High throughput identification of clinical isolates of Staphylococcus aureus using MALDI-TOF-MS of intact cells

Infection, Genetics and Evolution 9 (2009) 507–513

Contents lists available at ScienceDirect

Infection, Genetics and Evolution journal homepage: www.elsevier.com/locate/meegid

High throughput identification of clinical isolates of Staphylococcus aureus using MALDI-TOF-MS of intact cells Lakshani Rajakaruna a, Gillian Hallas a, Linda Molenaar b, Diane Dare c, Helen Sutton c, Vesela Encheva a, Renata Culak a, Ingrid Innes a, Graham Ball d, Armine M. Sefton e, Melvin Eydmann f, Angela M. Kearns g, Haroun N. Shah a,* a

Department for Bioanalysis and Horizon Technologies, Health Protection Agency Centre for Infections, 61 Colindale Avenue, London NW9 5EQ, United Kingdom Zaans Medisch Centrum Hospital, Regionaal Microbiologisch Laboratorium, Koningin Julianaplein 58, 1502 DV Zaandam, The Netherlands Manchester Metropolitan University, All Saints Building, Oxford Road, Manchester M15 6BH, United Kingdom d School of Biomedical and Natural Sciences, The Nottingham Trent University, Clifton Campus, Clifton Lane, Nottingham NG11 8NS, United Kingdom e Institute of Cell and Molecular Science, Centre for Infectious Disease, Barts and The London School of Medicine and Dentistry, 4 Newark Street, London E1 2AT, United Kingdom f Dept. of Microbiology, Pathology & Pharmacy Building, The Royal London Hospital, 80 Newark Street, London E1 2ES, United Kingdom g Staphylococcal Reference Unit, Health Protection Agency, Centre for Infections, London NW9 5EQ, United Kingdom b c

A R T I C L E I N F O

A B S T R A C T

Article history: Received 16 October 2008 Received in revised form 26 January 2009 Accepted 27 January 2009 Available online 6 February 2009

Staphylococcus aureus remains an important human pathogen responsible for a high burden of disease in healthcare and community settings. The emergence of multidrug-resistant strains is of increasing concern world-wide. The identification of S. aureus is currently based upon phenotypic and genotypic methods. Here, an alternative approach involving mass spectral analysis of surface-associated proteins of intact bacterial cells by matrix-assisted laser desorption/ionisation time of flight mass spectrometry (MALDI-TOF-MS) was investigated using 95 isolates obtained directly from a clinical laboratory at The Royal London Hospital and 39 isolates from the Staphylococcal Reference Unit, Health Protection Agency, London. Results obtained indicate that clinical isolates share many common mass ions with-type/ reference strains which allowed their correct identification when searched against a comprehensive database that has been in the process of development for several years. The existing database contains more than 5000 profiles of various bacterial pathogens, but comprises mainly type or reference strains. The MicrobeLynx software successfully identified all isolates to the correct genus and all but four to the correct species. These were misidentified in the first instance due to contamination or low mass ion intensity but once the cultures were purified and re-analysed they were confirmed as S. aureus by both MALDI-TOF-MS and 16S rRNA sequence analysis. The high percentage of correct identifications coupled with the high speed and the minimal sample preparation required, indicate that MALDI-TOF-MS has the potential to perform high throughput identification of clinical isolates of S. aureus despite the inherent diversity of this species. The method is, however, only reproducible if variable parameters such as sample preparation, media, growth condition, etc. are standardised. ß 2009 Elsevier B.V. All rights reserved.

Keywords: MALDI-TOF-MS Intact cell MALDI Staphylococcus aureus Microbial identification 16S rRNA

1. Introduction The incidence of Staphylococcus aureus related infections has increased dramatically since the emergence of methicillinresistant strains and high rates of mortality and morbidity are occurring world-wide (Oliveira et al., 2002). In the UK there has been a significant increase from 1990s with 9854 cases of MRSA bloodstream infections reported between April 2006 and December 2007 (http://www.hpa.org.uk/webw/) while in the USA it is

* Corresponding author. Tel.: +44 20 8327 6749; fax: +44 20 8327 7870. E-mail address: [email protected] (H.N. Shah). 1567-1348/$ – see front matter ß 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.meegid.2009.01.012

estimated that more than 94,000 life thretning infections and 18,650 deaths occured in 2005 (Klevens et al., 2007; Matthew et al., 2005). Isolates sent to diagnostic laboratories have long been identified by traditional biochemical tests and then subtyped by a variety of methods (Fang and Hedin, 2003; Ip et al., 2003; Jury et al., 2006). However, criteria for the identification of species are still equivocal, some strains being misidentified with closely related species. Given the clinical importance of nosocomial infections, failure to unequivocally identify between species can have far reaching consequences clinically and epidemiologically. In Streptococcus alone, it is estimated that up to 13% of strains reaching clinical laboratories may be identified incorrectly (Kikuchi et al., 1995).

