Development of a Rapid Reverse Blot Hybridization Assay for Detection of Clinically Relevant Antibiotic Resistance Genes in Blood Cultures Testing Positive for Gram-Negative Bacteria

Rapid and accurate identification of the causative pathogens of bloodstream infections is crucial for the prompt initiation of appropriate antimicrobial therapy to decrease the related morbidity and mortality rates. The aim of this study was to evaluate the performance of a newly developed PCR-reverse blot hybridization assay (REBA) for the rapid detection of Gram-negative bacteria (GNB) and their extended-spectrum β-lactamase (ESBL), AmpC β-lactamase, and carbapenemase resistance genes directly from the blood culture bottles. The REBA-EAC (ESBL, AmpC β-lactamase, carbapenemase) assay was performed on 327 isolates that were confirmed to have an ESBL producer phenotype, 200 positive blood culture (PBCs) specimens, and 200 negative blood culture specimens. The concordance rate between the results of REBA-EAC assay and ESBL phenotypic test was 94.2%. The sensitivity, specificity, positive predictive value, and negative predictive value of the REBA-EAC assay for GNB identification in blood culture specimens were 100% (95% CI 0.938–1.000, P < 0.001), 100% (95% CI 0.986–1.000, P < 0.001), 100% (95% CI 0.938–1.000, P < 0.001), and 100% (95% CI 0.986–1.000, P < 0.001), respectively. All 17 EAC-producing GNB isolates from the 73 PBCs were detected by the REBA-EAC assay. The REBA-EAC assay allowed easy differentiation between EAC and non-EAC genes in all isolates. Moreover, the REBA-EAC assay was a rapid and reliable method for identifying GNB and their β-lactamase resistance genes in PBCs. Thus, this assay may provide essential information for accelerating therapeutic decisions to achieve earlier appropriate antibiotic treatment during the acute phase of bloodstream infection.


INTRODUCTION
Antibiotic resistance among pathogenic organisms has become a major healthcare problem worldwide, with serious consequences for the treatment of infectious diseases (Gottlieb and Nimmo, 2011). Early targeted antibiotic treatment is an important prognostic factor, especially for seriously ill patients (Cosgrove, 2006). Among the GNB, the Enterobacteriaceae family includes the most widespread pathogens causing infectious diseases; its members are among the most common causes of nosocomial infection. Beta-lactamase production is the major resistance mechanism of the Enterobacteriaceae. Specifically, extended-spectrum β-lactamases (ESBLs), plasmid-mediated cephalosporinases (pAmpCs), and carbapenemases pose major therapeutic challenges in the treatment of hospitalized and community-based patients. Rapid and accurate identification of β-lactamase-producing Enterobacteriaceae is crucial for reducing clinical failure due to the use of inappropriate antimicrobial therapy and for preventing nosocomial outbreaks.
The CLSI has released recommendations for the detection of ESBLs in clinical isolates of Klebsiella pneumoniae, Klebsiella oxytoca, Escherichia coli, and Proteus mirabilis using clavulanic acid (CLSI, 2015). However, no guidelines have yet been issued for the detection of AmpCs. Moreover, the occurrence of bacteria that co-produce ESBLs and AmpCs has become a serious issue in the context of Enterobacteriaceae β-lactam resistance. Furthermore, overproduction of AmpCs in association with a porin defect can lead to carbapenem resistance (Jacoby, 2009). Although the modified Hodge test for carbapenemases has been found to be a useful tool for the detection of carbapenemase-producing Enterobacteriaceae, this test lacks sensitivity for NDM carbapenemases. Moreover, false positive results can occur in isolates that produce ESBL or AmpC enzymes coupled with porin loss.

Clinical Isolates
To assess the diagnostic performance of the REBA-EAC assay, 327 ESBL phenotype clinical isolates obtained from 2010 to 2014 at Wonju Severance Christian Hospital (WSCH) were used. Isolates were identified by conventional identification method including phenotypic characteristics such as colony morphology, Gram staining result, and commercial identification kit results. MicroScan (Beckman Coulter, Brea, CA, USA) overnight Pos BP Combo 28, MICroSTREP Plus, overnight Neg Combo 53 and Neg Combo 54 panels were used for the identification of GPB, streptococci and GNB. For the identification of Candida spp., VITEK 2 (bioMérieux, Durham, NC, USA) yeast identification card was used. The ESBL phenotype of each isolate was confirmed by the microbroth dilution method, which is the test recommended by the CLSI. AmpC β-lactamase and carbapenemase production was confirmed using qPCR and by sequencing analyses with EAC primers ( Table 2). This study was approved by the Institutional Ethics Committee at Yonsei University Wonju College of Medicine (approval no. CR312055-002) and samples were de-identified.

