bdrF₂ of Lyme Disease Spirochetes Is Coexpressed with a Series of Cytoplasmic Proteins and Is Produced Specifically during Early Infection

Abstract

The Bdr proteins are polymorphic inner membrane proteins produced by most Borrelia species. In Borrelia burgdorferi B31MI, the18 bdr genes form three subfamilies, bdrD, bdrE, and bdrF. The production of at least one of the Bdr paralogs, BdrF₂, is up-regulated in host-adapted spirochetes, suggesting a role for the protein in the mammalian environment. Here, we demonstrate using reverse transcriptase (RT) PCR that BBG29, BBG30, BBG31, and BBG32, which reside upstream of bdrF₂, are cotranscribed with bdrF₂ as a five-gene operon. While the functions of most of these proteins are unknown, BBG32 encodes a putative DNA helicase. Real-time RT-PCR analyses demonstrated higher levels of bdrF₂ transcript relative to other genes of the operon, suggesting that bdrF₂ may also be transcribed independently from an internal promoter. Internal promoters were detected using the 5′ rapid amplification of cDNA ends system. The putative promoter associated with bdrF₂ was found to be highly similar in sequence to the multiple promoters associated with the ospC gene. Real-time RT-PCR analyses, performed to assess the expression of these genes in infected mice, revealed that genes of the bdrF₂ locus are expressed only during early infection, suggesting a role in the establishment of infection. To further characterize the proteins encoded by the bdrF₂ locus, which have unknown functions, the cellular localizations of these proteins were determined by Triton X-114 extraction and phase partitioning. BBG29 and BBG31 were found to be cytoplasmic. To determine if these proteins elicit an antibody (Ab) response during infection, immunoblot analyses were performed. Abs to these proteins were not detected. Based on the analyses presented here, we offer the hypothesis that BdrF₂ and other proteins encoded by the operon form an inner-membrane-associated protein complex that may interact with DNA and which carries out its functional role during transmission or the early stages of infection.

The members of the genus Borrelia are causative agents of human and veterinary diseases, including Lyme disease, relapsing fever, epizootic bovine abortion, and avian borreliosis (2). Lyme disease is the most prevalent arthropod-borne disease in North America, with ∼24,000 cases reported to the Centers for Disease Control and Prevention in 2003. However, due to low compliance with reporting requirements, the incidence of this infection is certainly much greater (30). Borrelia spp. possess a segmented genome comprised of a linear chromosome and a variable collection of linear and circular plasmids (1, 3). The plasmids exhibit extensive sequence redundancy. Approximately 175 paralogous gene families, most of which encode proteins with unknown functions, have been delineated in Borrelia burgdorferi (10). The biological rationale for the maintenance of these gene families, an energy-expensive process, remains unclear. The Borrelia-specific paralogous protein families likely define the unique aspects of Borrelia biology and pathogenesis.

The bdr genes, which are carried on both linear and circular plasmids, encode a family of proteins ranging in size from 20 to 31 kDa (5, 22, 32). Size differences among Bdr paralogs are due to varying numbers of repeats (6) that harbor putative Ser-Thr phosphorylation motifs. We have demonstrated that the Bdr proteins are anchored to the inner membrane via a highly hydrophobic transmembrane-spanning C-terminal domain (26). Carboxy-terminal residues of the Bdr proteins are thought to extend into the periplasm, where they are potentially linked to the peptidoglycan, with the remainder of the protein residing in the cytoplasm. The importance of the Bdr proteins in Borrelia biology is highlighted by the fact that they are unique to this genus and exhibit genuswide distribution (25). The bdr genes form six distinct subfamilies referred to as bdrA through bdrF (6). Individual paralogs of each subfamily are differentiated within an isolate by a numerical subscript. For example B. burgdorferi B31MI, which harbors 18 different bdr alleles, carries three bdrF genes designated bdrF₁, bdrF₂, and bdrF₃. All Borrelia isolates maintain and express members of at least two different Bdr subfamilies, suggesting that individual subfamilies or paralogs may play different functional roles or be differentially expressed in different environments. To date, Bdr expression has been assessed only at the protein level (24). Using the dialysis membrane chamber rat implant model, the production of some Bdr proteins was up-regulated in spirochetes that were implanted for 8 days, while others were down-regulated (24). Most notably, the amount of BdrF₂ protein was significantly greater in host adapted bacteria than in bacteria temperature shifted from 25 to 37°C.

