Investigating genome reduction of Bordetella pertussis using a multiplex PCR-based reverse line blot assay (mPCR/RLB)
© Lam et al.; licensee BioMed Central Ltd. 2014
Received: 15 July 2014
Accepted: 2 October 2014
Published: 15 October 2014
The genetic composition of the bacterium causing whooping cough, Bordetella pertussis, has been investigated using microarray studies in order to examine potential genetic contributors to the disease re-emergence in the past decade. Regions of difference (RDs) have been previously identified as clusters of genes flanked by insertion sequences which are variably present in different sets of isolates, and have also been shown to be potential markers of B. pertussis evolution.
This study used microarray data to identify and select a panel of RDs; primers and probes for these RDs were then designed to test for the presence or absence of these regions in a novel and less expensive multiplex PCR-based reverse line blot (mPCR/RLB) assay. By comparing the presence or absence of RDs, we aimed to determine the genomic variability of a diverse collection of B. pertussis strains and how they have changed over time.
A B. pertussis specific mPCR/RLB using 43 genes representing 30 RDs, was developed and used to characterise a set of 42 B. pertussis isolates. When mapped against the previously identified evolutionary relationships of the strains, the losses of two RDs - BP0910A - BP00930 and BP1948-BP1962 - were found to be associated with significant events in B. pertussis history: the loss of BP0910A - BP00930 coincided with introduction of whole cell vaccines in the 1950s while that of BP1948-BP1962 occurred after the introduction of acellular vaccines. The loss of BP1948-BP1962 also coincided with expansion of the most recent B. pertussis strains.
The mPCR/RLB assay offers an inexpensive and fast method of determining the gene content of B. pertussis strains and also confirms that gene losses are an ongoing feature of B. pertussis evolution.
Bordetella pertussis is a highly homogenous organism which diverged from a Bordetella bronchiseptica-like ancestor 0.7-3.5 million years ago . Although a high degree of sequence similarity exists between B. pertussis, B. bronchiseptica and Bordetella parapertussis, it has been recognised that significant losses of groups of genes, termed regions of difference (RDs), have occurred during B. pertussis evolution . Comprehensive studies using comparative genome hybridisation (CGH) microarray experiments have investigated such changes to the B. pertussis genome and correlated the emergence of currently circulating isolates with the absence of specific RDs [3–5].
While CGH is a method that allows genome comparison amongst large numbers of isolates, it is costly and labour-intensive when screening large numbers of samples. In contrast, multiplex PCR (mPCR)-based reverse line blot (RLB) assay is an established method with diverse range of applications in pathogen detection and typing . The RLB used in this assay is a DNA macroarray, which relies on attachment of biotinylated PCR-amplified products to specific amine-labeled DNA probes. The assay allows multiple isolates to be processed in a single blot and, by using a multiplex approach, more that 40 targets can be probed simultaneously without increasing the number of individual PCRs . This technique has been successfully applied to genotyping of bacterial pathogens of public health importance, such as Streptococcus agalactiae, Streptococcus pneumoniae, uropathogenic Escherichia coli, Staphylococcus aureus, methicillin resistant Staphylococcus aureus as well as different viral pathogens including human papilloma virus (HPV)  and human adenoviruses .
We previously used single nucleotide polymorphisms (SNPs) to characterise an international collection of over 300 B. pertussis isolates. These isolates were differentiated into 42 unique SNP profiles (SPs) and six distinct clusters (clusters I to VI) , with recently emerging isolates in cluster I. In this study, we used representative isolates of each of the 42 SPs to develop an inexpensive specific mPCR/RLB to further determine the genome variation in B. pertussis. The presence or absence of previously determined RDs as revealed by the mPCR/RLB was compared against and used in conjunction with previous SNP typing to observe genomic variability of B. pertussis through time.
