Duplex PCR Assay Simultaneously Detecting and Differentiating Bartonella quintana, B. henselae, and Coxiella burnetii in Surgical Heart Valve Specimens

Clinical specimens and standards.

A total of 17 cases diagnosed as IE collected between 2000 and 2008 were included in the study. Patient charts were reviewed by members of the infectious disease team to confirm the clinical diagnosis of IE. No institutional research board approval was sought, and other laboratory data were unavailable. The B. quintana-, B. henselae-, and C. burnetii-specific fragments were generated by PCR amplification (see below) and were subsequently cloned into the pCR2.1 vector (Invitrogen, Carlsbad, CA). The DNA concentration of the three recombinant plasmid standards (pCT-Bq for B. quintana, pCT-Bh for B. henselae, and pcT-Cb for C. burnetii) was calibrated by spectrophotometry at 260 nm (17). The inserted plasmids were adjusted to 10,000 copies/μl and stored at −70°C prior to testing.

Specimen processing and DNA extraction.

Approximately 20 mg from surgical heart valve tissue specimens was used for processing and extraction. Nucleic acids were extracted using the QIAquick blood/tissue kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. Formalin-fixed and paraffin-embedded (FFPE) heart valve tissue blocks were marked to localize the areas most suspected to be infected according to the corresponding slides stained with hematoxylin and eosin. The FFPE specimens were treated with xylene in order to remove the paraffin before tissue lysis and DNA extraction were performed (12). The total extracted nucleic acids were eluted into 200 μl of water.

PCR amplification.

Primer sets were designed to target the citrate synthase gene (gltA) for the Bartonella species (1240F and 1497Rm) (19) and an insertion sequence IS111 gene (10) for C. burnetii (Cox-F [5′-ACT CAA CGC ACT GGA ACC-3′] and Cox-R [5′-TAG CTG AAG CCA ATT CGC-3′]). PCR amplification was performed in an ABI 9700 thermal cycler (Applied Biosystems, Foster City, CA) programmed for a three-step PCR procedure as previously described (23). In brief, 50 μl of the PCR mixture contained the following: 1× buffer; 1.5 mM MgCl₂; 10% glycerol; 200 μM (each) dATP, dCTP, and dGTP; 100 μM dTTP; 90 μM dUTP; 10 μM digoxigenin-11-dUTP (Roche Biochemicals, Indianapolis, IN); 1 μM of each primer; 0.01 unit/μl uracil N-glycosylase (Epicentre Technologies, Madison, WI); 0.025 unit/μl AmpliTaq gold DNA polymerase (Applied Biosystems); and 25 μl of nucleic acid extract. Two monoplex PCRs were used to separately amplify the Bartonella species and C. burnetii. For the dPCR procedure, the two primer sets were added to one single tube to simultaneously amplify the three bacterial pathogens.

Enzyme immunoassay (EIA) detection and differentiation.

Amplification products were then identified by detecting digoxigenin-labeled PCR products with a PCR enzyme-linked immunosorbent assay kit (Roche Biochemicals) in a 96-well microtiter plate as previously described (23). Three species-specific 5′-biotinylated capture probes (5′-ACG CTT GAA AAT ATT GCT CTA-3′ for B. henselae, 5′-GCA CTT GAA AAT ACT GCT CTG-3′ for B. quintana, and 5′-ATC ACC ACG GAA AAC ACC-3′ for C. burnetii) were applied to detect and differentiate the amplification product. Output signals were measured at an optical density of 450 nm (OD₄₅₀) and 490 (OD₄₉₀). A positive result was defined as an OD₄₅₀ − OD₄₉₀ value greater than or equal to 0.1 as previously described (23).

Human β-actin gene real-time TaqMan PCR.

The human β-actin gene was amplified as a “housekeeping” gene for each sample extract as an internal amplification control. The real-time TaqMan PCR assay was performed on the 7700 ABI Prism sequence detector (Applied Biosystems, Foster City, CA) as described previously (17). The primers and fluorophore hydrolysis probes for human β-actin gene and real-time PCR protocol were published previously (17).

Evaluation standard.

Results from the 16S rRNA gene amplification and sequencing of extracted DNA from the FFPE blocks were used as the evaluation standard. PCR amplification of the first 500 bp of the 16S rRNA gene was performed using the MicroSeq 500 16S bacterial sequencing kit according to procedures described previously (22, 24). Bidirectional sequences of the PCR amplification product were determined, and a phylogenetic analysis was performed by online analysis at the Ribosomal Database Project II site (http://wdem.nig.ac.jp/RDP/html/index.html ) and the MicroSeq Database Library (Applied Biosystems, Foster City, CA).

