Oligonucleotide Microarray for the Identification of Carbapenemase Genes of Molecular Classes A, B, and D

Full Text

INTRODUCTION During the last 20 years P-lactam antibiotics have been one of the drugs of choice for the treatment of a wide range of severe infectious diseases caused by gram-negative microorganisms. Among these antibiotics, carbapenems are one of the most successful drug groups (chemical structure shown in Fig. 1). This group of antibiotics is characterized by a wide spectrum of sensitive organisms, low toxicity, and good pharmacokinetics [1]. However, their efficiency has recently begun to show limitations with the emergence of drug-resistant strains. Carbapenem-resistant strains are most often found among nosocomial (in-hospital) infections belonging to the Pseudomonas spp. and Acinetobacter spp. genera. Healthcare facilities in the Russian Federation are, for example, noting a considerable increase in the number of pathogens resistant to carbapenem treatment. 38% of the observed Pseudomonas aeruginosa strains are now resistant to these drugs [2]. The emergence of P-lactam resistance in gram-negative bacteria can happen through several mechanisms, which include alteration of membrane permeability due to defective porine channels [3, 4], or activation of efflux systems [5]. However, the most clinically and epidemiologically important mechanism is the production of bacterial enzymes P-lactamases which hydrolyze the P-lactam ring of the antibiotic drug - P-lactamases [6, 7]. The P-lactamases are currently divided into 4 molecular classes - A, B, C, and D, based on their primary structure. The A-, C- and D-class enzymes are serine-type hydrolases, while the B-class enzymes are metal- lo-hydrolases, which bear one or two zinc atoms in their active site [8]. Several molecular classes of P-lactamase possess carbapenemase activity: however, the most of- Fig. 1. Chemical structure of carbapenemes. P-Lactam ring in dashed lines. ten found and clinically important carpbapenemases are KPC-type A-class enzymes [9], 5 groups of metallo-P-lactamases (VIM, IMP, SPM, GIM, SIM) [10], and a number of OXA-type D-class enzymes (subgroups OXA-23, OXA-40, OXA-51, OXA-58) [11]. Of all the numerous P-lactamases, carbapenemases are the most dangerous; they display high catalytic activity and wide substrate specificity, which includes practically every class of P-lactam antibiotics. Since carbapenemase-encoding genes are located on a plasmid, they can spread among pathogenic microorganisms at a rapid pace. Because of the variety of carbapenemases and the danger of their spreading, there is a need for robust methods for the detection of enzyme production, which can then be used for choosing the best suited treatment and for epidemiological control over the spreading of specific drug-resistant types. This is currently accomplished by using microbiological tests [6, 12, 13]. However, these tests take time and are ineffective for the identification of carba-penemase types. Identification of OXA-type enzymes by phenotype-based tests is virtually impossible [14]. Several PCR-based methods have been suggested for the identification of metallo-P-lactamase genes of the most spread VIM and IMP types [15, 16], and the main OXA-type carbapenemase subgroups [17]. A method for the identification of 5 groups of metallo-P-lactamases, involving multiplex real-time PCR with subsequent analysis of the melting curves of the obtained amplicons, has recently been developed [18]. However, the multiplex capacity of PCR is usually limited, which makes simultaneous detection of a large number of genes impossible. Hybridization analysis based on a microarray-tech-nology is a promising method of identification