Evaluation of species-specific score cut-off values for various Staphylococcus species using a MALDI Biotyper-based identification

Transcription

1 Journal of Medical Microbiology (2012), 61, DOI /jmm Evaluation of species-specific score cut-off values for various Staphylococcus species using a MALDI Biotyper-based identification Cindy Richter, Stephan Hollstein, Jaroslaw Woloszyn, Martin Kaase, Sören G. Gatermann and Florian Szabados Correspondence Florian Szabados florian.szabados@rub.de Received 23 January 2012 Accepted 22 June 2012 Institute for Hygiene and Microbiology, Department of Medical Microbiology, Ruhr-University Bochum, Germany Although previous studies investigating the MALDI Biotyper database (matrix-assisted laser desorption/ionization time-of-flight mass spectrometry-based identification) have proven its high accuracy for bacterial identification, the studies differed in sample preparation, number of replicates, quantity of shots and target types used. In particular, the score cut-off values of special importance for reliable species identification varied. The aim of the present study was to identify species-specific differences in the mean score values for staphylococci. Cut-off values recommended by the manufacturer were adapted using the 20th percentile to rule out unknown score-modifying factors, even though the specificity was high and the lowest cut-off values would also yield an accurate result. Whilst correct species diagnosis was obtained in % of samples (1382/1420), only 220 of all duplicates (15.49 %) revealed a score of 2.3, whilst 968 (68.17 %) had a score between 2.0 and 2.299, and 194 (13.66 %) had a score of,2.0. Ten of 21 species had a calculated 20th percentile of,2.0 and one species of,1.7. In conclusion, the use of species-specific cut-off values improves the relative sensitivity of species identification in staphylococci. INTRODUCTION The first descriptions of matrix-assisted laser desorption/ ionization time of flight mass spectrometry (MALDI-TOF MS) for bacterial identification were published more than 15 years ago (Claydon et al., 1996; Holland et al., 1996). A variety of studies declared this method to be cost-effective and rapid for bacterial identification (Eigner et al., 2009; Marklein et al., 2009; Seng et al., 2009; Szabados et al., 2010; Carpaij et al., 2011), which is a proven benefit for patient care (Barenfanger et al., 1999; Neville et al., 2011). MALDI-TOF MS is based on soft laser ionization on intact bacteria or bacterial extracts. Specific spectra of protein peaks are yielded by detecting the ribosomal peptide and protein ions according to their relative masses and charges. Distinct protein peaks can be used for genus and species identification. Quality-controlled databases of reference spectra including many relevant micro-organisms were established from the first proof of principle experiments to integrate this method into clinical laboratories. To date, several commercial databases and non-commercial approaches for bacterial identification have been evaluated (Carbonnelle et al., 2007; Friedrichs et al., 2007; Rajakaruna Abbreviations: CoNS, coagulase-negative staphylococci; MALDI-TOF MS, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. et al., 2009; Seng et al., 2009). Nevertheless, these experiments have varied, particularly with regard to the sample preparation protocols: either formic acid extraction or direct smear preparation. Moreover, the samples were measured with different mass spectrometer instruments and the spectra generated were compared with varying ranges of proteomic fingerprints (Friedrichs et al., 2007; Rajakaruna et al., 2009). In this study, the MALDI Biotyper 2.0 database with a mass-to-charge ratio of Da was used, most of which are thought to represent ribosomal protein peaks. This database contains.3000 well-characterized microbial species, including 124 spectra representing 36 different staphylococcal species and subspecies. Although the accuracy of the MALDI Biotyper 2.0 database has been described in a variety of studies (Seng et al., 2009; Bizzini et al., 2010; Harris et al., 2010; Risch et al., 2010; Szabados et al., 2010; Neville et al., 2011), these experiments also varied in many details. A divergent number of tested strains, number of spectra to generate a sum spectrum and number of replicates tested and the use of different mass spectrometry instruments and target types make it difficult to compare the published data. Furthermore, the score values to define probable or secure species identification were altered in many studies compared with the manufacturer s suggestions. In all these studies, the same score classification was G 2012 SGM Printed in Great Britain 1409

