ORIGINAL ARTICLE

10.1111/j.1469-0691.2011.03696.x

Rapid and accurate identication of genomic species from the Acinetobacter baumannii (Ab) group by MALDI-TOF MS
P. Espinal1, H. Seifert2, L. Dijkshoorn3, J. Vila1 and I. Roca1 1) Department of Clinical Microbiology, Hospital Clinic of Barcelona, School of Medicine, University of Barcelona, IDIBAPS, and Barcelona Centre for International Health Research (CRESIB), Barcelona, Spain, 2) Institute for Medical Microbiology, Immunology and Hygiene, University of Cologne, Cologne, Germany and 3) Department of Infectious Diseases, Leiden University Medical Center, Leiden, the Netherlands

Abstract
The closely related members of the Acinetobacter baumannii (Ab) group (A. baumannii, A. pittii and A. nosocomialis) are difcult to identify with phenotypic tests in diagnostic laboratories. Genotypic identication methods require special skills and most do not provide rapid results. The aim of this study was to investigate the ability of MALDI-TOF MS to identify members of the Ab group. Sixty epidemiologically unrelated Acinetobacter spp. isolates were investigated by MALDI-TOF MS: 18 A. baumannii, 17 A. pittii, 18 A. nosocomialis and seven additional isolates representing other Acinetobacter spp. All strains were veried by ARDRA, rRNA intergenic spacer (ITS), recA sequencing and blaOXA-51. MALDI-TOF MS correctly identied all the genomic strains but erroneously identied A. nosocomialis as A. baumannii because there was no reference strain within the Bruker database. Peak analysis of individual spectra from representative strains of each member of A. baumannii, A. pittii and A. nosocomialis suggested enough differences between their protein signatures to allow accurate identication using MALDI-TOF MS. Inclusion of specic signature proles for A. nosocomialis within the Bruker database allowed the correct identication of this genomic species. MALDI-TOF MS spectra can be used as a fast, simple and reliable method to identify members of the Ab group. The rapid and accurate identication of clinically signicant Acinetobacter strains will improve insight into their epidemiology and allow for targeted therapeutic and infection control measures against clinically important strains. Keywords: Acinetobacter, ARDRA, ITS, MALDI-TOF MS, recA, sequence typing Original Submission: 8 July 2011; Revised Submission: 27 September 2011; Accepted: 8 October 2011 Editor: G. Greub Clin Microbiol Infect
Corresponding author: J. Vila, Department of Clinical Microbiology, Hospital Clnic, Villaroel, 170, Barcelona, Spain E-mail: [email protected]

Introduction
The Acinetobacter genus comprises Gram-negative non-fermenting coccobacilli with 25 validly named species and nine genomic species dened by genomic DNA-DNA hybridization [1]. Among these, Acinetobacter baumannii constitutes the most important species causing nosocomial infections, particularly in intensive care units (ICU) [2], although A. pittii and A. nosocomialis (formerly Acinetobacter genomic species 3 and gen. sp. 13TU, respectively [3]) are emerging as important pathogens and have been involved in a number of outbreaks in ICUs [1]. A. baumannii, A. pittii and A. nosocomialis

as well as the environmental species A. calcoaceticus are highly similar from a phenotypic point of view as well as by DNA-DNA hybridization, which has led to their inclusion in the A. calcoaceticus-A. baumannii (Acb) complex [4]. In fact, the three clinically important members of this group, also known as the A. baumannii (Ab) group [2], are so much alike that they cannot be differentiated by currently available identication systems and A. pittii and A. nosocomialis are often erroneously identied as A. baumannii by routine commercial systems such as API and VITEK (bioMerieux, Marcy lEtoile, France) or PHOENIX (Becton Dickinson, Franklin Lakes, NJ, USA.) [5,6]. Genotypic methodologies to distinguish individual genomic species have been developed, including amplied 16S ribosomal DNA restriction analysis (ARDRA) [7], tRNA spacer ngerprinting [8] and selective amplication of restriction fragments (AFLP) [9]. Specic gene sequences can also be used, including the intergenic spacer (ITS) region between