Results and discussion
Development and application of multiplex PCR/ reverse line blot to detect selected regions of difference
Details of genes within selected regions of difference (RD) for this study and comparison to RD in other studies
Genes within RDs*
Representative gene for this study
King et al.(2010)
Brinig et al.(2006)
Heikkinen et al.(2007)
Caro et al.(2006)
MaoC family protein
Putative succinate-semialdehyde dehydrogenase [NADP+]
Putative CoA ligase
Heme uptake regulator
Putative exported protein
Putative exported protein
Branched-chain amino acid-binding protein
Putative ferrisiderophore receptor
lysR family transcriptional regulator
Putative integral membrane protein
Putative periplasmic protein
Sarcosine oxidase beta subunit
FolD bifunctional protein
GntR family transcriptional regulator
Putative gamma-glutamyltranspeptidase (exported protein)
Putative IclR-family transcriptional regulator
Putative binding-protein-dependent transport protein
Putative phage terminase
Conserved hypothetical protein
Relationships between RD absences and B. pertussis SNP evolution
To determine whether the presence of RD or absence was linked with the evolution of B. pertussis, the results of the mPCR/RLB assay were plotted against the SNP-based phylogenetic tree as previously determined by Octavia et al. and shown in Figure 2. BP0910A- BP0930 was found only in isolates with SPs in cluster VI, which consists of isolates from the pre-vaccination era. In contrast, BP1948-BP1966 was detected in all isolates with SPs in all clusters except cluster I.
The association of these two RDs and SNP clusters could, potentially, be explained by carry-over (or hitchhiking) of gene losses in fitter B. pertussis strains . Genetic hitchhiking, also known as selective sweeps, refer to genetic changes in neighboring genes that get fixed as a result of linkage with genes carrying advantageous mutations which are selected for, but these genetic changes may not necessarily contribute to the fitness of the organism. In B. pertussis, genetic hitchhiking is most clearly observed with the two RD losses. The loss of BP910A-BP930 is associated with the potentially advantageous change to ptxA2 which arose when the whole cell vaccine was introduced, while the loss of BP1948-1966 is associated with changes to prn2 and ptxP3 in cluster I, which may be driven by selection pressure from the acellular vaccine as discussed previously . The loss of BP1948-1966 may also be driven by immune selection pressure against BP1948 which encodes an immunogenic protein as discussed below.
Apart from these two major RD absences that correlated with specific SNP clustering, the majority of gene losses occurred independently in a ‘mosaic’ pattern, and were not associated with the introduction of vaccination or change in antigen alleles . Even within clusters I and II, which are generally highly homogeneous, patterns of gene loss differed; BP1553 and BP3322 each absent from 2 of 7 isolates in cluster I and similarly, BP2825 and BP1553 absent from 1 and 2 isolates respectively, in cluster II.
A limitation of this study is that only single representatives of each SP were used and interpretation of sporadic gene losses from this collection of isolates is difficult without performing additional investigations with more isolates from SP clusters. However, in-depth analyses of French, Finnish, Dutch and Swedish B. pertussis isolates demonstrated that correlations could be also made between RD losses, specific pulse-field gel electrophoresis (PFGE) profiles and antigen alleles [3, 4, 17, 18]. The absence of BP1948-BP1966 in cluster I isolates in this study, in particular, can be likened to the same RD loss in Finnish isolates collected since 1999 which belong to PFGE Group IVβ . As these isolates have only emerged recently, a strong temporal relationship with BP1948-BP1966 could be made. In addition, given the highly homogeneous nature of B. pertussis populations, it is likely that isolates within this PFGE profile are identical to cluster I isolates in this study. Previous reports of RD losses and virulence-associated alleles by King et al.[4, 5] showed a definitive relationship between BP1948-BP1966 loss and isolates carrying the ptxP3 allele. All cluster I isolates in this study and additional isolates not included in this study carried the ptxP3 allele in addition to prn2, a non-ACV allele . Thus, the RD was lost in current, globally circulating isolates that have expanded over time from a single clone.
Functional significance of gene losses from regions of difference
The RD encompassing BP910A-BP934 was absent from most isolates analysed by several other studies [4, 5, 19, 22] as well as this one. The defining feature of isolates, which had retained this RD, was the fact that all were collected before the introduction of whole cell vaccines. Within this RD, the majority of genes in BP910A-BP934 encode hypothetical proteins or transcriptional regulators. Genes of known function include a putative succinate-demialdehyde dehydrogenase and a citrate utilization protein.
