|Home | About | Journals | Submit | Contact Us | Français|
Borrelia burgdorferi sensu stricto is the major causative agent of Lyme disease in the United States, while B. garinii and B. afzelii are more prevalent in Europe. The highly complex genome of B. burgdorferi is comprised of a linear chromosome and a large number of variably sized linear and circular plasmids. Many plasmids of this spirochete are unstable during its culture in vitro. Given that many of the B. burgdorferi virulence factors identified to date are plasmid encoded, spirochetal plasmid content determination is essential for genetic analysis of Lyme pathogenesis. Although PCR-based assays facilitate plasmid profiling of sequenced B. burgdorferi strains, a rapid genetic content determination strategy for nonsequenced strains has not yet been described. In this study, we combined pulsed-field gel electrophoresis (PFGE) and Southern hybridization for detection of genes encoding known virulence factors, ribosomal RNA gene spacer restriction fragment length polymorphism types (RSTs), ospC group determination, and sequencing of the variable dbpA and ospC genes. We show that two strains isolated from the same tick and both originally named N40 are in fact very distinct. Furthermore, we failed to detect bbk32, which encodes a fibronectin-binding adhesin, in one “N40” strain. Thus, two distinct strains that show different plasmid profiles, as determined by PFGE and PCR, were isolated from the same tick and vary in their ospC and dbpA sequences. However, both belong to group RST3B.
Borrelia burgdorferi sensu stricto is the primary causative agent of Lyme disease in the United States, while B. garinii and B. afzelii also commonly cause this disease in Europe (96). The genome size of the sequenced B. burgdorferi sensu stricto strain B31 is approximately 1.52 Mb, and it is a typical example of the relatively small genomes of Lyme spirochetes (40). Although Lyme spirochetes lack major biosynthetic pathways and are dependent upon the vertebrate host and tick vector to fulfill their nutritional requirements, they produce a significant number of virulence factors. Consequently, despite their small genome, these spirochetes can cause multisystemic disease. A hallmark of B. burgdorferi bacteria is the presence of a large number of linear and circular plasmids that make up about a third of their total genomes (40). Interestingly, a majority of the important currently known virulence factor genes of this spirochete are located on various plasmids (69).
B. burgdorferi strains contain 10 to 20 or even more linear and circular plasmids. Due to the sequence availability of the B31 strain of B. burgdorferi, researchers often use it as a reference for comparative analysis of different Borrelia strains. Before the completion of the sequence of B31, pulsed-field gel electrophoresis (PFGE) was primarily employed in determining the plasmid content of various spirochete strains, even though large circular plasmids are not resolved well in such gels and different linear plasmids often have very similar sizes in the 25- to 30-kb range (8, 63, 111, 112). PFGE is still used to discriminate the plasmid profiles of B. burgdorferi clinical isolates (45, 46, 84, 85), although plasmid loss in culture (see below) can make it an unreliable indicator of relatedness.
Many of the plasmids are easily lost during in vitro culture (90); however, lp54 and cp26 have been found to be the most stable (15, 78, 80, 100). Therefore, it is imperative that methods be available to determine the full complement of the plasmids in various strains. Purser and Norris (80) devised a simple PCR-based method that can detect and differentiate all 21 plasmids of strain B31. They and others have used this method to determine correlations between the presence of specific plasmids and B. burgdorferi infectivity (47, 48, 52, 80). However, the endogenous plasmid contents of other B. burgdorferi strains are different from those of the B31 strain, and reorganization of the plasmids appears to occur frequently in nature (71). Thus, at present a similar PCR-based technique for plasmid identification can be employed for other B. burgdorferi strains only after their completed genome sequences become available. This remains a major problem for the researchers who are studying the pathogenesis of Lyme disease using B. burgdorferi strains other than B31. This difficulty has broad consequences since it limits genetic and epidemiological studies to a single pathogenic strain that may not possess a full repertoire of the virulence genes of Lyme spirochetes. Although genome sequencing has become easier, understanding and annotating B. burgdorferi sequences are difficult due to gene duplication and the presence of virtually identical sequences on similar-sized plasmids, such as cp32s (1, 25, 41, 45, 98). Thus, determination of the plasmid content of different B. burgdorferi strains remains rather tedious. Therefore, there is a real need for a strategy to determine the plasmid profile of various B. burgdorferi strains in a streamlined manner that is applicable to different strains.
