Novel Duplex Real-Time PCR Assay Detects Bordetella holmesii in Specimens from Patients with Pertussis-Like Symptoms in Ontario, Canada

Abstract

Bordetella holmesii is a human pathogen found mainly in immunocompromised patients. A specific real-time PCR assay was developed and successfully used to identify specimens from which B. holmesii was misidentified as Bordetella pertussis and to establish the prevalence of B. holmesii in Ontario patients with pertussis-like symptoms.

The Gram-negative bacterium Bordetella holmesii was first described in 1995. It is known to cause disease (e.g., septicemia) in patients with serious underlying medical conditions (2, 5, 7, 12, 14) and has also been detected in nasopharyngeal specimens from patients with a pertussis-like illness (6, 15). Two insertion sequences, IS481 and IS1001, are commonly used PCR targets for the diagnosis of Bordetella pertussis and Bordetella parapertussis, respectively (4, 11). Since both targets are also present in B. holmesii, disease due to this pathogen may be mistakenly attributed to other Bordetella species (8, 9). The aim of this study was to develop a duplex real-time PCR assay able to detect and discriminate between B. pertussis and B. holmesii. Additionally, 1,775 specimens positive for IS481 and/or IS1001 collected during 2007 and 2008 were retrospectively tested with the new assay to determine the prevalence of B. holmesii in Ontario, Canada.

To date, the B. holmesii genome has not been fully sequenced, and thus, it is challenging to identify novel targets for specific PCR identification of this organism. The housekeeping gene recA has been used in previous studies for detection of B. holmesii (1, 13) and was selected as a PCR target for this study. Specific detection of B. holmesii was accomplished by designing a real-time PCR assay targeting a 50-bp segment of the recA gene (GenBank accession no. AF399664) that is polymorphic between B. holmesii, B. pertussis, B. parapertussis, Bordetella bronchiseptica, Bordetella hinzii, and Bordetella avium (Fig. 1).

FIG. 1.

Multiple sequence alignment of a 50-bp segment of the recA gene for several Bordetella species. Real-time PCR primers designed for this study are underlined, and the TaqMan MGB probe is highlighted.

To confirm the conservation of B. holmesii recA (BHrecA) among clinical isolates, the full gene from 8 B. holmesii strains cultured from patients was amplified and sequenced. Sequencing revealed that BHrecA was 100% conserved in the 8 isolates tested.

Analytical sensitivity of the assay was determined using a dilution series (10-fold dilutions, 5.5 ng to 5.5 × 10⁻¹⁰ ng) of BHrecA cloned into pCR2.1 (Invitrogen, Carlsbad, CA). Each dilution was tested in triplicate, and the experiment was performed 3 separate times to ensure accuracy of the results. The assay found a 100% probability that detection of 55 ag of cloned material (equal to 2 copies of recA) was possible at a threshold cycle (C_T) of 38.3 ± 1.0. Furthermore, specificity of the real-time PCR assay was verified by testing purified genomic DNA from 32 different bacterial, fungal, and viral pathogens, including B. pertussis (n = 54), B. parapertussis (n = 32), B. bronchiseptica (n = 1), B. hinzii (n = 1), Acinetobacter sp., Alcaligenes faecalis, Alcaligenes xylosoxidans, Achromobacter piechaudii, Corynebacterium diphtheriae, Enterococcus faecalis, Enterococcus faecium, Enterococcus durans, Enterococcus hirae, Enterococcus gallinarum, Enterococcus avium, Enterococcus raffinosus, Escherichia coli, Haemophilus influenzae, Haemophilus parainfluenzae, Helicobacter pylori, Legionella pneumophila, Mycobacterium tuberculosis, Mycoplasma pneumoniae, Neisseria meningitidis (A, B, C, W, X, Y, and Z), Pseudomonas aeroginosa, Ralstonia eutropha, Ralstonia sp., Staphylococcus aureus, Streptococcus pyogenes, Candida albicans, Herpes simplex virus 1, and Herpes simplex virus 2. The BHrecA primers and probe successfully detected DNA from B. holmesii and not from any other Bordetella species or other pathogens tested.

To develop our novel duplex real-time PCR assay for the simultaneous detection of B. pertussis and B. holmesii, the BHrecA primers and probe were combined with a currently used IS481 real-time PCR. To ensure that sensitivity was not significantly affected by combining the targets, simplex IS481 and simplex BHrecA reactions were carried out and validated alongside the duplex assay. IS481 real-time PCR was carried out using the published primers and probe of Kösters et al. (4), modified by using 3′ BHQ-1 for the probe (IDT, Coralville, IA). The oligonucleotide primers BHrecA_fwd (5′-CGGTTCGCTGGGTCTCG-3′) and BHrecA_rev (5′-CCCGCGGCAGACCAC-3′) were used in this study for real-time PCR detection of B. holmesii. The TaqMan MGB probe was designed with a NED fluorescent reporter on the 5′ end and nonfluorescent black hole quencher at the 3′ end with the sequence 5′-CATCGCATTGGGCG-3′ (Applied Biosystems, Foster City, CA). All primers were purchased from Invitrogen.