The second major RD loss was BP1948-BP1966, which was absent from seven, all very recent, isolates of the 42 selected. RD BP1948-BP1966 is 22.7 kb in size and consists of 18 genes, the majority of which are involved in energy metabolism, transport or binding or are pseudogenes . Interestingly, BP1948, which encodes a 44 kDa branched chain amino acid binding protein involved in membrane transport, has been identified as immunogenic by Tefon et al. and the deletion of this region may potentially be advantageous to B. pertussis if its loss decreases overall immunogenicity and herd immunity in a population .
Other genes of interest within variably absent RDs included copper resistance proteins within the RD BP3314-BP3322 (represented by BP3319 and BP3322) . BP3322 was absent from three B. pertussis strains, two of which belong to the predominant SPs (SP13 and SP14) in SNP cluster I, whereas BP3319 was absent from L669 and both BP3319 and BP3322 were absent from L1022- two cluster VI strains. The regulation of copper concentrations is tightly controlled to prevent toxicity within the cell , although the exact effects of these gene losses have not been determined. The absence of the whole RD from earlier strains compared with individual gene losses from SP13 and SP14 indicate that some of these more recent losses occurred independently.
The losses of RDs or RD-associated genes shown in this study demonstrate that genome changes, particularly genome reduction, are an ongoing process in B. pertussis. Most of the individual RD-associated gene losses have occurred randomly, except that the loss of two RDs was each shown to be temporally associated with an evolutionary event, as shown by a change in SNP cluster. However, these events are not likely to have been directly due to selection pressure but instead could be explained by “genetic hitchhiking”  in fitter B. pertussis variants as a result of selection somewhere else in the genome. We also demonstrated that an mPCR/RLB method can be used as a rapid method to detect presence or absence of RDs.
Based on previous SNP typing of a diverse range of B. pertussis by Octavia et al., one B. pertussis isolate from each of the 42 unique SNP profiles was selected to represent the evolution of B. pertussis since the 1920s. Where possible, isolates were selected to be most representative of their SP in terms of year and region of isolation. Bacterial isolates were grown on Bordet-Gengou agar (BD) supplemented with 10% defibrinated horse blood (Oxoid) for 3–5 days at 37°C. DNA was extracted using the phenol/chloroform method and used for mPCR.
Identification and selection of regions of difference
RDs were previously defined by Cummings et al. as two or more adjacent array elements not detected in at least 3 strains of B. pertussis. However, further studies identified additional RDs consisting of either single genes or two consecutive genes, which were variably present in different isolate collections [3–5, 16, 19]. For this study, previously identified RDs were considered suitable for analysis if there was at least one gene, which had been shown to be variably absent from previously studied B. pertussis populations. The prevalence of each RD loss was also calculated and RDs present in more than 98% of the B. pertussis population from each study were not studied. To avoid confusion, specific RDs are referred to in this study by the corresponding genes on either side of the RDs. A total of 35 RDs met the criteria and were used for further analysis. Larger RDs and RDs which had been shown to be most variable, had more than one gene selected to represent the RD. Six RDs met these criteria and therefore had more than one representative gene (Table 1).
Primer and probe design
B. pertussis Tohama I, a completely sequenced reference strain (NC_002929) , was used as the basis for designing PCR primers and probes for each of the selected RDs. All primers were 18–24 bp in length and amplified regions with similar G + C content for efficient multiplexing. Primer interactions and dimer formation within mPCRs were analysed using Autodimer software . Amplicons were 150–300 bp with annealing temperatures between 50-60°C, from regions within selected RDs, in order to allow simultaneous amplification in a multiplex reaction.
Corresponding RLB probes for each of the selected genes were also designed using Tohama I genome as a reference to ensure specific binding. Each probe was designed to be 20–22 bp in length and complementary to individual PCR products. Each probe was labelled at the 5’ end with an amine group to allow binding to the nylon membrane. The sequences of forward and reverse primers, expected PCR product size, corresponding RD probes and the function of each gene of interest are listed in Additional file 1: Table S1. The selected genes from each RD were each individually tested by single-plex PCR on a subset of strains and confirmed on 2% agarose gel to determine whether the selected regions/genes were absent before combining primer sets into a multiplex-PCR.