The dPCR assay was designed to amplify two targets, a 258-bp Bartonella species-specific gltA gene and a 256-bp C. burnetii-specific IS111 gene, in one PCR tube. Three microwells were used in the subsequent EIA detection procedure to detect and differentiate the three pathogens, B. quintana, B. henselae, and C. burnetii. The entire procedure, from specimen processing to result reporting, can be completed within 6 hours.

Experiments were performed to determine the assay analytical sensitivity by testing three recombinant plasmid standards (pCT-Bq, pCT-Bh, and pcT-Cb) spiked with pooled pathogen-free FFPE tissues. Plasmids from 0 to 10,000 copies/reaction were included in the experiment, and each dilution was tested three times. The analytic sensitivities of the dPCR assay for all plasmids into which a B. quintana, B. henselae, or C. burnetii amplification fragment was inserted were the same at 10 copies per reaction, which was equivalent to 80 copies/gram of tissue (Table 1) . The analytical sensitivities of the dPCR assay were equivalent to the test run in a monoplex PCR format (Table 1).

A total of 17 surgical specimens collected from patients (10 males and 7 females, ages 11 to 81 years) with a clinical diagnosis of culture-negative endocarditis who received valve replacements were included in the study. The specimens included samples from three prosthetic, eight FFPE, and six fresh/frozen heart valves. The human β-actin gene was amplified in each nucleic acid extract, indicating that total inhibition had not occurred in the nucleic acid amplification reaction.

Among 17 specimens collected from patients with a clinical diagnosis of culture-negative endocarditis, 2 (11.8%), 4 (23.5%), and 2 (11.8%) were positive for B. quintana, B. henselae, and C. burnetii, respectively, by the dPCR assay. The results matched those obtained by bacterial 16S rRNA gene amplification and sequencing, which was used as the validation standard. In nine dPCR-negative specimens, 16S rRNA gene amplification and sequencing identified four additional bacterial pathogens, including Haemophilus parainfluenzae, Streptococcus dysgalactiae, Streptococcus gordonii, and Streptococcus sanguinis (Table 2). The dPCR assay provided 100% sensitivities/specificities in the detection of B. quintana, B. henselae, and C. burnetii in surgical heart valve specimens.

In this study, a dPCR assay was developed to simultaneously detect and differentiate between B. quintana, B. henselae and C. burnetii. The assay possesses the following characteristics: (i) involvement of microscopic findings to localize the most suspicious specimens for processing; (ii) amplification of two targets simultaneously in a single reaction; (iii) incorporation of an additional signal amplification in the EIA detection procedure, providing higher sensitivities than other test formats, such as real-time TaqMan PCR; (iv) maintenance of species-level specificity with an internal probe incorporated into the EIA detection procedure; (v) nonbinding of the reporter molecule to the microwell with the signal-to-noise ratio of the assay up to 1,000, providing an objective qualitative cutoff value; and (vi) avoidance of carryover contamination by uracil N-glycosylase-based chemical modification.

The previous molecular protocols that detected bacterial pathogens in surgical heart valves have been based mainly on amplification and sequencing, if detected, of the universal bacterial 16S rRNA gene (1, 3, 6, 12, 14, 15, 18). However, this broad-range procedure suffered from high numbers of false-positive results and indeterminate results due to indigenous and exogenous bacterial genome contamination, limiting its use in clinical diagnostic services. Besides airborne and carryover contamination during the specimen collection/processing and reaction setup, the assays may be severely compromised by DNA backgrounds in PCR reagents, including the polymerizing enzymes, nucleotides, primers, buffer, and water (2, 7, 11). This time-consuming, multistep amplification and sequencing procedure makes quality assurance and quality control extremely difficult when used in routine diagnostic services. In addition, the coexistence of multiple bacterial pathogens in heart valve specimens may result in indeterminate results (mixed sequencing) by the universal process. In this study, I focused on targeting organism-specific genomes (10, 19) and developed a dPCR assay to simultaneously detect B. quintana, B. henselae, and C. burnetii, the three most encountered bacterial organisms that cause culture-negative endocarditis (4, 21).

Serologic testing remains the test of choice for laboratory confirmation of culture-negative endocarditis, including those caused by B. quintana, B. henselae, and C. burnetii. While clinical presentations of Bartonella- and Coxiella-caused endocarditis may be similar, distinguishing between these organisms is of importance, because optimal therapy and duration of treatment are different for these organisms. Serological cross-reactions between B. quintana, B. henselae, and C. burnetii have been reported (8, 16). We reported a case of B. quintana endocarditis with positive serology for C. burnetii, demonstrating the difficulty in using routine serology alone to make a diagnosis of Bartonella- or Coxiella-caused endocarditis (20). The dPCR assay described here detects and differentiates Bartonella or Coxiella species simultaneously, providing a useful distinguishing/confirmatory test, especially for anatomic pathology consultation services.