点击显示 收起
Applied Genetics, Institut de Biologie et de Médecine Moléculaires, Université Libre de Bruxelles, B-6041 Gosselies, Belgium,1 Institut de Zoologie, Université de NeuchÂtel, CH-2007 NeuchÂtel, Switzerland2
Received 27 February 2003/ Returned for modification 12 May 2003/ Accepted 27 May 2003
ABSTRACT |
---|
Top Abstract Introduction Materials and Methods Results Discussion References |
---|
INTRODUCTION |
---|
Top Abstract Introduction Materials and Methods Results Discussion References |
---|
A large number of isolates were recovered from various geographic areas and biological sources (patients, animals, and ticks). Molecular analyses such as protein and plasmid profiles, reactivity with monoclonal antibodies, DNA-DNA relatedness, 16 rRNA, flagellin and ospA gene sequence analyses, and finally multilocus enzyme electrophoresis demonstrated that these B. burgdorferi isolates are phenotypically and genotypically divergent (11, 12, 28, 35, 38, 40). These studies also established that all isolates are gathered in several species and genomic groups that include B. burgdorferi sensu stricto, B. garinii, B. afzelii, B. japonica, B. andersonii, B. valaisiana, B. lusitaniae, B. tanukii, B. turdi, B. bissettii, and B. sinica species (3, 13, 21, 26, 33). The term B. burgdorferi sensu lato is generally used to refer to all isolates belonging to these species or genomic groups.
To date only B. burgdorferi sensu stricto, B. garinii, and B. afzelii are known to be responsible for causing human disease (2, 7, 27, 30). In fact, it seems that infections with B. burgdorferi sensu stricto tend to lead to arthritic symptoms, while those with B. garinii appear to cause neurological complications. Late cutaneous manifestations (acrodermatitis) of Lyme disease seem to be associated with strains belonging to B. afzelii. However, it also appeared that B. valaisiana may be pathogenic for humans, since one serum sample from a patient with Lyme arthritis was more reactive by immunoblotting to this species (36). Moreover, DNA specific for this species was detected by PCR in skin biopsies of patients with erythema migrans and acrodermatitis chronica atrophicans (34). In the case of the seven other B. burgdorferi sensu lato species—B. japonica, B. andersonii, B. turdi, B. tanukii, B. bissettii, B. sinica, and B. lusitaniae—they were isolated from tick species in the United States (B. andersonii and B. bissettii); Japan (B. japonica, B. turdi, and B. tanukii); China (B. sinica); and Portugal, Czech Republic, Slovakia, Byelorussia, and North Africa (B. lusitaniae), and they are considered nonpathogenic for humans (12, 14, 18, 20, 24, 33). Nevertheless, a recent study has pointed out that a low passage of B. lusitaniae (strain PotiB2) is able to induce disease in susceptible mice (42).
Molecular techniques are commonly used for the identification and typing of B. burgdorferi sensu lato isolates. The isolates can therefore be categorized into phenotypic or genomic groups on the basis of the molecular targets used for the analysis. Many of the currently used molecular techniques (DNA-DNA reassociation analyses, ribotyping, pulsed-field gel electrophoresis, PCR-based restriction fragment length polymorphism analysis, random amplified polymorphic DNA analysis, and DNA sequencing) for typing rely on electrophoresis separation of DNA fragments of different molecular lengths (9, 10, 21, 22, 29, 31, 32). The electrophoretic result is represented by a pattern of bands on a gel. Since these patterns may be extremely complex, the ease with which patterns are interpreted is an important factor in evaluating the utility of particular typing method. Another important factor is its ease of use. Finally, the technical difficulty and time to obtain a result must also be evaluated in assessing the utility of a particular typing method.
In this way, we set up a colorimetric post-PCR detection system (PCR-reverse line blot assay [PCR-RLBA]) that is able to distinguish each known European B. burgdorferi sensu lato species. This method relies on the use of a specific capture oligonucleotide covalently linked to nylon membrane strips and of a specific polybiotinylated detection probe. Amplified DNA, sandwiched between these two molecules, is then detected via streptavidin-conjugated enzyme and a colorimetric substrate. This post-PCR system has been evaluated on amplified DNA from a large number of B. burgdorferi sensu lato strains, and from related or unrelated species. The target of the PCR and the post-PCR detection system is a DNA fragment located in the ospA gene that encodes one of the major outer membrane lipoproteins. The ospA gene was chosen because the clustering of Borrelia strains in the phylogenetic tree based on ospA DNA sequences is in agreement with the classification based on the sequence analysis of conserved chromosomal genes as well as with data obtained by pulsed-field gel electrophoresis and random amplified polymorphic DNA fingerprinting (3).