508

L. Rajakaruna et al. / Infection, Genetics and Evolution 9 (2009) 507–513

Diagnostic laboratories are usually required to process large numbers of specimens; hence high throughput methods are favoured. During the last 5 years we have utilised long chain cellular fatty acid (LCFA) profiles in tandem with 16S rDNA sequence analysis to characterise clinical isolates referred to our laboratory for identification. However, due largely to the cumbersome nature of the method, LCFAs have ceased in favour of 16S rDNA analysis and, while such a method is reliable for the identification of S. aureus, in general a polyphasic approach provides a more robust system of identification. As a high throughput method, we have been pursuing matrixassisted laser desorption/ionisation time of flight mass spectrometry (MALDI-TOF-MS) of intact cells because of the minimal sample preparation required, the negligible biomass used and the ease with which this method can be applied. The method involves gentle laser ionisation of the cells taken directly off an agar plate and yields a mass spectrum of intact cells; the cells may even be cultured following analysis (Claydon et al., 1996). The mass spectral profiles obtained are surface-associated molecules and data so far indicate that they are characteristic of each species (see Keys et al., 2004). We therefore set out some years ago to develop an extensive database of type and reference strains that are quality controlled with the aim of using this database subsequently to rapidly identify clinical isolates as a high throughput system (Keys et al., 2004). However, while it is anticipated that clinical isolates will deviate somewhat from the reference spectra held in the database, it is hypothesised that there will be sufficient conserved mass ions in the spectrum of each isolate to enable its mass spectral profile to be matched to its counterpart in the database. The present study seeks to challenge the currently held database of over 5000 mass spectral profiles against a large number of clinical isolates that have been taken directly from a hospital diagnostic laboratory and a Staphylococcal Reference Unit following presumptive morphological identification on blood agar plates. 2. Materials and methods

subtypes of current epidemic MRSA strains (EMRSA-15, -16 and -17) in addition to representatives of three pandemic CommunityAssociated MRSA lineages (so-called USA300, South West pacific and European clones). These were sub-cultured to Columbia blood agar plates for analysis by MALDI-TOF-MS. 2.3. Presumptive and confirmatory tests 2.3.1. Morphology All The Royal London Hospital isolates of S aureus were presumptively identified by an experienced biomedical scientist using gram staining, colonial morphology (golden to white colonies on blood agar), a commercial Staphylococcal latex kit (Pro-Lab Diagnostics ‘Prolex Staph Latex’ agglutination kit which detects Clumping Factor or Protein A in S. aureus isolates), and DNAase production (Oxoid DNAase agar). 2.3.2. Identification of S. aureus by 16S ribosomal DNA sequencing Isolates were grown on CBA (Media Department, HPA) for 24 h at 37 8C. DNA was extracted by resuspending the cells in 60 ml of Prepman Ultra (Applied Biosystems) and heating the suspension for 10 min at 99 8C followed by incubation at 4 8C. The sample was centrifuged at 2500  g for 10 min and 1 ml was used for PCR. The primers (MWG) used were ANT1F, 50 -AGA GTT TGA TCC TGG CTC AG-30 , and 1392R, 50 -ACG GGC GGT GTG TAC AAG-30 giving a product size of 1300 bp. PCR cycle conditions using PCR Ready Mix (Sigma, UK) were as follows: initial denaturation 95 8C for 2 min, followed by 35 cycles of 95 8C for 45 s, 56 8C for 45 s and 72 8C for 60 s. Final extension was carried out at 75 8C for 5 min. Products were cleaned using AMPure PCR Purification Magnetic Beads Kit (Beckman Coulter, UK), following the manufacturer’s instructions. The sequencing primers used were 357F, 50 -CTC CTA CGG GAG GCA GCA G-30 , and 3R, 50 -GTT GCG CTC GTT GCG GGA CT-30 . Sequences were analysed on a Positive Beckman Capillary sequencer (CEQ 8000). The bacteria detected by 16S rDNA PCR were identified by sequence comparison to the GenBank database using BLAST (http://www.ncbi.nlm.nih.gov).