2 C. Richter and others applied to all measurements, independent of putative species-specific variations. Misidentifications were not described in detail concerning calculated score values. The MALDI Biotyper 2.0 software compares the peak profile of tested strains with database entries and calculates a logarithmic score between 0 and 3, which quantifies the similarity to the database entries. High values in this score represent high similarity with respective database entries. The exact algorithm by which this score is generated has not been disclosed by the company. The quality of species identification is divided into several categories according to the score: as specified by the manufacturer s instructions, identification scores of 2.3 are required for a reliable ( highly probable ) identification to species and genus level, score values 2.0 and,2.3 represent probable identification to species level, score values 1.7 and,2.0 represent a reliable genus level, and scores of,1.7 are regarded as unreliable. Due to the log dependence, even small differences in the score cut-off value represent larger differences in the peak profile. In a recently published study, differences in the mean specific score values in a variety of species were described (Szabados et al., 2012). Nevertheless, only a few species within the genus Staphylococcus were included in this study. In order to characterize these described differences in the mean score values in more detail, we used the ethanol/formic acid extraction procedure as suggested by the manufacturer. For definition of speciesspecific score cut-off values, the 20th percentiles of 691 concordantly identified Staphylococcus isolates were applied (19 isolates of species with low sample numbers were excluded). In addition, the accuracy of MALDI-TOF MSbased identification on score values was analysed using reference strains. METHODS A total of 697 routinely isolated clinical strains of staphylococci and 13 type strains were included in this study. This collection originated from hospitals in and around Bochum, Germany. All bacteria were isolated from human sources from various origins. Table 1 summarizes the tested staphylococci. This study comprised species of coagulase-negative staphylococci (CoNS) of relevance to humans. As isolates were collected randomly and non-consecutively, some species might have been over- or under-represented compared with their prevalence in humans. Reference strains. Thirteen reference strains were used: Staphylococcus caprae (DSM 20608) (Kawamura et al., 1998); Staphylococcus cohnii (DSM 20260), Staphylococcus haemolyticus (DSM 20263), Staphylococcus saprophyticus (DSM 20229) and Staphylococcus xylosus (DSM 20267) (Schleifer & Kloos, 1975); Staphylococcus hominis (DSM 20328), Staphylococcus simulans (DSM 20322) and Staphylococcus warneri (DSM 20316) (Kloos & Schleifer, 1975); Staphylococcus epidermidis (DSM 1798) (Winslow, 1908); Staphylococcus hyicus (DSM 20459) (Schleifer & Kocur, 1973); Staphylococcus intermedius (DSM 20373) (Hajek, 1976); and Staphylococcus lugdunensis (DSM 4804) and Staphylococcus schleiferi (DSM 4807) (Freney et al., 1988). Identification of Staphylococcus aureus. A total of 105 consecutive clinical single-copy strains of S. aureus were recovered from January 2006 to March All isolates were cultured on Columbia blood agar plates (biomérieux) and incubated at 37 uc in a humidified atmosphere containing 5 % CO 2. S. aureus was identified initially on the basis of colony characteristics and a Slidex Staph Plus agglutination test (biomérieux) or the GPI card of the Vitek 2 automated identification system (biomérieux) and confirmed by molecular methods (Martineau et al., 1998). All strains were SA442- positive and were included as molecularly defined strains in this study. Identification of CoNS. All CoNS isolates, excluding the reference strains, were collected between 2000 and 2008 and identified initially using the GPI card of the Vitek 2 automated system (biomérieux). In this study, 592 single-copy isolates of molecularly defined CoNS were included (Table 1). Species-specific PCRs were used as the reference method in case of discrepant identification results (Table 2). Forty-six strains of 19 species (excluding meticillin-sensitive S. aureus, S. hyicus and S. xylosus) were identified using amplification