MATERIALS AND METHODS |
---|
Top Abstract Introduction Materials and Methods Results Discussion References |
---|
|
DNA isolation from Borrelia strains. Genomic DNA from all bacterial strains was extracted as previously described (7). DNA concentrations were determined spectrophotometrically by measuring the absorbance at 260 nm. The genomic DNA was heated for 10 min at boiling point just before the PCR assay. For DNA amplifications, 100 ng of the preparations was used as template DNA.
Preparation of clinical specimens for PCR amplification. Three ml of biological fluids (urine, serum, CSF, or synovial fluid) was centrifuged for 10 min at 12,000 x g to pellet spirochetes. The pellets were washed twice in PBS (NaCl [140 mM], KCl [2.68 mM], Na2HPO4 · 2H2O [8.09 mM], KH2PO4 [1.47 mM] [pH 7.2]), resuspended in 30 µl of distilled water, and boiled for 10 min. Samples (10 µl) were removed and processed for PCR amplification.
Synthetic primers and probes. Oligonucleotide primers and probes were synthesized via the solid-phase phosphoramidite method (5) on an Applied Biosystems (Rotterdam, The Netherlands) synthesizer (model 394). The first set of primers (consensus direct PCR primer and consensus reverse PCR primer 1) was used to amplify a 475-bp DNA fragment located in the ospA gene of B. burgdorferi sensu stricto, B. garinii, B. afzelii, and B. lusitaniae, respectively, while the second set of primers (consensus direct PCR primer and consensus reverse PCR primer 2) was used to specifically amplify a 505-bp DNA fragment located in the ospA gene of B. valaisiana strains (Table 1). The amplified DNA fragments present sufficient diversity and homology to design genospecific capture probes that allow the typing of each Borrelia genospecies and one detection probe that is able to recognize all tested B. burgdorferi genospecies. The detection probe carries a multifork-like structure on its 5' end with eight biotin groups (8). The sequences of both capture and detection probes and their positions within the amplified sequences are shown in Table 1. These probes were used in the reverse blot hybridization procedure (see "RLBA" below).
|
Chemical condensation of capture oligonucleotides onto nylon membranes. The condensation of 5' phosphate oligonucleotides onto primary amino groups has already been described by De Vos and coworkers (8). Briefly, the 5' phosphate oligonucleotides (200 ng/µl) were diluted into 10 mM 1-methyl-N-imidazole containing N'-(3-dimethylaminopropyl)-N-ethylcarbodiimide hydrochloride (80 ng/µl) and applied to a nylon membrane strip (Biodyne C; Pall Corporation, Brussels, Belgium) with a pump mechanism which delivers controlled amounts of probe to this membrane. The nylon membrane was then incubated for 15 min at room temperature, and unlinked oligonucleotides were removed by three washing steps with a solution of NaOH (0.4 M)-sodium dodecyl sulfate (0.25%) at 50°C. The Borrelia genotyping nylon membrane strip contained five capture probes laid out in five lines, each specific to B. burgdorferi sensu stricto, B. garinii, B. afzelii, B. valaisiana, and B. lusitaniae, respectively.