2.1. Phase I: Sample collection—The Royal London Hospital isolates 2.4. Preparation of bacterial cultures Samples were collected from the clinical laboratory of The London Hospital following five visits between September and October 2005. At each visit, approximately 100–150 cultures representative of each working bench of the laboratory were selected randomly for analysis. Each isolate was identified by standard methods used at The Royal London Hospital. Each isolate was given a batch number and for each the culture conditions, gram stain, and the morphology were recorded together with any presumptive identification. Cultures were then assigned an HPA number and sub-cultured onto blood agar slopes for transport to HPA Centre for Infections, London. Upon arrival there, they were incubated at 37 8C in air if presumptively identified as Staphylococcus sp. Isolates were plated for purity onto Columbia blood agar and stored using the Microbank preservation system (Pro-Lab Diagnostics) at 80 8C. A total of 580 isolates were stored. S. aureus isolates comprised 18.4% of the total (n = 107) and constituted the largest single group of isolates collected. This group was therefore selected to assess the potential of the MALDI-TOF-MS approach. 2.2. Phase II: Staphylococcal Reference Unit isolates Thirty-nine isolates were provided by the HPA’s Staphylococcal Reference Unit and included cultures that were sent to the laboratory from geographically diverse centres throughout England and Wales for identification and characterisation. All except two were MRSA strains isolated during 2007 and included

In order to identify a suitable medium required for the optimal growth for S. aureus, and assess any differences in the culture medium on MALDI-TOF mass spectral profiles, isolates were maintained on Colombia blood agar (CBA), nutrient agar (NA), chocolate agar (CHOC) and mannitol salt agar (MSA). Also to determine the optimal incubation time required, isolates were incubated for different time periods: 24, 48 and 72 h. Growth conditions were aerobic and at 37 8C. All isolates were subcultured on three successive culture plates prior to MALDI-TOF-MS analysis. 2.5. Target plate preparation Cultures were grown as described above at different incubation periods on CBA, NA, CHOC and MSA. Chemically cleaned target plates (Waters Corporation, UK) were wiped with methanol and allowed to air-dry prior to use. A small amount of growth was removed from the culture plate using a 1 ml loop and transferred to the target plate. One ml matrix solution containing acetonitrile, water and methanol (1:1:1), 0.01 M 18-crown-6 ether, 0.1% formic acid (v/v), saturated with 5-chloro-2-mercaptobenzothiazole (CMBT) for Gram-positive organisms at a concentration of 3.0 mg/ml was added to the plate. Matrix solutions were sonicated in a sonic bath for 10 min prior to application. Lock mass wells were spotted with 1 ml of a 1:1 mixture of alpha-cyano-4-hydroxycinnamic acid at a concentration of 14.0 mg/ml and peptide

L. Rajakaruna et al. / Infection, Genetics and Evolution 9 (2009) 507–513

mixture. The peptide mixture contained bradykinin, angiotensin I, glu-fibrinopeptide B, rennin, ACTH (18-39 clip), all at 1 pmol/ml, bovine insulin 2 pmol/ml and ubiquitin 10 pmol/ml. All reagents unless otherwise stated were from Sigma, UK. As the control, Micrococcus lylae (NCTC 11037) was used. Twelve replicates were analysed for each sample. 2.6. Data acquisition and processing This was done using MALDI-TOF Mass Spectrometer (Waters/ Micromass Ltd., UK) and data processing was performed using MicrobeLynxTM software as described previously (Keys et al., 2004; Shah et al., 2008). Where necessary, the data were validated in the laboratory of the co-applicant Dr. D. Dare in Manchester. The instrument was fitted with a 337 nm nitrogen laser. Fifteen spectra per sample well and 10 spectra per lock mass well were collected for each strain in the mass range of 500–10,000 Da. Individual spectral profiles were lock mass corrected with the exact mass of the rennin peak, which is 1760 Da and then the 15 spectral profiles were combined in order to improve the mass accuracy and to produce a reproducible bacterial spectrum for each replicate. The spectra obtained for the clinical isolates of S. aureus were searched against the database MicrobeLynxTM 2005 and the search was based upon an estimation of the probability of the mass spectral peaks in the test spectrum to be comparable with the database spectrum. A list of top eight matches was provided together with Root Mean Square (RMS) value. A high absolute probability and low RMS value indicated a good match.

509

3. Results and discussion Traditionally bacterial identification has been based upon phenotypic traits, in particular biochemical tests. However, despite continual refinement, these methods remain time consuming and results maybe equivocal especially for non-fermentative taxa. Increasingly, diagnostic laboratories must process more samples accurately in less time and, given the throughput required, many laboratories are turning towards new and emerging technologies to provide solutions. MALDI-TOF-MS is an attractive alternative because the method is accurate, rapid, requires minimal sample and preparation time while analysis and results is achievable in minutes. The method requires no pre-extraction nor purification steps as cells are directly analysed in an intact form and, in the method used here, cells may even be sub-cultured post-MS analysis for additional tests such as 16S rRNA analysis. However, there are numerous methods of sample preparation involving fractionation steps for example for peptide and protein biomarker signatures (Williams et al., 2006) whole or intact cells using a Linear, Reflectron TOF or Fourier Transform mass spectrometry (Lay and Liyange, 2006; Jones et al., 2006). Lay and Liyange (2006) use the term ‘intact’ to refer to the fact that the culture suspension is examined directly and that the cells are not treated or processed to isolate cellular components, but point out that the cells may not necessarily remain whole. This is in direct contrast to methods that utilise additional steps to isolate proteins or other analytes prior to analysis and which have additional advantages (see e.g. Lay, 2002). Which ever method utilised, the ultimate aim is to obtain highly reproducible mass spectral profiles that may be used to compare