In this study, we analyzed the transcriptional expression patterns of bdrF₂ and its upstream genes with the underlying rationale that the functions of the upstream genes may be linked to those of the Bdr proteins and that information obtained from their study will facilitate efforts to define Bdr function. Here, we demonstrate cotranscription of BBG29, BBG30, BBG31, BBG32, and bdrF₂ and independent transcription of bdrF₂ from an internal ospC-like promoter. Like that seen for ospC, real-time reverse transcriptase (RT) PCR demonstrated that these genes are expressed only during early infection and that their expression is responsive to temperature. The similarity between the promoter elements and expression patterns of bdrF₂ and ospC suggest that these genes may be regulated through a common mechanism. At the protein level, immunoblot analyses demonstrated that proteins encoded by the bdrF₂ locus do not elicit an antibody (Ab) response, and consistent with this, Triton X-114 extraction and phase partitioning analyses, coupled with the deduced properties of these proteins, indicate a cytoplasmic localization for BBG29 through BBG32. The data obtained in this study will provide insight into the molecular mechanisms associated with the differential and temporal expression of Borrelia genes and will assist future analyses designed to identify the functional role of the Bdr protein family in Borrelia biology and pathogenesis.

MATERIALS AND METHODS

Cultivation of Borrelia isolates and animal studies.

B. burgdorferi B31MI was cultivated at 25, 33, or 37°C in BSK-H complete medium (Sigma), harvested by centrifugation, and washed twice with phosphate-buffered saline (PBS). To establish B. burgdorferi infections in mice, C3H/HeJ mice were needle inoculated intradermally between the shoulder blades with 10⁴ spirochetes (in PBS). To confirm infection and to isolate RNA for transcriptional analyses, ear punch biopsy specimens were collected 2 and 4 weeks postinoculation from three mice (and one control). A portion of the biopsy specimen was placed in BSK-H complete medium to allow growth of the bacteria. The remaining sample was used for RNA isolation as described below. Blood was collected from the infected mice at 2 and 4 weeks, and sera (henceforth referred to as infection sera) were obtained using standard methods. Infection of the mice was also confirmed through immunoblot analyses of the infection sera using membrane-immobilized B. burgdorferi B31MI cell lysate as a test antigen.

RNA isolation methods.

For RNA isolation from bacterial cultures, bacteria harvested as described above were lysed in a diethylpyrocarbonate-treated 1% solution of sodium dodecyl sulfate (SDS), and then RNA was recovered using the RNeasy Midi kit (QIAGEN) as instructed by the manufacturer. The integrity of the RNA was assessed by electrophoresis in a 1.5% agarose gel using Tris-acetate-EDTA (TAE) buffer. To isolate RNA from the tissues of infected mice, ear punch biopsy specimens were quick frozen on dry ice in 1.5-ml Eppendorf tubes, and macerated using a mortar and pestle, and RNA was purified using the RNeasy Mini kit (QIAGEN). Residual DNA was removed by treatment with DNase I (37°C; 1 h) in DNase buffer (Invitrogen). The DNase was inactivated with 2.5 mM EDTA (70°C; 10 min). The RNA was then used as a template in RT-PCR as described below. DNA was also extracted from infected mouse tissue using the DNeasy Tissue kit (QIAGEN).

PCR, RT-PCR, and real-time quantitative RT-PCR.

To generate a template for PCR, spirochetes were harvested from actively growing cultures by centrifugation, washed with PBS, and boiled for 10 min in H₂O. Standard PCR was performed as previously described (18), and the PCR products were analyzed by electrophoresis in 1.8% agarose or 2.5% MetaPhor agarose gels in TAE buffer. The amplicons were visualized by ethidium bromide staining. To determine the optimal parameters for PCR with each primer set, an annealing temperature gradient of 45 to 74°C was tested. The optimal magnesium concentration for each primer set was determined by testing magnesium concentrations ranging from 1.5 to 4 mM.

