spread of the strains involved. This requires the use of reliable typing techniques for monitoring both epidemic and sporadic outbreaks. The rapid development of molecular approaches during the last 20 years has provided the opportunity to explore the genetic characteristics of individual strains of S. aureus (18). Consequently, a variety of molecular techniques have been introduced for studying the epidemiology of MRSA, including pulsed-field gel electrophoresis (PFGE), multilocus sequence typing (MLST), staphylococcal cassette chromosome mec (SCCmec) typing, Staphylococcus protein A gene (spa) typing, and multiple-locus variable number tandem repeat (VNTR) analysis (MLVA). PFGE, spa typing, and MLST are the most widely used typing techniques, with the former regarded as the gold standard due to its high discriminatory power and reproducibility (7, 17, 20). However, technical limitations restrict the use of PFGE, spa typing, and MLST in the majority of routine hospital laboratories: they are time-consuming and too expensive for routine screening, and they require specific expertise and specialist equipment. Although PFGE is usually reproducible in the context of a single-specialist laboratory, reproducibility between laboratories can sometimes be challenging (3). Currently, more than 20 sequenced S. aureus genomes are in the public domain (http://www.ebi.ac.uk/genomes/bacteria.html). The similarities between the core genomes of S. aureus are very high, with the nucleotide identities between orthologues ranging from 97.7% and 99.8% (16). The observed differences are largely due to single-nucleotide polymorphisms (SNPs). The aim of the current study was to exploit SNP variations between S. aureus genomes in and around SmaI restriction sites (CCCGGG) to develop a genotyping technique that can be performed in a few hours using technology that is available in many routine clinical microbiology laboratories.

MATERIALS AND METHODS

Strains and growth conditions.

Well-characterized strains of S. aureus and Staphylococcus epidermidis strains were obtained from various sources. Sequenced S. aureus strains CA-629 (NRS745), MRSA252 (Sanger 252/NRS71), MSSA476 (Sanger 476/NRS72, MW2 (NRS123), COL (NRS100), Mu3 (ATCC 700698/NRS2), Mu50 (ATCC 700699/NRS1), N315 (NRS70), USA300 (FPR3757/NRS482), USA1000 (AIS2006061/NRS483), and USA1100 (HIP12899/NRS484) and S. epidermidis strain RP62A (ATCC 35984/NRS101) were obtained from the Network on Antimicrobial Resistance in Staphylococcus aureus (NARS) program, S. aureus JH9 from Herminia de Lencastre (Instituto de Tecnologia Química e Biológica, Portugal), S. aureus epidemic MRSA strain EMRSA-15 (sequenced strain H050960412) from Mark Enright (Imperial College, London, United Kingdom.), S. aureus NCTC8325 from the National Collection of Type Cultures (Health Protection Agency, London, United Kingdom), and S. epidermidis ATCC 12228 from Jan Maarten van Dijl (University of Groningen, The Netherlands). Ninety methicillin-susceptible S. aureus (MSSA)/MRSA clinical strains were obtained from three National Health Service (NHS) hospitals in the North East of England, the Freeman Hospital (FH), the Royal Victoria infirmary (RVI), and the Newcastle General Hospital (NGH), all belonging to the Newcastle upon Tyne Hospitals NHS Foundation Trust. The strains were isolated from clinical and screening samples. Initial identification was performed using conventional methods (selective media, coagulase production, latex agglutination [PBP2a], and antibiotic susceptibility). The presence of the diagnostic coagulase (coa-Fd, 5′-GAAGAGAAGAAAGTTGAAGAACC-3′; coa-Rd, 5′-CTTAAAGATGGTCTGTTTTGTTCC-3′) and methicillin resistance (mecA-Fd, 5′-GCAATCGCTAAAGAACTAAGTA-3′; mecA-Rd, 5′-TGACACGATAGCCATCTTC-3′) genes was confirmed by diagnostic PCR. Bacterial strains received on blood agar plates were tested directly by the SmaI restriction site-based multiplex PCR (SmaI-multiplex PCR) technique, whereas those received on agar slopes were subcultured on brain heart infusion (BHI) agar prior to typing.

Extraction of the chromosomal DNA.