Fig. 1. Mass spectral profiles of S. aureus (clinical isolate HPA 30), when grown on four different types of media viz. mannitol salt agar (MSA), nutrient agar (NA), chocolate agar (CHOC) and Columbia blood agar (CBA). The greatest peak density of mass ions in the m/z 500 to m/z 4000 range was detected when S. aureus was grown on MSA and included a number of mass ions with m/z above 3000 Da. CBA and CHOC profiles also contained a broad range of ions in the m/z range of 500–3200. Expanded portions of both spectra (not shown) show consistent spectral features. The least number of m/z peaks were detected from cells grown on NA and concentrated within a narrow section of the spectrum (m/z: between 2239, 2638 and 3047).

510

L. Rajakaruna et al. / Infection, Genetics and Evolution 9 (2009) 507–513

unknown isolates with those of reference spectra for example for species identification (Welham et al., 1998; Dai et al., 1999; Chong et al., 1997; Liang et al., 1996; Holland et al., 1996; Claydon et al., 1996; Hillenkamp et al., 1991). Parameters such as the culture medium (CBA) and culture time (24 h) were hitherto standardised for development of the MALDI database using a panel of Gram-negative and Gram-positive taxa (see e.g. Shah et al., 2000, 2002). The profiles generated are derived from surface-associated molecules which in turn are affected by environmental pressure. Consequently, some of these parameters such as culture media constituents and growth times were briefly re-examined here using a randomly selected clinical isolate of S. aureus (HPA 30). Strain HPA 30 was cultured on four different types of solid media which included Columbia blood agar, chocolate agar (CHOC), nutrient agar and mannitol salt agar to assess the comparative peak density of the ions in each spectrum (Fig. 1). There were striking differences in MS profiles of cells grown on different media except between CBA and CHOC which are similar in composition except for the presence of lysed blood in the latter. Cells cultured on CBA and CHOC possessed mass ions in the m/z range of 500–3250 (Fig. 1). Several high intensity peaks (563, 618, 787, 796 and 825 Da) in the low range were characteristic for CBA while a 3012 Da peak had the highest intensity on CHOC grown cells. By contrast the mass ions of cells grown on NA were markedly suppressed and only apparent in m/z range 2000–3000. MSA is used for the selective isolation and differentiation of S. aureus that ferment mannitol and can grow on at a high salt concentration. The mass spectral pattern obtained from cells grown on MSA contained the highest number of peaks and spanned the m/z range from 500 to 5000. Mass ions with m/z higher than 3000 were detected and were unique for this type of medium. Consequently, in future studies based solely on S. aureus, it may be prudent to utilise MSA as a culture medium for comparative analysis of isolates. However, the mass ion density of cells grown on CBA were still significant compared to MSA and because the latter was used to develop the database, CBA was also retained as

the culture medium for the present study. To confirm that a clinical isolate would react similarly to a type/reference strain by yielding its maximum mass ions in 24 h, strain HPA 30 was cultured on CBA for 24, 48 and 72 h (Fig. 2). With increasing incubation time, the overall quality of the mass spectral traces decreased. The largest density of peaks was present after 24 h; with a slight decrease in intensity in some peaks after 48 h. However, after 72 h very poor profiles were obtained containing a relatively small number of mass ions (Fig. 2). Therefore, for the rest of the study all isolates of S. aureus were cultured for 24 h on CBA prior to MALDI-TOF-MS analysis. The 24 h incubation time is also beneficial if using this approach for identification in a clinical laboratory where short turn-around times are essential. The 95 clinical isolates examined in Phase I of this study were presumptively identified as S. aureus at The Royal London Hospital in October 2005 and further confirmed as authentic S. aureus by the Staphylococcus Reference Unit at HPA (see Section 2). A further 39 clinical isolates from the Staphylococcus Reference Unit were then analysed in Phase II of the study to ascertain the performance of the database. When analysed by MALDI-TOF-MS, all The London Hospital isolates, except four, were correctly identified to genus and species levels using the 2005 database. Of the aberrant strains, three isolates (HPA 80, HPA 547 and HPA 549) were incorrectly identified by the software as Streptococcus pyogenes, S. haemolyticus and S. epidermidis respectively (Table 1a). The remaining strain HPA 260 was identified as S. warneri using the more comprehensive database of 2006 (Table 1a). When assigning identification status of a strain, several factors are taken into account such as the top eight matches in the database, RMS and absolute probability. For the majority of the isolates the top eight hits were all S. aureus thus permitting unambiguous identification. However, when the top eight matches included one or two different Staphylococcus spp., the entry with the top RMS score and absolute probability was selected. All strains tested showed similar profiles and consistent mass ions at 2200 Da and 2600 Da and appears to be specific