As a first step in conducting the RT-PCR analyses, residual DNA was removed from the RNA preparations by treatment with DNase I (37°C; 2 h; Invitrogen). The complete elimination of DNA was confirmed through test PCR analyses using the RNA as the template, a FlaB primer set, and Taq DNA polymerase. No product was detected except in the positive controls, indicating that the RNA preparations were free of DNA. RT-PCR was performed using Superscript II reverse transcriptase (Invitrogen) with 1 μg of RNA, 2.5 pmol of each specific primer (Table 1), and other reaction mixture components supplied by the manufacturer. The reaction mixtures were incubated at 42°C for 50 min, and the RT was inactivated by incubation at 70°C for 15 min. As a negative control, reactions were also run without RT. After reverse transcription, the cDNA was used as the template for amplification in PCR with Taq polymerase with the following cycle parameters: 40 cycles of 94°C for 45 s, 50°C for 45 s, and 72°C for 1.5 min. All RT-PCR and real-time RT-PCR analyses were performed using an MJ Research Opticon II Thermal Cycler.

TABLE 1.

Primers for PCR, RT-PCR, real-time RT-PCR, and ligase-independent cloning

Oligonucleotide name	Oligonucleotide sequence (5′-3′)a	Target sequence
G29IF	GACAGTCGCACCACAATGTC	G29 (739-758)
G29IR	CCATCCCTACGCTTCATCAT	G29 (1084-1103)
G30IF	GGGACAAGATTTGGCATGTT	G30 (65-84)
G30IR	GCACCAGCAACATTCTTTTG	G30 (349-368)
G31IF	TGTACCCCATAAAGGAGGGAGA	G31 (95-116)
G31IR	CAGACTGCGAAGCAAACTAGGA	G31 (471-492)
G32IF	GCTTTTGCCCTCAATATTGCTC	G32 (148-168)
G32IR	AGTTCAAGTGCAAGCGAACGTA	G32 (536-557)
G33IF	GAAAAGCAATTTGGCATAAAGTTTGA	G33 (148-173)
G33IR	ATTGGCTTCCAGCTTTATTTCTAGCC	G33 (548-573)
G29-30F	ACGCCGCCGGAGAGAATTTACTTA	G29 (1327-1350)
G29-30R	CTTCTGAATCCACTCATCCACGAAC	G30 (186-210)
G30-31F	TGGCTAGGATTGTCAAAGACTATAAAGA	G30 (305-332)
G30-31R	AGGCCCGACAACCCTTATAAAT	G31 (9-31)
G31-32F	CATATGCGTCTGATGGAGCTAGA	G31 (340-362)
G31-32R	ACTGGGCCGTGCTCCAATTACTA	G32 (113-135)
G32-33F	CAACTTTGGGAGAATCGGCAGCAT	G32 (623-646)
