SGM Journals
DIAGNOSTICS, TYPING AND IDENTIFICATION

Identification of molecularly defined Staphylococcus aureus strains using matrix-assisted laser desorption/ionization time of flight mass spectrometry and the Biotyper 2.0 database

Florian Szabados, Jaroslaw Woloszyn, Cindy Richter, Martin Kaase, Sören Gatermann

  • Institute for Hygiene and Microbiology, Department of Medical Microbiology, University of Bochum, Bochum, Germany
  • Correspondence
    Florian Szabados
    Florian.szabados{at}rub.de

    • Journal of Medical Microbiology 2010; 59(7):787–790 · https://doi.org/10.1099/jmm.0.016733-0

    View at publisher PubMed

    Abstract

    Matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF MS) has recently been introduced for bacterial identification. To our knowledge, this is the first study where the Biotyper 2.0 database (Bruker Daltonics) has been applied for bacterial identification in a local strain collection of molecularly defined Staphylococcus aureus. We showed that the accuracy of the Biotyper 2.0-based identification for 602 molecularly defined strains of S. aureus, irrespective of meticillin resistance, was equivalent to that of the molecularly defined reference even at a score cut-off value of 2. Also, 412 isolates of 20 different species of non-S. aureus staphylococci were all correctly identified to species level compared to the molecularly defined reference. Moreover, the MALDI-TOF MS-based S. aureus identification approach was clearly faster than more time-consuming methods such as a molecular identification approach.

    INTRODUCTION

    Matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF MS) was introduced several years ago as a new method for bacterial identification (Bernardo et al., 2002; Du et al., 2002; Edwards-Jones et al., 2000; Walker et al., 2002). This technology was applied to selected isolates of Staphylococcus aureus and it was shown that bacterial identification using distinct protein peaks is possible. These experiments differed with regard to the matrix used for MALDI-TOF MS, the protein-extraction protocols and the bacterial culture conditions. Moreover, the use of different proteomic fingerprint ranges makes it difficult to compare the published data. For the Biotyper 2.0 identification approach, protein peaks in the mass-to-charge ratio of 3000–15 000 Da were used, most of them thought to represent ribosomal protein peaks.

    Three studies with differences with regard to the study design have been described applying MALDI-TOF MS technology to more than 68 strains of S. aureus (Carbonnelle et al., 2007; Rajakaruna et al., 2009; Seng et al., 2009). In these studies, the rate of correct identification seemed to be dependent on the underlying database and algorithm used for the identification, characterization of the strain collection as well as the investigated proteomic fingerprint range. In the first study (Carbonnelle et al., 2007), about 70 strains were tested at a mass-to-charge ratio of 1000–11 000 Da (m/z). Instead of a commercial database, a limited number of peaks were used for an identification approach to identify S. aureus amongst other coagulase-negative staphylococci. In another study (Rajakaruna et al., 2009), about 134 strains were tested at a mass-to-charge ratio of 500–10 000 Da using MicrobeLynx software (Waters). Four isolates were not identified (3 % mis-identification) to genus or species level (Rajakaruna et al., 2009). In a third study (Seng et al., 2009), the Biotyper 2.0 database (Bruker Daltonics) was applied. A rate of correct species-level identification of only 94.4 % was achieved for S. aureus (n=347 strains). However, the reference identification of S. aureus in this study was done with the Vitek 2 (bioMérieux) automated identification system. To our knowledge, ours is the first study of the evaluation of the Biotyper 2.0 database (Bruker Daltonics) where a large number of 602 consecutively collected, single-copy strains of molecularly defined S. aureus and 412 molecularly characterized non-S. aureus strains were included.

    METHODS

    Identification of S. aureus.

    In the first collection, 363 single-copy strains of meticillin-susceptible S. aureus (MSSA) were consecutively collected from January 2006 to March 2006.

    In the second collection, 243 single-copy strains of meticillin-resistant S. aureus (MRSA) were consecutively collected from January 2007 to September 2007.

    S. aureus was primarily identified by typical colony morphology and the Slidex Staph Plus agglutination test (bioMérieux) or the GPI card of the Vitek 2 automated identification system (bioMérieux). The species identification of meticillin-susceptible strains was confirmed using S. aureus specific primer SA442, as described previously (Martineau et al., 1998). All MSSA strains from the first collection were SA442-positive and were included as molecularly defined strains in this study.