2011 The Authors Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases

Clinical Microbiology and Infection

CMI

the 16S23S rRNA genes [10], recA [11], rpoB [12] and gyrB [13], although not all of these methods successfully discriminate among members of the Ab group. In addition, most of these methodologies are laborious, time-consuming, require special skills and are unsuitable for use in routine clinical identication. Failure to correctly identify clinical isolates below the Ab group level may be misleading because infections caused by the different species within this group may differ in their clinical outcome [14]. Organisms belonging to these three species display different characteristics regarding colonization of human skin, antimicrobial susceptibility and mortality rates [14,15]. Therefore, a more rapid and accurate identication strategy is needed. Matrix-assisted laser desorption ionization-time of ight (MALDI-TOF) mass spectrometry (MS) has been adapted for the identication of different microorganisms at the genus, species or subspecies level and has proven to be a very user-friendly technique requiring tiny amounts of sample yet providing accurate results within minutes. MALDI-TOF MS generates protein ngerprint signatures that can rapidly be compared with those in a database of reference spectra [16]. Although MALDI-TOF MS has been applied to the identication of non-fermenting bacteria, including some Acinetobacter species [1719], it has never been tested with different members of the Ab group. The present study was undertaken to evaluate the use of MALDI-TOF MS in the identication of members of the Ab group and to compare its discriminatory power with that of ARDRA, recA and ITS sequencing.

analysed by gel electrophoresis or sent for sequencing and compared among all sequences and reference strains.
ARDRA

The complete 16S rRNA gene was amplied as described in [7]. Aliquots of the PCR product were independently digested with AluI, CfoI, MboI, MspI and RsaI (Promega Bio tech Iberica, Madrid, Spain). Restriction patterns were analysed by gel electrophoresis in 2.5% (w/v) agarose and patterns were compared with the library proles described in [7].
MALDI-TOF MS

MALDI-TOF was conducted on a Microex LT (Bruker Daltonics GmbH, Leipzig, Germany) benchtop instrument operated in linear positive mode under control of the FlexControl 3.0 software (Bruker Daltonics) at a laser frequency of 20 Hz within a range mass from 2000 to 20000 Da. Mass spectra were processed using FlexAnalysis 3.0 software (Bruker Daltonics). External calibration was performed using the Bruker Daltonics Bacterial Test standard according to the manufacturers instructions. Bacterial extracts. Pure cultures were grown on Columbia sheep blood agar (Becton Dickinson) at 37C for 24 h. One full 1 lL sterile loop of bacterial sample was suspended in 300 lL of sterile water and mixed with 900 lL of absolute ethanol. Samples were centrifuged at 12 000 g for 2 min and the supernatant was discarded. The pellet was resuspended in 50 lL of 70% formic acid (Sigma chemical Co., St Louis, MO, USA) and 50 lL of acetonitrile (Sigma) and centrifuged at 12 000 g for 2 min. The supernatant was collected and stored at )20C. Generation of a local database. One microlitre of each bacterial extract from three strains each of A. baumannii (RUH 875, RUH 134, RUH 5875), A. pittii (52, 60, 69) and A. nosocomialis (95, 192, 212), was spotted eight times onto a ground steel target and air-dried. Each sample was overlaid with 1 lL of a saturated matrix solution of a-cyano-4hydroxy-cinnamic acid (Bruker Daltonics) in 50% acetonitrile (Sigma), 2.5% triuoroacetic acid (Sigma) and air-dried. Each spot was measured three times. Every measurement was the sum spectrum accumulated from 250 laser shots (5 50 laser shots on different locations according to a predened lattice raster). The resulting 24 spectra were carefully analysed using the FlexAnalysis software to yield a minimum of 20 accurate spectra that were uploaded onto the MALDI BioTyper 2.0 software (Bruker Daltonics) to create a single

Materials and Methods

Bacterial isolates

Sixty epidemiologically unrelated Acinetobacter spp. isolates were included in the study: a set of known clinical isolates from our collection, 16 A. pittii, 16 A. nosocomialis and 13 A. baumannii; and a set of reference strains, RUH 204 (A. junii), RUH 509 (A. pittii), RUH 503 (A. nosocomialis), RUH 44 (A. haemolyticus), RUH 45 (A. lwofi), RUH 3517 (A. radioresistens), RUH 584 (A. calcoaceticus), RUH 875 (A. baumannii, European clone I), RUH 134 (A. baumannii, European clone II), RUH 5875 (A. baumannii, European clone III), 17BJ-209 (Acinetobacter gen. sp. 17), A. haemolyticus-JV, A. baumannii AYE strain, A. baumannii ATCC 19606 and A. nosocomialis ATCC 17903.
Amplication of the ITS region, recA and blaOXA-51

Amplication of the ITS region, recA and blaOXA-51 was performed according to Chang et al. [10], Nowak et al. [20] and Ruz et al. [21], respectively. Puried PCR products were

2011 The Authors Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases, CMI

CMI

Espinal et al.