G32-33R	GTGGATCCCGGGTGGTGTCGAGTTGTTCCATACCTAGTCGCAGA	G33 (57-80)
FlaB-F	GCAGTTCAATCAGGTAACGG	FlaB (280-299)
FlaB-R	AGGTTTTCAATAGCATACTC	FlaB (844-863)
G29QF	AAGGCCTTAGACGAGCACAA	G29 (991-1010)
G29QR	CCATCCCTACGCTTCATCAT	G29 (1084-1103)
G30QF	TGGCATGTTTACAGAAGTACGG	G30 (76-87)
G30QR	TGAATCCACTCATCCACGAA	G30 (184-203)
G31QF	GGTTGTCGGGCCTATAAACA	G31 (19-38)
G31QR	AAAAACGCATCTCCCTCCTT	G31 (106-125)
G32GF	TTCGCTTGCACTTGAACTTG	G32 (540-559)
G32QR	TGTAATGCTGCCGATTCTCC	G32 (631-650)
G33QF	AGCTGGAAGCCAATAGCAAA	G33 (626-645)
G33QR	TGGGCACAACTACTACTGCAA	G33 (722-742)
FlaF	TTCATGTTGGAGCAAACCAA	Flagellin (524-543)
FlaR	CTGAGCAGTTTGAGCTCCCT	Flagellin (602-621)
G29P1	GCGACATTGTGGTGCGA	G29 (744-760)
G29P2	GGTGCGACTGTCAACATCCTGATT	G29 (727-750)
G29P3	CTGCACAGTGCGCACACTAACAACCTT	G29 (295-321)
G30P1	CTTGTGCACCAGCAACAT	G30 (356-373)
G30P2	TGTATTTCTTCTGAATCCACTCATCCACG	G30 (186-214)
G30P3	TGTATTTCTTCTGAATCCACTCATCCACG	G30 (179-207)
G31P1	CCTTATAAATATTCATCAAC	G31 (−1-+19)
G31P2	CCTTATAAATATTCATCAACCTTGTATGGT	G31 (−11-+19)
G31P3	AATATTCATCAACCTTGTATGGTAATTGG	G31 (−17-+12)
G32P1	TCAGCGCTTCTGGAAACT	G32 (315-332)
G32P2	GGAATTCCAAGTTCAAGTGCAAGCGA	G32 (279-306)
G32P3	AGTGCAAGCGAACGTATATTACGACTCA	G32 (198-225)
G33P1	GCAGTGGGCACAACTACT	G33 (729-746)
G33P2	AGAACTTTGCTATTGGCTTCCAGCTT	G33 (625-650)
G33P3	GGCTTCCAGCTTTATTTCTAGCCTTTGT	G33 (543-570)
AAP	GGCCACGCGTCGACTAGTACGGGIIGGGIIGGGIIG	Abridged anchor primer
AUAP	GGCCACGCGTCGACTAGTAC	Abridged universal primer
G29EKF	GACGACGACAAGATGAATAAAATATTATTAAATAATGCTAGA	G29 (1-30)
G29EKR	GAGGAGAAGCCCGGTTTAAATTGCTTTTTTATCAATTCCAAG	G29 (1441-1467)
G30EKF	GACGACGACAAGATGGCTTTAGGAGGTAAAATGAGAATAAGA	G30 (1-30)
G30EKR	GAGGAGAAGCCCGGTTTATCCTCCTGTTCCCGCCG	G30 (410-429)
G31EKF	GACGACGACAAGATGATGAATATTTATAAGGGTTGTCG	G31 (1-26)
G31EKR	GAGGAGAAGCCCGGTTGTAAATTCAAAATCTACGCAGTCGA	G31(501-526)
G32EKF	GACGACGACAAGATGAAAGTCGCTAGTCTTATAAGGTCAA	G32 (1-28)
G32EKR	GAGGAGAAGCCCGGTCTCACCTACTAAATCATTGTAACACA	G32 (886-912)
G33EKF	GACGACGAAAGATGAAATCATCAGTAGTGACAACAAGTA	G33 (1-28)
G33EKR	GAGGAGAAGCCCGGTCTTGCAAATTACAAAAGCGCATTATATC	G33 (846-872)

