European Journal of Clinical Microbiology & Infectious Diseases

, Volume 27, Issue 9, pp 791–796

Detection of meticillin-resistant Staphylococcus aureus and Panton-Valentine leukocidin directly from clinical samples and the development of a multiplex assay using real-time polymerase chain reaction

Authors

  • L. Renwick
    • Department of MicrobiologyRoyal Infirmary of Edinburgh
  • A. Hardie
    • Department of MicrobiologyRoyal Infirmary of Edinburgh
  • E. K. Girvan
    • Scottish MRSA Reference Laboratory, Department of MicrobiologyStobhill Hospital
  • M. Smith
    • Department of MicrobiologyRoyal Infirmary of Edinburgh
  • G. Leadbetter
    • Department of MicrobiologyRoyal Infirmary of Edinburgh
  • E. Claas
    • Department of Medical Microbiology, Center of Infectious DiseasesLeiden University Medical Center
  • D. Morrison
    • Scottish MRSA Reference Laboratory, Department of MicrobiologyStobhill Hospital
  • A. P. Gibb
    • Department of MicrobiologyRoyal Infirmary of Edinburgh
  • J. Dave
    • Department of MicrobiologyRoyal Infirmary of Edinburgh
    • Department of MicrobiologyRoyal Infirmary of Edinburgh
Article

DOI: 10.1007/s10096-008-0503-9

Cite this article as:
Renwick, L., Hardie, A., Girvan, E.K. et al. Eur J Clin Microbiol Infect Dis (2008) 27: 791. doi:10.1007/s10096-008-0503-9

Abstract

Meticillin-resistant Staphylococcus aureus (MRSA) is a major pathogen responsible for significant numbers of healthcare-associated infections and isolates containing Panton-Valentine leukocidin (PVL) that cause severe skin infections are emerging as a serious problem. The rapid detection of MRSA would be an invaluable tool in a diagnostic laboratory. The aim of this study is to develop real-time polymerase chain reaction (PCR) assays for the detection of MRSA and PVL directly from clinical samples, and then combining these assays. Individual assays for MRSA (SCCmec) and PVL (lukF and lukS) were optimised and evaluated with screening and wound swabs, respectively. MRSA- and PVL-positive isolates were detected by the assays with an analytical sensitivity of 100 cfu per reaction. No other bacterial species were amplified. Fifty of 402 (12.4%) nasal swabs were positive by culture and PCR. Four of the 402 (1.0%) swabs were PCR-positive/culture-negative. Three of the 402 (0.7%) swabs were PCR-negative/culture-positive. The sensitivity of the MRSA assay is 95% and the specificity is 99% using conventional culture as the gold standard. Five of 240 wound swabs (2.1%) were positive for PVL. Three of the PVL-positive swabs were meticillin-sensitive Staphylococcus aureus (MSSA) and two were MRSA. The MRSA assay is a powerful and sensitive diagnostic tool, giving rapid results and could allow more timely treatment and infection control decisions to be taken. It can also, when combined with the PVL assay, provide valuable epidemiological information.

Introduction

Meticillin-resistant Staphylococcus aureus (MRSA) is a major nosocomial pathogen, first described in 1961 [1]. The meticillin-resistance gene, mecA, is carried on a mobile genetic element called the staphylococcal cassette chromosome (SCCmec) and encodes a penicillin binding protein (PBP2a). In addition, the SCCmec is located near the 5′ end of the open reading frame of a gene of unknown function, orfX. MRSA incidence varies between countries; in the United Kingdom, around 40% of Staphylococcus aureus blood culture isolates are meticillin-resistant [2]. Risk factors for MRSA include advanced age, previous infection, frequent admissions to hospital, residence in a long-term care facility and transfer between hospitals [3]. The treatment for MRSA depends on the site and severity of infection. Vancomycin is recommended for the treatment of systemic and invasive infections in the United Kingdom [4].