and sequencing of the soda gene (Martineau et al., 2000; Poyart et al., 2001; Shrestha et al., 2002; Stuhlmeier & Stuhlmeier, 2003; Morot-Bizot et al., 2004; Iwase et al., 2007; Hauschild & Stepanovic, 2008; Noguchi et al., 2010): Staphylococcus arlettae (n51), Staphylococcus capitis (n54), S. cohnii (n51), S. epidermidis (n53), S. haemolyticus (n56), S. hominis (n516), S. lugdunensis (n55), S. saprophyticus (n51), S. schleiferi (n52), Staphylococcus succinus (n55) and S. warneri (n52). MALDI-TOF MS-based identification. All strains were stored in frozen stock culture and subcultured on Columbia agar plates (biomérieux) for 24 h at 37 uc twice before they were examined by MALDI-TOF MS using a Microflex LT instrument (Bruker Daltonics), Flexcontrol 3.0 software and the MALDI Biotyper 2.0 database version (Bruker Daltonics). The instrument was used to calculate and process the analytical data according to the manufacturer s instructions. An ethanol/formic acid preparation was used in accordance with the manufacturer s specifications and as described previously (Szabados et al., 2011a). Each sample was measured in duplicate. To eliminate a suspected preparation error, the identification was repeated once, if the initially obtained score was,1.7. For implementation of new species-specific score cut-off values, all scores of correctly identified measurements were included in this study. After retesting of respective strains, all measured scores were compared and the first result was chosen if there was no error of preparation recognizable. A no peaks found result was interpreted as insufficient identification by MALDI-TOF MS, even when the second sample was identified correctly. The 20th percentile score (Fig. 1), depending on the number of samples, was calculated to define new species-specific score cut-off values for secure species identification. Statistical analysis. The data obtained were analysed using SPSS 19.0 for Windows. To describe the distributions of score values, mean, SD, median and extreme values were used. Additionally, the 20th percentile was calculated. RESULTS A total of 710 strains were measured in duplicate (n5 1420). According to the manufacturer s score classification, 220 (15.49 %) had a score of 2.3, 968 (68.17 %) had a score 2.0 and,2.3, and 194 (13.66 %) had a score of,2.0. A no peaks found result was generated by 2.25 % (32/ 1420) where the second sample was identified correctly. In 1410 Journal of Medical Microbiology 61

3 MALDI Biotyper species-specific score values Table 1. Distribution of the tested staphylococci Staphylococcus species No. routinely isolated strains No. reference strains Total S. arlettae S. aureus S. capitis S. caprae S. carnosus S. chromogenes S. cohnii S. epidermidis S. haemolyticus S. hominis S. hyicus S. intermedius S. lugdunensis S. pasteuri S. saprophyticus S. schleiferi S. sciuri S. simulans S. succinus S. warneri S. xylosus Total % (6/1420), MALDI-TOF MS misidentified the first sample, although the second sample was identified correctly. Four samples of different isolates of S. saprophyticus were falsely identified as Staphylococcus equorum (score of 1.687), Staphylococcus piscifermentans (2.040), S. xylosus (1.842) and Pseudomonas balearica (1.318). The single isolate of S. carnosus was misidentified once as S. piscifermentans (1.767), and one sample of S. cohnii was misidentified as S. xylosus (1.842). Although 97.3 % of the isolates were identified correctly to the species level by the MALDI Biotyper database (irrespective of the score value), only % of the isolates were interpreted as highly probable species identification (score 2.3), and an additional % were interpreted as secure genus and probable species identification (score 2.0 and,2.3), if the score cut-off values suggested by the manufacturer were applied. The distribution of scores for each species is summarized in Fig. 1, with horizontal dashed lines indicating the manufacturer s score cut-off values. According to the manufacturer s recommendations, isolates would have been misidentified due to score values under the threshold. Even though the specificity was very high (100 % in all staphylococci, except for S. xylosus), a species-specific 20th percentile was chosen arbitrarily as the score cut-off value for identification to improve security. Only species with more than three isolates in duplicate were included to evaluate the 20th percentile. This led to exclusion of six of 21 species (S. arlettae, S. carnosus, S. chromogenes, S. hyicus, S. pasteuri and