^aThe tails added to allow for ligase-independent cloning into the pET32-Ek/LIC vector are indicated by underlining.

Real-time RT-PCR was performed using the SYBR Green PCR Master Mix (Applied Biosystems) as instructed by the manufacturer. Primer pairs (final concentration, 0.5 μM) were designed to amplify a 100- to 150-bp fragment of each gene. All reactions were performed in triplicate. A standard curve was generated using serial dilutions of a known amount of genomic DNA as the template. PCR was performed using the following cycling parameters: 1 cycle of 10 min at 95°C and 40 cycles of 94°C for 10 s, 60°C for 20 s, and 72°C for 20 s. Melting curves were generated over a temperature range of 55 to 95°C. PCR products were quantified, and the data were analyzed using software provided by the thermocycler manufacturer (MJ Research).

Identification of promoter elements using the 5′ RACE approach.

To identify the promoters of the bdrF₂ operon and its individual genes, the 5′ rapid amplification of cDNA ends (RACE) (Invitrogen) approach was employed as described by the manufacturer. In brief, primers (2.5 pmol) specific for each gene (Table 1) were annealed with 3 μg of total cellular RNA, and cDNA was generated using SuperScript II reverse transcriptase (42°C; 50 min). The RNA was degraded using RNase, and the cDNA was purified using the GlassMAX spin cartridge (Invitrogen). A poly(C) tail was added to the 3′ end of the cDNA using terminal nucleotidyltransferase. The manufacturer-supplied Abridged Anchor primer, which hybridizes to the 3′ poly(C) tail of the cDNA, was then used in conjunction with a nested gene-specific primer to reamplify the cDNA. A final PCR was performed using the Abridged universal amplification primer in conjunction with a third nested gene-specific primer. The resulting PCR products were analyzed by agarose gel electrophoresis and cloned into the TOPO pCR2.1 vector (Invitrogen), a TA cloning vector, as instructed by the manufacturer. DNA sequence analysis was performed using automated methods, and the transcriptional start site (TSS) was inferred from the sequence.

SDS-PAGE, immunoblotting methods, and generation of recombinant protein and antisera.

To generate recombinant BBG29, BBG30, BBG31, BBG32, and BdrF₂, primers (Table 1) were designed to amplify the entire coding sequence with tails to allow annealing of the PCR product into the pET32-Ek/LIC vector (Novagen) in a ligase-independent fashion. PCR was performed using the Expand High Fidelity PCR system (Roche) as described by the manufacturer. The PCR products were then treated with T4 DNA polymerase to generate single-stranded overhangs and annealed into the pET32-Ek/LIC vector as instructed by the supplier. The recombinant plasmid was transformed into and propagated in Escherichia coli NOVABlue cells. Plasmid purified from the NOVABlue cells was then transformed into E. coli BL21(DE3) cells. Colonies found to be carrying the desired insert through PCR screening were cultivated overnight at 37°C in Luria broth with ampicillin (50 μg ml⁻¹). IPTG (isopropyl-β-d-thiogalactopyranoside) induction was performed as previously described (19). All recombinant proteins expressed from the pET32-Ek/LIC vector were generated with an N-terminal fusion that contains both S and His tags. The N-terminal fusion adds ∼17 kDa to the molecular mass of the recombinant protein. All SDS-polyacrylamide gel electrophoresis (PAGE) and immunoblotting procedures employed in these analyses were as previously described (14). To verify expression of the recombinant proteins, E. coli cells that had been induced with IPTG were lysed, fractionated by SDS-PAGE, immunoblotted, and screened with S protein-horseradish peroxidase conjugate (Novagen). In this and all other immunoblot analyses, the Super Signal substrate (Pierce) was employed for chemiluminescent detection of S protein and Ab binding.

To generate antisera to the full-length recombinant proteins, 25 μg of recombinant protein in Freund's complete adjuvant was injected into C3H/HeJ mice (4 to 6 weeks of age), and then boosts were given at 2 and 4 weeks in incomplete Freund's adjuvant. The mice were sacrificed at week 8, and the specificities of the antisera were confirmed by immunoblot analyses. In all subsequent analyses, the antisera were used at a dilution of 1:1,500.

Triton X-114 extraction and phase partitioning.

Triton X-114 extraction and phase partitioning were performed as described by Cunningham et al. (8). Briefly, exponential-phase cells (from a 2-ml culture) were harvested by centrifugation, resuspended in 1% Triton X-114 (in PBS), and incubated at 4°C overnight with gentle rocking. The detergent-insoluble phase was collected by centrifugation at 4°C at 15,000 × g. The supernatant was incubated at 37°C for 15 min and then centrifuged (15,000 × g) at room temperature to separate the aqueous and detergent-soluble phases. Each sample was extracted twice to ensure complete partitioning. The resulting samples were prepared for SDS-PAGE by resuspension in SDS solubilizing solution, fractionated in SDS-12.5% PAGE gels, and immunoblotted. A series of identical blots were then screened with anti-DbpA (1:1,000), anti-FlaB (1:1,000), anti-Bdr (1:,1000), or antiserum generated with peptides specific for BBG29, BBG30, BBG31, and BBG32 (1:200). The generation of the antisera is described below. Goat anti-mouse immunoglobulin G (IgG) antiserum served as the secondary Ab for all antisera except anti-Bdr, for which goat anti-rabbit IgG was used. Detection was accomplished using the Pierce SuperSignal substrate.