Six milliliters of culture grown overnight in 10 ml of BHI broth at 37°C was transferred into four 1.5-ml microcentrifuge tubes and centrifuged (3 min, 13,000 × g at room temperature). The supernatants were discarded and the pellets combined by resuspending in 550 μl of lysis buffer lacking lysozyme (25 mM Tris, 25 mM EDTA, 0.3 M sucrose). Approximately 500 μl of glass beads was added in a 2-ml screw-cap tube, and the mixture was shaken in a Mikro-dismembrator (Braun, Germany) for 5 min at 2,600 rpm. After cell disruption, the mixture was centrifuged for 10 min at 13,000 × g, and 180 μl of supernatant was transferred to a 1.5-ml microcentrifuge tube. The remaining extraction steps were carried out using a Qiagen DNeasy extraction kit according to the manufacturer's instructions.

For the rapid extraction of chromosomal DNA, two colonies from an overnight culture were transferred into 20 μl of sterile Milli-Q water containing 3 μl lysostaphin (1 mg/ml; Sigma) in a microcentrifuge tube and mixed thoroughly. The cell suspension was incubated at 37°C for 15 to 20 min, heated to 100°C for 10 min, and then centrifuged at 13,000 × g for 5 min. Three microliters of the resulting supernatant was used in the multiplex PCR.

Uniplex and multiplex PCR.

Visual oligonucleotide modeling platform (OMP) 6.0v DNA software (BioGene, United Kingdom) was used to design primer pairs (Table 1). The primers used in the multiplex PCR were optimized with respect to secondary structure, homology, annealing temperature, and assay conditions. Forward primers contained the SmaI site at their 3′ ends and the reverse primers was designed from a conserved region between approximately 80 and 300 bp downstream of the target SmaI site. The primer pairs were designed to be compatible in a single 10-fold multiplex PCR and to conform to predetermined parameters, including the optimal length of the primers and PCR products, isothermal melting temperature, and the avoidance of primer dimers and secondary structures. Primers were obtained from MWG (Germany) and BioGene (United Kingdom).

Table 1.

Oligonucleotide primers for SmaI-multiplex PCR typing

Group name or no.	Primera	Oligonucleotide sequences (5′ to 3′)	GC (%)	Temp (°C)	Amplicon no.	Amplicon size (bp)
mecA	mecA_F	GTGGAATTGGCCAATACAGGAAC	48	60.6	1	502
	mecA_R	GTTAGTTGAATATCTTTGCCATC	35	55.3
1	SmaI-1_F1*	TTCATATTTTCCAATCGCCCG	42.9	55.9	2	169
	SmaI-1_F2*	TTCATATTTTTCAATCGCCCG	38.1	54.0
	SmaI-1_R	CTATAATTCCCTATTTTGACATCTTC	30.8	56.9
2	SmaI-2_F	GCAGGTAATGCAGAAGACCCG	57.1	61.8	3	268
	SmaI-2_R	CATTAGTTCCGTATGTAATCTTATC	32	56.4
3	SmaI-3_F	GACCTGGTGATCCCG	61.1	58.2	4	94
	SmaI-3_R	GGTAATTCTGTGCCACTATTAGC	43.5	58.9
4	SmaI-4_F	TTGTAGGCCTATTCCCG	53	58.6	5	120
	SmaI-4_R1*	CAGTAATCTTAACTTCAATTGTGTCACC	35.7	60.7
	SmaI-4_R2*	CAGTAATTTTAACTTCAATTGTGTCACC	32.1	59.3
5	SmaI-5_F	GAACGAAAATGCGCCCCG	61.1	58.2	6	Variable
	SmaI-5_R	AGAAAACGTATTTCTTTATCTTTTCG	26.9	55.3
6	SmaI-6_F	GAACTGGTGATCCTGAATTCCCG	52	62.4	7	223
	SmaI-6_R	GTCGAACGTGGACCAAAAGGC	57	61.8
7	SmaI-7_F	CGCTTTTGATTTTTTTGACCCG	40.9	56.5	8	146
	SmaI-7_R	CATCCGGAGTTTGCTCATGAC	52.4	59.8
8	SmaI-8_F	AGCAGGCAGTGCACCCG	70.6	60.0	9	315
	SmaI-8_R	CGATTTATGAGGTATGAAGGAAC	39.1	57.1
9	SmaI-9_F	CGGCCTTGTTTATTAGATCCCG	50	60.3	10	293
	SmaI-9_R	CTAAAAAGTATGTTAGGCACGTG	39.1	57.1

^aF, forward; R, reverse; *, two primers were used to due to the presence of SNPs.

