Rapid Molecular Characterization of Acinetobacter baumannii Clones with rep-PCR and Evaluation of Carbapenemase Genes by New Multiplex PCR in Hospital District of Helsinki and Uusimaa

Multidrug-resistant Acinetobacter baumannii (MDRAB) is an increasing problem worldwide. Prevalence of carbapenem resistance in Acinetobacter spp. due to acquired carbapenemase genes is not known in Finland. The purpose of this study was to examine prevalence and clonal spread of multiresistant A. baumannii group species, and their carbapenemase genes. A total of 55 Acinetobacter isolates were evaluated with repetitive PCR (DiversiLab) to analyse clonality of isolates, in conjunction with antimicrobial susceptibility profile for ampicillin/sulbactam, colistin, imipenem, meropenem, rifampicin and tigecycline. In addition, a new real-time PCR assay, detecting most clinically important carbapenemase genes just in two multiplex reactions, was developed. The assay detects genes for KPC, VIM, IMP, GES-1/-10, OXA-48, NDM, GIM-1, SPM-1, IMI/NMC-A, SME, CMY-10, SFC-1, SIM-1, OXA-23-like, OXA-24/40-like, OXA-58 and ISAbaI-OXA-51-like junction, and allows confident detection of isolates harbouring acquired carbapenemase genes. There was a time-dependent, clonal spread of multiresistant A. baumannii strongly correlating with carbapenamase gene profile, at least in this geographically restricted study material. The new carbapenemase screening assay was able to detect all the genes correctly suggesting it might be suitable for epidemiologic screening purposes in clinical laboratories.

To determine genetic and epidemiological relatedness, genomic fingerprinting of clinical isolates is required. One of the most effective method is the repetitive extragenic palindromic sequencebased polymerase chain reaction (rep-PCR), which is commercially available known as the DiversiLab microbial typing system (bioMérieux, Marcy L'Etoile, France) [8]. This system has been proven useful in the typing of A. baumannii and has demonstrated good discriminatory ability, comparable with pulsed-field gel electroproresis (PFGE) and multilocus sequence typing (MLST) [9,10]. Recently this rep-PCR typing system, DiversiLab, has identified eight carbapenem-resistant A. baumannii clonal lineages (WW1 to WW8) that are distributed worldwide [4]. DiversiLab fingerprints between laboratories were recently tested and clustering was found to be conserved [11].
The carbapenem resistance has recently attracted new interest as a subset among tens of gene families has spread to Enterobacteriaceae [12][13][14], despite a much longer history among Pseudomonas and Acinetobacter species. A. baumannii may harbour most of the acquired carbapenemase genes within Enterobacteriaceae, and Pseudomonas in addition to their characteristics CDHL genes [7].
Recently, new molecular assays have been described to detect most prevalent carbapenemase genes [15], or a subset of A. baumannii selective carbapenemase genes. Due to limited gene set, or technical limitations, most new tests are not suitable for clinical routine monitoring in low prevalence settings [16]. In addition, combinations of other resistance mechanisms, such as reduced permeability due porin mutations, or defect, and efflux pumps in conjunction with ampC b-lactamases are the most common cause of carbapenem resistance in low prevalence areas [14]. Therefore, an imipenem hydrolysis test or dedicated MALDI-TOF [17] and more extensive screening of resistance mechanisms in a reference laboratory are often needed to reliably exclude carbapenemase genes.
The aim of this study was to investigate the carbapenemase genes of A. baumannii and the correlation between these genes and clonal lineages. The feasibility of a new real-time PCR assay was tested for screening of most important carbapenemase genes detected among A. baumannii, Enterobacteriaceae, and Pseudomonas species.

Bacterial strains and culture conditions
A total of 55 Acinetobacter isolates from 44 patients were detected. 51 isolates with reduced susceptibility to carbapenem from HUSLAB (Laboratory of Helsinki University Central Hospital) between Jun 18 th 1993 and Jan 18 th 2008 were collected and four Acinetobacter isolates suscebtible to carbapenems were included as controls. Helsinki University Hospital is responsible for the secondary and tertiary care of app. 1.5 million people. The culture samples from this area received by HUSLAB are both from these hospitals as well as from outpatients of this geographical area, the Helsinki and Uusimaa district in southern Finland. The culture samples in this study were from patients treated in nine different hospitals (Table S1).
Acinetobacter isolates were cultured in aerobic atmosphere on chocolate and cysteine lactose electrolyte deficient (CLED) agar and incubated at 35uC for 18 h. Colonies with typical morphology and biochemistry were identified as A. baumannii complex. Identification with the VITEK 2 (bioMérieux, Marcy L'Etoile, France) system with GN card was performed, as well. 16S rRNA gene sequencing was performed when biochemical identification was equivocal. In addition a house-keeping OXA-51-like (bla OXA-51-like ) gene was detected separately within all the clinical isolates with reduced susceptibility to carbapenems, whereas carbapenem susceptible control strains did not harbour OXA-51-like genes.
Antimicrobial susceptibility testing was performed by the disk diffusion method according to the CLSI guidelines (http://www. clsi.org). MICs for ampicillin/sulbactam, colistin, imipenem, meropenem, rifampicin and tigecycline by E-test (AB BIODISC, Solna, Sweden) were determined on Mueller-Hinton agar according to manufacturer's instructions.

