A real-time PCR assay with improved specificity for detection and discrimination of all clinically relevant Bordetella species by the presence and distribution of three Insertion Sequence elements

Background In Dutch laboratories molecular detection of B. pertussis and B. parapertussis is commonly based on insertion sequences IS481 and IS1001, respectively. Both IS elements are more widely spread among Bordetella species. Both Bordetella holmesii, and B. bronchiseptica can harbour IS481. Also, IS1001 is found among B. bronchiseptica. IS481, and IS1001 based PCR thus lacks specificity when used for detection of specific Bordetella spp. Findings We designed a PCR based on IS1002, another IS element that is present among Bordetella species, and exploited it as a template in combination with PCR for IS481, and IS1001. In combining the PCRs for IS481, IS1001, and IS1002, and including an inhibition control, we were able to detect and discriminate all clinically relevant Bordetella species. Conclusions We developed an improved PCR method for specific detection of B. pertussis, B. parapertussis, B. holmesii, and B. bronchiseptica.


Background
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][3][4][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][9][10][11][12][13][14]. Most PCRs are based on detection of the insertion sequence elements IS481, and IS1001, because they exist in multiple copies in the chromosome. A high copy number of PCR target, contributes to the sensitivity of detection. IS481 (1053 bp) shows a high degree of homology between different members of Bordetella species of 96%. All IS1001 (1306 bp) sequences known in Genbank are 100% homologues, as is IS1002 (1040 bp). The degree of homology between IS481, IS1001, and IS1002 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, IS1001 is found in all B. parapertussis, but IS1002 is found only in B. parapertussis that infect humans and is absent from B. parapertussis sheep isolates [15]. B. pertussis harbors both IS481 and IS1002 while some B. bronchiseptica strains may have either IS481 or IS1001 [15]. B. holmesii only has IS481.
Due to the distribution of IS481, and IS1001, 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, IS1002, as target in PCR. IS1002, in addition to IS481, and IS1001, which enables discrimination of B. pertussis, B. holmesii, B. parapertussis, and B. bronchiseptica.

Methods
IS1002 has not been exploited before as a template in PCR detection. We developed a specific IS1002 PCR (Table 1) to improve our ability to recognize the correct Bordetella species, and to combine it with IS481, and IS1001 specific PCRs. Addition of Phocine Herpes Virus (PhHV) as internal control acts to monitor the extraction as well as the efficiency of amplification [17].
We investigated standard laboratory strains of Bordetella (kindly provided by Dr. Frits Mooi and Kees Heuvelman, Laboratory for Vaccine Preventable Diseases, National Institute of Public Health and the Environment, Bilthoven, The Netherlands), which are shown in Table 2. Since detection of B. pertussis is of highest concern, we investigated the performance of the newly developed PCR on 100 clinical samples that were previously positive. To verify the specificity of the PCR we investigated 20 clinical respiratory tract samples that were suspect for other pathogens than Bordetella. Prior to PCR, laboratory strains were diluted and boiled to release DNA (equivalent to approximately 5 cells/μl). Clinical samples were extracted using EasyMAG (Biomerieux, Grenoble, France).
PCRs for detection of IS481, IS1001, and IS1002 were performed in a reaction mixture of 25 μl containing 0.5 μM of IS481 and IS1001 primers, 0.8 μM and 0.6 μM of IS1002 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 IS481, IS1001, IS1002, 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.

Evaluation of sensitivity
In order to evaluate the performance of the combination of all four PCRs in a multiplex format performed on standard laboratory strains (Table 2): IS481, IS1001, IS1002, and the internal control, we compared Ct values to each single PCR. No significant differences were found (not shown). Since IS481 is present with a much higher copy number in B. pertussis than IS1002 (approximately 200 and 10 copies, respectively), Ct values of IS481 were on average 3.7 lower compared to Ct values of IS1002 detection. Of a 10 CFU/ml dilution of B. pertussis cells both IS481 and IS1002 were positive with Ct values of respectively 30.6 and 34.4. The sensitivity of B. parapertussis detection was comparable with that using only IS1001 as target.

Evaluation of specificity
To evaluate the specificity of each IS-based PCR, laboratory strains of most known Bordetella species (with exception of B. avium and B. trematum; Table 2), were subjected to the IS-based PCRs. Results were as was expected (Table 2).
To investigate whether all targets were amplified with high fidelity, and would not confound the specificity of detection, we made serial dilutions and subjected these to PCR. With limiting concentrations of B. pertussis bacteria, detection of IS1002 is lost before detection of IS481, because of the difference in copy number.

Discussion
In this study we included IS1002 as target in PCR based detection of Bordetella, in addition to IS481 and IS1001, which is the commonly used PCR in The Netherlands. We aimed to improve the specificity of PCR because the addition of IS1002 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 IS481 and this accounts for the large proportion of false positive results. Indeed, IS481 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 IS481 is very rarely found among B. bronchiseptica (1% of strains), in contrast to IS1001 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 IS481 cannot be discriminated from B. holmesii using IS-based PCRs. If PCR presents a single IS481 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 (IS481) and B. parapertussis (IS1001), we did not find more than one Bordetella species present in clinical specimens.
During the evaluation, we sometimes observed a weak positive signal for IS481 (Ct > 37) from B. hinzii, although high concentrations of this organism were negative in PCR. This may indicate that B. hinzii might contain an IS481-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 IS481, IS1001 and IS1002 for detection and discrimination of the clinically relevant Bordetella species.

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