A Rapid MALDI-TOF MS Identification Database at Genospecies Level for Clinical and Environmental Aeromonas Strains

The genus Aeromonas has undergone a number of taxonomic and nomenclature revisions over the past 20 years, and new (sub)species and biogroups are continuously described. Standard identification methods such as biochemical characterization have deficiencies and do not allow clarification of the taxonomic position. This report describes the development of a matrix-assisted laser desorption/ionisation–time of flight mass spectrometry (MALDI-TOF MS) identification database for a rapid identification of clinical and environmental Aeromonas isolates.


Introduction
Bacteria belonging to the genus Aeromonas are widely distributed in freshwater and brackish environments, and have long been recognized as etiologic agents for fish diseases [1]. They are included into the class Gammaproteobacteria, comprising Gramnegative, non-spore-forming rod-shaped bacteria, are facultative anaerobic oxidase-and catalase-positive, glucose-fermenting, resistant to the vibriostatic agent O/129, and generally motile [2].
Aeromonas play also a significant role as opportunistic pathogens for humans causing gastroenteritis, septicemia, pneumonia, meningitis, and wound infections in immunocompetent as well as in compromised patients. A. hydrophila, A. caviae and A. veronii (biovar sobria and biovar veronii), are clinically the most significant species [3].
Several phylogenetic studies on Aeromonas allowed the elevation of the genus name to the rank of family [2,6,7,8]. Nevertheless the taxonomy of this genus is rather complex and has been submitted to ongoing changes due to newly described species [9,10,11,12,13,14] and rearrangements of existing taxa [15,16,17,18,19,20]. One major problem in Aeromonas identification relies on the fact that some species are phenotypically very similar (e.g. A. caviae and A. media, A. veronii and A. sobria). Several molecular methods have been therefore applied as an alternative to the laborious DNA-DNA hybridization technique for resolving the Aeromonas taxonomy and even though the sequence analysis of ribosomal RNA genes allowed for the discrimination of the genospecies [6,21,22], other more discriminating housekeeping genes such as gyrB and rpoD are now increasingly used [8,23,24,25,26]. Nevertheless, sequencing and phylogenetic methods are costly, time consuming and therefore not appropriate for a rapid species identification in the diagnostic laboratory. A valid alternative to conventional methods of bacterial identification and classification, based on the characterization of biomarker molecules, but definitely more rapid and reliable is the mass spectrometry technique [27]; MALDI-TOF MS (matrix assisted laser desorption ionization mass spectrometry -time of flight) combined with a reliable database is a powerful method for the identification and comparison of microbial isolates based on protein fingerprints analysis of whole cells [28]. MALDI-TOF MS applications in microbiology are important for proteomic and natural product analyses [29]. This technique can be used to detect non-volatile and thermally unstable molecules from a few to several hundred kDa, the most applicable range used for the analysis is 2-20 kDa. The identification of microorganisms by MALDI-TOF MS is based on the detection of mass signals from biomarkers that are specific at genus, species or sub-group level.
All mass spectra were generated in positive linear mode by scanning the sample spot with the laser beam, and after signal acquisition, the raw mass spectra are processed automatically by smoothing, baseline correction and peak recognition [30]. The essential information used for microbial identification is contained in a peak list containing m/z values and intensities. This list is analysed by comparison to the database SARAMIS TM (Spectral Archive And Microbial Identification System), in which the identification at the species level is based on a percentage of confidence referred to reference spectra (SuperSpectra TM ) that contain family, genus and species specific m/z biomarkers, as described in the SARAMIS TM user manual. For the generation of one SuperSpectra TM some representatives isolates of one species from different locations (hospitals, reference centers and strain culture collections) are needed. Beside the FingerprintSpectra every isolate will be determined by accredited and published microorganism identification procedures. The SuperSpectra TM are generated based on measurements of well known microorganisms and contain sets of genus, species and strain biomarkers which are characteristic for the respective group of microorganisms. Superspecta TM are computed from typical strains covering more than 90% of the intraspecific diversity in most species.
Accuracy of the identification strongly relies upon the robustness of the database and the choice of reference isolates. This is especially important when considering genera comprising species of clinical and environmental origin presenting a high genetic diversity.
There are excellent precedents for the application of MALDI-TOF MS for taxonomic studies [31,32,33,34], as well as for routine diagnostic [35].
Previous studies proved the applicability of this technique for the identification of the Aeromonas species [36,37,38]. The major aim of this study was to establish a rapid and reliable species identification tool for the genus Aeromonas using the SARAMIS TM identification system based on a relatively high number of phylogenetically well characterized isolates of clinical and environmental origin.

Bacterial Strains
92 morphologically and genetically well characterized strains (see supporting information Table S1) belonging to all known genospecies of the genus Aeromonas were used to create the m/z reference library system using the SARAMIS TM software. All strains were phylogenetically typed and assigned to the respective genetic species using the housekeeping gene gyrB. The obtained sequences were deposited in GenBank and accession numbers are listed in Table S1. The mass fingerprinting identification database produced was then evaluated on 741 clinical and environmental isolates. All strains were grown on Blood Agar at 30uC for 24 hours previous to the protein fingerprinting mass spectrometry analysis.

DNA Extraction
Genomic DNA was extracted from colonies grown on blood agar according to Demarta et al. [39], and resuspended in TrisEDTA buffer (10 mM Tris/HCl, 1 mM EDTA, pH 8.0).

PCR Amplification and Sequencing
The sets of primers used for amplification and sequencing of the gyrB gene have been reported elsewhere [40,41].

