Detection of cfxA and cfxA2, the β-Lactamase Genes of Prevotella spp., in Clinical Samples from Dentoalveolar Infection by Real-Time PCR

Clinical samples.A total of 87 clinical samples were obtained from patients with dentoalveolar infection (53 males, 34 females; mean age, 48.1 years) attending the Oral and Maxillofacial Surgery Clinic of Kanazawa University Hospital between September 2001 and March 2005. Each of the samples obtained was recovered from a distinct and individual patient. A total of 72 pus samples (yellow thick pus, 50 samples; blood-like pus, 22 samples) were taken by aspiration from the abscess with a disposable syringe with an 18-gauge needle. Each sample was stored at −80°C prior to analysis by real-time PCR, with a portion immediately used for cultural bacterial examination. The remaining 15 samples were pus-like exudates from the root canal of the tooth involved. These samples were obtained using three sterilized paper points (size 25). Each sample held within the paper points was suspended in 100 μl of saline solution by vigorous vortexing for 20 s. A 20-μl volume of the suspension was cultured, and the remainder was stored at −80°C prior to DNA extraction.

Bacteriological culture examination.Samples were inoculated aerobically and microaerophilically on brucella HK agar (Kyokuto, Tokyo, Japan) containing 5% (vol/vol) sheep blood for 48 h at 37°C. Culture of strict anaerobes was on brucella HK agar with 5% (vol/vol) sheep blood, which was incubated in an anaerobic atmosphere for up to 7 days at 37°C. Duplicate plates were prepared containing paromomycin (Pfizer, Tokyo, Japan) (75 mg/liter) and vancomycin (Shionogi, Osaka, Japan) (2.5 mg/liter) to selectively isolate strictly anaerobic gram-negative bacilli. Isolates were identified by conventional methods (20). Bacterial growth was recorded to determine the approximate number of bacteria in each sample. After identification, a colony of each isolate was stored at −80°C.

β-lactamase production and antimicrobial susceptibility.A nitrocefin disk (Cefinase; Becton Dickinson, Cockeysville, Md.) was used to determine whether strains were positive for β-lactamase production (12-14). Susceptibility of 242 randomly chosen clinical isolates (Prevotella spp., 139 isolates; Fusobacterium, 33; Peptostreptococcus, 23; Campylobacter, 13; Porphyromonas, 12; Corynebacterium, 4; Veillonella, 7; Streptococcus, 5; Bacteroides, 3; Bifidobacterium, 1; Gemella, 1; Lactobacillus, 1) to penicillin G was determined by a disk diffusion method approved by the Clinical and Laboratory Standards Institute (CLSI; formerly NCCLS) (22). In addition, MICs of Prevotella isolates for amoxicillin (Astellas, Tokyo, Japan) and amoxicillin-clavulanate (GlaxoSmithKline, Middlesex, United Kingdom) were determined by the agar dilution method recommended by CLSI (21). Bacteroides fragilis ATCC 25285 and Bacteroides thetaiotaomicron ATCC 29741 were used as quality-control strains in each test. The resistance breakpoints for amoxicillin and amoxicillin-clavulanate were determined to be 2 and 16 μg/ml, respectively, based on the CLSI guidelines (21).

Extraction of DNA from bacteria.Tested strains were subcultured on the brucella HK agar containing 5% (vol/vol) sheep blood for 48 h at 37°C, and resulting colonies were suspended in distilled water (100 μl) at a concentration equivalent to a MacFarland 1.5 standard. The suspension was heated to 96°C for 10 min and chilled to 4°C for 5 min, and the cell debris was removed by centrifugation. The supernatant was used as a template for PCR.

Extraction of DNA from clinical samples.A 50-μl volume of pus was suspended in 50 μl of distilled water and treated as described above. For root canal exudates, the suspension (100 μl) was centrifuged at 12,000×g for 10 min and the supernatant (80 μl) was discarded. The concentrate (20 μl of the remainder) was then used for DNA extraction. This concentration procedure was also employed for blood-like pus samples.

Real-time PCR.The primers and the TaqMan probe were designed using cfxA and cfxA2 gene sequences (GenBank accession no. U38243 and AF11810, respectively) and Primer Express software version 2 (Applied Biosystems, Foster, CA). The nucleotide sequence of the forward primer was 5′-GCGCAAATCCTCCTTTAACAA-3′ (KAG309), and the reverse primer sequence was 5′-ACCGCCACACCAATTTCG-3′ (KAG310). The sequence of the real-time PCR probe was 5′-TGATAGCATTTCTCAAATTGTCTCAGCTTGTCC-3′ (cfxA Taq Pr). The cfxA Taq Pr was 5′ end labeled with 6-carboxyfluoscein (FAM) as the reporter dye and 3′ end labeled with 6-carboxytetramethylrhodamine (TAMRA) as the quencher. The primers and probe were selected from a region with 100% nucleotide identity between cfxA and cfxA2. Although amino acid substitutions of cfxA and cfxA2 have been reported (6), the target regions for PCR amplification have not been found to differ in sequence. The real-time PCR was performed in a 25-μl final volume containing 12.5 μl of Premix Ex Taq (Perfect Real Time; Takara, Kyoto, Japan), 2 μl of DNA template, 0.5 μl of 10 μM KAG309 and KAG310 primers and 0.5 μl of 10 μM cfxA Taq Pr probe at a final concentration of 0.2 μM, 0.5 μl of ROX reference dye (50×), and 8.5 μl of distilled water. All reactions were run on an ABI Prism 7000 sequence detection system (Applied Biosystems) in triplicate with the following cycling parameters: 95°C for 10 s followed by 40 cycles of 95°C for 10 s and 60°C for 31 s.