Uniplex PCR was carried out in a reaction mixture volume of 50 μl, containing 200 μM dNTP, 5 μl of 10× reaction buffer, 1.5 mM MgCl₂, 0.8 μM concentrations of the primers, 1 μl of 5 U/μl Taq polymerase (Fermentas, Germany), and 3 μl of DNA template. The reaction conditions were as follows: initial denaturation for 5 min at 94°C, followed by 30 cycles of 94°C for 1 min, 52 to 50°C for 1 min, and 72°C for 1 min, with a final extension at 72°C for 5 min. The resulting amplicons were separated on 1 to 1.5% agarose at 90 V for 50 to 130 min. The gels were stained with ethidium bromide (0.5 μg/ml in 1× Tris-borate-EDTA [TBE]) and visualized with a gel documentation system.

Multiplex PCR was performed in a reaction mixture volume of 50 μl, containing 1× DNA polymerase buffer, 200 μM dNTP, 1.5 mM MgCl₂ (Fermentas, Germany), 8× bovine serum albumin (BSA; New England BioLabs), 0.8 μM concentrations of primers except mecA primers, which were 1.6 μM, 1 μl of 5 U/μl Taq polymerase (Fermentas, Germany), and 3 μl of DNA template. The reaction conditions were as follows: initial denaturation for 2 min at 96°C, followed by 35 cycles of 96°C for 40 s, 56°C for 40 s, and 72°C for 40 s, with a final extension at 72°C for 2 min. The resulting amplicons were separated on 4% low-melting-point agarose (LMP; BioGene, United Kingdom) in 1 × TBE at 110 V for 3 h and stained as above.

PFGE, MLST, and SCCmec typing.

For PFGE, chromosomal DNA was prepared in agarose plugs using a contour-clamped homogeneous electric field (CHEF) genomic DNA plug kit (Bio-Rad, United Kingdom) according to the manufacturer's instructions excepting for the addition of 4 μl lysostaphin. Restriction fragments were separated in a 1% agarose gel using a CHEF-DRII apparatus (Bio-Rad, United Kingdom), at a field strength of 200 V for 26 h with initial and final switching times of 5 s and 60 s, respectively. MLST was performed according to the protocol of Enright et al. (7). Nucleotide sequences were compared with the MLST database through the MLST website (http://www.mlst.net) for assignment of allelic profiles and sequence type (ST). Multiplex PCR to determine the SCCmec types was carried out according to Boye et al. (2).

Bioinformatics tools.

The DNA sequences of the following 18 S. aureus strains were obtained from the European Bioinformatics Institute (EMBL-EBI; http://www.ebi.ac.uk/genomes/bacteria.html) and the Sanger Institute (http://www.sanger.ac.uk): NCTC8325, MRSA252, MSSA476, COL, RF122, JH1, JH9, N315, Mu3, Mu50, MW2, Newman, USA300FPR3757, USA300TCH1516, ED98, LGA251, TW20, and EMRSA-15. The sequences were analyzed using the restriction digest tool at the Comprehensive Microbial Resource (http://cmr.jcvi.org) at the J. Craig Venter Institute. Gel images were saved as TIFF files and analyzed using BioNumerics software (v3.50; Applied Maths, United Kingdom). The analysis of multiplex PCR and PFGE gels was based on band differences and a clustering dendrogram was generated by the unweighted-pair group method with arithmetic mean (UPGMA) using a Dice coefficient of similarity of 1.5%. The neighbor-joining tree of the MLST data was constructed from allelic profiles and ST using Phylodendron (http://pubmlst.org/analysis/).

RESULTS

In this study, a novel multiplex PCR was developed for use in routine microbiology laboratories to screen for relationships between S. aureus strains isolated in clinical settings. The primers were designed around the SmaI restrictions sites (CCC↓GGG) used for PFGE electropherograms. The sequenced genomes of a number of S. aureus strains were subject to a bioinformatics analysis to determine the gene neighborhoods that contain SmaI restrictions sites. This analysis revealed that the number of SmaI restrictions sites varied from 24 to 29 among the S. aureus strains included in the analysis (Table 2).

Table 2.