The B. burgdorferi strain N40 was originally isolated from an Ixodes scapularis tick from Westchester County of New York by Durland Fish (12). This uncloned culture was found to be highly infectious in the mouse model and was later cloned independently by different laboratories. N40 has been important in the study of Lyme disease pathogenesis for the past 3 decades, and its use in animal infection models has led to a number of important findings. For example, it was used to determine the differential response of various strains of mice and primates to B. burgdorferi (2, 5, 9–11, 13, 17, 18, 23, 24, 37, 38, 44, 68). In addition, many of the discoveries on the roles of different virulence factors during mammalian and tick infection were initially made using N40, genetic stability of the culture was reported in this strain when recovered from chronically infected immunocompetent mice (78), and, in particular, immunological responses to B. burgdorferi have also been thoroughly investigated using this strain. We have been studying N40 clone D10/E9 and have examined the roles of several virulence factors in this clone (32, 39, 56, 72, 75, 76, 86). However, we recently began to suspect that published work from various laboratories using the strain name “N40” may in reality refer to several different strains or at least different clones of the same strain. For example, very different sequences have previously been reported for the same N40 genes (3) and also observed by us (see below), and various genes, including bbk32, have been reported to be present and absent from N40 (38, 79). We report here that it is likely that the original “N40”-carrying tick was infected with multiple B. burgdorferi strains, which led to independent cloning and selection of different clonal isolates that were given the same N40 designation.
In order to clarify the past literature on this important strain and to avoid more confusion in the future, we obtained four N40 cultures from different laboratories and compared them in detail in this report. We anticipate that this study will also help other researchers evaluate the published literature on the pathogenesis of different N40 cultures. Furthermore, the set of techniques described herein could be useful for characterization of newly isolated B. burgdorferi strains during epidemiological studies. They can also be exploited for developing a PCR assay for the established virulence factors to detect the presence of specific plasmids, characterize new B. burgdorferi strains, and, thus, expand genetic studies to a large repertoire of unsequenced spirochete strains.
The clone cN40 culture was cloned in the laboratory of Stephen Barthold (University of California, Davis) and obtained from Janis Weis (University of Utah). The clone N40B and N40C cultures were the kind gift of Benjamin Luft (SUNY, Stony Brook, NY); the former was originally from Stephen Barthold, and the latter was from Martin Schriefer (CDC). The N40 clone D10/E9 (N40D10/E9) was isolated in John Leong's laboratory (Tufts Medical School, Boston, MA) from the uncloned culture provided by Allan Steere. Strain B31 clone A3 was provided by Patricia Rosa (Rocky Mountain Laboratories, NIAID, Hamilton, MT). B. burgdorferi cultures were grown at 33°C in BSKII medium containing 6% rabbit serum (Sigma-Aldrich, St. Louis, MO). The postcloning in vitro passage numbers for cN40, N40D10/E9, and B31A3 are passage 3 (P3), P2, and P4, respectively, and the passage numbers for N40B and N40C were not available.
B. burgdorferi genomic DNA was isolated using a previously described procedure (72). Primers to determine the plasmid profiles were designed on the basis of the sequenced genome of N40B for plasmid analysis by PCR (see Tables S1 and S2 in the supplemental material) using the programs previously described (77). PCR products were resolved by 1% agarose gel electrophoresis.
Approximately 109 cells of each B. burgdorferi strain were washed twice, resuspended in 20 μl of buffer (10 mM Tris, pH 7.6, 1 M NaCl), mixed with 100 μl of 1.5% low-melting-temperature SeaPlaque agarose (Lonza, Rockland, ME), and placed in the insert mold (Bio-Rad Life Sciences, CA). The agarose gel inserts (final concentration, 1.25% agarose) were processed for the release of intact genomic and plasmid DNA from bacteria while the DNA is still embedded in the agarose plug. Briefly, inserts were treated with 1 mg/ml lysozyme in the Tris-NaCl buffer for 3 h, followed by overnight treatment at 37°C with 0.5 mg/ml proteinase K solution prepared in the same buffer containing 5 mM MgCl2 and 1% SDS. After washing several times, 1-mm-thick pieces of the inserts were loaded in a 1.2% agarose gel and the wells were sealed with 1.2% agarose and subjected to contour-clamped homogeneous electric field (CHEF) electrophoresis using a CHEF-DR III unit (Bio-Rad Life Sciences, CA) in 1× TAE (Tris-acetate-EDTA) buffer. CHEF electrophoresis was conducted at 10°C at a ramped pulse time of 0.1 s to 8 s at a 120° angle and 6 V for 42 to 48 h. Plasmid patterns were visualized after ethidium bromide staining. PCR products of genes encoding known proteins with important roles in pathogenesis, such as dbpB, bbd14, bbe22 (pncA), bbk32, and vlsE, were amplified using the genomic DNA as template and the primers listed in Table S2 in the supplemental material. PCR products were purified using a PCR purification kit (Qiagen Inc., CA) and were used for preparation of the 32P-labeled probes using standard protocols. Southern hybridization was conducted using these [32P]dATP-labeled probes following a previously described procedure (87).