that yields quick results of high informative value. This technique has considerable advantages over traditional methods, since it allows a multi-parametric analysis and also uses a miniscule sample, which reduces cost and the time needed to obtain results [19, 20]. The goal of this work was to develop a method for the identification of A-, B-, and D-class carbapenemase genes involving hybridization analysis on membrane-based DNA microarrays which could be visualized by colorimetric detection. experimental procedures The collection of primers for the amplification of car-bapenemase genes and the amino-modified oligonucleotide probes were synthesized by Synthol (Moscow, Russia). Samples of the bacterial DNA extracted from control strains of A. baumanii, Ps. aeruginosa, Escherichia coli, and Klebsiella pneumonia were provided by the Institute of Antimicrobial Chemotherapy of the Smolensk State Medical Academy. These samples produced carbopenemases VIM-1, VIM-2, VIM-4, VIM-7, IMP-1, IMP-2, SPM-1, OXA-23, OXA-40, OXA-51, OXA-58, and KPC-3. Cell cultures of microorganisms from the Enterobacteriaceae family and A. baumanii and Ps. aeruginosa strains, either sensitive or resistant to carbapenems, according to phenotypic tests performed on a VITEK automatic analyzer (BioMerieux, France), were provided by the Burdenko Institute of Neurosurgery. Bacterial DNA extraction Extraction of bacterial DNA from a cell suspension with no less than 105 CFU/ml was performed using temperature lysis in a buffer. 500 jil of suspension was placed in centrifuge tubes, and the cells were pelleted by centrifugation at 10,000 g for 1 minute. After removing the supernatant, 100 jil of buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.5) was added to the pellet, which was then re-suspended with a shaker. The tubes were then incubated on a solid-medium thermostat for 20 minutes at a temperature of 990 C. After heating, the samples were centrifuged for 1 minute at 10,000 g. A PCR sample used 1 jil of the obtained supernatant. Amplification of carbapenemase genes using multiplex PCR with simultaneous biotin labeling Amplification of carbapenemase (A-, B-, and D-class) gene-fragments, with simultaneous biotin labeling, was performed in two multiplex PCR reactions (one reaction amplified the genes of all the metallo-p-lactamases and the other amplified the OXA- and KPC-type p-lactamase genes). Each multiplex PCR sample was 25 jil in volume and contained the following: 10 mM Tris-HCl-buffer with 2.5 mM of magnesium acetate, 50 mM KCl pH 8.3, 2.5 units of Taq-DNA-polymerase, 100 iM dATP, dGTP, dCTP, 60 ^M dTTP, 40 jiM dUTP- 11-biotin (Fermentas, Germany), 0.4 jj.M each of the direct and reverse primer for each group of carbap-enemase, and 1 jil of the template DNA solution. Amplification was performed in a Mastercycler gradient amplifier (Eppendorf, Germany) according to the following protocol: initial denaturation at 94°C (2 min), 25 cycles of amplification (20 sec - denaturation at 94°C, 30 sec - annealing of the primers at 65°C, 1 minute -elongation at 72°C), and a final elongation step at 72°C (6 min). Horizontal electrophoresis of the PCR products was performed in a 1% agarose gel with TAE buffer (40 mM Tris, 20 mM acetic acid, 1 mM EDTA, pH 8.5) with etidium bromide added to a final concentration of 1.6 ^g/ml. Visualization was performed on a UV-tran-silluminator at a wavelength of 260 nm. Fragmentation of the PCR-products The DNA was fragmented at room temperature for 5 minutes. Amplified DNA was diluted to a concentration of 30 ng/microliter with the reaction buffer (40 mM Tris-HCl, 10 mM MgSO4, 1 mM CaCl2, pH 8.0), and the mixture was then supplemented with DNAase I (Pro-mega, Germany). The