The duplex PCR master mix consisted of 1× TaqMan gene expression master mix (Applied Biosystems), 0.8 μM (each) BHrecA forward and reverse primers, 0.4 μM BHrecA probe, 1 μM (each) IS481 forward and reverse primers, 0.25 μM IS481 probe, 5 μl of template, and enough sterile nuclease-free water to bring the total reaction volume to 20 μl. The simplex BHrecA PCR master mix followed the same recipe with the elimination of IS481 primers and probe. The currently used simplex IS481 PCR master mix used 0.5 μM (each) IS481 forward and reverse primers and 0.125 μM IS481 probe. The samples were subjected to an initial amplification cycle of 50°C for 2 min and 95°C for 15 min, followed by 45 cycles at 94°C for 15 s and 60°C for 1 min. Five microliters of cloned BHrecA at 5 fg/μl was used as a positive PCR control; the negative control was 5 μl of sterile H₂O. Amplification, detection, and data analysis were performed with an Applied Biosystems 7900HT real-time PCR system and the SDS v2.3 software program. A blinded set of 117 patient specimens, previously confirmed by culture, was used to validate the IS481/BHrecA duplex PCR assay, with 100% concordance found in detecting B. pertussis and B. holmesii. It was found that the duplex values for IS481 averaged 1.1 C_Ts lower than those for simplex IS481 reactions. Furthermore, the duplex values for BHrecA were on average 2.4 C_Ts lower than those for simplex BHrecA reactions. Therefore, duplex reactions were more sensitive than simplex real-time PCR and could be used for routine diagnostics.

For detection of B. holmesii in clinical patient specimens, nasopharyngeal samples were obtained using Dacron tipped swabs (Bio Nuclear Diagnostics Inc., Toronto, CA) and transported in 500 μl of phosphate-buffered saline (PBS). Patient swabs from suspected B. pertussis cases were handled by the Respiratory & Legionnaires' section at the Central Public Health Laboratory (CPHL), where they were all processed for culture and then subjected to rapid boiling extractions for real-time PCR amplification using the targets IS481 and IS1001, as well as an extraction and PCR control (human RNase P gene), the results of which can be seen in Table 1 . PBS was found to have no inhibitory effect on the PCRs. A total of 1,775 Bordetella-positive patient specimens collected in 2007 and 2008 were retrospectively tested using the duplex real-time PCR assay to determine the prevalence of Bordetella holmesii in patients with suspected pertussis infections. All B. holmesii culture-positive specimens (n = 8) were positive using the real-time PCR duplex assay, and 4 additional specimens were identified as B. holmesii based on the duplex assay.

TABLE 1.

Results of real-time PCR using BHrecA, IS481, and IS1001 for patient specimens collected over a 2-year period^a

Target	Organism detected	2007 specimens			2008 specimens
		No. of specimens	% of total Bordetella-positive specimens	% of total specimens	No. of specimens	% of total Bordetella-positive specimens	% of total specimens
IS481	B. pertussis	883	95.4	17.7	824	97.1	15.6
IS1001	B. parapertussis	26	2.8	0.5	18	2.1	0.3
IS481 and IS1001	B. pertussis and B. parapertussisb	8	0.9	0.2	4	0.5	0.5
IS481 and BHrecA	B. holmesii	9c	1.0	0.2	3d	0.4	0.1
Total Bordetella positive		926		18.6	849		16.5
Total specimens received		4,983			5,271

^aBordetella species were differentiated based on the detection of IS481 without detection of BHrecA (B. pertussis) or detection of both IS481 and BHrecA (B. holmesii).

^bCoinfection.

^cSix were culture positive.

^dTwo were culture positive.

To further confirm that specimens positive for BHrecA were not dual infections with B. pertussis, a second real-time PCR was run simultaneously using the BP3385 target, which is specific for B. pertussis (3). It was found that none of these BHrecA-positive specimens were detected by BP3385.

Consistent with a study performed in the United States (15), the prevalence of B. holmesii in Ontario appears to be relatively low, with positivity in less than 1% of all nasopharyngeal swabs collected. However, B. holmesii is indeed present and associated with pertussis-like symptoms in patients. A key point for clinicians and public health workers elucidated in this study is that use of IS481 for the detection of B. pertussis should be done with the awareness of its cross-reactivity, since four previously misidentified specimens were detected using the B. holmesii-specific real-time PCR assay. At present, the clinical impact of misdiagnosing a B. holmesii infection as B. pertussis is unknown. The limited data available indicate B. holmesii is susceptible to a number of antibiotics (5, 12). However, it has not been established if erythromycin, the most commonly used drug for treatment of pertussis, is effective against B. holmesii, and MIC testing has suggested that erythromycin is less active against B. holmesii than against B. pertussis (10). Therefore, correct identification of this pathogen is important for active surveillance of Bordetella holmesii as an infectious agent and for greater determination of its role as a respiratory pathogen.