    All isolates from the second MRSA collection were primarily identified as described above. In addition, meticillin resistance was confirmed using a PCR for the mecA gene as described previously (Shrestha et al., 2002). Most MRSA strains from the second collection were mecA- and SA442-positive (n=239) and were included as molecularly defined MRSA strains in this study. Four isolates were identified as non-S. aureus staphylococci and were not included. Moreover, the MRSA collection was characterized by PFGE as described earlier (Goering & Tenover, 1997). One isolate from each PFGE group was analysed by multilocus sequence typing as described previously (Enright et al., 2000).

    Identification of staphylococci other than S. aureus.

    All isolates were primarily identified by the Vitek 2 automated system (bioMérieux) and collected between 2000 and 2008. The isolates were later molecularly confirmed. The molecular identification result was used as reference method. In our study, 412 single-copy isolates of molecularly defined staphylococci other than S. aureus were included. Sixty-four isolates of 20 different species were identified using amplification and sequencing of the sodA (Poyart et al., 2001) or rpoB (Drancourt & Raoult, 2002) gene. The following species were included: Staphylococcus hominis (n=14), Staphylococcus lugdunensis (n=12), Staphylococcus saprophyticus (n=7), Staphylococcus warneri (n=5), Staphylococcus capitis (n=4), Staphylococcus haemolyticus (n=4), Staphylococcus epidermidis (n=3), Staphylococcus succinus (n=2), Staphylococcus sciuri (n=2), Staphylococcus arlettae (n=1), Staphylococcus simulans (n=1), Staphylococcus cohnii (n=1), Staphylococcus lentus (n=1), Staphylococcus pasteuri (n=1), Staphylococcus piscifermentans (n=1), Staphylococcus hyicus (n=1), Staphylococcus intermedius (n=1), Staphylococcus schleiferi (n=1), Staphylococcus xylosus (n=1) and Staphylococcus caprae (n=1). An additional 348 isolates, 277 isolates of S. saprophyticus (Martineau et al., 2001), 34 isolates of S. haemolyticus (Iwase et al., 2007) and 37 isolates of S. hominis (Hauschild & Stepanovic, 2008), were also included as strains that had been molecularly defined by species-specific PCR.

    MALDI-TOF MS.

    All strains were examined by MALDI-TOF MS using a Microflex LT instrument (Bruker Daltonics), Flexcontrol 3.0 software and the Biotyper 2.0 database (Bruker Daltonics). These instruments were used to calculate and process the analytical data according to the manufacturer's instructions. For sample preparation, one colony was suspended in 500 μl 70 % ethanol, and the suspension was centrifuged for 5 min at 13 000 g. The supernatant was discarded and 25 μl 70 % formic acid was added. Bacteria were gently mixed and 25 μl acetonitrile was added. The suspension was centrifuged for another 2 min at 13 000 g and 1 μl was transferred onto an MSP 96 polished steel target (Bruker Daltonics). The samples were covered with 1 μl matrix solution (a saturated solution of α-cyano-4-hydroxycinnamic acid in 50 % acetonitrile, 2.5 % trifluoroacetic acid). The analysed mass range of spectra was 2000–20 000 m/z. Each spectrum was obtained after 300 shots in an automatic acquisition mode. For the identification approach, a mass-to-charge range of 3000–15 000 Da was used.

    Identification was performed in duplicate and the higher score was retained. The identification score cut-off values were applied on each measurement according to the manufacturer's instructions. According to this score system, a score of ≥2 is recommended for a probable species and a score of greater than 2.3 is recommended for a secure species identification. Two MSSA isolates with an initial score slightly below 2.0 were retested.

    RESULTS AND DISCUSSION

    For the identification of S. aureus, fast agglutination tests are often used. The accuracy of most of these has been questioned over time (Personne et al., 1997), and so many of them have been increased by supplementing with more ingredients. Highly accurate molecular identification approaches have been described (Drancourt & Raoult, 2002; Martineau et al., 1998) but these are time-intensive compared to fast agglutination tests.

