Rapid Typing of Bordetella pertussis Pertussis Toxin Gene Variants by LightCycler Real-Time PCR and Fluorescence Resonance Energy Transfer Hybridization Probe Melting Curve Analysis

Abstract

A LightCycler real-time PCR hybridization probe assay was developed for rapid typing of gene variants of the Bordetella pertussis virulence factor pertussis toxin. The assay correctly identified the ptxS1 alleles of all strains tested, comprising 57 Finnish clinical isolates and 2 vaccine strains. The method is simple, reliable, and suitable for large-scale screening of B. pertussis strains.

Pertussis (whooping cough) is one of the 10 most common causes of death from infectious disease worldwide (15). Vaccination is the most effective method for prevention and control of pertussis and has been used successfully for decades. Bordetella pertussis, the causative agent of pertussis, produces various virulence factors, of which fimbriae (2, 23), pertactin (PRN) (7, 8), and pertussis toxin (PT) (24, 25) induce protective immunity and are included in the new acellular pertussis vaccines.

During the past 10 years, resurgence of pertussis has been seen in several countries with high vaccination coverage (5, 13, 17). One explanation for the increasing incidence is the antigenic divergence between the vaccine strains and the circulating strains. Recently, antigenic variation of PRN and PT has been reported in Europe and the United States (5, 10, 12, 19, 20). PT is composed of five subunits; the toxic functions are located in the S1 subunit, which consists of 269 amino acids (21, 22). The polymorphism in the PT operon is mainly found in the gene encoding the S1 subunit. To date, four alleles have been described, namely, ptxS1A, ptxS1B, ptxS1D, and ptxS1E (Fig. 1) (19, 20). The strains used for whole-cell or acellular vaccines harbor either the ptxS1B or the ptxS1D allele, whereas most of the strains currently circulating harbor the ptxS1A allele. Only a few strains have been found to harbor the ptxS1E allele. The biological significance of the observed variation is not known and remains to be elucidated. In the United States, three ptxS1 alleles were identified when the ptxS1 genes of 152 B. pertussis strains isolated from 1935 through 1999 were sequenced. The vaccine alleles ptxS1B and ptxS1D were prevalent in the United States from 1935 through 1974. However, the ptxS1A allele replaced the vaccine alleles and has been the most prevalent allele from 1975 to 1987 and 1989 to 1999 (64 and 78%, respectively) (5). Similar trends with respect to the ptxS1 gene have also been observed in Europe (19, 20).

FIG. 1.

Polymorphic nucleotide sequences of the ptxS1 subunit. For ptxS1A, ptxS1B, and ptxS1E, only differences from ptxS1D are indicated. Dots indicate identical bases. Numbers refer to positions in the ptxS1D allele. The nucleotide differences that are used as targets for hybridization probes are boldfaced.

So far, the only method for determining PT variants has been PCR-based sequencing, a relatively time-consuming and expensive method (18). To combat pertussis and to design more-effective vaccines, the variation of virulence factors of the B. pertussis strains circulating in the population has to be monitored. In this study, we developed a LightCycler real-time PCR assay combined with fluorescence resonance energy transfer (FRET) hybridization probe melting curve analysis to characterize the ptxS1 alleles. Strains with the prevalent allele ptxS1A were first differentiated from strains harboring the ptxS1B, ptxS1D, or ptxS1E allele (Fig. 2). Then the strains with the ptxS1E allele were further distinguished from the strains with the vaccine type alleles ptxS1B and ptxS1D.

FIG. 2.

Schematic diagram of the work flow used in this study to determine the ptxS1 alleles.