MALDI-TOF MS identication of Acinetobacter spp.

mean spectrum (MSP) for each strain with the BioTyper MSP creation standard method. Identication of sample bacterial extracts. MALDI-TOF MS analysis was performed from bacterial extracts or directly from grown colonies. Two bacterial extracts from independent cultures were generated for each tested strain; 1 lL of each extract was spotted onto a ground steel target and overlaid with 1 lL of matrix. To perform direct colony analysis a small fraction of a single colony was spread over two spots and overlaid with 1 lL of matrix. This procedure was performed twice. Each sample spot was measured as before and analysed with the MALDI BioTyper against either the default Bruker database or the association of the Bruker database and our own reference spectra. Accuracy of the identication was determined by a logarithmic score value resulting from the alignment of peaks to the best matching reference spectrum. According to the manufacturer: a log score >2.3 indicated highly probable species identication; a score value between 2.0 and 2.3 indicated secure genus identication and probable species identication; a score value between 1.7 and 2.0 indicated probable genus identication; and a score value <1.7 was considered as a non-reliable identication (Table 1). Optimal identication was granted when two independent spectra were within the same log score range. Data analysis. The MALDI-TOF MS-generated dendogram was constructed using the correlation distance measure with the average linkage algorithm settings of the BioTyper 2.0 software. Sequence analysis of ITS and recA was performed by similarity searches against nucleotide databases (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Identication to the species level was conrmed when the matching sequence displayed at least 96% identity within a 99% coverage to the query sequence. Phylogenetic analysis was performed with MEGA version 5 [22]. MUSCLE was used for sequence alignment [23] and phylogenetic trees were constructed using the neighbour-joining method with genetic distances computed by Kimuras two-parameter model [24] with a bootstrap value of 1000 replicates.

Results and Discussion

Sixty previously characterized Acinetobacter isolates including 18 A. nosocomialis strains, 17 A. pittii strains, 18 A. baumannii isolates as well as seven reference strains representing various other Acinetobacter spp. were analysed by genotypic methods and MALDI-TOF MS. Bacterial extracts were analysed by MALDI-TOF MS and compared against the default Bruker database. Accurate identication at the species level was achieved for all reference strains but one, all A. baumannii strains and all but one of the A. pittii isolates. However, the MALDI BioTyper software erroneously identied the 18 A. nosocomialis strains as A. baumannii with MALDI scores <2, which was already expected due to the lack of A. nosocomialis-specic signatures within the Bruker database. Peak analysis of individual spectra from representative strains of A. baumannii, A. pittii and A. nosocomialis, however, suggested enough differences between their protein signatures to allow accurate identication using MALDI-TOF MS (Fig. 1). Hence, three strains each of A. baumannii, A. pittii and A. nosocomialis were chosen to create a set of reference spectra to complement the Bruker database. Cluster analysis showed that the protein signatures from A. nosocomialis form a separate cluster closely related to that of A. baumannii (Fig. 2). Spectra from all 60 isolates were re-analysed against a local database that incorporated the reference signatures for A. nosocomialis and this time MALDI-TOF MS analysis provided unambiguous identication for all A. nosocomialis strains. Overall, 98.3% of all isolates were identied at the species level: 86.7% (52/60) showed a log score >2.3; 11.6% (7/60) were identied with a log score between 2.0 and 2.3; and only 1.7% (1/60) were identied at the probable genus level with a log score between 1.7 and 2.0. None of the isolates reported a log score <1.7 (Table 1). All A. baumannii and A. nosocomialis strains produced log scores >2.3 while four A. pittii isolates were identied with log scores between 2.0 and 2.3. Hence, only 76.5% of the A. pittii strains were identied with log scores >2.3, a fact that might reect the need to include more diverse representative signatures of A. pittii strains within the Bruker database.

TABLE 1. MALDI-TOF MS identication of Acinetobacter spp. relative to log score values

Log score 2.33.0 2.02.3 1.72.0 <1.7 Total A. baumannii (%) 18 (100) 0 0 0 18 A. pittii (%) 13 (76.5) 4 (23.5) 0 0 17 A. nosocomialis (%) 18 (100) 0 0 0 18 Other Acinetobacter species (%) 3 (42.9) 3 (42.9) 1 (14.2) 0 7 Total (%) 52 (86.7) 7 (11.6) 1 (1.7) 0 60/60

2011 The Authors Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases, CMI

Clinical Microbiology and Infection

CMI

FIG. 1. Comparison of representative A. baumannii, A. pittii and A. nosocomialis peak proles (m/z) generated by MALDI-TOF MS. a.u., arbitrary intensity given by the software.