S. xylosus). These species are indicated by asterisks in Fig. 1. Seven of 15 species (S. caprae, S. intermedius, S. warneri, S. sciuri, S. schleiferi, S. succinus and S. epidermidis) had a calculated score cut-off value of,2, and one species (S. cohnii) had a score of,1.7. Score values for S. lugdunensis produced the highest 20th percentile, with a value of (Fig. 1). Except for S. xylosus (99.8 %), all tested species had a specificity of 100 % in the tested collective. The accuracy of MALDI-TOF MS at the isolate level was also analysed by the inclusion of 13 type strains in the MALDI Biotyper 2.0 database. If the strain tested is clonally identical to the strain used for the database entry and a strain-specific influence on the score value is therefore ruled out, the highest theoretical score (a score of 3) should be able to be obtained. All but one of the type strains (S. epidermidis DSM 1798 gained a no peaks result) were identified correctly to the species level (96.15 %). However, the best matching score was not necessarily with the respective identical DSM reference strain but with other DSM reference strains of the same species used in the MALDI Biotyper database. Seven reference strains (DSM 4804, DSM 20229, DSM 20322, DSM 20328, DSM 20373, DMS and DSM 20608) were detected twice with identical DSM strains (53.85 %). Two reference strains (DSM and DSM 20316) were only identified in one of two measurements (15.38 %). In four strains (DSM 1798, DSM 4807, DSM and DSM 20263), identification failed twice with regard to the DSM reference strain (30.77 %). In the case of incorrect DSM identification, the lowest score for the correct DSM number at the

4 C. Richter and others Table 2. Primers used for PCR in this study F, Forward; R, reverse. Gene Strain(s) Sequence (5 A3 ) Source soda S. arlettae, S. caprae, S. carnosus, S. chromogenes, S. cohnii, S. hominis, S. intermedius, S. pasteuri, S. schleiferi, S. sciuri, S. simulans, S. succinus F: CCITAYICITAYGAYGCIYTIGARCC Poyart et al. (2001) R: ARRTARTAIGCRTGYTCCCAIACRTC dnaj S. warneri F: GCCAAAAGAGACTATTATGA Hauschild & Stepanovic (2008) R: ATTGYTTACCYGTTTGTGTACC soda S. capitis F: GCTAATTTAGATAGCGTACCTTCA Iwase et al. (2007) R: CAGATCCAAAGCGTGCA soda S. haemolyticus F: GTTGAGGGAACAGAT Iwase et al. (2007) R: CAGCTGTTTGAATATCTT soda S. warneri F: GTAGCTAACTTAGATAGTGTTCCTTCT Iwase et al. (2007) R: CCGCCACCGTTATTTCTT SA442 S. aureus F: GTCGGTACACGATATTCTTCACG Shrestha et al. (2002) R: CTCTCGTATGACCAGCTTCGGTAC Chromosomal fragment S. epidermidis F: ATCAAAAAGTTGGCGAACCTTTTCA Stuhlmeier & Stuhlmeier (2003) of S. epidermidis R: CAAAAGAGCGTGGAGAAAAGTATCA Chromosomal fragment S. saprophyticus F: TCAAAAAGTTTTCTAAAAAATTTAC Martineau et al. (2000) of S. saprophyticus R: ACGGGCGTCCACAAAATCAATAGGA tana S. lugdunensis F: GCTGCGCCAATTTGTTCTAAATAT Noguchi et al. (2010) R: AGCATGGGCAATAACAGCAGTAA Chromosomal fragment of S. xylosus S. xylosus F: AACGCGCAACGTGATAAAATTAATG Morot-Bizot et al. (2004) R: AACGCGCAACAGCAATTACG measuring sheet was yielded by S. cohnii (DSM 20260) with a score value of The highest scoring DSM reference strain was S. lugdunensis DSM 4804 with a score of (Table 3). DISCUSSION In recent years, staphylococci have increasingly come to attention due to the clinical importance of hospitalacquired infections (von Eiff et al., 2006), particularly with the upward trend of antibiotic resistance in the case of meticillin-resistant S. aureus. CoNS are increasingly observed as causes of infections of intravascular devices as well as of prosthetic infections (Szabados et al., 2011c). These bacteria must be identified at the species level using reliable and reproducible methods. Bacteria are usually identified by their biochemical characteristics, which is moderately reliable for many species but time-consuming. Other highly reliable methods such as molecular methods are expensive, technically demanding and also time-consuming. Clinical microbiology laboratories are increasingly challenged to deal with the numerous microbial samples within an adequate time. In contrast, MALDI-TOF MS requires simple sample preparation and the analysis can be performed within minutes. Whether ethanol/formic acid extraction is preferable to direct smear preparation is controversial, particularly for Gram-positive bacteria and yeasts (Bizzini et al., 2010; Szabados et al., 2012). During ethanol/formic acid extraction, the organisms are killed, in contrast to direct smear preparation. This may lower the risk of laboratory-acquired infections caused by pathogenic