    MALDI-TOF MS-based identification of bacteria could be an alternative to molecular tests if the test accuracy is proven. MALDI-TOF MS-based identification has been shown to be a fast and accurate technology in the identification of a variety of bacteria (Hsieh et al., 2008; Rajakaruna et al., 2009; Sauer & Kliem, 2010; Seng et al., 2009). A MALDI-TOF MS-based fingerprint analysis using the Biotyper 2.0 database was applied in our study on 363 molecularly defined MSSA strains. A correct species diagnosis was calculated in 361 of the 363 S. aureus strains with a mean score of 2.31 (sd 0.064) according to a score cut-off value of 2.0. The minimum score was 1.986 and the maximum score was 2.498. In 2 of the 363 MALDI-TOF-based identifications, a score of below 2.0 (1.986 and 1.999) was obtained. These isolates were retested and a score of >2.0 (2.101 and 2.2) was obtained in the second measurement. The lower initial scores were interpreted as sample preparation problems, e.g. the proportion of matrix and sample in the first measurement was not optimal. No discrepant results regarding species identification were obtained between duplicates. In 144 of the 363 MSSA strains, a score of 2–2.3 was achieved. In 219 of the 363 MSSA strains, a score greater than 2.3 was achieved. The rate of correct identification to species level (relative sensitivity) of the MALDI-TOF-based method was 100 % (239 out of 239) in the MRSA strain collection with a mean score of 2.323 (sd 0.07). The minimum score was 2.054 and the maximum score was 2.461. Eighty-three of the 239 MRSA strains achieved a score of between 2.0 and 2.3, and 156 of the 239 MRSA strains achieved a score of greater than 2.3. Our data indicate that the accuracy (relative sensitivity) of the MALDI-TOF MS-based identification is equivalent to that of SA442-PCR. The epidemiological distribution of the MRSA isolates was 49 % in PFGE group 35 (ST022), 13 % in PFGE group 11 (ST225), 13 % in PFGE group 13 (ST008), 2 % in PFGE groups 1 and 7 (ST228), 4 % in PFGE group 16 (ST045) and 6 % in the other PFGE groups. The distribution of the sequence types was similar to that in previous reports.

    A total of 412 molecularly defined staphylococci other than S. aureus were included to test the accuracy (relative specificity) of the MALDI-TOF MS-based S. aureus identification (Table 1⇓). Sixty-four strains from 20 different species were identified using amplification and sequencing of the sodA or rpoB gene (Drancourt & Raoult, 2002; Poyart et al., 2001). In addition, 348 strains of S. haemolyticus, S. hominis and S. saprophyticus were molecularly characterized by species-specific PCRs. In particular, non-S. aureus strains that can be misidentified as S. aureus with regard to certain agglutination tests, such as S. hominis, S. saprophyticus, S. simulans, S. hyicus, S. schleiferi, S. xylosus and S. intermedius, were herein included. In the MALDI-TOF MS-based identification using the Biotyper 2.0 database, none of the 412 molecularly defined staphylococci other than S. aureus were identified as S. aureus. Moreover, all of the non-S. aureus staphylococci were identified correctly to species level compared to the molecularly defined reference. A similar result of high accuracy of the identification of non-S. aureus strains has been shown previously by others (Dupont et al., 2009). The mean time to result for our species-specific PCR was 2 h and 32 min, similar to the time taken for an automated identification, e.g. the GPI card (bioMérieux). In contrast to these identification approaches, the MALDI-TOF MS-based identification was much faster with a mean time to result of 22 min (n=180 identifications).

    Table 1.

    Biotyper-based identification of S. aureus and non-S. aureus staphylococci

    MALDI-TOF MS-based identification of two molecularly defined MSSA and MRSA strain collections using the Biotyper 2.0 database and the S. aureus specific SA442-PCR (Martineau et al., 2001) as a molecularly defined reference. The accuracy was 100 % for all 602 S. aureus isolates and 412 isolates of 20 different non-S. aureus staphylococcus species compared to the molecularly defined reference identification. nd, Not determined.

    The differences in the test accuracy in published studies using a molecularly defined reference (Carbonnelle et al., 2007; Rajakaruna et al., 2009) indicate that the accuracy depends on the applied databases and algorithms, more so than on the MALDI-TOF MS technology itself. The lower rate of 94.4 % correct identifications in a previous study using the Biotyper 2.0 database (Seng et al., 2009) could be explained by differences in the sample preparation and acquisition, reference method and in the study design. Similar to another study (Carbonnelle et al., 2007), the identification of S. aureus to species level was possible in all cases within the tested group of staphylococci. In the aforementioned study, a non-commercial database was used and only a limited number of peaks were picked manually. In a further study (Rajakaruna et al., 2009), in which a molecularly defined reference was also applied, the rate of correct species-level identification was lower than in our study. In particular, a binary classification to investigate the relative specificity in this study was missing. The Biotyper 2.0 database contains characterized isolates. Altogether four isolates of S. aureus (ATCC 33862, ATCC 33591, ATCC 29213 and ATCC 25923), seven isolates of S. aureus subsp. aureus (DSM 4910, DSM 11822, DSM 3463, DSM 20491, DSM 346, DSM 20652 and DMS 20232) as well as one isolate of S. aureus subsp. anaerobius (DSM 20714) are included in the database. For all S. aureus strains, irrespective of meticillin resistance, we achieved 62.3 % (375 out of 602 strains) identifications with a Biotyper 2.0 score of greater than 2.3. In addition, in 37.7 % (227 out of 602 strains) of the identifications, a Biotyper 2.0 score of between 2.0 and 2.3 was achieved.