Despite its spread in hospitals and nursing homes, MRSA has not disseminated in the wider community until recently, and this has coincided with increased reports of Panton-Valentine leukocidin (PVL)-carrying strains [5, 6]. PVL-positive MRSA and meticillin-sensitive Staphylococcus aureus (MSSA) are associated with severe skin infections and necrotising pneumonia [7, 8]. PVL is a non-specific pore-forming toxin that destroys human monocytes, macrophages and polymorphonuclear leukocytes [9]. A pair of co-transcribed genes, lukF and lukS, carried on a bacteriophage encode the toxin [10]. Patients with recurrent boils or abscesses and without risk factors for MRSA, playing contact sports or living in close proximity to others, e.g. on military barracks, are likely to be infected with PVL-positive strains [11, 12]. There is increasing evidence that the presence of PVL is associated with more invasive disease [1315].

Current guidelines for the identification of MRSA recommend culture [16]. Chromogenic media yield reliable negative results at 20 h [17], but isolation, sensitivity testing and the confirmation of MRSA takes 48–72 h. The rapid diagnosis of MRSA, in less than one day, by the use of molecular approaches could lead to the early implementation of appropriate antibiotics and reduced hospital bed stays. As molecular methods become commonplace for MRSA, screening all patients for clean surgery becomes a viable option. Real-time polymerase chain reaction (PCR) with fluorescent-labelled probes allows the simultaneous detection of up to four targets in the same reaction at minimal extra cost and the detection of PVL in the same assay might improve clinical management.

The aim of this study was to design real-time PCR assays targetting the SCCmec for the detection of MRSA and lukF and lukS genes for PVL directly from clinical samples, and then multiplex the assays and assess the feasibility for their use in a diagnostic laboratory.

Materials and methods

Bacterial strains

A selection of staphylococci (26 coagulase-negative staphylococci [CNS], 26 MRSA, 25 MSSA) and a panel of other organisms commonly isolated from clinical samples, including Streptococcus pyogenes, Alpha haemolytic streptococci, Propionibacterium acnes, Enterococcus faecalis ATCC 29212, Corynebacterium striatum, Escherichia coli ATCC 35218, Proteus mirabilis NCTC 10975, Pseudomonas aeruginosa ATCC 27853, Clostridium perfringens NCTC 8237, Haemophilus influenzae ATCC 49766, Streptococcus pneumoniae ATCC 49619 and Candida albicans ATCC 76615, were obtained from culture collections to assess the specificity. The Scottish MRSA Reference Laboratory (SMRSARL) supplied PVL-positive MRSA and MSSA controls. All organisms were cultured on Columbia agar with horse blood (Oxoid, Basingstoke, United Kingdom), incubated at 37°C in air plus 5% CO2 or anaerobically, as appropriate, and stored at 4°C prior to nucleic acid extraction and PCR.

Clinical specimens: MRSA assay

Four hundred and two nasal swabs were collected from inpatients during March and April 2006 at the Royal Infirmary of Edinburgh (RIE). Swabs were cultured on Oxacillin Resistance Screening Agar (Oxoid, Basingstoke, United Kingdom) and inoculated into 2 ml nutrient broth (E&O Laboratories Ltd., Bonnybridge, United Kingdom). Culture plates were incubated at 37°C for 24–48 h, positive colonies picked and a latex agglutination test for Staphylococcus aureus surface antigens was carried out (Pastorex®, Biostat, Stockport, United Kingdom). Antibiotic susceptibility was determined by VITEK (bioMérieux, Basingstoke, United Kingdom). Broth suspensions were stored at 4°C prior to nucleic acid extraction and PCR, and was used to inoculate the enrichment cultures, if required. Discrepant samples were retested following salt broth enrichment culture and any PCR-positive/culture-negative samples after enrichment were sequenced to confirm MRSA.

Clinical specimens: PVL assay

Two hundred and forty swabs from 231 patients with Staphylococcus aureus infections during May and June 2006 at the RIE were collected. MRSA screening swabs were not included, as this sample type is likely to have an extremely low prevalence of PVL-carrying Staphylococcus aureus [18]. The isolates were identified according to standard laboratory practice by morphology and biochemical tests. Briefly, swabs were cultured on blood agar and examined after 24 h as above. Isolates on blood agar and swabs in Amies transport medium were stored at 4°C prior to nucleic acid extraction and PCR. The patient information and clinical details were collected for all 240 samples.