Fig. 2. Mass spectral profiles of S. aureus (clinical isolate HPA 30), obtained after 24, 48 or 72 h of growth on CBA. The greatest number of mass ions was detected on cells grown for 24 h. Consistent mass spectral profiles were obtained after 48 h growth but there was a reduction in mass intensity for some of the characteristic mass ions such as m/z 3046 and 2649. Only a few detectable ions were observed after 72 h emphasising the need to analyse cells within the early to late exponential phase of growth (ca. 24–48 h) to obtain high quality MALDI-TOF-MS profiles.

L. Rajakaruna et al. / Infection, Genetics and Evolution 9 (2009) 507–513

511

Table 1a Key features of the clinical isolates obtained from The Royal London Hospital and their corresponding identifications obtained by different methods. The first MALDI analysis was performed against the 2005 version of the database at the HPA while the second analysis was against the 2006 version in Manchester. The four isolates, HPA80, HPA260, HPA547 and HPA549 were initially not identified as S. aureus by the Microbe Lynx software on either of run but subsequently confirmed as S. aureus by sequence analysis of the 16S ribosomal RNA gene. HPA strain designation of isolates used in this study

Hospital identification

MALDI identification of strains using 2005 database

MALDI identification of strains using 2006 database

26, 30, 39, 40, 41, 75, 76, 77, 78, 79, 81, 82, 83, 84, 85, 86, 133, 134, 139, 140, 144, 160, 222, 229, 230, 242, 247, 258, 261, 262, 346, 347, 353, 401, 408, 412, 444, 523, 524, 545, 573 233, 239, 248, 249, 250, 256, 257, 259, 273, 279, 280, 281, 284, 285, 287, 293, 294, 318, 319, 322, 323, 324, 334, 339, 344, 345, 348, 356, 358, 389, 404, 405, 409, 410, 441, 442, 443, 489, 496, 497, 499, 500, 501, 546, 550, 556, 563, 569, 571 80 260 547 549

MRSA

S. aureus

S. aureus

S. aureus

S. aureus

S. aureus

MRSA S. aureus S. aureus S. aureus

Streptococcus pyogenes S. aureus S. haemolyticus S. epidermidis

S. aureus S. warneri S. aureus –

Table 1b Key features of the clinical isolates obtained from the Staphylococcal Reference Unit and their corresponding identifications obtained by MALDI analysis. A total of 39 isolates were analysed using 2006 database version. Thirty-five isolates were correctly identified from the first run and 4 as S. epidermidis by the Microbe Lynx software. The four isolates which were incorrectly identified were re-run and were confirmed as S. aureus. No. isolates

Staphylococcal Reference Unit identification

2006 MALDI database identification

2nd MALDI Run for incorrect samples

35 4

S. aureus S. aureus

S. aureus S. epidermidis

– S. aureus

biomarkers for the genus Staphylococcus (Fig. 3a). For some of the clinical isolates, unusual peaks were detected in the mass range of 2000–3400 Da which were absent from reference/type strains indicating possible phenotypic differences (Fig. 3b). The three isolates incorrectly identified using the 2005 database were further characterised using sequence analysis of the 16S ribosomal gene. Following PCR amplification and sequence analysis, BLAST searches using the NCBI database confirmed that nearly all strains were S. aureus (98–100% sequence homology) but there were indications of contamination in these four samples (Tables 1a and 1b) that may have contributed to their incorrect assignment. Stocks of the cells were revived and the identity of the contaminants ascertained as S. pyogenes, S. warneri, S. haemolyticus and S. epidermidis (see Table 1a). To validate the methodology and assess the inter-laboratory reproducibility of these results, a parallel study was undertaken in collaboration with Manchester Metropolitan University (MMU) and included four purified cultures of S. aureus. This involved reanalysis of all isolates on a different MALDI-TOF-MS instrument in Manchester using the method described here. The resulting data were searched against MicrobeLynxTM 2006. M. lylae (NCTC 13377), EMRSA (NCTC 13134), MRSA (NCTC 11940), S. aureus (NCTC 7727), S. epidermidis (NCTC 11047) and S. epidermidis (NCTC 11407) were included as ‘blind’ controls. All strains were correctly identified at MMU to the species level using the updated 2006 database which contained a further 700 spectra. In a few instances samples became contaminated during transport as revealed from their RMS and probability values. These were again purified by