RLBA. RLBA is based on the hybridization of a target DNA, previously amplified by PCR, to a capture probe covalently linked onto the surface of a nylon membrane strip. The hybridized DNA is then recognized by a multifork-like oligonucleotide probe carrying 8 biotin groups that usually increases five times the sensitivity of the assay (8). For one reaction, 20 µl of a PCR mixture was mixed with 80 µl of NaOH (0.25 M) and 400 µl of water for 10 min. The denatured PCR products were then placed onto a Borrelia genotyping strip and mixed with 500 µl of a Tris-buffered solution (Tris-HCl [0.1 M], NaCl [1 M], MgCl2 [2 mM], Triton X-100 [0.05%] [pH 7.5]) containing 5% nonfat dried milk, 0.2 M acetic acid, and 40 pmol of the detection probe. To permit the capture of the PCR products, the nylon strip was incubated for 90 min at 37°C under slight agitation. The nylon strip was then washed five times with 1 ml of a Tris-buffered solution containing 5% nonfat dried milk. To detect hybridized target DNA, 1 ml of diluted (2,500-fold) streptavidin-alkaline phosphatase (Roche Molecular Biochemicals), in the Tris-buffered solution (pH 7.5) containing 150 mM NaCl was added, and incubated for 30 min at room temperature. To reveal the hybridized product associated with streptavidin-alkaline phosphatase, the strip was washed five times with 1 ml of Tris-buffered solution (pH 7.5) containing 150 mM NaCl, washed twice with 1 ml of a Tris-buffered solution (pH 9.5) containing 100 mM NaCl and 50 mM MgCl2, and then incubated for 20 min with 1 ml of a solution containing nitroblue tetrazolium (165 µg/ml) and BCIP (5-bromo-4-chloro-3-indolylphosphate) (825 µg/ml). After color development, the strip was rinsed in deionized water, dried, and photographed. Results of this hybridization process were compared to a reference strip, which indicates the exact position of the capture probes specific to each Borrelia genospecies.
DNA sequencing and computer analyses. Nucleotide sequencing of some PCR-amplified fragments was determined by the dideoxy chain termination method on denatured double-stranded DNA using the BigDye Terminator Cycle Sequencing kit (Applied Biosystems).
All algorithms used in this study are part of the Genetics Computer Group package (version 10.3; GCG release 10.3). In particular, the sequence comparisons were achieved with the program GAP, using the default parameters (gap creation penalty of 8 and gap extension penalty of 2). The phylogenetic trees were constructed from PILEUP files, with the programs DISTANCES and GROWTREE, by using the unweighted pair-group method using arithmetic averages and all additional default parameters. The database searches were realized either with the FASTA and TFASTA programs of the GCG 10.3 package or with BLASTN, BLASTP, and BLASTX programs, available at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/BLAST/).
RESULTS |
---|
Top Abstract Introduction Materials and Methods Reults Discussion References |
---|
Detection of B. burgdorferi sensu lato PCR DNA fragments with the RLBA. To examine the efficiency of the reverse line blot system, the PCR mixtures were independently incubated with nylon membrane strips containing five specific capture probes to the five targeted Borrelia species, laid out in five lines (for the principle of this method, see "RLBA" in Materials and Methods). An example of this detection assay is presented in Fig. 1. It was shown that the B31 B. burgdorferi sensu stricto strain is specifically recognized by the B. burgdorferi capture probe and not by the other. The same interpretation can be applied to strains belonging to other species indicating the specificity of this hybridization system. The results were compared to those obtained with an RFLP method (32) that is based on the MseI restriction pattern of the PCR-amplified 5S-23S intergenic region (Table 2) (14, 16, 17). The results indicated that there is an excellent correlation between the two methods. From the 120 PCR mixtures specific to Borrelia species, 117 were positive with the PCR-RLBA, and 106 correlated with the results obtained with the reference RFLP method. Therefore, 11 isolates did not correlate completely with the RFLP results (Tables 2 and 3). One isolate (NE219) was a mixture of two genospecies, B. garinii and B. valaisiana, according to RFLP patterns, but probably due to subcultures, only one genospecies was detected by the PCR-RLBA. Seven isolates (NE229, NE230, NE246, NE248, NE249, NE253, and SLNE901) out of these 11, which were considered by the RFLP method as belonging to the B. valaisiana species, showed up in a complex response (strongly positive for B. valaisiana; positive for B. garinii; and in some cases—isolates NE248, NE249, NE253, and SLNE901—weakly positive for B. lusitaniae) by using the PCR-RLBA (Tables 2 and 3). To solve this discrepancy, the ospA gene of these seven isolates was cloned and sequenced. Their ospA DNA sequence was then compared among themselves and to known ospA sequences. The results of these analyses indicated that these sequences were greatly homologous (70 to 96%) among them and were clustered into B. valaisiana groups I and II (Fig. 2). The unweighted pair-group method using arithmetic averages was used to construct a phylogenetic tree on the basis of the DNA sequence similarity matrix of ospA sequences (Fig. 2). The isolates NE229, NE230, NE246 and NE248 constitute a tight cluster with the M53 strain (B. valaisiana group I), whereas isolates NE249, NE253 and SLNE901 are closely related to NE231 and M7 strains (B. valaisiana group II). This genomic diversity was taken into account by designing more specific B. garinii and B. lusitaniae capture probes. This was done by removing one base either at the 5' end or at the 3' end of the capture B. garinii or the capture B. lusitaniae oligonucleotide, respectively. These modifications were integrated into an improved version of the PCR-RLBA. In this case, the seven divergent strains were only recognized by B. valaisiana capture probe (data not shown).