Nucleic acid extraction

Two extraction methods were evaluated with the SCCmec PCR; the QIAamp DNA mini kit (QIAGEN Ltd., Crawley, United Kingdom) and the NucliSens® easyMAG™ system (bioMérieux, Basingstoke, United Kingdom). easyMAG™ is an automated platform based on the Boom method [19] that utilises magnetic silica with a large surface binding area, enabling sample pooling, and is highly reproducible [20]. Nutrient broth suspensions were pre-treated with Proteinase K (QIAGEN Ltd., Crawley, United Kingdom) and then 200 μl was extracted according to the manufacturer’s instructions. The best method was used for the PVL and multiplex PCRs.

Isolates for PVL and multiplex PCRs were suspended in 2 ml of 0.9% NaCl to 0.5 McFarland equivalent and 200 μl of this suspension was extracted using the NucliSens® easyMAG™ system (bioMérieux, Basingstoke, United Kingdom) according to the manufacturers’ instructions. Purified nucleic acid was eluted in 110 μl of the suspension. Swabs were agitated for 10 s in 2 ml of 0.9% NaCl, and then 200 μl was extracted as before.

Primers and probes

SCCmec

Primers designed by Huletsky et al. [21] were synthesised (Operon, Cologne, Germany) (see Table 1 for all of the primer and probe sequences). The PCR utilises five forward primers to amplify the most common SCCmec types and a reverse primer in the orfX region with three molecular beacons for detection. A sequence alignment was carried out using Simmonics version 1.4 [22] and a degenerate TaqMan probe based on the three published molecular beacon sequences designed.
Table 1

Oligonucleotides used in this study

Assay

Oligo

Sequence (5′–3′)

MRSA

mecii

GTCAAAAATCATGAACCTCATTACTTATG

meciii

ATTTCATATATGTAATTCCTCCACATCTC

meciv

CAAATATTATCTCGTAATTTACCTTGTTC

mecv

CTCTGCTTTATATTATAAAATTACGGCTG

mecvii

CACTTTTTATTCTTCAAAGATTTGAGC

orfX

GGATCAAACGGCCTGCACA

Probe

FAM-CRTAGTTACTRCGTTGTAAGACGTC-BHQ

PVL

Forward

ACACACTATGGCAATAGTTATTT

Reverse

AAAGCAATGCAATTGATGTA

Probe

HEX-ATTTGTAAACAGAAATTACACAGTTAAATATGA-BHQ

FAM and HEX=fluorescent labels; BHQ=black-hole quencher

PVL

The primers and probes described by McDonald et al. [23] that target a conserved area across the lukF and lukS genes were synthesised (Eurogentec, Southampton, United Kingdom). The PVL probe was labelled with a different fluorophore to the SCCmec probe to allow multiplexing (the sequences are shown in Table 1). PVL confirmations were carried out by the SMRSARL according to the method described by Lina et al. [24]. Briefly, following nucleic acid extraction, a conventional PCR is carried out with lukS and lukF primers distinct from the McDonald primers and the products visualised on a gel.

PCR amplification

SCCmec and PVL assays were optimised individually for primer and probe concentration, magnesium concentration and annealing temperature. The final reaction volume for the multiplex PCR was 25 μl, consisting of 10 μl extracted nucleic acid, 2.5 μl 10 × real-time PCR buffer (QIAGEN Ltd., Crawley, United Kingdom), 1.25 Units HotStarTaq (QIAGEN Ltd., Crawley, United Kingdom), 3.0 mM MgCl2 (final concentration 4.5 mM), 0.5 μM each SCCmec primer, 0.35 μM SCCmec probe, 0.2 μM PVL forward primer, 0.4 μM PVL reverse primer and 0.1 μM PVL probe. Amplification, detection and analysis were performed in an ABI 7500 real-time PCR system (Applied Biosystems, Warrington, United Kingdom) under the following conditions: 1 × 95°C for 15 min followed by 50 × 95°c for 15 s, 60°C for 40 s and 72°C for 30 s.

SCCmec sequencing

Products of the SCCmec PCR were sequenced. Sequence analysis was set up using the BigDye™ Terminator Cycle Sequencing Kit (Applied Biosystems, Warrington, United Kingdom) according to the manufacturer’s instructions. The products were analysed on the ABI 3730 DNA Analyser (Applied Biosystems, Warrington, United Kingdom) and the sequences were analysed using Simmonics version 1.4 [22].