FIG. 2. Dendogram derived from the MALDITOF MS-specic signatures for all Acinetobacter spp. included within the Bruker database plus specic signatures for A. baumannii, A. pittii and A. nosocomialis created in this study. Distance values are relative and normalized to a maximal value of 1000.

All reference Acinetobacter spp. included in this study were correctly identied with log scores ranging between 2.2 and 2.5, except for strain Acinetobacter gen. sp. 17BJ-209, which was incorrectly identied as A. junii with a log score of 1.7. MALDI-TOF MS analysis of all isolates directly from grown colonies instead of processed bacterial extracts yielded very similar results provided that freshly grown cultures (<24 h) were used and that the amount of colony spotted onto the agar plates was optimized (data not shown). To assess the accuracy of species identication by MALDITOF MS we compared the above-mentioned results with

those obtained by molecular identication methods such as ARDRA and recA/ITS sequencing. The presence of blaOXA-51 was also analysed because it has been suggested to be highly specic for A. baumannii [25]. ARDRA was used as our reference standard because it is likely to be the most reliable method for the identication of members of the Ab group. In general, MALDI-TOF MS correlated well with all three methods. ITS sequencing correctly identied 98.3% (59/60) of the strains, which is in good agreement with that reported by Chang et al. [10], and recA analysis showed an identication rate of 100% at the species level.

2011 The Authors Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases, CMI

CMI

Espinal et al.

MALDI-TOF MS identication of Acinetobacter spp.

FIG. 3. Dendograms derived from the ITS (a) and recA (b) sequences. GeneBank accession numbers for all reference ITS and recA sequences are included within the corresponding labelling of each taxon. The scale bar indicates a genetic distance of 0.01 and the numbers shown next to each node represent the percentage bootstrap value of 1000 replicates.

2011 The Authors Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases, CMI

Clinical Microbiology and Infection

CMI

PCR ngerprinting of ITS and particularly recA provide an accurate identication of strains belonging to the Ab group, but both methods are laborious and time-consuming. Only two discrepancies were found among the four methods. (i) Strain Acinetobacter gen. sp. 17BJ-209 [26] was identied as Acinetobacter gen. sp. 17 by ARDRA, ITS and recA sequencing, but as A. junii by MALDI-TOF (albeit with a log score of 1.7). The protein signature of Acinetobacter gen. sp. 17, however, is missing in the Bruker database, which would account for the failure of MALDI-TOF MS to correctly identify this isolate. It is worth mentioning that ARDRA analysis of A. junii and Acinetobacter gen. sp. 17 with the initial ve enzymes generates identical patterns of bands and, therefore, additional restriction with both BfaI and BsmAI is needed for accurate identication, which indicates the closeness of these strains and may explain misidentication by MALDI-TOF MS. (ii) Strain A. pittii-91, presumably representing A. pittii, was identied as Acinetobacter gen. sp. 14BJ by ARDRA and recA sequencing but as A. pittii by both MALDI-TOF (with a log score of 2.0) and ITS sequencing. Of note, the protein signature for Acinetobacter gen. sp. 14BJ is also missing in the Bruker database. Strain A. pittii-91 was shown to be haemolytic when grown on blood agar plates, a trait shared among genomic species 1317 from Bouvet and Jeanjean (BJ) [27] but not by A. pittii, thus conrming the reliability of the reference standard. Isolates belonging to Acinetobacter gen. sp. 13BJ to 15BJ and Acinetobacter gen. sp. 16 are rarely found in clinical samples, providing a very limited number of available reference strains and, therefore, these groups are poorly dened with methodologies other than DNA-DNA hybridization [8]. The recA and ITS sequences from all isolates as well as from a few reference strains retrieved from GeneBank were used to construct representative dendograms in order to analyse their correct clustering (Fig. 3). recA and ITS-generated dendograms provided four dened clusters, one for each genomic species included within the Acb complex. A. calcoaceticus and A. pittii were also grouped in a larger monophyletic group, suggesting a close genetic relatedness, while A. baumannii and A. nosocomialis are apparently more closely related to each other than to the other members of the Acb complex, in agreement with the clustering of genomic species by MALDI-TOF MS-specic signatures and DNADNA hybridization [28]. PCR detection of the blaOXA-51 gene only generated positive amplication in those strains correctly identied as A. baumannii, thus reinforcing the screening value of this method to rapidly differentiate A. baumannii from other Acinetobacter spp. [25] despite a few exceptions described in the literature [29].