Phylogenetic Analyses
Nucleotide sequences of gyrB gene (fragment of 1100 bp) was aligned and phylogenetically analysed using MEGA version 3.1 [42].
Phylogenetic tree was constructed using the Neighbour-Joining method with genetic distances computed by employing Kimura's 2-parameter method [41].

MALDI-TOF MS
Strains were transferred from the colony directly on a 48position stainless steel FlexiMass TM target plate (Shimadzu Biotech, Kyoto, Japan) using a plastic loop. The transferred colony material was then overlaid with 0.5 ml of Matrix (DHB 75%) solution containing 75 mg/ml 2, 5-dihydroxybenzoic acid in acetonitrile/ethanol/water (1:1:1) supplemented with 3% trifluoroacetic acid. All mass spectra were acquired using an AXIMA Confidence TM (Shimadzu Biotech, Kyoto, Japan) mass spectrometer, equipped with a nitrogen laser (pulse width: 3 ns) operated in positive linear mode. The measured mass range of spectra was 2000-20,000 Da. A minimum of 20 laser shots per sample was used to generate each ion spectrum. For each bacterial sample, 50 protein mass fingerprints were averaged and processed.
All spectra were processed by the MALDI-TOF MS Launchpad 2.8 software (Shimadzu Biotech, Kyoto, Japan).

Data Analysis
A database identification system was established analyzing 92 morphologically and genetically well characterized Aeromonas strains belonging to all known species of the genus. The resulting peak lists of these samples were exported to the SARAMIS TM software package (bioMérieux, France) and submitted to singlelinkage cluster analysis to produce taxonomic trees. These trees were compared to a gyrB phylogenetic tree (Neighbour-Joining). Specific biomarkers containing sets of genus, species and strain characteristic masses were used for the creation of species-specific SuperSpectra TM recognizing the most frequently encountered species. 11 different SuperSpectra TM were created that allow identifications of: A. hydrophila, A. caviae, A. media, A. tecta, A. popoffi, A. eucrenophila, A. encheleia, A. bestiarum, A. salmonicida, A. sobria and A. veronii).

Results and Discussion
The protein mass fingerprint analysis emerging from the MALDI-TOF MS data of 92 genetically well characterized Aeromonas strains provided a good separation at genospecies (Fig. 1) level comparable with the phylogenetic tree obtained by gyrB gene sequencing.
In fact both trees clustered the species A. veronii (A. veronii biovar sobria, A. veronii biovar sobria), A. culicicola, and A. allosaccarophila together, confirming the hypothesis that this group in fact represents only one genospecies [18].
Interesting the m/z profiles analysis allowed to separate the two biovars veronii and sobria, furthermore the profile of the strain ATCC 51106 A. veronii biovar sobria was more closely related to that of A. allosaccarophila ATCC 51208 than to that of A. veronii biovar veronii, confirming the results obtained with the gyrB sequences.
Moreover MALDI-TOF MS analysis categorized in a single cluster A. encheleia and the unnamed Aeromonas sp. HG11 [23] and allowed the segregation in the different genospecies of the A. salmonicida/A. bestiarum/A. popoffii group.
A. salmonicida and A. bestiarum are difficult to separate on the basis of 16S rRNA (differ in only 2 nucleotide positions) [2] but they could be separated using gyrB as well as other housekeeping genes such as rpoB or rpoD.
At the subspecies level, A. salmonicida formed a very uniform group, with respective intraspecies substitution rates of 1.3 and 0.8% for gyrB and rpoB, rendering very difficult to classify strains at the subspecies level [41]. MALDI-TOF MS seemed to allow a better differentiation of the strains in study. The type strains of each subspecies were well differentiated and formed a defined group in the MALDI-TOF MS dendrogram (Fig. 1).
A branch in the MALDI-TOF MS dendrogram groups in one single cluster strains assigned to the species A. aquariorum and A. hydrophila subsp. dhakensis (Fig.1). Data based on phylogenetic analysis by sequencing gyrB, rpoD and 16S rRNA [43], strongly suggested that strains of A. hydrophila subsp. dhakensis belongs in fact to the species A. aquariorum, confirming the results obtained with MALDI-TOF MS (Fig. 1).
Due to the reliable identification at species level, it was possible to create 11 different SuperSpectra TM for A. hydrophila, A. caviae,     Table 2).
These results demonstrate that the mass spectral data of the strains contained sufficient protein information to distinguish between genera, species, and strains ( Table 2).
Another mass spectrometry study of intact-cell with Aeromonas strains [37] also confirmed that the signals generated from the analysis of the protein masses could be used as specific biomarkers for the differentiation below the species level. For the the majority of the species analysed the identification was successful.
With A. tecta and A. sobria we obtained a correct identification for all the strains, whereas for A. eucrenophila, A. salmonicida, and A. hydrophila only 1 strains for the first two and 2 strains for the last species could not be identified.
Identification of A. popoffii with the created SuperSpectra TM was possible only in 46% of the cases. These failure could be due to insufficient coverage of the specific SuperSpectra TM or lack of performance of the last.
The approach presented in this paper uses the technique MALDI-TOF MS to develop a rapid, sensitive and specific method to detect isolates of the genus Aeromonas.
Our work highlighted the importance of testing well characterized strains of different origins for producing high quality MALDI-TOF MS databases as rapid identification tools. In conclusion, we can affirm that MALDI-TOF MS is a rapid and relatively inexpensive method for the identification of Aeromonas species and constitutes a valid alternative to conventional methods of identification and classification.