reaction was stopped by the addition of 3 mM EDTA and 10-minute incubation at 65°C. Immobilization of the oligonucleotide probes on a membrane-based DNA-microarray BioTrace NT nitrocellulose (Pall Corporation, USA) was used as a support for DNA-microarray. Modification of the membranes was performed according to [21], using 1-ethyl-3-(3-dimethylaminopropyl)carboimide (Sigma, USA). The oligonucleotides were diluted in a buffer (160 mM Na2SO4, 130 mM Na2HPO4) to a final concentration of 20 |^M and then applied onto the membranes by an XactIITM Arrayer robot (LabNEXT Inc., USA) using 300 ^m pins. After the procedure, the membranes were incubated at 60°C for 30 minutes. Hybridization on the DNA microarray Prior to hybridization, the microarrays were washed with PBST buffer (0.01 M K2HP04, 0.15 M NaCl, 0.05% Tween-20, pH 7.0) twice, 10 minutes each time at room temperature, and then blocked in a solution of 1% bovine serum albumine (BSA) and 1% casein (Sigma, USA) in PBS buffer (0.01 M K2HP04, 0.15 M NaCl, pH 7.0) at 37°C for 30 minutes. 500 ng of fragmented and labeled DNA was then diluted in the hybridization buffer - 2x SSPE (0.3 M NaCl, 0.02 M NaH2PO4, 2 mM EDTA, pH 7.4), which also included 1.6 pmol/ml of the control biotin-labeled oligonucleotide (positive hybridization control). The microarray was then placed into the hybridization mixture (300 |^l per 1 array) and incubated at 45°C for 1 hour in a Thermomixer comfort apparatus (Eppendorf, Germany). After hybridization, the membranes were washed with PBST twice for 15 minutes at room temperature. Detection and hybridization data analysis The microarrays were incubated in a solution of streptavidine-peroxidase conjugate (Imtek, Russia) (0.2 ^g/ml) in PBST for 30 minutes at 37°C. Then they were washed in PBST for 10 minutes and placed into a substrate solution containing 3,3',5,5'-tetramethylben-zydene (TMB), H202 (NVO Immunotech, Russia) and sodium dextransulphate (M = 8000, Pharmacia, Sweden) (final concentration - 0.5% (by mass)) for 10 minutes, after which the arrays were washed in distilled water and air-dried. The membrane microarrays were scanned on a Perfection V750 Pro (Epson, Germany) scanner at a resolution of 4,800 dpi. The obtained images (in TIFF format) were analyzed using Scan Array Express (PerkinElmer, version 3.0, Germany) software, and the intensity values of the analytic signals at various spots of the microarray were determined. The absolute values of the signals were then recalculated into relative signals, using the mean intensity of the positive hybridization control used in each array. RESULTS AND DISCUSSION Molecular design of the oligonucleotide probes The Genbank database currently has information on 10 KPC-type enzymes, 52 metallo- P-lactamases (23 from the IMP group, 26 from the VIM group and one member of each of the following groups - SPM, SIM and GIM), and also 70 carbapenemases from the OXA group. Alignment of the amino acid and encoding sequences of these enzymes shows that only enzymes from the KPC group display a high degree of similarity within their group (differ by 1-2 amino acid substitutions), while the numerous members of the IMP, VIM, and OXA groups differ considerably from their group members. Because of this, each group was split into separate subgroups, which included enzymes whose genes were highly similar. Thus, the VIM group was divided into 3 subgroups (VIM-1, VIM-2, and VIM-7); the IMP group, into 6 subgroups (IMP-1, IMP-2, IMP-5, IMP- 11, IMP-12, and IMP-14); while carbapenemases from the OXA groups were divided into 4 subgroups (OXA-23, OXA-40, OXA-51 and OXA-58). Group and subgroup classification of carbapenemases based on their amino acid sequence alignments and their