The correct identication of the different species within the Ab group is of great interest because infections caused by these organisms have a different impact on clinical outcome and, hence, different treatment and management approaches are needed [14]. Adjustment of the default MALDI-TOF MS database allowed the identication of all members of the Ab group as well as other Acinetobacter spp. with similar accuracy compared with ARDRA, recA and ITS sequencing and highlighted the strong dependency of MALDI-TOF MS, ITS and recA sequencing on a validated reference database that dramatically inuences the accuracy of identication. When compared with conventional methods in the diagnostic routine laboratories for Acinetobacter spp., MALDITOF MS becomes a robust method capable of identifying over 98% of all isolates to the species level in <5 min per sample, at a lower overall cost and with similar or better accuracy. Further database renements will improve the rapid identication of additional species and might even allow the characterization of resistance and virulence determinants when specic protein peaks are produced [30], making MALDI-TOF MS the ideal technique for routine clinical microbiology testing.

Acknowledgements
The authors would like to thank Thomas Maier and Markus Kostrzewaba, from Bruker Daltonics, for excellence technical assistance.

Transparency Declaration
This study was supported by the Spanish Ministry of Health (FIS 08/00195), by grant 2009SGR1256 from the Generalitat de Catalunya and by Ministerio de Sanidad y Consumo, Instituto de Salud Carlos III -FEDER, Spanish Network for the Research in Infectious Diseases (REIPI RD06/0008). This work has also been supported by funding from the European Community (TROCAR contract HEALTH-F3-2008-223031 and AntiPathoGN contract HEALTH-F3-2008-223101). The authors declare that they have no conicts of interest.

References
1. Espinal P, Roca I, Vila J. Clinical impact and molecular basis of antimicrobial resistance in non-baumannii Acinetobacter. Future Microbiol 2011; 6: 495511. 2. Peleg AY, Seifert H, Paterson DL. Acinetobacter baumannii: emergence of a successful pathogen. Clin Microbiol Rev 2008; 21: 538582.

2011 The Authors Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases, CMI

CMI

Espinal et al.

MALDI-TOF MS identication of Acinetobacter spp.

3. Nemec A, Krizova L, Maixnerova M et al. Genotypic and phenotypic characterization of the Acinetobacter calcoaceticus-Acinetobacter baumannii complex with the proposal of Acinetobacter pittii sp. nov. (formerly Acinetobacter genomic species 3) and Acinetobacter nosocomialis sp. nov. (formerly Acinetobacter genomic species 13TU). Res Microbiol 2011; 162: 393404. 4. Dijkshoorn L, Nemec A, Seifert H. An increasing threat in hospitals: multidrug-resistant Acinetobacter baumannii. Nat Rev Microbiol 2007; 5: 939951. 5. Bernards AT, van der Toorn J, van Boven CP, Dijkshoorn L. Evaluation of the ability of a commercial system to identify Acinetobacter genomic species. Eur J Clin Microbiol Infect Dis 1996; 15: 303308. 6. Ferreira L, Vega S, Sanchez-Juanes F et al. Identifying bacteria using a matrix-assisted laser desorption ionization time-of-ight (MALDITOF) mass spectrometer. Comparison with routine methods used in clinical microbiology laboratories. Enferm Infecc Microbiol Clin 2010; 28: 492497. 7. Dijkshoorn L, Van Harsselaar B, Tjernberg I, Bouvet PJ, Vaneechoutte M. Evaluation of amplied ribosomal DNA restriction analysis for identication of Acinetobacter genomic species. Syst Appl Microbiol 1998; 21: 3339. 8. Ehrenstein B, Bernards AT, Dijkshoorn L et al. Acinetobacter species identication by using tRNA spacer ngerprinting. J Clin Microbiol 1996; 34: 24142420. 9. Janssen P, Maquelin K, Coopman R et al. Discrimination of Acinetobacter genomic species by AFLP ngerprinting. Int J Syst Bacteriol 1997; 47: 11791187. 10. Chang HC, Wei YF, Dijkshoorn L et al. Species-level identication of isolates of the Acinetobacter calcoaceticus-Acinetobacter baumannii complex by sequence analysis of the 16S-23S rRNA gene spacer region. J Clin Microbiol 2005; 43: 16321639. 11. Krawczyk B, Lewandowski K, Kur J. Comparative studies of the Acinetobacter genus and the species identication method based on the recA sequences. Mol Cell Probes 2002; 16: 111. 12. Gundi VA, Dijkshoorn L, Burignat S, Raoult D, La Scola B. Validation of partial rpoB gene sequence analysis for the identication of clinically important and emerging Acinetobacter species. Microbiology 2009; 155: 23332341. 13. Higgins PG, Lehmann M, Wisplinghoff H, Seifert H. gyrB multiplex PCR to differentiate between Acinetobacter calcoaceticus and Acinetobacter genomic species 3. J Clin Microbiol 2010; 48: 45924594. 14. Chuang YC, Sheng WH, Li SY et al. Inuence of genospecies of Acinetobacter baumannii complex on clinical outcomes of patients with acinetobacter bacteremia. Clin Infect Dis 2011; 52: 352360. 15. Lim YM, Shin KS, Kim J. Distinct antimicrobial resistance patterns and antimicrobial resistance-harboring genes according to genomic species of Acinetobacter isolates. J Clin Microbiol 2007; 45: 902905.