S. aureus reference strains and typing data and number of SmaI sites^a

Strain	PFGE type	SmaI-multiplex PCR		MLST		SCCmec type	spa type	No. of SmaI sites
		Type	Profile	ST	CC
COL	B	2	1-0-0-1-0-1a-1-0-0-0	250	8	I	t008	27
USA300-FPR3757	C	3	1-0-0-1-0-1c-1-1-0-0	8	8	IV	ND	26
MW2	D2	5	1-1-0-1-1-1d-1-0-0-0	1	1	IV	t128	28
Mu50	E	6	1-0-0-0-1-1d-1-0-0-0	5	5	II	t002	25
Mu3	ND	6	1-0-0-0-1-1d-1-0-0-0	5	5	II	t002	26
N315	E1	6	1-0-0-0-1-1d-1-0-0-0	5	5	II	t002	25
JH9	E2	7	1-0-0-0-1-1e-1-0-0-0	5	5	II	ND	28
EMRSA-15	F	8	1-0-1-0-1-0-0-1-0-1	22	22	IV	ND	24
MRSA252	G	9	1-1-0-0-1-0-0-1-1-1	36	30	II	t018	26
USA1100	J	10	1-1-1-0-1-0-0-0-1-1	30	30	IV	ND	ND
CA-629	I	11	1-1-0-0-1-0-0-1-0-1	87	59	V	ND	ND
USA1000	H	12	1-1-0-0-1-0-0-0-1-0	59	59	IV	ND	ND
NCTC8325	A	1	0-0-0-1-0-1b-1-0-0-0	8	8	MSSA	t211	25
MSSA476	D1	4	0-1-0-1-1-1d-1-0-0-0	1	1	MSSA	t607	28

^aThe order of the amplicons in the profile is given in Table 1.

Abbreviations: CC, clonal complex; MLST, multilocus sequence typing; ND, not determined; PFGE, pulsed-field gel electrophoresis; spa type, Ridom spa type; SCCmec, staphylococcal cassette chromosome mec; ST, sequence type.

No. of SmaI sites refers to the total number of chromosomal SmaI sites.

The alignment of the gene neighborhoods around syntenic SmaI sites showed that more than 50 SmaI groups were identified among the 18 genomes analyzed. S. aureus has between five and six copies of the rRNA operons, each usually with two SmaI sites (9, 19). However, since gene neighborhoods associated with rRNA genes are less discriminatory, these sites were excluded from the analysis. A matrix of the remaining gene neighborhoods was used to identify SmaI restriction profiles for each strain. More than 30 profiles were identified, and the 10 most discriminatory SmaI groups were selected for analysis. The corresponding restriction sites were used to generate a typing system based on the output from the PFGE SmaI macrorestriction fragments (10 to 30 fragments of between 10 to 800 kb in length). However, instead of the time-consuming DNA extraction and digestion technique associated with PFGE, we devised a multiplex PCR technique based on SNPs and rapidly extracted chromosomal DNA.

Primer pairs were designed for each of the nine discriminatory SmaI groups. The locations of the gene neighborhoods used for the multiplex PCR are shown in Fig. 1. Most the amplicons were located within a single gene, the exceptions being SmaI groups 7 and 5. The group 7 amplicon was located at an intergenic location between genes encoding putative phage proteins, while the group 5 amplicon was located in an intergenic region downstream of the dipeptidase (pepV) gene (Fig. 1, Table 3).

Fig. 1.

Positions and distribution of high-discriminatory SmaI restriction sites on the S. aureus genome. Blue rectangles (indicated by black arrows) refer to PCR amplicons of SmaI groups. Mb, megabase pair.

Table 3.