Identification based upon sequence homology of the 16S rRNA gene

Final MALDI identification of strains that was initially misidentified

S. aureus – S. aureus S. aureus

S. S. S. S.

aureus aureus aureus aureus

sub-culture, re-run and the results were concordant with other isolates. The presence of occasional contaminants in a sample that leads to altered RMS and probability values emphasises the high sensitivity of the technology which allowed for detection of mass ions from the contaminants even before they were visually detected on the agar plates. In a similar manner, in Phase II of the study, the 39 clinical isolates obtained from the Staphylococcal Reference Unit were also correctly identified by the software to species level. These comprised a collection of temporally, geographically and genotypically diverse strains of S. aureus including MSSAs along with representatives of epidemic HA-MRSA (EMRSA-15, -16 and -17), CA-MRSA (so-called USA300, South West pacific and European clones) and two ‘aberrant’ strains (one coagulase-negative S. aureus and one small colony variant). Their successful identification to the species level by MALDI-TOF-MS gives further support to this approach. The results of this study demonstrate clearly that isolates of S. aureus may be identified with a high degree of confidence in a clinical laboratory. However, despite good representation of MRSA strains analysed in the present study (44%), we were unable to identify any consistent mass ions that distinguished MRSA strains from other S. aureus using the same technique reported by Edwards-Jones et al. (2000) and must cast doubt on the validity of this approach to distinguish such important human pathogens. A subsequent study by Carbonelle et al. (2007) showed clear separation of coagulase-negative staphylococci from other closely related taxa. Where there are clear biochemical differences between strains as in the case of UK EMRSA-15 and -16, these may be apparent, however, in such studies it is necessary to preestablish a database and be able to separate stable from inconsistent biomarkers prior to analysis to apply this to a particular situation as shown by Carbonelle et al. (2007). Reliance on a few distinguishing biomarker peaks to differentiate populations of S. aureus into biological types nearly always begin deviating as more wild-type isolates are studied (unpublished observations), therefore the entire mass spectrum should be taken into consideration when devising a classification scheme. The taxonomic level to which this technology may be applied is contentious with reports showing applications to strain level (see e.g. Arnold et al., 2006) and for the differentiation of biotypes (e.g. MRSA and MSSA; Edwards-Jones et al., 2000). The method we have described over the years and used here (see e.g. Keys et al., 2004;

512

L. Rajakaruna et al. / Infection, Genetics and Evolution 9 (2009) 507–513

Fig. 3. MALDI-TOF MS spectrum of the type strain of S. aureus (NCTC 10035) and MS spectra of closely related species. The lack of consistent biomarkers for this group suggests that the entire range of mass ions should be used for comparative analysis to the species level. While most of the MALDI spectra are consistent for isolates, the diversity within the species S. aureus may be gauged through meticulous comparative analysis of the spectrum of each strain. Here, the partial MS spectra within the m/z range 2000–3400 shows the diversity among four clinical isolates. The arrows indicate some of the prominent mass ions.

Shah et al., 2008) was aimed at rapid characterisation of microorganisms to the species level and was intentionally devised to facilitate application in a high throughput clinical laboratory. To date MALDI-TOF-MS of intact cells has been widely used to characterise bacteria, however, its potential as a frontline tool in the diagnostic laboratory has not been reported. Such an application necessitates development of a microbial database in which MS spectra are attained under highly controlled experimental parameters using type and reference strains from an accredited culture collection such as NCTC. The database used in the present study currently holds MALDI-TOF MS profiles of more than 5000 bacterial strains. Here, the performance of this database and the MicrobeLynx search engine to match the spectra of clinical isolates (which may differ) to the profiles of NCTC strains was tested for the first time. At present it is difficult to predict which areas of microbiology MALDI may be used to subtype isolates. This will ultimately depend on the method used, the inherent diversity within a species and the ability to identify unique and consistent biomarkers for a given set of isolates. There are numerous parameters that may be assessed to develop a method for a particular study but given the sensitivity of the technique and variable output between instruments, it is doubtful whether these will have universal application and is more likely to be specific for a given laboratory. For example, as shown here, cells grown on MSA may be useful for subtyping

isolates for a particular study but it seems implausible that these will be reproducible in another environment using a different instrument and parameters. However, it is our conviction that, at the species level, there is now sufficient evidence available from several studies to predict that a general database, encompassing species across the microbial kingdom, is achievable. Indeed the mass spectral database developed by AnagnosTec is now equivalent in size to the 16S rRNA database, contains mainly clinical isolates and therefore offers for the first time in microbiology the potential to develop a genomic–proteomic polyphasic system for the identification of microorganisms. We have recently tested this database and preliminary data from our studies and other groups (unpublished) indicate that after several years of development, a universal technique is fast approaching maturity and will soon be used alongside other well-established methods as a frontline approach for diagnostic microbiology. Success of such a system will depend on global users adopting a unified approach and a lead group taking responsibility to curate and interact with users to continually validate results as the system is used across the broad range of microbial kingdom. 4. Conclusion The present study assessed the potential of MALDI-TOF-MS of intact S. aureus cells as a diagnostic tool in the clinical laboratory