The bbi06 gene from B31 was PCR amplified using high-fidelity Herculase enzyme (Agilent Technologies Inc., CA), genomic DNA as template, and specific primers (see Table S2 in the supplemental material) with the restriction enzyme site sequence underlined. The PCR product was digested, cloned into the pET30a vector (Merck4Biosciences, Darmstadt, Germany), and sequenced to confirm mtnN insertion and to ascertain the accuracy of the sequence. The plasmid clone was used to transform E. coli strain BL21(DE3), and after culture, an optical density at 600 nm of 0.4 to 0.6 was obtained and bacteria were induced overnight at 37°C with 1 mM isopropyl-β-d-1-thiogalactopyranoside to produce recombinant MtnN protein. Polyhistidine-tagged MtnN protein was purified from the filtered lysates of induced bacteria using a cobalt-affinity column according to the manufacturer's instructions (Clontech Laboratories, Inc., CA). Antibodies directed against the recombinant His-MtnN protein were raised in BALB/c mice according to an immunization schedule described previously (75) and under the protocol approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Medicine and Dentistry of New Jersey.
Bacterial cell pellets were resuspended in Laemmli loading solution and analyzed on duplicate 10% SDS-polyacrylamide gels (53). The proteins separated in one gel were stained using a Silver Stain Plus kit (Bio-Rad Life Sciences, CA). Proteins in the replicate gel were transferred to an Immobilon-P polyvinylidene difluoride (PVDF) membrane (Millipore) (87), followed by immunoblotting using MtnN-specific polyclonal mouse antiserum and treatment with antimouse alkaline phosphatase-conjugated secondary antibodies. The presence of MtnN protein was detected using a substrate containing the mixture of 5-bromo-4-chloro-3 indolylphosphate (BCIP) and nitroblue tetrazolium (NBT).
RST determinations were conducted using the nested PCR product obtained by using the procedure previously described (61). Briefly, the primers PA (5′-GGTATGTTTAGTGAGGG-3′) and P95 (5′-GGTTAGAGCGCAGGTCTG-3′) were used in the first step, and the PCR product obtained was then used as template with the PB (5′-CGTACTGGAAAGTGCGGCTG-3′) and P97 (5′-GATGTTCAACTCATCCTGGTCCC-3′) primers in the second PCR. The same amplification program for both first- and second-round PCRs consisted of 35 cycles of denaturation at 94°C for 30 s, annealing at 52°C for 30 s, and extension at 72°C for 30 s. The PCR products were digested using restriction enzymes MseI and HinfI, and the agarose gel electrophoresis band patterns were compared.
To determine the nucleotide sequences of ospC and dbpA genes, primers flanking these genes, listed in Table S2 in the supplemental material, were used for PCR amplification, and the resulting products were sequenced at the Tufts Core Facility in Boston, MA.
Fifty micrograms of total proteins extracted from B. burgdorferi cells was lyophilized and dissolved to 1 mg/ml in 1:1-diluted SDS boiling buffer/urea sample buffer before loading. Two-dimensional (2D) electrophoresis was performed using the carrier Ampholine method of isoelectric focusing (IEF) (21, 65) by Kendrick Labs, Inc. (Madison, WI). Isoelectric focusing was carried out in a glass tube of inner diameter 2.3 mm using 2% pH 4 to 8 mix Servalytes (Serva, Heidelberg, Germany) for 9,600 V-h. Fifty nanograms of an IEF internal standard, tropomyosin, was added to the sample. This protein migrates as a doublet with a lower polypeptide spot of molecular weight 33,000 and pI 5.2 (marked in the gels). After equilibration for 10 min in buffer containing 10% glycerol, 50 mM dithiothreitol, 2.3% SDS, and 0.0625 M Tris, pH 6.8, each tube gel was sealed to the top of a stacking gel that overlaid a 10% acrylamide slab gel and SDS-gel electrophoresis was carried out. The silver-stained gels were dried between sheets of cellophane with the acid edge to the left. Duplicate gels were obtained from each sample and were scanned with a laser densitometer (model PDSI; Molecular Dynamics Inc., Sunnyvale, CA). The scanner was checked for linearity prior to scanning with a calibrated neutral density filter set (Melles Griot, Irvine, CA). The images were analyzed using Progenesis Same Spots software (version 4.0, 2010; Nonlinear Dynamics) and Progenesis PG240 software (version 2006; Nonlinear Dynamics, Durham, NC). Selected spots were cut out and limited matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) analyses were conducted at the Protein Core Facility of Columbia University in New York, NY.
Gel spots were transferred to clean tubes, water was added to completely hydrate the gels, and the plastic coating was removed with clean tweezers. Gel spots were prepared for digestion by washing twice with 100 μl of 0.05 M Tris, pH 8.5, 30% acetonitrile for 20 min with shaking and then with 100% acetonitrile for 1 to 2 min. After removing the washes, the gel pieces were dried for 30 min in a Speed-Vac concentrator.