SmaI gene neighborhoods and their locus tags on the S. aureus genomes

SmaI group	Strain	PCR product size (bp)	Gene neighborhood (name) (locus tag)	Primer 1 (L coordinate)	Primer 2 (R coordinate)
1	MRSA252	169	Capsular polysaccharide synthesis enzyme (cap8J) (SAR0160)	178,613	178,781
2	EMRSA15	268	Ser-Asp-rich fibrinogen/bone sialoprotein-binding protein (sdrD) (SAEMRSA15_04890)	583,595	583,862
3	NCTC8325	94	Conserved hypothetical protein (SAOUHSC_01174)	1,124,339	1,124,432
4	MRSA252	120	Polynucleotide phosphorylase/polyadenylase (pnpA) (SAR1250)	1,310,202	1,310,321
5a	COL	408	241 bp at 5′ end of dipeptidase (pepV) (SACOL1801)	1,850,334	1,850,741
			272 bp at 3′ end of hypothetical protein
5b	NCTC8325	351	241 bp at 5′ end of dipeptidase (pepV) (SAOUHSC_01868)	1,776,376	1,776,726
			272 bp at 3′ end of hypothetical protein
5c	USA300-FPR3757	238	241 bp at 5′ end of dipeptidase (pepV) (SAUSA300_1697)	1,872,815	1,873,052
			272 bp at 3′ end of hypothetical protein
5d	Mu50	179	241 bp at 5′ end of dipeptidase (pepV) (SAV1751)	1,881,973	1,882,151
			139 bp at 3′ end of hypothetical protein
5e	JH9	65	241 bp at 5′ end of dipeptidase (pepV) (SaurJH9_1806)	1,927,295	1,927,359
			139 bp at 3′ end of hypothetical protein
6	NCTC8325	223	Serine protease (splC) (SAOUHSC_01939)	1,845,158	1,845,380
7	MRSA252	146	79 bp at 5′ end of hypothetical phage protein (SAR2065a)	2,150,907	2,151,052
			117 bp at 3′ end of hypothetical phage protein (SAR2066)
8	MRSA252	315	Hypothetical protein	2,385,807	2,386,121
9	MRSA252	293	Nitrate reductase alpha chain (narG) (SAR2486)	2,558,310	2,558,602

Verification of primers and optimization of the multiplex PCR.

The specificity and discriminatory power of the primers were initially analyzed by uniplex PCR. In most cases, the nine primer pairs generated amplicons of the expected sizes in the expected strains. The exception was primer pair 5 that generated amplicons of variable sizes (e.g., ranging from 65 bp in JH9 to 408 bp in NCTC8325) (Table 3). This variability was subsequently shown to be due to the presence of a variable number of Staphylococcus aureus repeat (STAR) elements in the intergenic region between a dipeptidase (pepV, SAOUHSC_01868) and a hypothetical protein gene (SAOUHSC_01869) as shown in Fig. 2.

Fig. 2.

SmaI group 5 primers that PCR amplify products of various sizes due to variable numbers of repeated STAR sequences. R1-5, repeated sequences 1 to 5.

After it was confirmed that the primer selection strategy was effective, Visual OMP software was used to redesign and optimize the primers for use in a single-multiplex PCR (Table 1). The PCR products sizes ranged from ∼65 bp to 408 bp. The nine optimized primer pairs, together with a universal primer pair which was designed to amplify a 502-bp fragment from methicillin-resistance (mecA) genes, were multiplexed against the DNA from nine of the sequenced S. aureus strains, NCTC8325, COL, MSSA476, MW2, Mu50, Mu3, N315, MRSA252, and EMRSA-15 and a clinical strain encoding Panton-Valentine leukocidin (MRSA-PVL⁺). In general, the specificity of the primers was excellent. However, the SmaI group primers 4 and 3 generated PCR products in some strains that lacked the respective SmaI sites. These products were found to be due to SNPs in the SmaI restriction sites themselves in which the “G” at nucleotide position 4 was replaced with an “A” in the case of group 4 primers (CCCAGG) and “T” in the case of group 3 primers (CCTGGG). To avoid such false-positive reactions, the forward primers for SmaI groups 3 and 4 were redesigned to increase their stringencies. When the multiplex PCR was repeated, the redesigned primers for groups 3 (94 bp) and 4 (120 bp) generated products of the expected size with no false positives.

The initial validation of the multiplex PCR was carried out using purified chromosomal DNA. However, to facilitate the use of SmaI-multiplex typing (SMT) in routine clinical laboratories, the method was validated on rapidly extracted DNA. DNA from strains NCTC 8325, EMRSA-15, and MRSA252 was extracted directly from a colony removed from an agar plate by boiling, with or without a short pretreatment with lysostaphin. The performance of SMT was excellent when the DNA was extracted following pretreatment with lysostaphin but variable without pretreatment. The specificity of the assay was determined using two sequenced coagulase-negative staphylococcus (CNS) strains, methicillin-resistant S. epidermidis RP62A and methicillin-sensitive S. epidermidis 12228. With the exception of the mecA primer pair in strain PP62A, the multiplex primers were shown to be specific to S. aureus, since no other PCR products were generated from either of the CNS strains (data not shown).