Statistical analysis

The sensitivity, specificity and positive and negative predictive values of the PCRs were calculated by comparing the PCR results with those of culture and reference laboratory identification.

Results

Extraction method

The two methods used to extract the DNA from all 402 nasal swabs and the SCCmec PCR results were compared. The easyMAG™ extraction was found to be more efficient and sensitive, with lower cycle threshold (Ct) values observed for identical samples (see Table 2). This method also detected more PCR-positive/culture-negative MRSAs than the QIAGEN kit and enabled further automation of the procedure.
Table 2

Cycle threshold (Ct) values across a ten-fold dilution series of the same meticillin-resistant Staphylococcus aureus (MRSA) isolate with different extraction methods

Bacteria (cfu/ml)

Ct value (QIAGEN extraction)

Ct value (easyMAG extraction)

1 × 107

17.4

17.4

1 × 106

24.0

20.5

1 × 105

28.5

24.2

1 × 104

31.5

28.0

1 × 103

35.7

31.3

1 × 102

41.3

34.8

1 × 101

42.5

Optimisation

The SCCmec PCR was optimised for primer and probe concentration, magnesium concentration and annealing temperature. Standard curves were generated from duplicate runs with ten-fold dilutions of the two dominant MRSA types in Scotland (EMRSA 15 and 16) to maximise the efficiency. From these standard curves, the sensitivity of the PCR was determined to be 100 cfu per reaction. The PVL assay was optimised in the same way using ten-fold dilutions of the isolate 06.3100.N, a PVL-positive MSSA obtained from the SMRSARL.

Clinical validation of the real-time SCCmec assay

Culture and PCR were performed on 402 nasal swabs to evaluate the SCCmec PCR. Fifty-seven of the 402 (15.7%) swabs were MRSA-positive by culture or PCR. Fifty of the 402 (12.4%) swabs were positive by both culture and PCR. Four of the 402 (1.0%) swabs were negative by culture and positive by PCR; these samples were sequenced. In all four cases, a homology search confirmed the sequence as SCCmec, suggesting that these samples can be considered as true positives. Three of the 402 (0.7%) swabs were positive by culture and negative by PCR. Compared to routine culture, this method has a sensitivity of 95%, a specificity of 99% and positive and negative predictive values of 93% and 99%, respectively.

Clinical validation of the real-time PVL assay

A collection of 240 Staphylococcus aureus wound isolates (86 MRSA, 156 MSSA) from 231 patients were pooled into groups of five and tested for PVL. For isolates from positive pools, the original swabs were re-extracted and tested individually. Five of the 240 (2.1%) isolates from five different patients were reproducibly positive (see Table 3). Three of the five PVL-positive swabs were MSSA and two of the five were MRSA. Each positive specimen was collected at an out-patient clinic or within 24 h of admission to the hospital from patients with no previous hospital admissions. All PVL-positive isolates and three negative isolates with high clinical suspicion of PVL were tested by a second PCR; five of five positives and three of three negatives were confirmed as such.
Table 3

Details of the specimens from which Panton-Valentine leukocidin (PVL)-positive isolates were detected

No.

Organism

Swab Type

DOB

Location

684913

MRSA

pus, breast abscess

25/06/1990

out-patient breast clinic

684496

MRSA

axilliary abscess

24/04/1978

accident and emergency

685806

MSSA

pus, axilliary abscess

21/02/1976

general surgery

686242

MSSA

cellulitis, leg

02/01/1954

infectious diseases

686748

MSSA

infected eczema, leg

13/07/2003

out-patient dermatology

Optimisation of the SCCmec and PVL multiplex PCR

The primer and probe concentrations determined for the individual assays were used without alteration. Magnesium and annealing temperature optimisation were carried out as before. Labelling the SCCmec and PVL probes with different fluorophores, SCCmec with FAM and PVL with HEX, meant that the targets could be differentiated when the assays were combined. The limit of detection of the assay was determined to be 100 cfu per reaction. There were no significant differences in Ct values between the individual and multiplex PCRs (see Table 4).
Table 4

Ct values across ten-fold dilution series for the same MRSA and PVL isolates in individual and multiplex PCRs

Bacteria (cfu/ml)

Ct value (individual PCR)

Ct value (multiplex PCR)

 