P-lactamase molecular classification are shown in Fig. 2. One of the main stages of DNA microarray development was the design of the oligonucleotide probe sequences required to detect various groups of carbapen-emase genes. The selection of an oligonucleotide probe for the identification of a group of genes is based on the alignment of the coding sequences of all the carbapene-mase genes in this group. What is needed is a sufficiently long fragment of the gene that is conserved in all of the members of this group and is no less than 18 nucleotides long. These regions were then analyzed in terms of melting temperature, G/C content and secondary structure formation. In order to perform microarray-based hybridization analysis, we selected the oligonucleotides that were unlikely to form secondary structures and Fig. 2. Classification of carbapenemases into groups and subgroups based on the alignment of amino acid sequences and their correspondence to molecular classes of P-lactamases. P-Lactamases IT Metalloenzymes (Zn2+) IT Serine I i Molecular classes c A D of enzymes B IT I FT Groups of carbapenemases OXA-23 Subgroups of IMP-1 OXA-40 carbapenemases IMP-2 OXA-51 IMP-5 OXA-58 IMP-11 IMP-12 IMP-14 VIM-1 VIM-2 VIM-7 KPC OXA IMP SPM VIM T SIM GIM whose melting temperature differed by no more than 10°C. Based on the selected sequences, we synthesized two oligonucleotide probes which were complementary to the direct and reverse strands of the gene. For additional typing of IMP, VIM, and OXA carbapenemase gene-subgroups, we chose regions with high similarity which bore no mutations inside a given subgroup and had low similarity with genes from other subgroups. In order to increase the specificity of the analytical procedure, we chose two oligonucleotide probe variants which corresponded to different regions of the gene. In this case, we synthesized the probes which were complementary to the reverse strand of the gene. The sequence of the chosen oligonucleotide probes and their characteristics are presented in Table 1. The length of the probes varied from 18 to 27 nucleotides, the G/C-content was 30- 60%, and the melting temperature was 63-72°C. Each probe was supplemented by additional spacers at the 5’-terminus, which helped to distance the probe from the array surface, thus removing steric barriers for hybridization and increasing the availability of the probe for the DNA-target. Optimization procedures showed that the best spacer contained 13 thymidine residues. The additional thymidines did not have any significant effects on the intensity of the hybridization signals, or on the specificity of hybridization (data not shown). Amplification of carbapenemase genes of various molecular classes with simultaneous labeling Amplification of A-, B-, and D-class carbapenemase genes utilized multiplex PCR with simultaneous biotin labeling. Biotin-labeled deoxyribouridine triphosphate (dUTP) was used as a labeling reagent and was incorporated into the DNA, along with unlabeled de-oxyribothymidine triphosphate (dTTP). The templates for the PCR reaction were bacterial DNA samples extracted from the control strains of microorganisms producing P-lactamases VIM-1, VIM-2, VIM-7, IMP-1, IMP-2, SPM-1, OXA-23, OXA-40, OXA-51, OXA-58, and KPC-3. Design of primers for the amplification of various carbapenemase genes was based on an alignment of the coding regions of these genes. Primers for amplification of the full-size carbapenemase genes from the KPC group and metallo-P-lactamase groups SPM, SIM, and GIM were chosen from regions which were conserved in this group, namely the gene termini. We could not find conserved regions longer than 20 bp for met-allo- P-lactamases from the IMP and VIM groups, nor