Gels were digested by adding 0.08 μg modified trypsin (sequencing grade; Roche Molecular Biochemicals) in 13 to 15 μl 0.025 M Tris, pH 8.5. The tubes were placed in a heating block at 32°C and left overnight. Peptides were extracted twice with 50 μl 50% acetonitrile–2% trifluoroacetic acid (TFA); the combined extracts were dried and resuspended in matrix solution.
Matrix solution was prepared by making a 10-mg/ml solution of 4-hydroxy-α-cyanocinnamic acid in 50% acetonitrile–0.1% TFA and adding two internal standards, angiotensin and ACTH 7-38 peptide, to the matrix solution.
The dried digest was dissolved in 3 μl matrix/standard solution, and 0.5 μl was spotted onto the sample plate. When the spot was completely dried, it was washed twice with water. MALDI-MS analysis was performed on the digest using an Applied Biosystems Voyager DE Pro mass spectrometer in the linear mode.
Average peptide masses were entered into search programs to search the NCBI and/or GenPept databases for a protein match. Programs used were Mascot from Matrix Science and MS-Fit at http://prospector.ucsf.edu. Cysteine residues were modified by acrylamide.
Parameters for the web-based search using Mascot were as follows: database, NCBI; taxonomy, bacteria; variable modifications, oxidation (M), carboxyamidomethyl (C); missed cleavages, 2; error tolerance for peptide average masses, 0.5 Da. Parameters for the web-based search using MS-FIT were as follows: database, NCBI; taxonomy, bacteria; constant mods, possible mods, oxidation of M; minimum number of peptides to match, 4.
The sequences of the dbpA gene from the N40C clone and the ospC gene from the N40D10/E9 clone genes were deposited in GenBank with accession nos. JN969070 and JN969069, respectively.
For this study, we obtained B. burgdorferi “N40” strains from four sources and designated them cN40, N40B and N40C, and N40D10/E9. The most frequently used strategies that have been utilized for classification of B. burgdorferi sensu stricto isolates involve DNA sequence variation at several loci. (i) A few ORFs show hypervariability, including the cp26-encoded ospC gene (bbb19 of stain B31), which is essential for mammalian infection (42, 70, 83). Therefore, B. burgdorferi strains are often grouped on the basis of ospC sequence, and the association of these groups with infectivity has been determined by different researchers (19, 20, 35, 54, 55, 81, 88). In addition, the lp54-enocded dbpA gene also shows significant variation among different spirochete strains and has been used in distinguishing isolates (14, 88). (ii) A restriction fragment length polymorphism (RFLP) between the 16S and 23S rRNA genes has been used to categorize isolates into four ribosomal spacer types (RSTs) called RST1, RST2, RST3A, and RST3B. The RST1, RST2, and RST3B groups have been shown to be significantly associated with invasiveness and disease in the mouse model of infection and in humans (46, 49, 57, 58, 60, 61, 103, 104, 107, 108). However, these methods alone cannot always be used to distinguish various B. burgdorferi strains, and these classifications do not contribute at all to determination of the overall plasmid content of the strains. In addition, PFGE display of linear plasmids has been used to differentiate various B. burgdorferi strains (6, 22, 26, 111). Below we describe the use of these methods as well as PCR amplification of strategically chosen genes to understand the potential differences among B. burgdorferi cultures named “N40.”