DNA Preparation
To prepare DNA templates for the REBA-EAC assay, one colony from each frozen stock and one colony from each clinical isolate was suspended in 100 µL of DNA extraction solution (Optipharm, Osong, Republic of Korea). The bacterial suspension was boiled for 10 min. After centrifugation at 13,000 × g for 10 min, the supernatant was used as a source of DNA template.

REBA-EAC Primers and Probes
The 26 specific primers and probes were used for the detection of GNB (10 probes) and their antibiotic resistance genes (16 EAC gene probes) and an example of the result is shown in Figure 1. The sequences of the primers and probes were compared by nucleotide-nucleotide NCBI BLAST (BLASTn) searches to determine their sequence homologies. Primers and probes were selected by multiple alignments using the MultAlign program (http://multalin.toulouse.inra.fr/multalin/). All publicly available nucleotide sequences of the TEM and SHV alleles related to the ESBL-production phenotype (http://www.lahey.  Stuart et al., 2010;Rubtsova et al., 2010). Based on these findings, 7 oligonucleotide probes (Figure 1) were designed to identify TEM-variants (one TEM-all probe and 6 probes for ESBL variants). Specifically, the probes were designed to detect the following ESBL-associated amino acid substitutions: Gln39Lys, Glu104Lys, Arg164Ser/His, Gly238Ser, and/or Glu240Lys. To identify SHV-variants, 6 probes ( Figure 1) were designed (one SHV-all probe and 5 probes for ESBL variants). Specifically, the probes were designed to detect the following ESBL-associated

PCR
PCR was performed in a final reaction volume of 20 µL. Each reaction contained 2× master mix (10 µL) (GeNet Bio, Daejeon, Republic of Korea), 5 µL primer mixture, and 5 µL sample DNA. The first 10 PCR cycles were comprised of denaturation at 95 • C for 30 s, followed by annealing and extension at 60 • C for 30 s. These 10 cycles were followed by 40 cycles of denaturation at 95 • C for 30 s and annealing/extension at 54 • C for 30 s. After the final cycle, samples were maintained at 72 • C for 10 min to complete the synthesis of all strands. The amplicon was visualized as a single 230 bp band using a ChemiDoc system (Bio-Rad, Hercules, CA, USA).

REBA-EAC Assay
To validate the efficiency of the selected probes, target DNA samples amplified from EAC-producing strains were applied to REBA membrane strips spotted with the selected probes. Two genes (FOX and SPM) were synthesized (Bioneer, Daejeon, Republic of Korea) and amplified with custom PCR primers (FOX, F-5 ′ -ATGCAACAACGACGTGCGTTCGC-3 ′ and R-5 ′ -TGGCAAGCTCGGCCATGGTCAC-3 ′ ; SPM, F-5 ′ -CGTTTTG TTTGTTGCTCGTTGCG-3 ′ and R-5 ′ -TCAGCTGCTTTTATC CGGTCTTTC-3 ′ ), resulting in amplicons of 395 and 400 bp, respectively. The resultant products were mutagenized after subcloning into the pBHA vector. Two plasmids were extracted from the transformants, and the mutated sequences were confirmed by sequence analysis (CosmoGenetech, Seoul, South Korea). For the REBA-EAC assay, hybridization and washing were performed according to the manufacturer's instructions. Each sample was tested in duplicate and all REBA experiments were performed twice. In brief, biotinylated PCR products were denatured at 25 • C for 5 min in denaturation solution. The resulting single-stranded PCR products were transferred to hybridization solution and then incubated with REBA-EAC membrane strips at 55 • C with shaking at 90 rpm for 30 min in the provided blotting tray. The strips were then washed twice with gentle shaking in 1 mL of washing solution for 10 min at 55 • C, incubated at 25 • C with a 1:2,000 dilution of streptavidin-alkaline phosphatase (AP) conjugate (Roche Diagnostics, Mannheim, Germany) in conjugate diluent solution (CDS) for 30 min, and finally, washed twice with 1 mL CDS at room temperature for 1 min. The colorimetric hybridization signals were visualized by the addition of a 1:50 dilution of the NBT/BCIP AP-mediated staining solution (Roche Diagnostics), and incubated until a color change was detected. Finally, the banding pattern was read and interpreted.