could we find any for OXA-type carbapenemases, since these groups showed a low degree of in-group similarity, which is why separate primers had to be selected for each subgroup. The length of the primers (20 - 28 nucleotides) was chosen so as to push their melting temperature to 62 -680C, which would allow simultaneous amplification of all the types of genes simultaneously and with equal efficiency. Primer selection also factored in G/C-content, and we chose structures that had a G/C content of 30 -60%. We also estimated the possibility of primer-dimer and secondary structure formation and favored the sequences which were least likely to do so. As a result, each group was fitted with several direct and reverse primers with various parameters. Various combinations of these primers were tested in PCR reactions in order a KPC-3 OXA-23 OXA-40 OXA-51 OXA-58 Markers 2000 1500 1000 750 500 250 b VIM-1 VIM-2 VIM-7 IMP-1 IMP-2 SPM-1 Markers 2000 1500 1000 750 500 250 Fig. 3. Electrophoretic separation of multiplex PCR products after amplification of OXA-, KPC-type carbapenemases (a) and metallo-P-lactamases (b). to test the specificity of gene amplification. The primers which showed the highest specificity with a good output were used for the multiplex procedure. The primer sequences are presented in Table 2. In order to determine the optimal primer-annealing temperature for the multiplex amplification of all the carbapenemase gene types, we first calculated the optimal annealing temperature for each pair of primers in a specific reaction and then selected the lowest temperature value for the multiplex PCR reaction. The range of the studied primer T ‘s was 52 - 680C. The optimal Ta turned out to be 60°C. Reactions at this temperature resulted in the efficient synthesis of specific PCR-products. After determining the optimal carbapenemase gene amplification conditions, we evaluated the possibility of performing multiplex PCR with 16 pairs of primers in order to amplify the genes of all of the studied groups simultaneously. However, under these conditions the specific product for most of the carbapenemase groups was either absent or the yield was very low, which made further microarray hybridization analysis impossible. Because of this, we decided to amplify the carbapenemase genes using a two step multiplex PCR procedure: the first used a mixture of primers specific to metallo-P-lactamases (11 pairs of primers), and the second used a mixture of primers specific to OXA and KPC carbapene-mases (5 pairs of primers). The results of electrophoretic analysis of the PCR-products obtained during the amplification of various carbapenemase genes from control microorganism strains are presented in Fig. 3. The yield of labeled specific products from the multiplex PCR procedure was approximately 40-50 ng/jal for each type of carbapenemases gene, which was sufficient for further microarray-based hybridization analysis. Amplification of nonspecific products was detected only for genes from the VIM-1 and VIM-2 subgroups: however, these products had low yields and, as it follows from later experiments, their presence did not affect the specificity of the hybridization analysis. Oligonucleotide microarray for detecting A-, B-, and D-class carbapenemase genes The DNA microarray for the identification of the major types of carbapenemases is made on a support of nitrocellulose and has a size of 6.0 x 9.5 mm. On its surface there are 40 immobilized oligonucleotide probes (14 probes for the identification of 7 distinct groups of carbapenemases and 26 probes for additional typing of these genes into subgroups). Each microarray also includes 3 types of control oligonucleotides: an immobilization control (a