bacteria. Bruker Daltonics suggests score cut-off values using ethanol/formic acid extraction for identification of bacteria to genus and species level. Although a score of 2.3 is recommended for highly probable species identification, previous studies (Bizzini et al., 2010; Dubois et al., 2010; Harris et al., 2010; Cherkaoui et al., 2011; Neville et al., 2011) modified this cut-off value to values of 1.7 and,2.0 for identification of solitary cultures on agar culture media. Concerning direct identification of bacteria from positive blood culture bottles, lower score cut-off values have been suggested (La Scola & Raoult, 2009; Stevenson et al., 2010; Szabados et al., 2011b). The currently available literature suggests that the manufacturer s recommendations are overly stringent (Neville et al., 2011). A recently published study analysed a variety of species with regard to differences in the mean specific score values for direct smear preparation (Szabados et al., 2012). In the latter study, only 105 isolates of 10 staphylococcal species were tested. In contrast, the present study evaluated 710 isolates of 21 staphylococcal species measured in duplicate. Both studies confirmed species-specific differences in the mean score values within the genus Staphylococcus using the MALDI 1412 Journal of Medical Microbiology 61

5 MALDI Biotyper species-specific score values Highly probable species identification Secure genus identification, probable species Score Probable genus identification No reliable identification S. arlettae S. aureus S. capitis S. carnosus S. chromogenes S. cohnii S. caprae S. epidermidis S. haemolyticus S. hominis S. hyicus S. intermedius S. lugdunensis S. pasteuri S. saprophyticus S. xylosus S. schleiferi S. sciuri S. simulans S. succinus S. warneri Min. Max Mean SD >2 (%) th per Cut-off Fig. 1. Distribution of mean score values of staphylococci displayed as box-and-whisker plots. The percentage of correctly identified isolates is also given for a score cut-off value of 2.0. An arbitrarily chosen 20th percentile was used to describe reasonable new species-specific cut-off values. Biotyper 2.0 database. To improve the accuracy of MALDI- TOF MS at the species level, new species-specific score cutoff values were defined by calculating the 20th percentiles (Fig. 1). These cut-off values can be applied directly to each measurement in a species-specific manner or could be added as a simplified classification scheme for conventional laboratory considerations by outlining matching groups. The 20th percentile was chosen as most of the misidentifications occurred in samples associated with score values in this range (Szabados et al., 2012). We suggest using a score of 2.0 for secure species identification of S. aureus, S. capitis, S. epidermidis, S. hominis, S. lugdunensis, S. saprophyticus, S. simulans and S. xylosus. According to the 20th percentile, a score of 1.7 is sufficient for identification at the species level for S. caprae, S. intermedius, S. schleiferi, S. sciuri, S. succinus and S. warneri. S. cohnii can be identified with a score cut-off value of 1.6. Apart from reducing the manufacturer s cut-off values, the implementation of constant variables for calculation of speciesspecific score values into the manufacturer s classification system is desirable. A lower mean species-specific score value could also indicate a certain species where the given database could be improved. MALDI-TOF MS has been shown to be accurate in both sensitivity and specificity at the species level for staphylococci (Seng et al., 2009; Dupont et al., 2010; Szabados et al., 2012). Except for S. xylosus, all tested species had a specificity of 100 %, which reinforces the utilization of low score cut-off values for species identification. S. equorum, S. piscifermentans (Dupont et al., 2010) and S. saprophyticus (Dubois et al., 2010) have been described previously to be incidentally not correctly identified by MALDI-TOF MS. Score values of up to 2.04 for misidentifications suggest that users should give more attention to critical species such as S. equorum, S. piscifermentans and S. xylosus. In the present study, the highest obtained score value was 2.580, with 3.0 being the maximum possible score. It can be assumed that clonally identical isolates, such as type strains, should be able to reach score values of 3.0. Interestingly, the score values of our tested reference strains were distinctly lower and did not even reach the score cut-off value for highly probable species identification. According to Harris et al. (2010), it is possible to detect score values of 3.0 if reference strains are used that are identical to the strains used for the database (Harris et al., 2010). This result also indicated that some reference strains used for this study were different to the respective peak profiles used