Real Time PCR (qPCR) Assay
All phenotypic ESBL producers were screened by qPCR (CFX-96 qPCR system; Bio-Rad) to identify their non-ESBL and ESBLcarrying genes using primers specific for detecting predominant subtypes of CTX-M, TEM, and SHV. Next, the AmpC-and carbapenemase-carrying genes were analyzed using qPCR (CFX-96 qPCR system; Bio-Rad). Specifically, the presence of genes encoding ACT, CMY-2-like, DHA, ACC-1, CMY-1-like/MOX, FOX, IMP, VIM, NDM, KPC, OXA-48-like, and SPM was analyzed ( Table 2). qPCR amplification was performed in a total volume of 20 µL, including 10 µL of 2× Thunderbird SYBR master mix (Toyobo, Osaka, Japan), 2 µL primer mixture, 5 µL template DNA, and 3 µL sterile distilled water. Each qPCR assay included internal control DNA that was used to simulate pure nucleic acid extraction and high sample quality, thus monitoring the effect of PCR inhibitors in the reaction. Positive and negative controls were also included throughout the procedure. Template controls containing sterile distilled water were also incorporated in each assay. The thermocycling conditions were as follows: 95 • C for 3 min, followed by 40 cycles of 95 • C for 3 s and 60 • C for 30 s. Each sample was tested in duplicate and all PCR assays were performed twice. The relative bacterial load was quantified by determining the cycle threshold (C T ), i.e., the number of PCR cycles required for the fluorescence to exceed a value significantly higher than the background fluorescence. A positive result was interpreted as when the C T value was less than 30 after fluorescence signal above background was observed.

Sequence Analysis
The PCR amplicons of all clinical isolates in this study were sequenced using an ABI 3730 automated DNA sequencer (Applied Biosystems, Foster City, CA, USA) and an ABI Prism BigDye Terminator (Applied Biosystems) system. The obtained sequences were compared with sequences in the NCBI GenBank database for species assignment.

Blood Culture Samples
To evaluate the diagnostic performance of the REBA-EAC assay on blood culture samples, 200 positive blood culture (PBC), and 200 negative blood culture (NBC) samples were collected at WSCH from April 2015 to June 2015. Blood samples were drawn from the patient at the bedside. Three or two pairs of culture bottles were inoculated with the blood samples and then incubated in a BacT/Alert 3D (bioMérieux, Durham, NC, USA) or BACTEC FX (Becton Dickinson Diagnostic Systems, Spark, MD, USA) blood culture system for 5 days. If no bacterial growth was detected within 5 days, the blood culture was considered to be negative. When bacterial growth was noted, the culture sample was inoculated onto blood and MacConkey agar plates (Becton Dickinson) and cultured overnight at 37 • C in a 5% CO 2 incubator.
For the preparation of DNA templates from the PBC and NBC samples, 0.5 mL of blood suspension was withdrawn directly from the culture bottle. Next, 1 mL of phosphate-buffered saline (pH 8.0) was added and the sample was spun by centrifugation at 13,000 × g for 1 min. The supernatant was then removed, after which the pellet was resuspended in 1 mL of erythrocyte lysis buffer (Sigma, St. Louis, MO, USA) and spun by centrifugation at 13,000 × g for 1 min. This wash step was repeated twice. The pellet was then resuspended in DNA extraction solution, as described above for the clinical isolates. To confirm the positive results from the REBA-EAC assay, qPCR and sequence analyses were performed using the same DNA samples.

Statistical Analysis
All statistical analyses were performed using Prism 5 software (GraphPad, La Jolla, CA, USA) and SPSS statistics software version 21.0 (IBM, Armonk, NY, USA). The sensitivity, specificity, negative predictive value, and positive predictive value of the qPCR and REBA-EAC assays were analyzed by calculating the kappa coefficient of the predictive ability, its corresponding P value, and the 95% confidence interval (CI).

Analytical Sensitivity and Specificity of the REBA-EAC Assay
Twenty-nine EAC-producing strains were identified using the designated target probes in the REBA-EAC assay, whereas 128 non-EAC-producing clinical isolates yielded negative results by the REBA-EAC assay (data not shown). No cross-genus or crossspecies reactivity was observed.
The REBA-EAC assay was able to detect single nucleotide polymorphisms found in the most prevalent TEM-and SHV-type ESBLs, including Gln39Lys, Glu104Lys, Arg164Ser, Gly238Ser, and/or Glu240Lys in the TEMs, and Asp179Ala/Asn/Gly, Gly238Ser/Ala, and/or Glu240Lys in the SHVs. TEM-52, TEM-66 (Glu104Lys and Gly238Ser mixed), and SHV-12 (Gly238Ser), which were identified as TEM/SHV-ESBLs by the REBA-EAC assay, were also confirmed ESBL-associated amino acid substitutions by sequence analyses. The target DNA samples hybridized strongly with the probes derived from their targets and showed no cross-reactivity (Figure 1). The analytical sensitivities of the 16 target probes (4 ESBL, 6 AmpC, and 6 carbapenemase probes) were determined using a standard curve of 10-fold dilutions (10 ng, 1 ng, 100 pg, 10 pg, 1 pg, 100 fg, and 10 fg) of DNA extracted from EAC-producing strains. The detection limit of the REBA-EAC assay for each probe ranged from 100 fg to 10 fg DNA per reaction (Figure 2).