To compare the four N40 cultures (see above), we first determined their plasmid profiles by PFGE. We observed that cN40 and N40B have the same plasmid profile and so could be derivatives of the same strain, while N40C and N40D10/E9 show very similar plasmid profiles that are different from cN40 and N40B (Fig. 1A). The lp36 and lp38 doublet seen in the B31 strain between the 33.5- and 48.5-kb marker bands is missing in both N40D10/E9 and N40C, and a different-sized singlet is present in N40D10/E9 and N40C at about 32 to 33 kb. Southern hybridization of PFGE blots with selected gene probes confirmed these results (Fig. 1B). For example, both bbk32 (on B31 lp36) and ospD (on B31 lp38) gene probes gave rise to clearly visible bands by Southern hybridization in B31, N40B, and cN40 but not in N40D10/E9 and N40C. There was weak cross-reactivity seen with the ospD gene probe, with a smaller band present in N40D10/E9 (data not shown). Thus, either the bbk32 and ospD genes are absent in N40D10/E9 and N40C or their DNA sequences are significantly different from those of B31. The successful amplification of the expected-sized DNA fragment from B31, cN40, and N40B strain DNA using bbk32 primers and failure to amplify this PCR product from N40D10/E9 and N40C genomic DNA indicate either that bbk32 is absent in N40D10/E9 and N40C or its sequence is very different in these cultures (Fig. 1C). The plasmid patterns of B31 and cN40 are further distinguishable from each other by the presence of the B31 lp17 bbd14 gene in cN40 and N40B on linear plasmids several kb larger than the lp17 plasmid of the B31 strain (Fig. 1B). This agrees with the sequence of the N40B equivalent plasmid (89), since it is approximately 21 kb in this strain, rather than 17 kb as in B31 (Fig. 1A and B; Table 1). Similarly, the vlsE1 gene is located on a plasmid that is significantly larger in both N40D10/E9 and N40C than in B31. The vlsE1-containing plasmid in cN40 and N40B is of intermediate size between the plasmids in the N40D10/E9 and B31 strains. This concurs with the sequencing results for the N40B strain, where vlsE1 is located on a 31.5-kb rather than 28-kb plasmid (Table 1) (S. Casjens, unpublished results). We also note a stained linear plasmid band at about 26 kb (just below the most intensely stained plasmid band which contains multiple lp28-type plasmids) in cN40 and N40B but not in the other two N40 cultures (Fig. 1A); however, its identity has not been determined. The other linear plasmids examined by Southern analysis in Fig. 1B either appear to be universally present in the four N40 cultures (lp25, lp28-2, lp28-4, and lp28-5; the last one gave no signal with B31) or universally gave a weak signal (lp21). In agreement with these observations, the genome sequences of B31 and N40B contain no lp28-5 and lp21 plasmid, respectively.
PFGE followed by multiple Southern hybridizations is too labor-intensive to routinely use to evaluate the plasmid content. The completed genome sequence of N40B (89) has recently made it possible to design primers that can be used for PCR detection of the linear and circular N40 plasmids, and these are described in Table S1 in the supplemental material. These primers produced PCR products of similar size from the linear plasmids of N40B and B31 clone A3, except that there was no cp9 or cp32-12 amplification from the B31 DNA. This agrees with the fact that these plasmids are known to be absent from B31 (36, 80, 98). Amplifications using these N40B-based primers distinguish among the four N40 cultures as follows (Fig. 2): (i) the cp9 primers amplify strongly from N40B and cN40 DNAs, while they do not amplify from N40C and amplify only weakly from N40D10/E9. (ii) The lp28-3 primers amplify strongly from cN40 but not from the other three N40 cultures. The N40B genome sequence did not contain an lp28-3, but it has been reported in other N40 cultures (51), so this failure is likely explained by the specific loss of lp28-3 from the N40B culture. (iii) The lp36 primers amplify strongly from cN40 and N40B but not from N40C and only weakly amplify a band of larger size from N40D10/E9. The N40B genome sequence did not contain an lp28-3, but it has been reported in other N40 cultures (51), so this failure is likely explained by the specific loss of lp28-3 from the N40B culture. These findings again support the idea that cN40 and N40B are similar to one another and N40C and N40D10/E9 are similar to each other but that these two pairs are quite different from each other. Interestingly, genetic diversity in lp28-3 and lp36 has also been reported previously, and these plasmids did not show an association with invasiveness of the spirochete strains (110). The following PCR amplification results do not simply agree with this hypothesis but can be explained by specific plasmid loss in particular cultures: the cp32-5 and cp32-10 primers fail to amplify DNA from only N40C, and the cp32-9 primer fails to amplify DNA from N40B and N40C. We note that all these culture differences could be explained by either (i) accumulated polymorphisms in the respective plasmid sequences or (ii) complete or partial loss of plasmids in the different cultures. We therefore analyzed the four N40 cultures in additional ways, as described below.
Schwartz and coworkers defined an RST-based grouping of B. burgdorferi sensu stricto strains (57). Strain N40 (probably cN40) has been reported to belong to the RST3B group (49, 50, 104). We used this method and conducted nested PCR with primers PA and P95, followed by PCR with primers PB and P97, as described in Materials and Methods. The resulting PCR products were digested with restriction enzymes MseI and HinfI. On the basis of the fragment patterns (Fig. 3), all four N40 cultures were determined to be of the RST3B group and the B31 culture was determined to belong to the RST1 group. These results agree with the previous RST determinations of the groups for N40 and B31 (49, 50, 57, 104).
To test our hypothesis that the location of important virulence factors is relatively conserved in various genetic elements, we also examined mtnN gene location and also determined its expression in the different N40 cultures. We have been studying the glycosaminoglycan (GAG)-mediated binding of B. burgdorferi strains to mammalian cells. Borrelia GAG-binding protein Bgp (product of chromosomal gene bb0588) facilitates spirochete binding to mammalian cells (75) and has methylthioadenosine/S-adenosylhomocysteine (MTA/SAH) nucleosidase activity (32, 77). Unlike other bacterial species (73), B. burgdorferi possesses three genes encoding homologous MTA/SAH nucleosidases; in addition to Bgp, it encodes two other MTA/SAH nucleosidases, cytoplasmic Pfs (product of chromosomal gene bb0375) and MtnN (encoded by bbi06 of B31 and N40B plasmid lp28-4). All three have nucleosidase enzymatic activity (32, 74, 77) (data not shown).