6 C. Richter and others Table 3. Identification of DSM reference strains Species Reference strain tested Records in database of respective species Highest scoring reference strain Score Score of identical DSM reference strain S. cohnii DSM DSM S. schleiferi DSM DSM S. epidermidis DSM No peaks found DSM S. haemolyticus DSM DSM S. warneri DSM DSM DSM S. xylosus DSM DSM DSM S. saprophyticus DSM DSM S. hyicus DSM DSM S. intermedius DSM DSM S. hominis DSM DSM S. simulans DSM DSM S. caprae DSM DSM S. lugdunensis DSM DSM All strains were measured in duplicate. in the database, which is also in line with the hypothesis that not all reference strains used in research are identical with regard to their clonal background. Protein expression depends on a variety of environmental factors, such as culture conditions, including different growth phases, agar plates, ph and temperature, and may also have contributed to this effect. Although several studies have pointed out that differences in peak profile are dependent on environmental conditions (Longo et al., 1999; Bernardo et al., 2002; Valentine et al., 2005; Wunschel et al., 2005), the identifying algorithm and detection range of the MALDI Biotyper database allow reliable identification from a variety of agar media. A large number of clonally different strains should be analysed to get a peak profile that is reliable and appropriate for one species. Lower mean species-specific score values also indicate certain species where the MALDI Biotyper database could be improved. Nevertheless, the lower score values within the subset of DSM isolates also indicate that other factors, such as variations in bacterial preparation, putative dependence on instruments or as yet unknown environmental factors, could reveal significantly lower score values. As the score calculation algorithm of the MALDI Biotyper has not been revealed by the manufacturer, such factors that lower attainable scores should be analysed in further studies to investigate their relevance in routine clinical applications, such as MALDI-TOF MS-based identification using the MALDI Biotyper database. Conclusions In summary, the recently introduced MALDI Biotyper database is suitable for highly accurate and rapid identification of staphylococci. Misidentifications to species level occurred only within the group of CoNS, and therefore discrimination of S. aureus from CoNS is highly reliable, even when lower score values are applied. Species-specific score values varied among the staphylococci, and these score cut-off values could be used to improve the identification scheme introduced by the manufacturer. REFERENCES Barenfanger, J., Drake, C. & Kacich, G. (1999). Clinical and financial benefits of rapid bacterial identification and antimicrobial susceptibility testing. J Clin Microbiol 37, Bernardo, K., Pakulat, N., Macht, M., Krut, O., Seifert, H., Fleer, S., Hünger, F. & Krönke, M. (2002). Identification and discrimination of 1414 Journal of Medical Microbiology 61

7 MALDI Biotyper species-specific score values Staphylococcus aureus strains using matrix-assisted laser desorption/ ionization-time of flight mass spectrometry. Proteomics 2, Bizzini, A., Durussel, C., Bille, J., Greub, G. & Prod hom, G. (2010). Performance of matrix-assisted laser desorption ionization-time of flight mass spectrometry for identification of bacterial strains routinely isolated in a clinical microbiology laboratory. J Clin Microbiol 48, Carbonnelle, 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 ionization-time of flight mass spectrometry. J Clin Microbiol 45, Carpaij, N., Willems, R. J., Bonten, M. J. & Fluit, A. C. (2011). Comparison of the identification of coagulase-negative staphylococci by matrix-assisted laser desorption ionization time-of-flight mass spectrometry and tuf sequencing. Eur J Clin Microbiol Infect Dis 30, Cherkaoui, A., Emonet, S., Fernandez, J., Schorderet, D. & Schrenzel, J. (2011). Evaluation of matrix-assisted laser desorption ionization-time of flight mass spectrometry for rapid identification of b-hemolytic streptococci. J Clin Microbiol 49, 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, Dubois,D.,Leyssene,D.,Chacornac,J.P.,Kostrzewa,M.,Schmit,P.O., Talon, R., Bonnet, R. & Delmas, J. (2010). Identification