Design of multiplex Real-Time carbapenemase gene screening assay
The assay was designed to detect most clinically relevant carbapenemase genes described within A. baumannii, Pseudomonas aeruginosa, and Enterobacteriaceae species. The design was performed using AlleleID software (http://www.premierbiosoft.com), taking into account all the globally known sub-variants in NCBI data base. For practical purposes, the assay was divided in two multiplex reactions consisting of nine and eight gene families, respectively. The assay was validated in vitro using 43 positive control strains (Table 1), which were confirmed at National Institute for Health and Welfare, Turku, Finland [14]. Since the target primer regions were fully conserved in silico, it was considered adequate to demonstrate PCR performance with one or more control species representing all the gene variants. In addition, synthetic gene constructs for SFC, CMY-1/10, SIM, SME, OXA-25, and OXA-58 genes containing a partial, nonfunctional resistance gene in E. coli plasmid (pIDTsmart), including the amplicon and app. 20 bp upstream and downstream sequence (Integrated DNA Technologies Inc, CA, USA). The plasmid was then transfected into the TOP10 strain according to manufacturer's instructions. The construct was ordered from IDT using pSMART plasmid, blunt-ended, containing a kanamycin resistance gene. The SFC, and SIM the control strains were obtained later (as a kind gift from Dr. Correia and Dr. Yunsop Chong and Kyungwon Lee, consequently). All the gene products were confirmed by sequencing with reference primers, or the gene

Validation of multiplex Real-Time PCR assay
Template DNA was extracted from a single colony on CLED plate grown overnight, and re-suspended in 100 ml TE-buffer (0,5 McF) and boiled 15 min. Each 20 ml real time PCR-reaction included 10 ml Maxima SYBR Green qPCR Master Mix (2X) (Scientific Fermentas, Schwerte, Germany), 6 ml Oligomix 1 or 2 ( Table 2), IDT (Integrated DNA Technologies, Inc.), 3 ml H 2 0, and 1 ml DNA template. Amplification was performed as follows: 95uC 10 min initial denaturation, 30 cycles with 95uC 20 sec denaturation, 58uC 30 sec annealing and extension, final extension 58uC 1 min and final denaturation 95uC 30 sec (MxPro Table 4. Primers used for sequencing of resistance genes by polymerase chain reaction (PCR). 3005P, Stratagene, La Jolla, CA, USA). Melting curve was determined between temperatures 58-95uC. Control strains are presented in Table 3. The PCR was run as a preformed oligonucleotide mixture with master mixture and template to avoid quality variations between the runs. A new oligonucleotide mixture was always tested with all the panel targets with set expected 19-25 Cq range in qPCR depending on the target ( Table 3). The oligonucleotide mixture was stored in stock concentrations in small aliquots, and a working dilution was formed for short term usage only. In addition, each PCR run including a representative negative and positive control for the given multiplex: KPC for multiplex 1 and NDM for multiplex 2. An acceptance range for positive controls (target +/2 3 Cq) was implemented to accept test series.

Rep-PCR
DNA was extracted from colonies on CLED plates using the UltraClean microbial DNA isolation kit (Mo Bio Laboratories, Solona Beach, CA, USA) and diluted to 35 ng/ml. The DNA was amplified using the DiversiLab Acinetobacter kit (Bacterial Barcodes, Inc. cat no DL-AB01, Athens, GA, USA) for DNA fingerprinting following the manufacturer's instructions. PCR was run on preheated thermal cycler (DNA Engine Tetrad 2, Peltier Thermal Cycler BioRad, Hercules, CA, USA) using the parameters according to manufacturer's recommendations. The kit specific positive and negative controls were run with each reaction set for the validation of amplification. The rep-PCR products were detected and the amplicons were separated using microfluidics labon-a-chip technology and analysed using the DiversiLab system (Bacterial Barcodes, Inc.). Further analysis was performed with the web-based DiversiLab software (version 3.4) using the band-based modified Kullback-Leibler distance for the calculation of percent similarities. The manufacturer provides guidelines for strain-level discrimination; similarity more than 97% is considered as indistinguishable (no differences in fingerprints), similarity more than 95% as similar (1-2 band difference in fingerprints) and similarity less than 95% as different. In this study optimal cut-off for clustering was 95%.

Ethics statement
The bacterial isolates analyzed in this study belong to the microbiological collections of HUSLAB (Laboratory of Helsinki University Central Hospital) and were obtained as part of routine clinical care in the past. Furthermore, all patient identifiers had been previously removed and data were analyzed anonymously. As the isolates were not clinical samples in the legal sense, no written or verbal consent was needed.