Diagnostic Performance of the REBA-EAC Assay for Clinical Isolates
Based on the genus-specific or species-specific REBA-EAC assay probes, the overall concordance rate of the assay for species identification of the 327 Enterobacteriaceae isolates between the conventional identification test and the REBA-EAC assay was 100%. Of the 327 ESBL-phenotypically confirmed isolates, 308 isolates (94.2%) were positive by the REBA-EAC assay for ESBL genes, 16 K. pneumoniae isolates had only AmpC plasmid-mediated DHA genes, and three E. coli isolates did not have any EAC genes by the REBA-EAC assay ( Table 3). Among the 324 isolates with at least one EAC gene, exclusive ESBL-producing isolates accounted for up to 70.0% of all EACproducers (one-type ESBL producer, 154 isolates; two-type ESBL co-producer, 73 isolates), followed by 23.1% for ESBL and AmpC β-lactamase; 4.9% for AmpC β-lactamase; 1.5% for ESBL and carbapenemase; and 0.3% for ESBL, AmpC, and carbapenemase (Table 3). Overall, 80.2% (247/308) of ESBL-producing isolates were the CTX-M type, with CTX-M-9 group (57.5%) being the most prevalent ESBL gene. For the TEM and SHV genes, the REBA-EAC assay was able to differentiate between non-ESBL TEM/SHV, and their ESBL derivatives for all isolates. Among the  144 pan-TEM-positive isolates tested by qPCR, 104 (72.2%) were positive for G104L , G238S,. Moreover, remaining 40 (27.8%) isolates, which were negative for G104L, G238S, and/or G240L, were considered to be non-ESBL producers by both the REBA-EAC assay and sequence analysis. The SHV-ESBL specific amino acid substitution, G238S, was detected in 31 (26.1%) of the 119 SHVproducing Klebsiella spp., and was confirmed in the SHV-2a, -5, and -12 genes by sequence analysis (Table 4).

DISCUSSION
Presently, antimicrobials are usually selected empirically for the initial treatment of bacteremia when a bloodstream infection is strongly suspected (Park et al., 2014). Furthermore, since the antimicrobial susceptibility results usually cannot be provided in <48-72 h, clinicians continue to rely on Gram stain results for selection of an appropriate antibiotic treatment for patients with PBCs (Uehara et al., 2009). Rapid and accurate EAC detection is crucial for effective infection control measures and for the selection of appropriate antimicrobial therapy, because the emergence of EAC-producing strains is a vital factor in the treatment of infections associated with sepsis. Like other mechanisms of antibiotic resistance, the genes for EACs are most often encoded on plasmids that can be readily transferred between bacteria (Haenni et al., 2014). The use of semiautomated systems for the identification and antimicrobial susceptibility testing of GNB is now routine practice in many laboratories. Compared to more recently developed molecular tests, traditional phenotypic susceptibility results usually take an additional 1-2 days after a PBC is reported. Since phenotypic detection of ESBLs, AmpCs, and carbapenemases is timeconsuming and the results may be difficult to interpret, a faster and more accurate detection method is highly desirable for routine use in clinical laboratories (Cohen Stuart et al., 2010). In general, commercially available molecular based tests have been developed for one of ESBL (CTX-M types, TEM, and SHV), AmpC β-lactamase (CMY-2-like, DHA, FOX, ACC-1, ACT/MIR, and CMY-1-like/MOX), and carbapenemase (VIM, IMP, NDM, KPC, and OXA-48-type) genes (Diekema and Pfaller, 2013). In comparison, REBA-EAC assay can provide information about GNB identification and broad range of clinically relevant antibiotic resistance genes in GNB.
One potential limitation of our study compared to previous studies is the limited number of samples, since the study was carried out at a single institution. Neither the AmpC βlactamase ACC-1, CMY-1-like/MOX, or FOX genes, nor the carbapenemase OXA-48-like gene, were detected in clinical isolates of this study. Moreover, since AmpC and carbapenemase genes are less prevalent than ESBL genes in Korea, we confirmed and included only ESBL-phenotype isolates in screening. Thus, there is a possibility that some of the non-ESBL isolates producing AmpC and/or carbapenemase could not be identified. Therefore, larger studies are required to confirm our results and limitations.
In conclusion, the recently developed REBA-EAC assay is an accurate, rapid, and convenient tool for identifying GNB.
This assay can also discriminate ESBL, AmpC β-lactamase, and carbapenemase genes and directly detect pathogen species from PBC samples in only 2 h. Therefore, the REBA-EAC assay can provide essential information that can accelerate therapeutic decisions for earlier appropriate antibiotic treatment in the acute phase of pathogen infection.

AUTHOR CONTRIBUTIONS
YU, WS, JBK, and HL designed the experimental strategy. HW and GY performed the experiments. HW, GY, and JK analyzed and interpreted the data. HW and GY wrote the paper with input from all co-authors. All authors have read and approved the final version of the manuscript.