biotin labeled oligonucleotide), a positive hybridization control (an oligonucleotide whose sequence is complementary to a biotin-labeled oligonucleotide which is added to the hybridization mixture), a negative hybridization control (an oligonucleotide with a random base sequence). In order to increase the reproducibility of this procedure, each oligonucleotide probe is present on the microarray in three copies. A schematic of the layout of the specific and control oligonucleotide probes on the surface of the DNA microarray is presented in Fig. 4. The carbapenemase gene identification procedure involved hybridization analysis on a DNA microarray and included the following stages: 1) amplification of the P-lactamase gene from DNA isolated from the clinical strain (by two multiplex PCR reactions); 2) hybridization of the biotin-labeled DNA with oligonucleotide probes on the surface of the microarray; 3) visualization of the hybridization results using a streptavidin-per-oxidase conjugate followed by colorimetric detection of the enzyme. We also optimized hybridization conditions. We analyzed hybridization efficiency at temperatures ranging from 40 to 50°C; the temperature could not be higher than the melting temperature (Tm) of any oligonucleotide probe and was limited by the high level of nonspecific hybridization at lower temperatures. Hybridization at 40°C displayed strong signals: however, most of the probes showed cross-hybridization with the genes of various carbapenemases. Hybridiza- Table 1. Sequences of the specific and control oligonucleotide probes Name Nucleotide sequence, 5’^3’ Length, nucleotides G/C, % Tm m °C Control oligonucleotides Immobilization control TCTAGACAGCCACTCATA-Biotin 18 44.4 60.4 Positive hybridization control GATTGGACGAGTCAGGAGC 19 57.9 66.1 Negative hybridization control TCTAGACAGCCACTCATA 18 44.4 60.4 Oligonucleotide probes for determining the carbapenemase group KPC_direct GCTTCCCACTGTGCAGCTCATTC 23 56.5 72.0 KPC reverse GAATGAGCTGCACAGTGGGAAGC 23 56.5 72.0 VIM_direct GGAGATTGAAAAGCAAATTGGACT 24 37.5 66.6 VIM reverse AGTCCAATTTGCTTTTCAATCTCC 24 37.5 66.6 IMP_direct GGAATAGAGTGGCTTAATTCTCG/A 23 41.3 64.7 IMP reverse C/TGAGAATTAAGCCACTCTATTCC 23 41.3 64.7 SPM_direct GATGGGACCGTTGTCATTG 19 52.6 64.9 SPM reverse CAATGACAACGGTCCCATC 19 52.6 64.9 SIM direct CCTTGGCAATCTAAGTGACGCAA 23 47.8 69.7 SIM reverse TTGCGTCACTTAGATTGCCAAGG 23 47.8 69.7 GIM_direct CACACTGGGAAATGGGCTTATA 22 45.5 66.7 GIM reverse TATAAGCCCATTTCCCAGTGTG 22 45.5 66.7 OXA_direct CCACAA/GGTG/AGGC/TTGGTTG/AAC 20 55.0 67.0 OXA reverse GTC /AAACCAG/ACCC/TACT/CTGTGG 20 55.0 67.0 Oligonucleotide probes for determining the carbapenemase sub-group VIM-1_568 TCAGCGAACGTGCTATACGG 20 55.0 68.3 VIM-1 590 GTTGTGCCGTTCATGAGTTGT 21 47.6 67.9 VIM-2 568 TCTGCGAGTGTGCTCTATGG 20 55.0 67.9 VIM-2_590 GTTGTGCGATTTATGAGTTGT 21 38.1 63.7 VIM-7 127 GTTCGGCTGTACAAGATTGGCG 22 54.5 70.0 VIM-7_181 CTCGGTGACACGGTGTAC 18 61.1 65.8 IMP-1_135 GTGGGGCGTTGTTCCTAAACATG 23 52.2 70.2 IMP-1_387 GGTTCAAGCCACAAATTCATTTAGC 25 40.0 67.8 IMP-2 264 TCAAAGGCACTATTTCCTCACATTTC 26 38.5 68.2 IMP-2_497 TACCTGAAAAGAAAATTTTATTCGGTG 27 29.6 65.7 IMP-5 506 AATAGAGTTTTGTTCGGTGGTT 22 36.4 65.0 IMP-5_459 TGGTCCAGGGCACACTCC 18 66.7 70.4 IMP-11_570 TGTTGAAGCATGGCCACATT 20 45.0 67.6 IMP-11_621 TGCAAAACTGGTTGTTCCAAGCC 23 47.8 70.9 IMP-12 226 AAATTAGTTGCTTGGTTTGTAGGG 24 37.5 66.4 IMP-12 495 GCTACCTGAAAACAAAATTTTATTCG 26 30.8 64.8 IMP-14 292 GGTGACAGTACGGCTGGAATAG 22 54.5 68.4 IMP-14_374 AAAAAGACAATAAGGTACAAGCTA 24 29.2 63.4 OXA-23_225 AAATACAGAATATGTGCCAGCCTCT 25 40.0 CO CO 6 OXA-23_309 GAAGGGCGAGAAAAGGTCATTTAC 24 45.8 68.0 OXA-40_225 AAATAAAGAATATGTCCCTGCATCA 25 32.0 65.6 OXA-40 329 GAACTTATCCTATGTGGGAGAAAG 24 41.7 64.8 OXA-51 225 