MRSA

PVL

MRSA

PVL

1 × 107

17.5

30.0

17.4

29.4

1 × 106

20.3

34.7

20.5

34.7

1 × 105

24.7

39.3

25.0

37.3

1 × 104

28.2

40.1

27.9

40.0

Multiplex PCR of the clinical samples

Multiplex PCR to detect both SCCmec and PVL was carried out on swab extracts from positive pools in the PVL validation with appropriate controls. Eighty-six percent of these swabs had grown commensals and other pathogens in addition to Staphylococcus aureus when cultured. The PCR reproducibly detected five of five PVL positives and 10 of 10 MRSA isolates with no reduction in sensitivity. Compared to culture and reference laboratory identification, this PCR assay had positive and negative predictive values of 100%.

Specificity

To assess the integrity of the primers and probes in the SCCmec, PVL and multiplex assays, each was tested for cross-reactivity with a panel of organisms commonly isolated in the bacteriology laboratory. In all cases, no product was observed.

Discussion

The identification of patients infected or colonised with MRSA is important for the implementation of infection control measures, and delays can result in increased transmission between patients [25]. A multiplex PCR was developed to provide a quick and reliable method of MRSA detection and a means with which to gather epidemiological data on the emergence of PVL-carrying strains.

Commercial assays are available for MRSA detection from clinical samples: BD GeneOhm™ MRSA assay (BD, Oxford, United Kingdom) is a real-time PCR approved for use with nasal swabs. This assay has a manual extraction protocol and is only recommended for use with the Cepheid SmartCycler® PCR system, meaning that it can only be used for screening swabs, and laboratories with other PCR systems will have to invest in new equipment. Many papers describe in-house PCR-based assays for the detection of MRSA from pure cultures, offering some advantage over culture on selective media, but, here, we have developed an in-house PCR assay that can detect MRSA directly from clinical samples. The multiplex assay performs well and gives clear, reproducible results. This assay benefits from a high level of automation and is compatible with any real-time PCR platform.

The primers and probe used in this assay are sufficient for detecting all current SCCmec types. The emergence of different clones could lead to a reduction in the sensitivity of the assay, but close links to reference labs carrying out surveillance and the ease with which primers can be modified should ensure that this does not happen.

The three culture-positive/PCR-negative results were due to poor inoculation of the nutrient broth; testing the isolates confirmed that the assay was capable of detecting these MRSAs. This might be avoided in future by using an improved swab collection system.

All PVL-positive specimens were confirmed by the SMRSARL. The finding that 2.1% of Staphylococcus aureus wound isolates carry PVL genes shows good concordance with the figure of 1.6% published for England and Wales [26], and the clinical details for the positive specimens were indicative of the types of infection, i.e. acute, purulent skin infections, associated with PVL [8].

The high throughput of real-time PCR, with automated extraction and no requirement for post-amplification analysis, where the processing time does not increase linearly with the number of samples, also makes it possible to expand current screening programmes. Studies have shown that, when only high-risk patients are screened, some cases are missed [27]. Pre-operative MRSA screening is known to be valuable and can lead to reductions in serious post-operative complications [28]. Detection allows for the possibility of eradication therapy and the fast implementation of infection control procedures over and above the standard methods, such as isolation, the use of disposable equipment and movement restriction.

Alongside positive and negative controls, it will be essential to add an internal control to each sample prior to nucleic acid extraction in order to monitor the quality of the extraction and check for PCR inhibition [29]. The extraction is generic, so the internal control need not be a bacterium. A DNA virus, such as phocine herpesvirus, for which primers and probes have been described [30], would work equally as well and will be easily and inexpensively worked into the multiplex with a third fluorophore.

Conventional culture will continue to have an important role in diagnostic bacteriology for the foreseeable future, due to its low cost and ability to detect a wide range of organisms. Molecular methods do, however, have the potential to supplement conventional culture where rapid tests for specific organisms and genes are needed. The assay described here for MRSA (and PVL) is an example of where molecular methods will have real clinical value in making the diagnosis and in directing therapy and infection control interventions.

Acknowledgements

This study was supported by a small project grant from the Chief Scientist Office and by Pfizer, AstraZeneca, bioMérieux and the Local Microbiology Endowment Fund.

Copyright information

© Springer-Verlag 2008