L. Rajakaruna et al. / Infection, Genetics and Evolution 9 (2009) 507–513

and evaluated the performance of an existing database developed for microbial identification (Keys et al., 2004). A major aim of this study was to collect field isolates of S. aureus from a clinical laboratory and, with minimal culturing time, to adapt to the laboratory conditions, challenge the database and the search engine to match their mass spectral profiles to reference spectra of S. aureus present in the database. All but four of the 134 clinical isolates presumptively identified as S. aureus identified correctly using MALDI-TOF-MS profiles. Four aberrant isolates gave discrepant results, but were later confirmed as S. aureus by 16S rRNA analysis. Notably, they were shown to have low level contamination which clearly affected the quality of their mass spectra. Spectral patterns of closely related species of S. aureus such as S. epidermidis, S. haemolyticus and S. saprophyticus showed a high degree of similarity across the genus but there were several genusspecific biomarkers (Fig. 3a). It was possible to ascertain the intraspecies phenotypic diversity among clinical isolates S. aureus and their similarity to type/reference strains by careful by comparative analysis of MALDI spectra. However, the overall similarity in the profiles (as illustrated by the RMS value) rather than specific mass ions were used for identification of strains. Despite the observed differences, there were enough stable characteristics in the profiles of S. aureus to achieve correct identification of isolates. References Arnold, R.J., Karty, J.A., Reilly, J.P., 2006. Bacterial strain differentiation by mass spectrometry. In: Wilkins, C.L., Lay, Jr., J.O. (Eds.), Identification of Microorganisms by Mass Spectrometry. John Wiley & Sons, Inc., Holboken, NJ, pp. 181–201. Carbonelle, E., Beretti, J.L., Cottyn, S., Quesne, G., Berche, P., Nassif, X., Ferroni, A., 2007. Rapid identification of Staphylococci isolated in clinical microbiology laboratories by matrix-assisted-laser-desorption ionisation time of flight mass spectrometry. J. Clin. Microbiol. 45 (7), 2156–2161. Chong, B.E., Wall, D.B., Lubman, D.M., Flynn, S.J., 1997. Rapid profiling of E. coli proteins up to 500 kDa from whole cell lysates using matrix-assisted laser desorption/ionisation mass spectrometry. Rapid Commun. Mass Spectrum. 11, 1900–1908. Claydon, M.A., Davey, S.N., Edwards-Jones, V., Gordon, D.B., 1996. The rapid identification of intact microorganisms using mass spectrometry. Nat. Biotechnol. 14, 1584–1586. Dai, Y., Li, L., Roser, D.C., Long, S.R., 1999. Detection and identification of low-mass peptides and proteins from solvent suspensions of Escherichia coli by high performance liquid chromatography fractionation and matrix-assisted laser desorption/ionisation mass spectrometry. Rapid Commun. Mass Spectrum. 13, 73–78. Edwards-Jones, V., Claydon, M.A., Evason, D.J., Walker, J., Fox, A.J., Gordon, D.B., 2000. Rapid discrimination between methicillin-sensitive and methicillinresistant Staphylococcus aureus by intact cell mass spectrometry. J. Med. Microbiol. 49, 295–300. Fang, H., Hedin, G., 2003. Rapid screening and identification of methicillin-resistant Staphylococcus aureus from clinical samples by selective-broth and real-time PCR assay. J. Clin. Microbiol. 41, 2894–2899.