|
|
|
Finally, there was a very good correlation (113 out of 117 B. burgdorferi sensu lato isolates, 96.5%) between the results obtained by using the improved version of the PCR-RLBA and those obtained with the reference RFLP method. For the four divergent isolates (except for isolate NE49), the PCR-RLBA results were confirmed by the analysis of the ospA gene sequence.
Analysis of clinical samples with the PCR-RLBA. In order to evaluate the reverse line blot system with patient samples, 44 clinical specimens (22 serum specimens, 18 urine specimens, and 4 CSF specimens) collected from both 18 neuroborreliosis patients (12 serum specimens, 10 urine specimens, and 2 CSF specimens) and 10 patients suffering from unrelated diseases (10 serum specimens, 8 urine specimens, and 2 CSF specimens) were subjected to PCR analysis using the mixture of primers (Table 1). As seen in Table 4, all the samples collected from the neuroborreliosis patients were positive by PCR, confirming the diagnosis of Lyme disease. In addition, the clinical specimens collected from patients with unrelated diseases were PCR negative. Each PCR mixture was then incubated with a nylon membrane strip containing the five capture probes specific to the five targeted Borrelia species. The results of these analyses indicated that 14 patients of the 18 tested appeared to have been infected with B. garinii, 9 appeared to have been infected with B. burgdorferi sensu stricto, 7 appeared to have been infected with B. afzelii, and interestingly, one patient appeared to have been infected with B. valaisiana. Indeed, five patients carried genetic material from two genospecies (B. garinii and B. afzelii, patients 3 and 12; B. burgdorferi sensu stricto and B. valaisiana, patient 6; B. burgdorferi sensu stricto and B. garinii, patients 7 and 13), and four others carried DNA from the three known Lyme disease-associated genospecies (B. burgdorferi sensu stricto, B. garinii, B. afzelii; patients 4 and 15 to 17).
|
DISCUSSION |
---|
Top Abstract Introduction Materials and Methods Results Discussion References |
---|
The first step of the study was to design oligonucleotide primers by comparing a large number of ospA DNA sequences, accessible in the GenBank/EMBL database. However, the high sequence diversity of the ospA gene in some Borrelia species required synthesizing a set of three primers, which allowed us to amplify the expected DNA fragment from 98% tested Borrelia isolates (n = 120). The combination of this PCR assay with the genospecies-specific RLBA (PCR-RLBA) enabled us to differentiate the five targeted genospecies, namely, B. burgdorferi sensu stricto, B. garinii, B. afzelii, B. valaisiana, and B. lusitaniae species. In the case of reference strains, the classification obtained by the PCR-RLBA method was in complete agreement with data obtained by other procedures such as restriction fragment length polymorphism analysis, DNA-DNA hybridization, multilocus enzyme electrophoresis, and 16S rRNA signature nucleotide analysis (9, 10, 21, 22, 29, 31, 32). For the isolates, some discrepancies (isolates NE49, NE219, NE228, and SLNE889 [Table 2]) appeared between the results obtained by the PCR-RLBA method and those obtained by the RFLP method. Nevertheless, DNA sequencing of the related PCR products confirmed the results obtained by PCR-RLBA, indicating that the combined use of the PCR and the RLBAs is more specific than the RFLP method. In addition, we also observed that some isolates (NE229, NE230, NE246, NE248, NE249, NE253, and SLNE901) that have an RFLP pattern similar to members of the B. valaisiana species exhibited a complex response (strongly positive for B. valaisiana; positive for B. garinii; and in some cases, weakly positive for B. lusitaniae) by using the PCR-RLBA method. DNA sequencing of the related PCR products indicated that this multifaceted response is probably due to the heterogeneity of the ospA sequence in the B. valaisiana species. Consequently, it induces a lower specificity of the B. garinii and B. lusitaniae capture probes. This problem was solved by reducing the length of these capture probes. The important ospA sequence heterogeneity within the B. valaisiana species confirmed the data described previously (25, 39), which indicated the clustering of this species into three clusters (Fig. 2). In this study, the phylogenetic analysis showed that the ospA-sequenced B. valaisiana isolates gather into two clusters, B. valaisiana groups I and II (Fig. 2). The larger cluster consists of two branches, one comprising the reference strains (UK, VS116 and M53) isolated from ticks in the United Kingdom, Switzerland, and The Netherlands, and the second containing both tick and human isolates. Two other tick isolates with strains NE231 and M7 represent a separated genomic group, B. valaisiana group II. None of the B. valaisiana isolates tested in this study seems to be a member of the third cluster, which was mainly discovered in Korea and Japan.