The discriminatory power of SMT against well-characterized S. aureus strains.

Fourteen reference strains (NCTC8325, MRSA252, MSSA476, COL, JH9, N315, Mu3, Mu50, MW2, USA300FPR3757, EMRSA-15, USA1000, USA1100, and CA-629) were typed by SMT. All of the strains were typeable, and the discriminatory power was both excellent and reproducible (Fig. 3A). SMT was able to distinguish between closely related community-acquired MSSA strains MSSA476 and MW2, community-acquired MRSA strains USA1000, USA1100, and CA-629, and hospital-acquired strains EMRSA-15 and MRSA252 (EMRSA-16) (Fig. 3A). It was not able to distinguish between the closely related hospital strains Mu50, Mu3, and N315. PFGE was performed with the same set of strains excepting Mu3 and CA-629. PFGE generated distinct patterns for all of the strains, including closely related strains Mu50 and N315 that had the same SMT type (Fig. 3B). The sequence (ST) and SCCmec types of all of the reference strains were previously established excepting that of CA-629, which was identified in this study. Nine STs in six clonal complexes (CCs) were found among the 14 S. aureus strains tested. The SMT types correlated well with the spa types of eight of the 14 reference strains, as identified from the Ridom database (http://spaserver.ridom.de).

Fig. 3.

(A) SmaI-multiplex PCR profiles of 11 reference S. aureus strains and three unrelated community-acquired MRSA strains with all 10 primer pairs groups as the last step of optimization of a novel typing technique. (B) SmaI-PFGE patterns of 10 reference S. aureus strains and three unrelated community-acquired MRSA strains that were typed by SmaI-multiplex PCR typing as shown in Fig. 6, left panel. MW, molecular size marker (lambda ladder); Kb, kilobase; MW 50bp, 50-bp molecular size markers.

In order to evaluate the discriminatory power of SMT, its performance for the 14 S. aureus reference strains was compared with that of PFGE and MLST typing (Table 2). The profiles from SMT and PFGE were analyzed using BioNumerics fingerprinting analysis software, while a neighbor-joining tree was constructed for the MLST data using the allelic profiles and STs of the reference strains. The UPGMA dendrograms of the SmaI and PFGE profiles clustered the reference strains into similar groups, although there were slight differences in the relationships between the groups (Fig. 4). Both methods grouped strains NCTC8325, USA300, and COL into a single contiguous cluster (Fig. 4), each with a unique profile. In contrast, strains NCTC8325 and USA300 exhibited the same MLST allelic profile, ST8 (3-3-1-1-4-4-3), while COL had a related profile, ST250 (3-3-1-1-4-4-16) that differs by a single SNP at the yqiL locus. Community-acquired strains MSSA476 and MRSA-MW2 belong to MLST allelic profile ST1 (1-1-1-1-1-1-1) but had distinct SMT and PFGE profiles (Fig. 4). Strains Mu50, Mu3, and N315 had the same MLST allelic (ST5; 1-4-1-4-12-1-10) and SmaI-typing PCR profiles, while JH9, also a member of ST5, was closely related to those strains and within the same SmaI-typing cluster. Significantly, all four of these strains had the same SCCmec type, namely, type II (Table 2). PFGE grouped Mu50, N315, and JH9 into a single cluster (Fig. 4); Mu3 was not analyzed by PFGE. SmaI typing clearly distinguished between United Kingdom hospital epidemic strains EMRSA-15 and MRSA252 (EMRSA-16). MRSA252 is in the same SmaI and MLST cluster as USA community-acquired MRSA strains USA1000, USA1100, and CA-629. Each of these strains had unique SMT and MLST types. Strains MRSA252, USA1000, and USA1100 had distinct PFGE profiles, with USA110 clustering with EMRSA-15 rather than MRSA252 (EMRSA-16).

Fig. 4.

UPGMA dendrogram of SmaI-multiplex PCR typing and SmaI-PFGE patterns based on the Dice similarity coefficient and the neighbor-joining tree of MLST and illustrating the genetic relationship among 14 reference S. aureus strains. A, B, C, and D represent the typing cluster groups of the reference S. aureus strains. CC, clonal complex.