513

Hillenkamp, F., Koras, M., Beavis, R.C., Chait, B.T., 1991. Matrix-assisted laser desorption/ionisation mass spectrometry of biopolymers. Anal. Chem. 63, 1193A–1203A. Holland, R.D., Wilkes, J.G., Refit, F., Sutherland, J.B., Persons, C.C., Voorhees, K.J., Lay Jr., J.O., 1996. Rapid identification of intact whole bacteria based on spectral patterns using matrix-assisted laser desorption/ionisation mass spectrometry. Rapid Commun. Mass Spectrum. 10, 1227–1232. Ip, M., Lyon, D.J., Chio, F., Enright, M.C., Cheng, A.F., 2003. Characterization of isolates of methicillin-resistant Staphylococcus aureus from Hong Kong by phage typing, pulsed-field gel electrophoresis, and fluorescent amplified-fragment length polymorphism analysis. J. Clin. Microbiol. 41 (11), 4980–4985. Jones, J.J., Stump, M.J., Wilkins, C.L., 2006. MALDI-FTMS of whole-cell bacteria. In: Wilkins, C.L., Lay, Jr., J.O. (Eds.), Identification of Microorganisms by Mass Spectrometry. John Wiley & Sons, Inc., Holboken, NJ, pp. 279–300. Jury, F., Al-Mahrous, M., Apostolou, M., Sandiford, S., Fox, A., Ollier, B.W., Upton, M., 2006. Rapid cost-effective subtyping of methicillin resistant Staphylococcus aureus by denaturing HPLC. J. Med. Microbiol. 55, 1053–1060. Keys, C.J., Dare, D.J., Sutton, H., Wells, G., Lunt, M., McKenna, T., McDowall, M., Shah, H.N., 2004. Compilation of a MALDI-TOF mass spectral database for the rapid screening and characterisation of bacteria implicated in human infectious diseases. Infect. Genet. Evol. 4, 221–242. Kikuchi, K., Enari, T., Totsuka, K., Shimizu, K., 1995. Comparison of phenotypic characteristics, DNA–DNA hybridization results, and results with a commercial rapid biochemical and enzymatic reaction system for identification of viridans group Streptococci. J. Clin. Microbiol. 33, 1215–1222. Klevens, R.M., Morrison, M.A., Nadle, J., Petit, S., Gershman, K., Ray, S., Harrison, L.H., Lynfield, R., Dumyati, G., Townes, J.M., Craig, A.S., Zell, E.R., Fosheim, G.E., McDougal, L.K., Carey, R.B., Fridkin, S.K., 2007. Invasive methicillin-resistant Staphylococcus aureus infections in the United States. J. Am. Med. Assoc. 298, 1763–1771. Lay Jr., J.O., 2002. MALDI-TOF mass spectrometry of bacteria. Mass Spectrom. Rev. 20, 172–194. Lay Jr., J.O., Liyange, R., 2006. MALDI-TOF mass spectrometry of intact bacteria. In: Wilkins, C.L., Lay, Jr., J.O. (Eds.), Identification of Microorganisms by Mass Spectrometry. John Wiley & Sons, Inc., Holboken, NJ, pp. 125–152. Liang, X., Zheng, K., Qian, M.G., Lubman, D.M., 1996. Determination of bacterial protein profiles by matrix-assisted laser desorption/ionisation mass spectrometry with high performance liquid chromatography. Rapid Commun. Mass Spectrum. 10, 1219–1226. Matthew, J., Kuehnert, Hill, H.A., Kupronis, B.A., Tokars, J.I., Solomon, S., Jernigan, D.B., 2005. Methicillin-resistant—Staphylococcus aureus Hospitalizations, United States. Emerg. Infect. Dis. 11, 868–872. Oliveira, D.C., Tomasz, A., De Lencastre, H., 2002. Secrets of success of a human pathogen: molecular evolution of pandemic clones of methicillin-resistant Staphylococcus aureus. Lancet Infect. Dis. 2, 180–189. Shah, H.N., Keys, C.J., Gharbia, S.E., Ralphson, K., Trundle, F., Brookhouse, I., Calydon, M., 2000. The application of MALDI-ToF mass spectrometry to profile the surface of intact bacterial cells. Microb. Ecol. Health Dis. 12, 241–246. Shah, H.N., Keys, C.J., Schmid, O., Gharbia, S.E., 2002. Matrix-assisted laser desorption/ionistaion time of flight mass spectrometry and proteomics; a new era in anaerobic microbiology. Clin. Infect. Dis. 35, 58–64. Shah, H.N., Molenaar, L., Rajakaruna, L., Russell, J.E., Dare, D., Sutton, H., Hallas, G., 2008. Assessment of the stability of cell-surface components of microorganisms by MALDI-TOF-MS following preservation on lenticule discs. FEMS Microbiol. Lett. 285 (2), 284–290. Welham, K.J., Domin, M.A., Scannell, D.E., Cohen, E., Ashton, D.S., 1998. The characterisation of micro-organisms by matrix-assisted laser desorption/ionisation mass spectrometry. Rapid Commun. Mass Spectrum. 12, 176–180. Williams, T.L., Monday, S.R., Musser, S.M., 2006. Bacterial protein biomarker discovery: a focused approach to developing molecular-based identification systems. In: Wilkins, C.L., Lay, Jr., J.O. (Eds.), Identification of Microorganisms by Mass Spectrometry. John Wiley & Sons, Inc., Holboken, NJ, pp. 203–227.