Each mPCR was optimised to contain at least 8–9 targets as it had not been possible to include all targets in one mPCR. Each reaction mixture consisted of 8–9 primer pairs at a concentration of 10 μmol each, ~30 ng genomic DNA, 2U Taq polymerase (Biotium), 0.25 mM each of dATP, dTTP, dGTP (Bioline) and 0.125 mM each of biotinylated/non-biotinylated dCTP (Roche), 3.5 mM MgCl2 and 1 M betaine (Sigma-Aldrich). Thermocycling conditions included a touchdown step and were as follows: initial denaturation at 96°C for 2 min; 35 cycles of 96°C for 30 sec, 55°C for 30 sec, 72°C for 15 sec; and final extension at 72°C for 7 min. Individual PCR confirmation was then carried out to determine the reliability of mPCR in amplifying desired products and whether any mis-amplification or primer dimer formation occurred.
Reverse line blot assay
The reverse line blot assay was carried out according to Kong et al.. Briefly, a Biodyne C membrane was activated with 16% (w/v) 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDAC) for 20 min and rinsed with sterile milliQ water. The membrane was placed in a Miniblotter® (Immunetics) and specific 5’ amine labeled oligonucleotide probes were then bound to the membrane and fixed with 0.1 M NaOH. The membrane was then removed, rotated 90° and placed back into the Miniblotter. Products from the five mPCR were combined and boiled for 5 min before being applied to the membrane perpendicular to the bound probes. The membrane was then incubated at 55°C for 1 hr before washing in 1× SSPE (150 mM sodium chloride, 10 mM sodium phosphate and 1 mM EDTA)/0.1% SDS solution. A streptavidin-peroxidase conjugate (Roche) was then added and the membrane further incubated at 42°C for 1 hr before washing and exposing with enhanced chemiluminescence (ECL) detection kit (GE LifeSciences). Chemiluminescence was detected using LAS3000 Imager (Fujifilm).
A clear hybridisation signal was interpreted as the presence of the corresponding gene, whereas the lack of signal was interpreted as the absence of a gene. Faint or indistinct hybridisations deemed ambiguous were confirmed using individual PCR and agarose gel electrophoresis.
This work did not involve human subjects, human material or human data and therefore did not require ethical approval.
We gratefully acknowledge the generous donations of isolates by Dr Kazunari Kamachi, Department of Bacterial Pathogenesis and Infection Control, National Institute of Infectious Diseases, Tokyo, Japan; Dr Shane Byrne, Sullivan Nicolaides Pathology, Brisbane, Queensland, Australia; the late Dr John Tapsall, Prince of Wales Hospital, Sydney, Australia; Dr Ian Carter, St George Hospital, Sydney, Australia; Dr David Andresen, Children’s Hospital at Westmead, Sydney, Australia; Dr Nicole Guiso, Institut Pasteur, Paris, France; Dr Raymond Tsang, National Microbiology Laboratory, Public Health Agency of Canada; and Dr Lucia Tondella, Centers for Disease Control and Prevention, USA. The funding for this project was provided by the National Health and Medical Research Council. Connie Lam was supported by an Australian Postgraduate Award.