of a variety of Staphylococcus species by matrix-assisted laser desorption ionizationtime of flight mass spectrometry. J Clin Microbiol 48, Dupont, C., Sivadon-Tardy, V., Bille, E., Dauphin, B., Beretti, J. L., Alvarez, A. S., Degand, N., Ferroni, A., Rottman, M. & other authors (2010). Identification of clinical coagulase-negative staphylococci, isolated in microbiology laboratories, by matrix-assisted laser desorption/ionization-time of flight mass spectrometry and two automated systems. Clin Microbiol Infect 16, Eigner, U., Holfelder, M., Oberdorfer, K., Betz-Wild, U., Bertsch, D. & Fahr, A. M. (2009). Performance of a matrix-assisted laser desorption ionization-time-of-flight mass spectrometry system for the identification of bacterial isolates in the clinical routine laboratory. Clin Lab 55, Freney, J., Brun, Y., Bes, M., Meugnier, H., Grimont, F., Grimont, P. A. D., Nervi, C. & Fleurette, J. (1988). Staphylococcus lugdunensis sp. nov. and Staphylococcus schleiferi sp. nov., two new species from human clinical specimens. Int J Syst Bacteriol 38, Friedrichs, C., Rodloff, A. C., Chhatwal, G. S., Schellenberger, W. & Eschrich, K. (2007). Rapid identification of viridans streptococci by mass spectrometric discrimination. J Clin Microbiol 45, Hajek, V. (1976). Staphylococcus intermedius, a new species isolated from animals. Int J Syst Bacteriol 26, Harris, L. G., El-Bouri, K., Johnston, S., Rees, E., Frommelt, L., Siemssen, N., Christner, M., Davies, A. P., Rohde, H. & Mack, D. (2010). Rapid identification of staphylococci from prosthetic joint infections using MALDI-TOF mass-spectrometry. Int J Artif Organs 33, Hauschild, T. & Stepanovic, S. (2008). Identification of Staphylococcus spp. by PCR-restriction fragment length polymorphism analysis of dnaj gene. J Clin Microbiol 46, Holland, R. D., Wilkes, J. G., Rafii, F., Sutherland, J. B., Persons, C. C., Voorhees, K. J. & Lay, J. O., Jr (1996). Rapid identification of intact whole bacteria based on spectral patterns using matrix-assisted laser desorption/ionization with time-of-flight mass spectrometry. Rapid Commun Mass Spectrom 10, Iwase, T., Seki, K., Shinji, H., Mizunoe, Y. & Masuda, S. (2007). Development of a real-time PCR assay for the detection and identification of Staphylococcus capitis, Staphylococcus haemolyticus and Staphylococcus warneri. J Med Microbiol 56, Kawamura, Y., Hou, X.-G., Sultana, F., Hirose, K., Miyake, M., Shu, S.-E. & Ezaki, T. (1998). Distribution of Staphylococcus species among human clinical specimens and emended description of Staphylococcus caprae. J Clin Microbiol 36, Kloos, W. E. & Schleifer, K. H. (1975). Isolation and characterisation of staphylococci from human skin. II. Description of four new species: Staphylococcus warneri, Staphylococcus capitis, Staphylococcus hominis, and Staphylococcus simulans. Int J Syst Bacteriol 25, La Scola, B. & Raoult, D. (2009). Direct identification of bacteria in positive blood culture bottles by matrix-assisted laser desorption ionisation time-of-flight mass spectrometry. PLoS ONE 4, e8041. Longo, M. A., Novella, I. S., Garcia, L. A. & Diaz, M. (1999). Comparison of Bacillus subtilis and Serratia marcescens as protease producers under different operating conditions. J Biosci Bioeng 88, Marklein, G., Josten, M., Klanke, U., Müller, E., Horré, R., Maier, T., Wenzel, T., Kostrzewa, M., Bierbaum, G. & other authors (2009). Matrix-assisted laser desorption ionization-time of flight mass spectrometry for fast and reliable identification of clinical yeast isolates. J Clin Microbiol 47, Martineau, F., Picard, F. J., Roy, P. H., Ouellette, M. & Bergeron, M. G. (1998). Species-specific and ubiquitous-dna-based assays for rapid identification of Staphylococcus aureus. J Clin Microbiol 36, Martineau, F., Picard, F. J., Ménard, C., Roy, P. H., Ouellette, M. & Bergeron, M. G. (2000). Development of a rapid PCR assay specific for Staphylococcus saprophyticus and application to direct detection from urine samples. J Clin Microbiol 38, Morot-Bizot, S. C., Talon, R. & Leroy, S. (2004). Development of a multiplex PCR for the identification of Staphylococcus genus and four staphylococcal species isolated from food. J ApplMicrobiol97, Neville, S. A., Lecordier, A., Ziochos, H., Chater, M. J., Gosbell, I. B., Maley, M. W. & van Hal, S. J. (2011). Utility of matrix-assisted laser desorption ionization-time of flight mass spectrometry following introduction for routine laboratory bacterial identification. J Clin Microbiol 