TTCGACCGAGTATGTACCTGCTTCG 25 52.0 71.7 OXA-51_578 GCCCAAAAGTCCAAGATGAAG 21 47.6 65.8 OXA-58 225 AAAAACAGCTTATATTCCTGCATCT 25 32.0 66.0 OXA-58_206 GCACGCATTTAGACCGAGC 19 57.9 67.7 Table 2. Primer sequences for the multiplex PCR-amplification of carbapenemase genes Type Sequence 5’^3’ Length, nucleotide G/C, % Tm, °C Length of PCR-product, bp KPC direct TTCTGCTGTCTTGTCTCTCATGG 23 47.8 64.7 801 reverse CCTCGCTGTGCTTGTCATCC 20 60.0 65.7 IMP-1 direct GGCGTTTATGTTCATACTTCGTTTG 25 40.0 64.4 584 reverse GTAAGTTTCAAGAGTGATGCGTCTCC 26 46.2 65.6 IMP-2 direct GGTGTTTATGTTCATACATCGTTCG 25 40.0 63.8 584 reverse GTACGTTTCAAGAGTGATGCGTCCCC 26 53.8 67.8 IMP-5 direct GGTGTTTATGTTCATACTTCGTTTG 25 36.0 62.5 584 reverse GTACGTTTCAAGAGTGATACATCTCC 26 42.3 63.4 IMP-11 direct GGTGTTTATGTTCATACATCGTTTG 25 36.0 62.6 584 reverse GTAAGCTTCAAGAGCGACGCATCTCC 26 53.8 67.8 IMP-12 direct GGTGTTTATCTTCATACATCTTTTG 25 32.0 60.5 584 reverse GTAAGTTTCAAGAGTGATGCGTTCCC 26 46.2 66.0 VIM-1 direct GTAGTTTATTGGTCTACATGACCGCGTC 28 46.4 66.9 743 reverse CGCTGTGTGCTGGAGCAAGTC 21 61.9 68.1 VIM-2 direct GTAAGTTATTGGTCTATTTGACCGCGTC 28 42.9 65.9 743 reverse CGTTGTGTGCTTGAGCAAGTC 21 52.4 64.7 VIM-7 direct AGCATATTCCGCACAGCCTGG 21 57.1 67.5 685 reverse CCGGGCGGTCTGGAATTGCTC 21 66.7 67.7 SPM direct CGTTTTGTTTGTTGCTCGTTGCGGG 25 52.0 67.4 648 reverse CCTTCACATTGGCATCTCCCAGATAAC 27 48.1 67.2 SIM direct GTTTGCGGAAGAAGCCCAGCC 21 61.9 68.6 613 reverse CTCCGATTTCACTGTGGCTTGGG 23 56.5 67.6 GIM direct CTTGTAGCGTTGCCAGCTTTAGCTC 25 52.0 67.8 638 reverse CTGAACTTCCAACTTTGCCATGCC 24 50.0 66.9 OXA-23 direct GAAACCCCGAGTCAGATTGTTCAAG 25 48.0 65.8 686 reverse GGCATTTCTGACCGCATTTCC 21 52.4 64.8 OXA-40 direct GTTTCTCTCAGTGCATGTTCATC 23 43.5 62.3 714 reverse CATTTCTAAGTTGAGCGAAAAGGGG 25 44.0 64.6 OXA-51 direct CGAAGCACACACTACGGGTG 20 60.0 65.4 649 reverse CTCTTTTCGAACAGAGCTAGGTATTC 26 42.3 63.4 OXA-58 direct CTTGTGCTGAGCATAGTATGAGTC 24 45.8 63.3 684 reverse CCACTTGCCCATCTGCCTTTTC 22 54.5 66.5 Table 3. Results on clinical sample testing on DNA microarrays Type of microorganism Carbapenemase sensitivity as determined by phenotypical tests Number of samples Detected carbapenemase types OXA-23 OXA-40 OXA-51 OXA-58 VIM-2 A. baumanii Resistant 10 1 5 10 4 - Ps.aeruginosa 11 - - - - 11 A. baumanii Sensitive 2 - - - - - K. pneumonia 3 - - - - - E. coli 2 - - - - - b a 0 Immobilization control 0 Positive hybridization control # Negative hybridization control # Probes for identification genes of metallo-P-lactamases O Probes for identification genes of OXA-type carbapenemases # Probes for identification genes of KPC-type carbapenemases • • • Fig. 4. Layout of specific and control oligonucleotide probes on the surface of the DNA microarray for the identification of the A-, B-, D-class carbapenemase genes. tion at 50°C displayed weak signal intensities for some of the probes. For these reasons, we chose 45°C as the optimal temperature. The hybridization buffer consisted of 2x SSPE supplemented by 0.2 % sodi dodecyl-sulphate (SDS) in order to improve the membrane’s wetting properties. The size of the labeled DNA-target proved to be a critical parameter for hybridization. Hybridization of labeled PCR-products whose size was 580-800 nucleotides with the appropriate oligonucleotide probes proved to yield weak signals. Additional fragmentation by DNAase (yielding fragments of 50-150 nucleotides) proved to increase hybridization signal intensity for most of the probes. Hybridization duration was assayed within a range of 0.5 to 4.0 hours. It was observed that the hybridization of biotin-labeled DNA onto immobilized probes reaches equilibrium after two hours of incubation with active mixing. We also noticed that conducting the reaction in kinetic conditions (1 hour) does