- Diavatopoulos DA, Cummings CA, Schouls LM, Brinig MM, Relman DA, Mooi FR: Bordetella pertussis, the causative agent of whooping cough, evolved from a distinct, human-associated lineage of B. bronchiseptica. PLoS Pathog. 2005, 1: e45-10.1371/journal.ppat.0010045.PubMedPubMed CentralView ArticleGoogle Scholar
- Parkhill J, Sebaihia M, Preston A, Murphy LD, Thomson N, Harris DE, Holden MT, Churcher CM, Bentley SD, Mungall KL, Cerdeno-Tarraga AM, Temple L, James K, Harris B, Quail MA, Achtman M, Atkin R, Baker S, Basham D, Bason N, Cherevach I, Chillingworth T, Collins M, Cronin A, Davis P, Doggett J, Feltwell T, Goble A, Hamlin N, Hauser H: Comparative analysis of the genome sequences of Bordetella pertussis, Bordetella parapertussis and Bordetella bronchiseptica. Nat Genet. 2003, 35: 32-40. 10.1038/ng1227.PubMedView ArticleGoogle Scholar
- Caro V, Hot D, Guigon G, Hubans C, Arrive M, Soubigou G, Renauld-Mongenie G, Antoine R, Locht C, Lemoine Y, Guiso N: Temporal analysis of French Bordetella pertussis isolates by comparative whole-genome hybridization. Emerg Infect Dis. 2006, 8: 2228-2235.Google Scholar
- King AJ, van Gorkom T, Pennings JL, van der Heide HG, He Q, Diavatopoulos D, Heuvelman K, van Gent M, van Leeuwen K, Mooi FR: Comparative genomic profiling of Dutch clinical Bordetella pertussis isolates using DNA microarrays: identification of genes absent from epidemic strains. BMC Genomics. 2008, 9: 311-10.1186/1471-2164-9-311.PubMedPubMed CentralView ArticleGoogle Scholar
- King AJ, van Gorkom T, van der Heide HG, Advani A, van der Lee S: Changes in the genomic content of circulating Bordetella pertussis strains isolated from the Netherlands, Sweden, Japan and Australia: adaptive evolution or drift?. BMC Genomics. 2010, 11: 64-10.1186/1471-2164-11-64.PubMedPubMed CentralView ArticleGoogle Scholar
- Kong F, Gilbert GL: Multiplex PCR-based reverse line blot hybridization assay (mPCR/RLB)–a practical epidemiological and diagnostic tool. Nat Protoc. 2006, 1: 2668-2680.PubMedView ArticleGoogle Scholar
- Kong F, Ma L, Gilbert GL: Simultaneous detection and serotype identification of Streptococcus agalactiae using multiplex PCR and reverse line blot hybridization. J Med Microbiol. 2005, 54: 1133-1138. 10.1099/jmm.0.46244-0.PubMedView ArticleGoogle Scholar
- Kong F, Brown M, Sabananthan A, Zeng X, Gilbert GL: Multiplex PCR-based reverse line blot hybridization assay to identify 23 Streptococcus pneumoniae polysaccharide vaccine serotypes. J Clin Microbiol. 2006, 44: 1887-1891. 10.1128/JCM.44.5.1887-1891.2006.PubMedPubMed CentralView ArticleGoogle Scholar
- Kudinha T, Kong F, Johnson JR, Andrew SD, Anderson P, Gilbert GL: Multiplex PCR-based reverse line blot assay for simultaneous detection of 22 virulence genes in uropathogenic Escherichia coli. Appl Environ Microbiol. 2012, 78: 1198-1202. 10.1128/AEM.06921-11.PubMedPubMed CentralView ArticleGoogle Scholar
- Cai Y, Kong F, Wang Q, Tong Z, Sintchenko V, Zeng X, Gilbert GL: Comparison of single- and multilocus sequence typing and toxin gene profiling for characterization of methicillin-resistant Staphylococcus aureus. J Clin Microbiol. 2007, 45: 3302-3308. 10.1128/JCM.01082-07.PubMedPubMed CentralView ArticleGoogle Scholar
- O’Sullivan MV, Zhou F, Sintchenko V, Gilbert GL: Prospective genotyping of hospital-acquired methicillin-resistant Staphylococcus aureus isolates by use of a novel, highly discriminatory binary typing system. J Clin Microbiol. 2012, 50: 3513-3519. 10.1128/JCM.01625-12.PubMedPubMed CentralView ArticleGoogle Scholar