Some or all of these enzymes are virulence factors, since disruption of the bgp gene results in both a 1,000-fold increase in the 50% infectious dose relative to the wild-type strain (our unpublished work) and a 10-fold reduction in colonization of various tissues by B. burgdorferi N40 clone D10/E9 (86). Furthermore, several inhibitors of the MTA/SAH nucleosidase kill B. burgdorferi, a phenotype more severe than that conferred by the same inhibitors on other bacteria (86), indicating the importance of nucleosidases for survival of these auxotrophic spirochetes. The enzymatically active and exported Bgp and MtnN proteins are present on or near the surface in Lyme disease spirochetes and so could help salvage and recycle the MTA and SAH that are generated as by-products of various metabolic pathways by both the bacteria and the host. The MtnN protein is predicted to be an MTA/SAH nucleosidase since its putative active-site residues for substrate binding and functional activity are conserved relative to the well-studied E. coli enzyme (see Fig. S1 in the supplemental material), and we have recently shown that the MtnN protein of N40D10/E9 exhibits nucleosidase activity using MTA and SAH substrates (data not shown). However, the presence and expression of the mtnN gene in other N40 cultures were not known.
Therefore, PFGE resolution of linear plasmids and Southern blot hybridization were conducted to determine the location of the mtnN gene in the different N40 cultures. A probe prepared from the nucleotide sequence encoding the signal peptide unique to mtnN localized the gene to an approximately 28-kb linear plasmid (lp28-4 in the sequenced N40B strain and likely to be lp28-4 in the others) in all tested B. burgdorferi cultures (Fig. 4A). To determine the expression of the MtnN protein in the B31 and N40 cultures grown under laboratory conditions, total cellular proteins were resolved by SDS-PAGE in duplicate. A silver-stained gel served as a loading control, and a Western blot of the duplicate gel indicated that MtnN was expressed to similar levels in all cultures (Fig. 4B).
OspC and DbpA lipoproteins are critical virulence factors of B. burgdorferi (16, 42, 70, 93, 99, 101, 106). These molecules show high variability, and hence, their sequences facilitate discrimination among B. burgdorferi isolates. The ospC and dbpA genes are always present in field isolates of B. burgdorferi and are located on the highly conserved cp26 and lp54 plasmids, respectively. Therefore, ospC and dbpA genes from all four “N40” cultures were PCR amplified using primers flanking these genes and sequenced. The identical dbpA gene sequences in cN40 and N40B further support our finding that these two cultures are very similar. The N40D10/E9 isolate was distinguished from the sequenced cN40 isolate by a single point difference in dbpA (see Fig. S2 in the supplemental material); however, surprisingly, the dbpA of N40C showed 16 and 17 base differences relative to cN40 and N40D10/E9, respectively. We have no obvious explanation for these unique changes in the dbpA gene of N40C. It is likely either that horizontal exchange or strong selection in the laboratory changed this gene after the original isolation of N40 or that N40C and N40D10/E9 actually represent two closely related strains but differentially evolved from the original isolate.
Sequence analysis of ospC genes of the four N40 cultures indicated that the cN40 and N40B genes differ by only 1 nucleotide, while N40C and N40D10/E9 differ by only 2 nucleotides (see Fig. S3 in the supplemental material). However, these two pairs are over 15% different from one another (96 to 98 differences in different pairwise gene sequence comparisons); cN40 and N40B are OspC type E, as previously reported for N40 (7, 105), whereas N40C and N40D10/E9 are type M (Table 2).
During this work we discovered that the GenBank nucleotide sequence database contains five different previous entries for the strain “N40” ospC gene sequence. The ospC gene sequences of four of these, those with accession nos. U04240, DQ437463, AY275221, and CP002239, are identical to the ospC sequence gene of our cN40 strain and only 1 bp different from that of strain N40B, whereas a partial sequence of ospC published with GenBank accession no. AF416430 (95) is only 1 and 2 bp different from the equivalent sequences of the ospC genes of N40C and N40D10/E9, respectively. The small sequence differences between cN40 and N40B, between N40C and N40D10/E9, and between these and the above-described GenBank sequences suggest that they diverged very recently, probably since their original isolation.