not dramatically weaken the signals as compared to the equilibrium state (about 10 - 20% depending on the probe). It also did not lower the specificity of the analytic procedure, which allowed the positive identification of all types of carbapenemase genes in the hybridization mixture. Figure 5 shows the results of an experiment in which control microorganism strains producing VIM-1 and IMP-1 metallo- P-lactamases and carbapenemases OXA-51 and 0XA-40 were tested on our DNA microarray. Identification of the P-lactamase group was assayed by the hybridization intensity with a group-specific probe, while additional typing was assayed by the hybridization intensity with subgroup-specific probes. The advantage of the microarray-based hybridization analysis is the possibility of simultaneously detecting several genes, which is demonstrated by testing the control A. baumanii strain for OXA-type carbapenemases. Testing clinical strains of microorganisms resistant to carbapenems The DNA microarray developed was tested on clinical strains of gram-negative microorganisms, either resistant or sensitive to carbapenems, as assayed by a phenotypical test. Table 3 shows the results for 28 clinical strains of Ps. aeruginosa, A. baumanii and En-terobacteriaceae spp., which display various levels of carbapenem sensitivity (strains were provided by the N.N. Burdenko Institute of Neurosurgery and Institute of Antimicrobial Chemotherapy of Smolensk State Medical Academy). All the A. baumanii strains which proved resistant to carbapenems (as assayed by pheno-typing) expressed two carbapenemase genes (OXA-51 and OXA-23 (1 strain), OXA-40 (5 strains), OXA-58 (4 strains)). All the carbapenem-resistant Ps. aeruginosa strains happened to possess a VIM-2-type metallo-P-lactamase gene. Testing of carbapenem-sensitive strains revealed no carbapenemase genes. Thus, the results of the microarray-based hybridization analysis are in accordance with the phenotyping tests. Moreover, the structure of the observed carbapenemases was confirmed by gene sequencing for two carbapen-em-resistant samples - one strain of A. baumanii and one strain of Ps. aeruginosa, which expressed 0XA-40, OXA-51 and VIM-2 genes. Thus, our method of hybridization analysis based on DNA microarray for the identification and typing of carbapenemase genes is highly accurate, productive, and can be used in clinical microbiological laboratories for the identification of carbapenemases and for studying their epidemiology. The phenotyping tests currently being used take time (from 24 to 48 hours) and are not always effective for determining carbapenemase types, such as OXA-type carbapenemases. Identification of carbapenemase genes on DNA microarrays allows rapid diagnostics, with the whole procedure taking only 4.5 hours, including 0.5 hours for bacterial DNA extraction, 1.5 hours for amplifying the carbapenemase genes and fragmenting the PCR-products, 1.5 hours for the hybridization and washing steps, and 1 hour for colorimetric detection of the hybridization results. An important feature of this method is the possibility of simultaneously identifying several genes in one sample.

About the authors

M M Ulyashova

Lomonosov Moscow State University

Chemistry Faculty

Yu I Khalilova

Lomonosov Moscow State University

Chemistry Faculty

M Yu Rubtsova

Lomonosov Moscow State University; ZAO 'NPP IMMUNOTEK', (NPP IMMUNOTECH, JSC)

Email: mrubtsova@gmail.com
Chemistry Faculty

M V Edelstein

Institute of Antimicrobial Chemotherapy, Smolensk State Medical Academy

I A Alexandrova

Burdenko Institute of Neurosurgery, Russian Academy of Medical Sciences

A M Egorov

Lomonosov Moscow State University

Chemistry Faculty

References

Supplementary files