16. Sogawa K, Watanabe M, Sato K et al. Use of the MALDI BioTyper system with MALDI-TOF mass spectrometry for rapid identication of microorganisms. Anal Bioanal Chem 2011; 400: 19051911. 17. Jacquier H, Carbonnelle E, Corvec S et al. Revisited distribution of nonfermenting Gram-negative bacilli clinical isolates. Eur J Clin Microbiol Infect Dis 2011; [Epub ahead of print]. doi: 10.1007/ s10096-011-1263-5. 18. Nemec A, Musilek M, Sedo O et al. Acinetobacter bereziniae sp. nov. and Acinetobacter guillouiae sp. nov., to accommodate Acinetobacter genomic species 10 and 11, respectively. Int J Syst Evol Microbiol 2010; 60: 896903. 19. Sedo O, Vorac A, Zdrahal Z. Optimization of mass spectral features in MALDI-TOF MS proling of Acinetobacter species. Syst Appl Microbiol 2011; 34: 3034. 20. Nowak A, Kur J. Genomic species typing of acinetobacters by polymerase chain reaction amplication of the recA gene. FEMS Microbiol Lett 1995; 130: 327332. 21. Ruz M, Mart S, Fernandez-Cuenca F, Pascual A, Vila J. High prevalence of carbapenem-hydrolysing oxacillinases in epidemiologically related and unrelated Acinetobacter baumannii clinical isolates in Spain. Clin Microbiol Infect 2007; 13: 11921198. 22. Tamura K, Peterson D, Peterson N et al. MEGA5: Molecular Evolutionary Genetics Analysis using Maximum Likelihood, Evolutionary Distance, and Maximum Parsimony Methods. Mol Biol Evol 2011; 28: 27312739. doi: 10.1093/molbev/msr121. 23. Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 2004; 32: 17921797. 24. Kimura M. A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J Mol Evol 1980; 16: 111120. 25. Turton JF, Woodford N, Glover J et al. Identication of Acinetobacter baumannii by detection of the blaOXA-51-like carbapenemase gene intrinsic to this species. J Clin Microbiol 2006; 44: 29742976. 26. Rodrguez-Bano J, Mart S, Ribera A et al. Nosocomial bacteremia due to an as yet unclassied Acinetobacter genomic species 17-like strain. J Clin Microbiol 2006; 44: 15871589. 27. Bouvet PJ, Jeanjean S. Delineation of new proteolytic genomic species in the genus Acinetobacter. Res Microbiol 1989; 140: 291299. 28. Tjernberg I, Ursing J. Clinical strains of Acinetobacter classied by DNA-DNA hybridization. APMIS 1989; 97: 595605. 29. Koh TH, Tan TT, Khoo CT et al. Acinetobacter calcoaceticus-Acinetobacter baumannii complex species in clinical specimens in Singapore. Epidemiol Infect 2011; [Epub ahead of print]. doi: 10.1017/ s0950268811001129. 30. Bittar F, Ouchenane Z, Smati F, Raoult D, Rolain JM. MALDI-TOFMS for rapid detection of staphylococcal Panton-Valentine leukocidin. Int J Antimicrob Agents 2009; 34: 467470.

2011 The Authors Clinical Microbiology and Infection 2011 European Society of Clinical Microbiology and Infectious Diseases, CMI