49, Noguchi, N., Goto, K., Ro, T., Narui, K., Ko, M., Nasu, Y., Utsumi, K., Takazawa, K., Moriyasu, F. & Sasatsu, M. (2010). Using the tannase gene to rapidly and simply identify Staphylococcus lugdunensis. Diagn Microbiol Infect Dis 66, Poyart, C., Quesne, G., Boumaila, C. & Trieu-Cuot, P. (2001). Rapid and accurate species-level identification of coagulase-negative staphylococci by using the soda gene as a target. J Clin Microbiol 39, Rajakaruna, L., Hallas, G., Molenaar, L., Dare, D., Sutton, H., Encheva, V., Culak, R., Innes, I., Ball, G. & Sefton, A. M. (2009). High throughput identification of clinical isolates of Staphylococcus aureus using MALDI- TOF-MS of intact cells. Infect Genet Evol 9, Risch, M., Radjenovic, D., Han, J. N., Wydler, M., Nydegger, U. & Risch, L. (2010). Comparison of MALDI TOF with conventional identification of clinically relevant bacteria. Swiss Med Wkly 140, w Schleifer, K. H. & Kloos, W. E. (1975). Isolation and characterization of staphylococci from human skin. I. Amended descriptions of Staphylococcus epidermidis and Staphylococcus saprophyticus, and description of three new species: Staphylococcus cohnii, Staphylococcus haemolyticus, and Staphylococcus xylosus. Int J Syst Bacteriol 25, Schleifer, K. H. & Kocur, M. (1973). Classification of staphylococci based on chemical and biochemical properties. Arch Mikrobiol 93,

8 C. Richter and others Seng, P., Drancourt, M., Gouriet, F., La Scola, B., Fournier, P. E., Rolain, J. M. & Raoult, D. (2009). Ongoing revolution in bacteriology: routine identification of bacteria by matrix-assisted laser desorption ionization time-of-flight mass spectrometry. Clin Infect Dis 49, Shrestha, N. K., Tuohy, M. J., Hall, G. S., Isada, C. M. & Procop, G. W. (2002). Rapid identification of Staphylococcus aureus and the meca gene from BacT/ALERT blood culture bottles by using the LightCycler system. J Clin Microbiol 40, Stevenson, L. G., Drake, S. K. & Murray, P. R. (2010). Rapid identification of bacteria in positive blood culture broths by matrixassisted laser desorption ionization-time of flight mass spectrometry. J Clin Microbiol 48, Stuhlmeier, R. & Stuhlmeier, K. M. (2003). Fast, simultaneous, and sensitive detection of staphylococci. J Clin Pathol 56, Szabados, F., Woloszyn, J., Richter, C., Kaase, M. & Gatermann, S. (2010). Identification of molecularly defined Staphylococcus aureus strains using matrix-assisted laser desorption/ionization time of flight mass spectrometry and the Biotyper 2.0 database. J Med Microbiol 59, Szabados, F., Woloszyn, J., Kaase, M. & Gatermann, S. G. (2011a). False-negative test results in the Slidex Staph Plus (biomérieux) agglutination test are mainly caused by spa-type t001 and t001-related strains. Eur J Clin Microbiol Infect Dis 30, Szabados, F., Michels, M., Kaase, M. & Gatermann, S. (2011b). The sensitivity of direct identification from positive BacT/ALERT2 (biomérieux) blood culture bottles by matrix-assisted laser desorption ionization time-of-flight mass spectrometry is low. Clin Microbiol Infect 17, Szabados, F., Anders, A., Kaase, M., Marlinghaus, L., Gatermann, S. G., Teske, W. & Lichtinger, T. (2011c). Late periprosthetic joint infection due to Staphylococcus lugdunensis identified by matrixassisted laser desorption/ionisation time of flight mass spectrometry: a case report and review of the literature. Case Rep Med 2011, Szabados, F., Tix, H., Anders, A., Kaase, M., Gatermann, S. G. & Geis, G. (2012). Evaluation of species-specific score cutoff values of routinely isolated clinically relevant bacteria using a direct smear preparation for matrix-assisted laser desorption/ionization time-offlight mass spectrometry-based bacterial identification. Eur J Clin Microbiol Infect Dis 31, Valentine, N., Wunschel, S., Wunschel, D., Petersen, C. & Wahl, K. (2005). Effect of culture conditions on microorganism identification by matrix-assisted laser desorption ionization mass spectrometry. Appl Environ Microbiol 71, von Eiff, C., Arciola, C. R., Montanaro, L., Becker, K. & Campoccia, D. (2006). Emerging Staphylococcus species as new pathogens in implant infections. Int J Artif Organs 29, Winslow, C. E. (1908). A new method of enumerating bacteria in air. Science 28, Wunschel, D. S., Hill, E. A., McLean, J. S., Jarman, K., Gorby, Y. A., Valentine, N. & Wahl, K. (2005). Effects of varied ph, growth rate and temperature using controlled fermentation and batch culture on matrix assisted laser desorption/ionization whole cell protein fingerprints. J Microbiol Methods 62, Journal of Medical Microbiology 61