To further examine differences between the cN40 and N40D10/E9 clones, 2D protein gel electrophoresis followed by sensitive silver staining was performed to observe the protein fingerprints of each clone. We selected cN40 rather than the sequenced N40B clone for this comparison because of the apparent loss of some plasmids in the latter. Even though the overall protein profiles of cN40 and N40D10/E9 grown in parallel in BSKII medium containing 6% rabbit serum appear to be somewhat similar, they showed several distinctions (Fig. 5). Among the 590 protein spots quantified, 41 showed a 10-fold or more difference in the level of expression, with several spots completely missing from one or the other culture (as marked in Fig. 5). A limited MALDI-MS analysis confirmed that the OspD protein spot in N40D10/E9 is present at extremely low levels when this culture is grown in BSKII medium containing rabbit serum (see Table S3 in the supplemental material). However, we previously showed that OspD is produced at significant levels when N40D10/E9 is grown in modified Kelly-Pettenkofer (MKP) medium supplemented with human or rabbit serum (75). We also note that there are differences in the protein concentrations of chromosomal genes between these two cultures; for example, potential breakdown products of GroEL and tentatively (only four peptides match) integrin binding membrane protein p66 (product of gene bb0603) are identified in cN40, but they are undetectable in N40D10/E9 (spot intensity, <0.05% against that of all measured polypeptide spots). The p66 protein was first recognized in N40D10/E9 as an adhesin that specifically recognizes the αIIbβ3 receptor on activated platelets and other mammalian cells (29, 30). Interestingly, this strain bound activated platelets efficiently only when the culture was grown in MKP medium supplemented with human serum and poorly and infrequently when the culture was grown in BSKII medium containing rabbit serum, conditions similar to those that we used in this study (31). However, the reason for this difference was not determined.
Lyme spirochetes can be isolated from Ixodes ticks in regions of endemicity, and several studies have shown that more than one strain often exists in individual ticks and that all of these strains are then transmitted to the mammalian hosts (33, 49). In the early days of the study of Lyme disease, this multistrain infection of ticks was not fully appreciated. Thus, mixed cultures were disseminated among laboratories, and when strain cloning became possible, different laboratories isolated different clones from the same culture, which were subsequently given the identical strain designation as the original culture, thus confusing the outcome of some research. We began to suspect that this was the scenario for the “N40” strain that has been studied extensively by various laboratories. Since our research primarily uses a clone, N40D10/E9, from the original N40 isolate, we decided to investigate whether we could differentiate various N40 cultures from one another using a series of tests, with the aim of using this information as a tool to reinterpret the results with “N40” that have been reported in the literature from a number of laboratories. The laboratory of Stephen Barthold cloned cN40 independently, and our results strongly indicate that N40D10/E9 and cN40 do indeed represent different strains cloned from the original N40 isolate. We suggest that the specific designations of various N40 clonal cultures, probably as defined here, be used in future studies to avoid confusion.
Using a combination of five techniques, PFGE, Southern hybridization, RST typing, PCR identification of the cN40 complete plasmid set, and sequence comparisons of ospC genes or dbpA genes, we compared four divergent “N40” cultures. These techniques have previously been employed to divide clinical and tick isolates of B. burgdorferi into different categories (8, 19, 27, 34, 35, 54, 55, 60, 61, 63, 88, 91, 107, 108, 111, 112), but all these techniques have not been used together in a single study, until now. The PFGE display and Southern analysis of linear plasmids, PCR identification of plasmids present, and sequence analysis of the ospC and dbpA genes all showed substantial differences among the four cultures and indicated that our cN40 and N40B clones were derivatives of the same strain and N40C and N40D10/E9 were similar and represent two similar strains that are rather different from cN40 and N40B. The difference in PFGE stained linear plasmid patterns in Fig. 1A, the PCR identification of plasmids in Fig. 2, and the nearly 20% difference in ospC gene sequences in Fig. S3 in the supplemental material demonstrate this most clearly.
PFGE followed by Southern hybridization of whole-cell DNA is a time-consuming and labor-intensive technique that can accurately determine the linear plasmid content of a B. burgdorferi isolate (and circular plasmids, if combined with restriction digestion); however, it does not distinguish plasmids of very similar size, which is not uncommon in Borrelia isolates. A PCR-based assay described by Norris and coworkers (80) allows the determination of the plasmid content of strain B31 after growth or mutagenesis, and this has resulted in real progress in understanding the molecular pathogenesis of this strain (42, 70, 92–94, 101, 106). In a recent study, the same group has described a high-throughput method using the Luminex multiplex technology to determine the plasmid content of this strain (64). Unfortunately, these techniques are currently limited to the sequenced strain B31. In the course of the work described here, we devised a set of tested and confirmed PCR primers from the reported N40 (N40B) genome sequence that can detect all 17 of the N40 plasmids (Table 1; Fig. 2; see Table S1 in the supplemental material). Poor amplification or a different-size PCR product likely indicates the differences in the sequence of that plasmid and/or partial loss of the plasmid from the culture. This primer set will also allow the careful study of the pathogenesis of this strain, which has not previously been possible due to the inability to determine whether all plasmids were present.