A total of 57 Finnish B. pertussis isolates, 2 Finnish vaccine strains (1772 and 18530), and 1 reference strain (18323) selected from the strain collection of the Finnish Pertussis Reference Laboratory, National Public Health Institute, Turku, Finland, were used to evaluate the assay. The Finnish clinical isolates harbored either ptxS1A (n = 50) or ptxS1B (n = 7). They were isolated from Finnish patients during the years 1956 to 1996. Finnish vaccine strains 1772 and 18530 harbored ptxS1B (n = 1) and ptxS1D (n = 1), respectively. Strain 18323 (AJ132095) harbored the ptxS1E allele. The ptxS1 genes of all the strains included in this study have been sequenced previously (19). The isolates had originally been identified by standardized methods as previously described (14). Briefly, culture plates were incubated in a humid atmosphere at 35°C and monitored daily for 7 days. Suspected colonies were Gram stained and tested by slide agglutination with antisera to B. pertussis (Murex Diagnostics, Dartford, England). The identities of the B. pertussis strains were confirmed by gas-liquid chromatography. For this study, the bacteria were cultivated on Regan-Lowe medium containing charcoal agar and defibrinated horse blood at 35°C for 3 days. Bacterial colonies on the plates were harvested for isolation of DNA.

DNA was extracted from bacterial colonies by using the DNA Isolation Kit for Blood/Bone Marrow/Tissue (Roche Diagnostics GmbH, Mannheim, Germany) according to the manufacturer's instructions. DNA concentrations in all samples were measured with a GeneQuant spectrophotometer (Pharmacia Biotech, Piscataway, N.J.) and were adjusted to 3 ng/μl. DNA preparations were stored at −20°C.

PCR-based sequencing was carried out as described previously (20). For the LightCycler real-time PCR, primers and probes (Table 1) were designed on the basis of the published sequences of the ptxS1 alleles (Fig. 1) and synthesized by TIB Molbiol, Berlin, Germany. Primer pairs BPS/BPRmt1 and BPS/BPRmt2 define 190- and 188-bp gene segments, respectively. The region to which probes QJ 3 BDE and QJ 4 BDE bind contains a stem-loop (GCGC-GCGC) structure. To destroy the interfering loop, reverse primers BP Rmt1 and BP Rmt2 contain a C-to-T modification corresponding to base 685 on the ptxS1D allele (Fig. 1 and Table 1). Probes QJ 3 BDE and QJ 4 BDE are specific for the ptxS1B, ptxS1D, and ptxS1E alleles (Fig. 1) and therefore are capable of differentiating the currently circulating strains with the ptxS1A allele from the strains with other alleles. Probes QJ 5 E and QJ 6 E further differentiate ptxS1E from ptxS1B or ptxS1D (Fig. 1). Probes QJ 4 BDE and QJ 6 E function as the anchor probes in FRET; they are labeled with LightCycler Red 640 and LightCycler Red 705, respectively, at their 5′ ends and are phosphorylated at their 3′ ends. Probes QJ 3 BDE and QJ 5 E function as sensor probes in FRET and carry fluorescein labels at their 3′ ends.

TABLE 1.

Primers and probes used in this study

Primer or probe	Position	Sequence (5′ → 3′)a
Primers
BP S	514-535	CCC GAA AAC ATC CGC AGG GTA A
BP Rmt1	703-686	AAG CGC CCA CCA CCG GTG
BP Rmt2	701-685	GCG CCC ACC ACC GGT GC
Probes
QJ 3 BDE	690-670	CGG TGC CAT GCG CAC CAA TGT-X
QJ 4 BDE	668-643	LC-Red640-CCG ACG ATC GAC GCT ACG GAC CTT CG-p
QJ 5 E	575-597	CAC GGA GTA TCC CAA CGC TCG C-X
QJ 6 E	600-622	LC-Red705-CGT CAG CCA GCA GAC TCG CGC CA-p

^aX, fluorescein; LC-Red640, LightCycler Red 640 nm; p, phosphate; LC-Red705, LightCycler Red 705 nm. The C-to-T modification used to destroy the interfering loop in the target is boldfaced.

The amplification reactions and the FRET hybridization probe melting curve analysis were carried out in a fluorescent thermal cycler (LightCycler; Roche Diagnostics GmbH) by using a LightCycler FastStart DNA Master Hybridization Probes kit (Roche Diagnostics). The reaction conditions were optimized according to the manufacturer's protocol.