Evaluation of the discriminatory power of SMT against clinical strains.

Ninety clinical MSSA/MRSA strains were used to evaluate the SMT technique. The vast majority of the strains were isolated from hospitalized patients, while a few were isolated from members of staff and the environment during periods of increased incidence of staphylococcal infection. The strains were typed by SMT, and the epidemiology data were sent to the infection control office on the same day. All of the tested strains were typeable by SMT with excellent discriminatory resolution. Figure 5 shows representative profiles obtained from such clinical strains. Sixty two percent (n = 56) of the tested strains were shown to be EMRSA-15 with two variants profiles. Nineteen distinct SmaI-multiplex PCR patterns were identified among the 90 clinical S. aureus strains, and these strains were subsequently analyzed by MLST typing and, in the case of the MRSA isolates, by SCCmec typing. Sixteen different sequence types were identified among the 19 SmaI-multiplex PCR profiles (Fig. 6).

Fig. 5.

SmaI-multiplex PCR profiles of S. aureus strains isolated during two periods of increased MRSA incidence. Cluster 1 consisted of 11 S. aureus isolates from the screening of patients and staff on a surgical ward. Cluster 2 consisted of five screening samples and one blood sample from a baby care unit.

Fig. 6.

The UPGMA dendrogram based on SmaI-multiplex PCR profiles and the neighbor-joining tree of MLST showing the relationship between 19 representative clinical MSSA and MRSA isolates. A, B, C, and D represent the cluster groups of the S. aureus isolates as determined by SMT and MLST. ST, sequence type; CC, clonal complex.

Comparative analysis of SMT and MLST data, the former by UGPMA and the latter by neighbor-joining tree analysis, showed that the two methods were in good agreement (Fig. 6). MSSA isolates (FOUTB1, FOUTB5, FOUTB6, and FOUTB10), which showed similar but distinct SMT profiles, belonged to sequence types ST5 and ST83, but were grouped in a single clonal complex, CC5. MRSA isolates FOUTB16 and -18 and MSSA isolate FOUTB11 were tightly clustered by SMT and were identified as sequence types ST45, ST46, and ST47, respectively, and they belong to clonal complex CC45. MRSA isolate FOUTB19 is a SMT variant of EMRSA-15 but shares the same sequence type, ST22.

DISCUSSION

Since their emergence in the early 1980s, EMRSA-15 and EMRSA-16 have become the most prevalent nosocomial epidemic MRSA clones in the United Kingdom. More recently, community-acquired MRSA strains have started to replace hospital-acquired MRSA strains in some health care settings (5). Understanding the epidemiology of MRSA strains has become a major challenge for health care institutions. A wide variety of molecular techniques are currently used to type MRSA strains, including PFGE, MLST, and spa and SCCmec typing. Because PFGE, MLST, and spa typing are relatively time-consuming and require specialist equipment and expertise, their use is often beyond the resources of routine clinical microbiology laboratories. To overcome these limitations, we have developed a typing technique with a discriminatory power that is similar to that of PFGE, but which is rapid and suitable for use in routine laboratories. Like PFGE, SMT is based around SNPs in discriminatory SmaI sites, and the presence or absence of a PCR product therefore reflects the presence or absence of a SNP in a specific SmaI restriction site: the size of the resulting product identifies the targeted SmaI site irrespective of whether it is methylated. The inability of SmaI to cleave methylated target sites is a potential limitation of PFGE.

Discriminatory power is the most important parameter of any typing technique. Therefore, a set of 10 primer pairs was optimized to function in a single-multiplex PCR using Visual OMP DNA analysis software. Attempts to design a primer pair for SmaI group 10 that did not give false-positive results in the multiplex reaction were unsuccessful, and consequently, this pair was replaced with a universal mecA primer pair. As a result, analysis of nine sequenced S. aureus strains gave distinct profiles with the exception of the closely related hospital strains Mu50, Mu3, and N315. Used in combination with a rapid DNA extraction protocol, the multiplex PCR was able to generate reproducible profiles in 4 to 6 h. It is possible to adapt this technology for an automated genotyping assay using real-time (RT)-multiplex PCR, reducing the processing time to less than 60 min. Since individual targets are identifiable on the basis of the size of their amplicons, the RT-PCR output could be processed directly via dedicated analytical software.