We also noticed during our studies of the N40 cultures that our primer set, which is heavily weighted toward virulence factor genes, amplified mostly fragments of the same size as those from the B31 and N40D10/E9 strains. This caused us to examine the plasmid sections that carry these genes in the other sequenced B. burgdorferi genomes (89). Although there appears to have been substantial genetic exchange among the linear plasmids in different isolates, we found that such exchange is especially prevalent near the ends of the plasmids. Thus, although exceptions do exist for specific virulence factors, especially if their genes lie near plasmid ends, as is the case with the especially variable B31 lp36 and lp28-1 plasmids and their homologs, these observations depict a limited genetic exchange in the middle region of the linear plasmids of B. burgdorferi (S. Casjens, unpublished data). We therefore suggest that virulence factor genes will be particularly useful to determine the plasmid content of different strains and that such an analysis can also be used for differentiation of B. burgdorferi new isolates and also for epidemiological analyses of the strains prevalent in the regions of endemicity of the United States and Europe.
Epidemiological studies often involve isolation and characterization of multiple B. burgdorferi strains from regions of endemicity (58, 59, 66, 67, 102). However, there is no simple test to determine whether a new isolate is indeed a very different strain or is similar to one of the strains previously isolated from the same region. In addition, there is a need for a method that can easily determine whether plasmids are lost during propagation or genetic manipulation of a particular B. burgdorferi strain. An infectious strain appears to be much more likely to have retained the important virulence factor genes on related plasmids rather than in random locations. Therefore, we suggest that in the beginning of a new strain characterization, PFGE of the endogenous plasmids could be conducted, especially if the clonal isolate will be used for the genetic studies later on. PCR-amplified products of virulence factors encoding genes located in each plasmid of B31 (or perhaps the new strain) can then be used as probes in Southern hybridization of the PFGE blots. Such a characterization will then help develop a PCR-based assay that can be used to determine the plasmid content of this strain in the follow-up studies. Some characteristics and virulence mechanisms are restricted to particular strains of any pathogen, and although such studies for Lyme disease have been severely limited in the past, B. burgdorferi studies involving multiple strains will be required to provide a complete picture of Lyme disease pathogenesis and will help determine the complete repertoire of the virulence factors contributing to this disease.
As a proof of this principle, we examined several known virulence factor-encoding genes as probes for Southern hybridization of the blot of the PFGE of the B. burgdorferi N40 and B31 strains (Fig. 1). In addition, examination of a new potentially important metabolic enzyme, MtnN, indicates that mtnN is located on a linear plasmid of 28 kb (probably lp28-4) in all tested strains. This is similar to the finding that the locations of important virulence factor-encoding genes, such as dbpA-dbpB, ospA-ospB, and ospC, are conserved in particular genome segments, even though sizes may be somewhat different due to genetic rearrangements, especially on the outer edges of the linear plasmids. Gene duplication followed by rearrangement and mutations appear to have resulted in the evolution and antigenic variation of a number of B. burgdorferi virulence factors and other proteins (4, 28, 43, 62, 82, 97, 109, 110). Often, duplicated genes that encode functionally redundant proteins that exhibit overlapping roles are present in the same operon, with each protein demonstrating altered specificities for host cell factors. Although mtnN and its homologs do not follow this pattern, their location is still conserved, further supporting our premise regarding the relatively conserved location of virulence factor-encoding genes. Hence, we believe that the use of the approach described here can help any researcher in designing a PCR-based assay to examine new strains and to examine the previously unstudied role of bacterial variation in the molecular pathogenesis of Lyme disease. Finally, a comparative analysis of the protein profiles can be used to determine association between the levels of protein production with the functional activity of a molecule. This analysis could further distinguish cN40 and N40D10/E9. Such analyses will be useful to differentiate other strains in the future.
A comprehensive study of four B. burgdorferi N40 isolates that have been used in the study of Lyme disease pathogenesis allowed us to differentiate these isolates and establish that they are derivatives of two different strains. This information will help to clarify the confusion that is due to the use of two different strains that were thought to be the same but that are in fact very different, and it will allow this research field to avoid future confusion on this point. We also propose that the use of PFGE and PCR of virulence factor genes followed by Southern hybridization is a useful strategy to discriminate different strains of B. burgdorferi sensu stricto. A PCR-based assay can then help determine the plasmid content of the strains.
We thank Kenneth Cornell and John Leong for critical reading of the manuscript and their constructive feedback.
This work was supported by grants AI089921 and AI49003 from the National Institute of Allergy and Infectious Diseases to N.P. and S.C., respectively.
Published ahead of print 30 January 2012
Supplemental material for this article may be found at http://iai.asm.org/.