Analytical specificity and sensitivity (detection limit).Ten CfxA and CfxA2 β-lactamase-positive and 10 β-lactamase-negative Prevotella strains, which had been stored in our laboratory, were examined to confirm specificity of real-time PCR. β-Lactamase production by all these strains had previously been confirmed using the nitrocefin disk, antimicrobial susceptibility tests, and the presence of cfxA and cfxA2 by PCR as described previously (4) using primers 5′-GCAAGTGCAGTTTAAGATT-3′ and 5′-GCTTTAGTTTGCATTTTCATC-3′. In addition to these Prevotella strains, clinical isolates of Porphyromonas gingivalis (n = 12), Peptostreptococcus micros (n = 20), Fusobacterium nucleatum (n = 36), Campylobacter gracilis (n = 7), and Veillonella sp. (n = 6) were also examined. All these strains were β-lactamase negative.

Quantified dilutions of four β-lactamase-positive Prevotella strains (P. intermedia, 2 strains; P. melaninogenica, 1; P. buccae, 1) were prepared by measuring the number of CFU. A randomly selected pus sample was confirmed to be cfxA and cfxA2 negative and β-lactamase negative using the real-time PCR, conventional PCR (4), and culture examination. A 20-μl portion of each bacterial dilution was mixed with same volume of the pus. Bacterial DNA extraction was done as described above.

Ethics for study.This study was approved by the Ethics Committee of Kanazawa University Graduate School of Medical Science.

Growth of bacteria from samples.Bacterial growth was found in all clinical samples. Half of the thick pus samples, 4 of 22 blood-like pus samples, and 14 of 15 pus-like exudates from root canal yielded heavy bacterial growth under anaerobic conditions. The remaining samples produced limited bacterial growth.

Bacteriology, β-lactamase production, and penicillin disk susceptibility.The results of cultural bacteriological examination are presented in Table 1. Prevotella, Peptostreptococcus, Fusobacterium, and Campylobacter spp. were predominantly isolated from the infections. Of the Prevotella isolates, 48 (31%) were found to be positive for β-lactamase; furthermore, there were no non-Prevotella β-lactamase-positive isolates detected. The penicillin disk susceptibility data was 100% concordant with the results of β-lactamase production testing (data not shown).

Analytical specificity and sensitivity of real-time PCR.Real-time PCR confirmed the presence of the cfxA and cfxA2 genes for all 10 β-lactamase-positive and none of the 10 β-lactamase-negative Prevotella strains. This gene was not detected by real-time PCR in any strains of Porphyromonas, Peptostreptococcus, Fusobacterium, Campylobacter, or Veillonella. Using “spiked” pus, real-time PCR revealed a detection limit of 5 × 10³ bacterial CFU/ml, corresponding to 5 CFU/reaction mixture, for three Prevotella strains tested. For one Prevotella strain, the detection limit was 1 ×10³ bacterial CFU/ml of pus.

Prevalence of cfxA and cfxA2 in Prevotella species and the MIC for amoxicillin-clavulanate.Although a total of 155 Prevotella isolates were recovered in this study, 16 strains that were stored by freezing were not subsequently recovered by culture for the PCR study. Therefore, to determine the prevalence of cfxA and cfxA2 in Prevotella isolates, a total of 139 clinical strains were examined. In this study, 48 (100%) of the β-lactamase-positive Prevotella strains were found to be positive for cfxA and cfxA2 by real-time PCR (Table 2). In contrast, all β-lactamase-negative strains were negative for the PCR. The β-lactamase-positive strains were highly susceptible to amoxicillin-clavulanate, although resistance to amoxicillin was recorded for all of the strains (Table 2). β-Lactamase-negative strain results showed low MICs for both amoxicillin and amoxicillin-clavulanate, with only one strain identified as having low-level resistance to amoxicillin (MIC, 2 μg/ml).

Direct detection of cfxA and cfxA2 from clinical pus samples by real-time PCR.Table 3 presents the agreement between cultural examination and PCR with respect to detection of β-lactamase-positive bacteria. Cultural bacteriological examination revealed that 31 (36%) of 87 clinical samples had β-lactamase-positive strains. Using the real-time PCR, cfxA and cfxA2 were detected for all these 31 β-lactamase-positive samples. Additionally, the real-time PCR also detected the resistance gene in 8 (14%) of the 56 samples deemed negative for β-lactamase by culture. Of these eight samples, six yielded β-lactamase-negative Prevotella results, and no Prevotella isolates were recovered from the remaining two samples.