    In conclusion, our data indicate that even a Biotyper 2.0 score cut-off value of greater than 2 is accurate and equivalent to a molecular S. aureus identification. The time to result of a MALDI-TOF MS-based identification was clearly decreased from that of molecular methods. Compared to the fast agglutination tests for S. aureus identification, the use of the MALDI-TOF MS-based approach for a routine S. aureus identification is questionable with regard to the time consumed. However, our data suggest the use of a MALDI-TOF MS-based identification in favour of a molecular identification in cases where the primarily gained S. aureus identification results are questionable.

    Acknowledgments

    We thank Gurpreet Khaira (Vancouver, Canada) for critically reading the manuscript.

    References

    1. Bernardo, K., Pakulat, N., Macht, M., Krut, O., Seifert, H., Fleer, S., Hunger, F. & Kronke, M. (2002). Identification and discrimination of Staphylococcus aureus strains using matrix-assisted laser desorption/ionization-time of flight mass spectrometry. Proteomics 2, 747–753.
    2. 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, 2156–2161.
    3. Drancourt, M. & Raoult, D. (2002). rpoB gene sequence-based identification of Staphylococcus species. J Clin Microbiol 40, 1333–1338.
    4. Du, Z., Yang, R., Guo, Z., Song, Y. & Wang, J. (2002). Identification of Staphylococcus aureus and determination of its methicillin resistance by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Anal Chem 74, 5487–5491.
    5. Dupont, C., Sivadon-Tardy, V., Bille, E., Dauphin, B., Beretti, J., Alvarez, A., Degand, N., Ferroni, A., Rottman, M. & other authors (2009). 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
    6. Edwards-Jones, V., Claydon, M. A., Evason, D. J., Walker, J., Fox, A. J. & Gordon, D. B. (2000). Rapid discrimination between methicillin-sensitive and methicillin-resistant Staphylococcus aureus by intact cell mass spectrometry. J Med Microbiol 49, 295–300.
    7. Enright, M. C., Day, N. P., Davies, C. E., Peacock, S. J. & Spratt, B. G. (2000). Multilocus sequence typing for characterization of methicillin-resistant and methicillin-susceptible clones of Staphylococcus aureus. J Clin Microbiol 38, 1008–1015.
    8. Goering, R. V. & Tenover, F. C. (1997). Epidemiological interpretation of chromosomal macro-restriction fragment patterns analyzed by pulsed-field gel electrophoresis. J Clin Microbiol 35, 2432–2433.
    9. Hauschild, T. & Stepanovic, S. (2008). Identification of Staphylococcus spp. by PCR-restriction fragment length polymorphism analysis of dnaJ gene. J Clin Microbiol 46, 3875–3879.
    10. Hsieh, S. Y., Tseng, C. L., Lee, Y. S., Kuo, A. J., Sun, C. F., Lin, Y. H. & Chen, J. K. (2008). Highly efficient classification and identification of human pathogenic bacteria by MALDI-TOF MS. Mol Cell Proteomics 7, 448–456.
    11. 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, 1346–1349.
    12. 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, 618–623.
    13. Martineau, F., Picard, F. J., Ke, D., Paradis, S., Roy, P. H., Ouellette, M. & Bergeron, M. G. (2001). Development of a PCR assay for identification of staphylococci at genus and species levels. J Clin Microbiol 39, 2541–2547.
    14. Personne, P., Bes, M., Lina, G., Vandenesch, F., Brun, Y. & Etienne, J. (1997). Comparative performances of six agglutination kits assessed by using typical and atypical strains of Staphylococcus aureus. J Clin Microbiol 35, 1138–1140.
    15. 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, 4296–4301.
    16. Rajakaruna, L., Hallas, G., Molenaar, L., Dare, D., Sutton, H., Encheva, V., Culak, R., Innes, I., Ball, G. & other authors (2009). High throughput identification of clinical isolates of Staphylococcus aureus using MALDI-TOF-MS of intact cells. Infect Genet Evol 9, 507–513.
    17. Sauer, S. & Kliem, M. (2010). Mass spectrometry tools for the classification and identification of bacteria. Nat Rev Microbiol 8, 74–82.
    18. 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, 543–551.
    19. 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, 2659–2661.
    20. Walker, J., Fox, A. J., Edwards-Jones, V. & Gordon, D. B. (2002). Intact cell mass spectrometry (ICMS) used to type methicillin-resistant Staphylococcus aureus: media effects and inter-laboratory reproducibility. J Microbiol Methods 48, 117–126.