Temporal variation of prevalent, endemic A. baumannii clones
A time dependent clonal variation among the analysed A. baumannii was observed. A predominant clone was detected during the follow-up period, typically lasting a few years, which was then substituted by a new clone (Figure 2). Briefly, first a few isolates, harbouring a mobile element with OXA-58 gene, appeared 1993-1996 and 2003-2006 (Clone 1, Figure 1), which was not detected in the following years, followed by a clone harbouring a mobile OXA-23-like gene (Clone 2, Figure 1). The results were consistent with DiversiLab typing, and characteristic antibiotic susceptibility profile associated with the OXA clones analyzed. Only five out of 55 species having OXA-23/-58 gene displayed a different rep-PCR profile. Based on rep-PCR analysis, two predominant clones were detected. One isolate having OXA-24-like gene was unique in DiversiLab analysis, as well. As expected, all the control isolates from patient with no known connection were unique in their rep-PCR profiles.

Association of antibiotic susceptibility with clonality and carbapenemase gene profile
In our study, OXA-58 isolates had lower MIC-values for to meropenem than OXA-23-like positive isolates that systematically had higher MIC-values ( Table 5). The isolates with non-acquired OXA-gene, displayed a marked variation and they included also some carbapenem resistant isolates. The control isolates (Figure 1) consisted of Acinetobacter spp not harbouring any of the OXA genes analyzed. These isolates were all carbapenem susceptible (Table 5).

Discussion
The carbapenemase producing multi-resistant gram negative rods are probably the most important challenge for hospital hygiene at the moment [13,19]. The great variety of underlying mechanisms, in contrast to simple mecA or mecC in MRSA, possesses a significant challenge to clinical screening process. Phenotypes are highly variable and many overlapping other resistance mechanisms complicate any simple screening approach. A straight-forward, economical method suitable for routine clinical diagnostics has not been available yet. In this paper we demonstrate the good performance of a new multiplex real-time PCR assay, detecting most important carbapenemases based on melting curve analysis, by applying it to an epidemiologically important set of clinical A. baumannii isolates. In a striking contrast to carbapenemase producing Enterobacteriaceae, which were first detected in Finland 2008 [14], the carbapenem resistant A. baumannii were detected in Finland already three decades ago. This study highlights the emergence of carbapenem-resistant A. baumannii isolates carrying the bla OXA-23-like gene (Clone 1), which replaced the bla OXA-58 gene (Clone 2) in three years ( Figure 2). These major clones might have been endemic.
The new carbapenemase detection assay was initially developed to detect carbapenemase producing Enterobacteriaceae isolates, but it also appeared to be a useful tool for P. aeruginosa and A. baumannii. After three years of clinical use, it has been proved to be sensitive and highly specific screening assay among more than 700 hundred isolates with reduced carbapenem susceptibility analysed to date [14]. One of the major problems related to molecular detection of many antibiotic resistance genes is the appearance of new genomic variants. For example, the variable regions of bla OXA-181 are up to 9% different from bla OXA-48 [20]. The new variants may not be detectable with the existing systems. To minimize the risk for false negative results, the primers were designed at conserved gene regions to achieve optimal amplification of all the current and forthcoming sub-variants. The SYBR Green chemistry was preferred to avoid false negative results due to minor mutations in the probe sequence. The probe based assays are often sensitive to just 1-2 mutations in probe sequence, whereas primers are usually less sensitive to minor target mutations. These design features were considered relevant to achieve a high exclusion power of clinically relevant, acquired carbapenemase genes among carbapenem resistant strains.
A. baumannii is a nosocomial pathogen, and epidemiological tools are important to develop effective strategies for better monitoring of MDRAB clinical isolates [21]. In this study we used rep-PCR because the method is suitable for comparison of isolate genetic profiles using standardized and automated format [22]. This method has previously demonstrated good discrimination ability of A. baumannii isolates [23,24]. We found two major clones with DiversiLab (Clone 1 and 2, Figure 1.) harbouring most of the isolates with bla OXA-23-like and bla OXA-58 genes. There were only few exceptions. The cases were mostly from departments of treating patients with severe burn trauma, or intensive care units.
In this study, a good correlation between the carbapenemase gene and DiversiLab typing suggested that they both could be effectively applied for epidemiological screening of A. baumannii species. The new carbapenemase gene screening assay has been in clinical use for more than three years, and it has been a highly suitable method for rapid unequivocal identification of isolates harbouring acquired carbapenemase genes among Acinetobacter, Pseudomonas aeruginosa, and Enterobacteriaceae species. This study suggests that the new molecular methods could be successfully applied in clinical diagnostics to monitor acquired carbapenemase genes, provided that they are user-friendly and cost-effective as well.  Supporting Information  Author Contributions