For differentiation of the ptxS1B, ptxS1D, and ptxS1E alleles from ptxS1A, the final reaction volume of 20 μl contained 2 μl of LightCycler-Fast Start DNA Master Hybridization Probes (containing Taq DNA polymerase, reaction buffer, and a mix of deoxynucleoside triphosphates) and 3 mM MgCl₂ (Roche); 4 pmol of the FRET probes QJ 3 BDE and QJ 4 BDE, 8 pmol of primers BP S and BP Rmt1, and 5% dimethyl sulfoxide (Merck, Darmstadt, Germany); and 2 μl of sample DNA. The amplification started with an initial denaturation step at 95°C for 10 min and continued with 40 cycles of denaturation at 95°C for 10 s, annealing at 55°C for 10 s, and extension at 72°C for 9 s. The temperature transition rate was 20°C/s. Fluorescence was measured at 640 nm (F2 channel) at the end of the annealing step of each cycle to monitor the accumulation of the PCR product. After amplification, a melting curve was acquired by heating the product 20°C/s to 95°C, cooling it to 50°C for 30 s, and then heating it 0.1°C/s to 94°C under continuous fluorescence measurement. Three positive controls with DNA isolated from strains with the ptxS1B, ptxS1D, and ptxS1E alleles and a negative control without template DNA were included in each run.

For differentiation of ptxS1E from ptxS1B or ptxS1D, the final reaction volume of 20 μl contained 2 μl of LightCycler-Fast Start DNA Master Hybridization Probes (containing Taq DNA polymerase, reaction buffer, and a mix of deoxynucleoside triphosphates) and 3 mM MgCl₂ (Roche); 4 pmol of the FRET probes QJ 5E and QJ 6E, 8 pmol of primers BP S and BP Rmt2, and 5% dimethyl sulfoxide (Merck); and 2 μl of sample DNA. The temperature profile of the real-time PCR included an initial denaturation step at 95°C for 6 min followed by 42 cycles of denaturation at 95°C for 10 s, annealing at 57°C for 10 s, and extension at 72°C for 9 s. The temperature transition rate was 20°C/s. Fluorescence was measured at 705 nm (F3 channel) at the end of the annealing step of each cycle to monitor the accumulation of the PCR product. After amplification, a melting curve was acquired by heating the product 20°C/s to 95°C, cooling it to 50°C for 35 s, and heating it 0.1°C/s to 94°C under continuous fluorescence measurement. A positive control with DNA isolated from the strain with the ptxS1E allele and a negative control without template DNA were included in each run.

Melting curve analysis was accomplished using Light Cycler Software, version 3.5.3, with automated fluorescence gain settings. When the temperature reached the melting temperature (T_m) of the probes, a rapid loss of fluorescence was observed as the two adjacently bound probes dissociated from their complementary targets and the FRET stopped. This was plotted as the negative derivative of fluorescence versus temperature to define the allele-specific melting curves (Fig. 3).

FIG. 3.

(A) Melting curves for the LightCycler QJ3 BDE and QJ4 BDE FRET hybridization probes. The specific melting peak can be observed only for the strains with the ptxS1B, ptxS1D, and ptxS1E alleles. (B) Melting curves for the LightCycler QJ5 E and QJ6 E FRET hybridization probes, demonstrating differentiation of the strains that harbor the ptxS1E allele from the vaccine type strains (ptxS1B and ptxS1D). A specific melting peak can be observed only for the strains with the ptxS1E allele.

A 20-μl volume of PCR product was removed from the capillary by reverse centrifugation into an empty Eppendorf tube. To confirm the specific amplification, a 10-μl volume was run (120 V for 1.5 h) in a 1.5% multipurpose agarose gel (Roche Diagnostics) together with a 100-bp DNA ladder (Amersham Pharmacia Biotech Inc., Piscataway, N.J.). After staining with ethidium bromide, the bands in the gel were visualized and photographed under UV light. To avoid PCR contamination, three separate rooms were used for preparing the PCR mixtures, performing the PCRs, and analyzing the PCR products.