The SmaI group 5 primers produced products of variable lengths, and subsequent analysis revealed this to be due to the presence of variable numbers of Staphylococcus aureus repeat (STAR) elements, CG-rich repeat sequences located in intergenic region of the S. aureus genome (4), between the group 5 forward and reverse primers. The presence of the STAR elements increased the discriminatory power of SMT.

The UPGMA dendrogram of SMT and PFGE and the neighbor-joining tree of MLST showed that the correlation between the three typing methods was excellent (Fig. 4), with the discriminatory power of SMT being intermediate between those of PFGE and MLST. The cluster analysis showed that SMT grouped most of the reference strains into the same four clusters as did PFGE and closely matched the MLST types (Fig. 4). For example, the clonal complex 8 (CC8) strains NCTC8325, USA300, and COL were grouped into a single cluster by both SMT and PFGE. Although NCTC8325 and USA300 have the same MLST type (ST8), they exhibited different SMT profiles.

In a comparative study, all orthologous gene pairs in the core genomes of MRSA-MW2 and MSSA476 showed 99.8% nucleotide identity (16). Although these strains had the same sequence type (ST1), they were distinguishable by SMT and PFGE profiling. In the case of SMT, their profiles were distinguished only by the amplicon generated from the mecA gene of MRSA-MW2. Similarly, HA-MRSA strain N315 and vancomycin intermediate-resistant strains Mu50, Mu3, and JH9 are closely related and share same sequence (ST5) and SCCmec (II) types (13). However, SMT is able to discriminate JH9 from strains Mu50, Mu3, and N315 (Fig. 4).

In the United Kingdom, the hospital-acquired epidemic clones EMRSA-15 and ERSA-16 were responsible for 95.6% (60.2 and 35.4%, respectively) of S. aureus bacteremia between 1998 and 2000 (15). More recently, community-acquired MRSA strains have been implicated in hospital-acquired infections, with CA-MRSA USA300 being responsible for 33% of HA-MRSA infections in a recent study (14). This makes the distinction between the most prevalent HA-MRSA and CA-MRSA strains an issue in many health care settings. In the current study, SMT discriminated between the most common United Kingdom hospital epidemic strains EMRSA-15, MRSA252 (EMRSA-16, also called USA200), and the well-characterized United States community-acquired MRSA strains USA1100, USA1000, and CA-629. Although those strains were grouped into a single cluster, each strain showed unique SMT, MLST, and PFGE profiles (Fig. 4).

SMT was used to type 90 S. aureus strains isolated from local hospitals. Nineteen distinct SMT profiles were observed among these isolates, and a representative strain of each profile was subsequently analyzed by MLST. This analysis confirmed that SMT had a better resolving power than MLST, since only 16 MLST types were found among the 19 SmaI-multiplex PCR profiles (Fig. 6). EMRSA-15 is the predominant epidemic MRSA clone in the United Kingdom (6), and most of the isolates (62%) were related to this strain. SMT distinguished between isolates that have the same sequence type (ST5) and between some variants of EMRSA-15 within the same MLST and SCCmec type (i.e., ST22/SCCmec-IV) (Fig. 6). In most cases, UPGMA clustering was in good agreement with the MLST phylogenetic tree (Fig. 6). The observation that SMT was able to identify phylogenies of clinical isolates (Fig. 6) supported the data obtained from the reference strains (Fig. 4), in which a high correlation was also observed between SMT and MLST.

Finally, we propose a simple nomenclatural system for SmaI type profiles based on the presence and absent of the nine SmaI sites according to their order on the genome (i.e., mecA-1 to SmaI-9). For example, SmaI-multiplex PCR type 2 of the COL strain, was given the SmaI profile 1-0-0-1-0-1a-1-0-0-0, according to the presence or absence of an amplicon at a specific allele. The letter for group 5 refers to the length of the variable PCR products generated by the SmaI group 5 primer pair (Table 3).

In conclusion, we described a new technique for typing both MSSA and MRSA strains. SMT is relatively cheap and provides reliable and comparable genotyping data. At the same time, SMT meets most of the criteria of a practical typing method: it is simple, relatively inexpensive, and highly discriminatory and does not require sophisticated equipment or expertise. Consequently, SMT could be used in many routine clinical microbiology laboratories to monitor ongoing S. aureus hospital outbreaks.