- Gravitt PE, Peyton CL, Apple RJ, Wheeler CM: Genotyping of 27 human papillomavirus types by using L1 consensus PCR products by a single-hybridization, reverse line blot detection method. J Clin Microbiol. 1998, 36: 3020-3027.PubMedPubMed CentralGoogle Scholar
- Cao Y, Kong F, Zhou F, Xiao M, Wang Q, Duan Y, Kesson AM, McPhie K, Gilbert GL, Dwyer DE: Genotyping of human adenoviruses using a PCR-based reverse line blot hybridisation assay. Pathology. 2011, 43: 488-494. 10.1097/PAT.0b013e328348.PubMedView ArticleGoogle Scholar
- Octavia S, Maharjan RP, Sintchenko V, Stevenson G, Reeves PR, Gilbert GL, Lan R: Insight into evolution of Bordetella pertussis from comparative genomic analysis: evidence of vaccine-driven selection. Mol Biol Evol. 2011, 28: 707-715. 10.1093/molbev/msq245.PubMedView ArticleGoogle Scholar
- van Gent M, Bart MJ, van der Heide HG, Heuvelman KJ, Mooi FR: Small mutations in Bordetella pertussis are associated with selective sweeps. PLoS One. 2012, 7: e46407-10.1371/journal.pone.0046407.PubMedPubMed CentralView ArticleGoogle Scholar
- Brinig MM, Cummings CA, Sanden GN, Stefanelli P, Lawrence A, Relman DA: Significant gene order and expression differences in Bordetella pertussis despite limited gene content variation. J Bacteriol. 2006, 188: 2375-2382. 10.1128/JB.188.7.2375-2382.2006.PubMedPubMed CentralView ArticleGoogle Scholar
- Heikkinen E, Xing DK, Olander RM, Hytonen J, Viljanen MK, Mertsola J, He Q: Bordetella pertussis isolates in Finland: serotype and fimbrial expression. BMC Microbiol. 2008, 8: 162-10.1186/1471-2180-8-162.PubMedPubMed CentralView ArticleGoogle Scholar
- Advani A, Van der Heide HG, Hallander HO, Mooi FR: Analysis of Swedish Bordetella pertussis isolates with three typing methods: characterization of an epidemic lineage. J Microbiol Methods. 2009, 78: 297-301. 10.1016/j.mimet.2009.06.019.PubMedView ArticleGoogle Scholar
- Heikkinen E, Kallonen T, Saarinen L, Sara R, King AJ, Mooi FR, Soini JT, Mertsola J, He Q: Comparative genomics of Bordetella pertussis reveals progressive gene loss in Finnish strains. PLoS One. 2007, 2: e904-10.1371/journal.pone.0000904.PubMedPubMed CentralView ArticleGoogle Scholar
- Caro V, Elomaa A, Brun D, Mertsola J, He Q, Guiso N: Bordetella pertussis, Finland and France. Emerg Infect Dis. 2006, 12: 987-989. 10.3201/eid1206.051283.PubMedPubMed CentralView ArticleGoogle Scholar
- Lam C, Octavia S, Bahrame Z, Sintchenko V, Gilbert GL, Lan R: Selection and emergence of pertussis toxin promoter ptxP3 allele in the evolution of Bordetella pertussis. Infect Genet Evol. 2012, 12: 492-495. 10.1016/j.meegid.2012.01.001.PubMedView ArticleGoogle Scholar
- Bouchez V, Caro V, Levillain E, Guigon G, Guiso N: Genomic content of Bordetella pertussis clinical isolates circulating in areas of intensive children vaccination. PLoS One. 2008, 3: e2437-10.1371/journal.pone.0002437.PubMedPubMed CentralView ArticleGoogle Scholar
- Tefon BE, Maass S, Ozcengiz E, Becher D, Hecker M, Ozcengiz G: A comprehensive analysis of Bordetella pertussis surface proteome and identification of new immunogenic proteins. Vaccine. 2011, 29: 3583-3595. 10.1016/j.vaccine.2011.02.086.PubMedView ArticleGoogle Scholar
- Cummings CA, Brinig MM, Lepp PW, van de Pas S, Relman DA: Bordetella species are distinguished by patterns of substantial gene loss and host adaptation. J Bacteriol. 2004, 186: 1484-1492. 10.1128/JB.186.5.1484-1492.2004.PubMedPubMed CentralView ArticleGoogle Scholar
- Vallone PM, Butler JM: AutoDimer: a screening tool for primer-dimer and hairpin structures. Biotechniques. 2004, 37: 226-231.PubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.