The genus Bordetella is comprised of 8 species, 4 of which are known to infect humans; B. pertussis, B. parapertussis, B. holmesii, and B. bronchiseptica. The most important cause for whooping cough is B. pertussis, followed by B. parapertussis. Bordetella holmesii was first described in 1995 [1], and has since been isolated from patients with a serious underlying disease [2–5]. B. bronchiseptica is usually restricted to animals but occasionally is isolated from immunocompromised patients [6, 7].

A large number of PCR (polymerase chain reaction) assays have been described for detection of Bordetella pertussis and B. parapertussis, and more recently B. holmesii[8–14]. Most PCRs are based on detection of the insertion sequence elements IS 481, and IS 1001, because they exist in multiple copies in the chromosome. A high copy number of PCR target, contributes to the sensitivity of detection. IS 481 (1053 bp) shows a high degree of homology between different members of Bordetella species of 96%. All IS 1001 (1306 bp) sequences known in Genbank are 100% homologues, as is IS 1002 (1040 bp). The degree of homology between IS 481, IS 1001, and IS 1002 is less than 5%.

The insertion sequences (IS) that are present in Bordetella spp., are distributed according to species and/or host specificity. For example, IS 1001 is found in all B. parapertussis, but IS 1002 is found only in B. parapertussis that infect humans and is absent from B. parapertussis sheep isolates [15]. B. pertussis harbors both IS 481 and IS 1002 while some B. bronchiseptica strains may have either IS 481 or IS 1001 [15]. B. holmesii only has IS 481.

Due to the distribution of IS 481, and IS 1001, PCRs for B. pertussis and B. parapertussis lack specificity [16]. In recent years many newly developed PCRs were introduced with improved specificity but often with a compromise to sensitivity [11, 13, 14].

The aim of this study was to improve the specificity of PCR by including another IS element, IS 1002, as target in PCR. IS 1002, in addition to IS 481, and IS 1001, which enables discrimination of B. pertussis, B. holmesii, B. parapertussis, and B. bronchiseptica.

PCRs for detection of IS 481, IS 1001, and IS 1002 were performed in a reaction mixture of 25 μl containing 0.5 μM of IS 481 and IS 1001 primers, 0.8 μM and 0.6 μM of IS 1002 Forward and Reverse primer, respectively, and 0.2 μM of PhHV primers. Probes (Table 1) were added with concentrations of 0.14, 0.14, 0.16, and 0.08 μM for respectively IS 481, IS 1001, IS 1002, and the internal control in PCR reaction mix (Sigma -Aldrich (E3004), Munich, Germany). Nine μl of template DNA was added. Amplification was carried out on an ABI 7500 Real-Time PCR system (Applied Biosystems (ABI), Nieuwerkerk a/d IJsel, The Netherlands). The temperature profile included initial denaturation of 4 min. at 94°C, followed by 50 cycles of 94°C for 15 sec., and 60°C for 1 min. Cycle treshold (Ct) values were determined automatically using the ABI SDS software.

In this study we included IS 1002 as target in PCR based detection of Bordetella, in addition to IS 481 and IS 1001, which is the commonly used PCR in The Netherlands. We aimed to improve the specificity of PCR because the addition of IS 1002 enables the discrimination of B. pertussis, B. parapertussis, B. holmesii, and B. bronchiseptica.

In the final report of QCMD (Pierard, D., O. Soetens, and G. Ieven. Bordetella pertussis (BPDNA09) EQA Pilot Study. February 2010. QCMD, Glasgow UK.), a high degree of false positivity was observed for detection of B. pertussis. It appeared that more than 80% of laboratories contributing to the study used a PCR that is based on IS 481 and this accounts for the large proportion of false positive results. Indeed, IS 481 positive samples could either be positive for B. pertussis, B. holmesii, or B. bronchiseptica. Here, we have shown that in our PCR assay these organisms can be discriminated from one another. There are however some limitations. In a previous study [15] it was shown that IS 481 is very rarely found among B. bronchiseptica (1% of strains), in contrast to IS 1001 that was found in approximately 50% of the studied strain collection. Consequently, approximately 50% of B. bronchiseptica carry no known IS elements and thus cannot be detected by IS-based PCR. B. bronchiseptica strains that contain a copy of IS 481 cannot be discriminated from B. holmesii using IS-based PCRs. If PCR presents a single IS 481 positive signal, one might re-investigate the sample using the B. holmesii specific PCR targeting the recA gene that was described earlier [18] to discriminate from B. bronchiseptica.

The performance of our assay may fail if clinical samples might contain more than one Bordetella species. However, after more than 15 years of experience with PCR detection of B. pertussis (IS 481) and B. parapertussis (IS 1001), we did not find more than one Bordetella species present in clinical specimens.

During the evaluation, we sometimes observed a weak positive signal for IS 481 (Ct > 37) from B. hinzii, although high concentrations of this organism were negative in PCR. This may indicate that B. hinzii might contain an IS 481-like sequence that does not exactly match our PCR. As B. hinzii is solely confined to birds, this finding will not confound PCR based on IS 481, IS 1001 and IS 1002 for detection and discrimination of the clinically relevant Bordetella species.

In conclusion, we have developed a real time PCR with improved specificity for detection and discrimination of all clinically relevant Bordetella species.