The ptxS1 alleles of all 57 clinical isolates tested (50 with the ptxS1A allele and 7 with the ptxS1B allele), the 2 Finnish vaccine strains (ptxS1B or ptxS1D), and the 1 reference strain (ptxS1E) were found to be correctly identified by FRET hybridization probe melting curve analysis. Probes QJ 3 BDE and QJ 4 BDE correctly hybridized to and gave signals for PCR products derived from strains harboring the ptxS1B, ptxS1D, or ptxS1E allele (Fig. 3A). Probes QJ 5 E and QJ 6 E hybridized to and gave signals for PCR products derived from strains harboring the ptxS1E allele (Fig. 3B). The specific mean T_m for QJ 3 BDE and QJ 4 BDE was 66.43°C (standard deviation, 0.41°C). The specific mean T_m for QJ 5 E and QJ 6E was 66.43°C (standard deviation, 0.17°C). The intra-assay variation was 0.03°C for both sets of probes. Interassay variations were 0.29°C for QJ 3 BDE and QJ 4 BDE and 0.17°C for QJ 5 E and QJ 6 E. Some nonspecific binding of the probes was exposed in the melting curve analysis, but the T_m at which the probes dissociated from their noncomplementary targets was significantly lower (<59°C) than that at which they dissociated from the specific targets. The presence of the PCR amplification product was also confirmed by gel electrophoresis.

Real-time PCR combined with FRET hybridization probes proved to be a good alternative to sequencing for identification of the ptxS1 gene alleles of B. pertussis. The assay correctly identified all the PT gene variants. The assay is simple, rapid (it can be carried out within 2 h), and reliable. Further, the intra- and interassay variations in T_m were extremely low. The nonspecific binding of the probes was differentiated from the specific binding based on the T_m. The disadvantages of this method are that novel ptxS1 genotypes can be missed and the ptxS1 gene alleles of the two vaccine subtypes cannot be differentiated from each other.

The B. pertussis isolates harboring the ptxS1A allele are first differentiated from the strains harboring ptxS1B, ptxS1D, and ptxS1E. Although the probes did not give a signal for the PCR products derived from strains with the ptxS1A allele, specific amplification products were seen with all of the B. pertussis strains studied (confirmed by gel electrophoresis).

The frequency of B. pertussis strains harboring the vaccine type ptxS1B or ptxS1D allele is decreasing with time in countries with high vaccination coverage. In the United States, the frequency of strains with vaccine alleles was 36% in 1975 to 1987 and 22% in 1989 to 1999 (5). A similar frequency trend has been observed in The Netherlands (20). The method described here is suitable for typing of PT alleles of B. pertussis clinical isolates. Strains with the one of the vaccine type alleles, ptxS1B and ptxS1D, or with ptxS1E can be detected by running the two hybridization probe assays, and the results can be obtained within 2 h.

The Centers for Disease Control and Prevention have classified pertussis as one of the reemerging infectious diseases (6). The incidence of pertussis has increased in many countries with high vaccination coverage, such as Australia, Canada, France, The Netherlands, and the United States (1, 3, 4, 9). There are several explanations for the resurgence of pertussis in vaccinated populations, such as improved surveillance, more-sensitive diagnostic methods, waning vaccine-induced immunity, and the adaptation of the bacterial population to vaccine-induced immunity. In the countries where pertussis notification rates have increased, an antigenic drift in PT and PRN has been observed and the vaccine type strains have been replaced by the antigenic variants (5, 11, 12, 19, 20). In Finland, 100% of the clinical strains isolated in the 1990s harbor the ptxS1A allele, whereas the strains isolated before the 1970s exclusively harbor the vaccine type allele ptxS1B (19). Whether this drift is caused by long-term pertussis vaccination remains to be elucidated. It is important not only to monitor the situation in highly vaccinated populations but also to study the strains currently circulating in countries with low vaccination coverage. For large-scale screening of B. pertussis isolates, tools that are more rapid and less laborious than sequencing are needed. We have previously used LightCycler real-time PCR and hybridization probes for typing of B. pertussis PRN gene variants (16). This assay, together with the method described in this study, makes possible rapid identification of the gene variants of both PT and PRN, two important antigens of B. pertussis.