In view of its sensitivity and specificity, the PCR-RLBA method thus offers the possibility to investigate, using a large number of samples, the possible correlation between the clinical outcome of Lyme disease and the genotype of the infecting B. burgdorferi sensu lato strains. The PCR-RLBA method was therefore used to analyze 18 cases of neuroborreliosis. The results showed that most patients had been infected with B. garinii (13 out of 18 patients [Table 4]), in agreement with previous observations showing the preferential association of this genospecies with neurological complications (8, 30, 27). Moreover, the PCR-RLBA analysis revealed several cases of multiple infection (Table 1). These data underscored the relatively high prevalence of multiple infection in Lyme disease (9 out of 18 studied cases). In fact, four patients were infected with three different Borrelia species, essentially by the well-recognized species involved in Lyme disease. Four patients were infected with two species (two patients were infected with B. garinii and B. afzelii, and two patients were infected with B. burgdorferi sensu stricto and B. garinii). In addition, the serum of one of these multiply-infected patients seems to be infected with both B. garinii and B. valaisiana isolates. Based on the DNA sequence of its ospA gene, this B. valaisiana isolate is a member of a subgroup of the B. valaisiana group I (Fig. 2). Until now, only two reports suggested evidence of a pathogenic potential of B. valaisiana. The first report demonstrated by PCR the presence of B. valaisiana genomic DNA in skin biopsy specimens of patients with erythema migrans and acrodermatitis chronica atrophicans (34). The second document showed by immunoblotting an association between clinical manifestations and presence of B. valaisiana in humans (36). Altogether, these data strongly suggest that B. valaisiana is infectious for humans and may play a role in the clinical outcome of Lyme disease in Europe.
In conclusion, we developed a PCR-RLBA method specific to five B. burgdorferi sensu lato species that have an important pathogenic potential. This methodology was evaluated using a large number of strains belonging to the five targeted Borrelia species. In terms of specificity and sensitivity, this method seems to be more specific and sensitive than the RFLP method, used as the referred method in this study. The PCR-RLBA method is also able to detect B. burgdorferi DNA in a variety of clinical specimens (urine, serum, CSF, and synovial fluid). In addition, we indicated by using both the PCR-RLBA method and DNA sequencing that B. valaisiana clusters into three genomic groups: two located in Europe and one in Asia. This stresses the possibility that B. valaisiana strains might be more heterogeneous than suspected up to now. This observation, together with that of multiple infection, might be of relevance in the evaluation of vaccines against B. burgdorferi sensu lato. Further analyses will also examine the usefulness of RLBA in the follow-up of Lyme disease patients.
Finally, RLBA can be performed in less than 3 h and does not require toxic chemical agents or radioisotopes and electrophoresis apparatus. In addition, RLBA can be adapted for automation, making its implementation in routine diagnostic laboratories a feasible prospect. Furthermore, RLBA can be adapted for the detection of amplified DNA originating from sources other than B. burgdorferi.
ACKNOWLEDGMENTS |
---|
This work has been supported by the Walloon Region of Belgium (Division Générale des Technologies et de la Recherche).
REFERENCES |
---|
Top Abstract Introduction Materials and Methods Results Dicussion References |
---|