Synthesis of peptides and MAPs and generation of antisera to each MAP construct.

Peptides specific for regions in the N- and C-terminal domains of BBG29, BBG30, BBG31, and BBG32 were synthesized (Table 2), and multiple antigenic peptide (MAP) constructs were generated. Each construct contained a T-cell epitope with the sequence ISQAVHAAHAEINEAGR-NH₂. After synthesis, a K residue was coupled to the N terminus of the T-cell epitope, and then two additional K residues were coupled to the lysine to generate a branched construct. The peptides listed in Table 2 were then coupled to the two epsilon and two alpha amino groups to generate the MAP construct. To generate antisera to each MAP construct, 50 BALB/c-CF1 female mice (7 to 10 weeks of age) were randomly assigned to 10 groups and immunized subcutaneously with 15 μg of each MAP construct in Montanide ISA720 (Seppic). Each mouse was boosted two times (at 2-week intervals) with the same material in Montanide ISA720. Sera were collected by tail snipping, and Ab titers were assessed by enzyme-linked immunosorbent assay. The specificities of the antisera were confirmed by screening immunoblots of whole-cell Borrelia lysates and lysates of E. coli that had been induced to express each of the recombinant proteins described above.

TABLE 2.

Peptides employed in the generation of MAP constructs

Peptide designation	Peptide sequence
OVA323-339 (T-cell epitope)	ISQAVHAAHAEINEAGR
BBG29-N	IDVKNQNYKNSCGVDRYSAC
BBG29-C	RSNIVKDSLGFKTIKGITLE
BBG30-N	TEIQNKAKDKSITNFPKLEP
BBG30-C	ILARIVKDYKDIFLGRLQKN
BBG31-N	LYPIKEGDAFLGIFYGYNKV
BBG31-C	GVFKFYGRDLNPQGEGIITK
BBG32-N	SFWINSVWGTDIHELEDKAR
BBG32-C	ERKKRKGRNTSKGNNTTKVK
BdrF2-N	DLSKRYYHNELTYRDLENLE
BdrF2-C	SSEKLKVSNRIVIIAVVVVP

RESULTS

Demonstration of the coexpression of BdrF₂ with upstream genes.

In B. burgdorferi B31MI, the BBG29, BBG30, BBG31, and BBG32 genes reside upstream of bdrF₂, have the same orientation, and exhibit relatively short intergenic spacers (Fig. 1 and Table 3). To determine if this gene cluster is cotranscribed, RT-PCR was employed. It is important to note that all of these genes except BBG30 belong to gene families (Table 3) (10). Hence, prior to performing the RT-PCR analyses, it was essential to establish the allele specificities of the primers. This was accomplished through PCR analysis using isolated B. burgdorferi B31MI genomic DNA as a template. All primer sets yielded a single amplicon of the expected size (Fig. 1). As a first step in the RT-PCR analyses, the expression of each individual gene in bacteria cultivated at 37°C was demonstrated using paralog-specific primer sets (Fig. 1). Duplicate reactions lacking reverse transcriptase did not yield products, indicating that the RNA preparations were free of contaminating genomic DNA. To determine if genes BBG29 through bdrF₂ are cotranscribed, primer sets designed to amplify across the intergenic spacer between each pair of genes were employed. A primer set designed to amplify from BBG29 through bdrF₂ was also tested. Products of the expected size were obtained in all reactions, indicating that BBG29, BBG30, BBG31, BBG32, and bdrF₂ are cotranscribed as a polycistronic mRNA (Fig. 1). Henceforth, we refer to this transcriptional unit as the bdrF₂ operon.

FIG. 1.

PCR and RT-PCR analyses of the bdrF₂ locus. A schematic of the bdrF₂ locus is depicted, with gene lengths (in basepairs) indicated above each gene and the regions amplified in the different PCR analyses indicated by brackets below the schematic. The small arrows above the schematic of the gene organization indicate the relative positions of transcriptional start sites identified as part of this study. The PCR panel on the upper left shows the PCR analyses that were performed to demonstrate the specificity of the primer sets. The panel on the upper right depicts the RT-PCR analyses of bdrF₂ and its upstream genes. RT-PCR was performed as described in the text using RNA extracted from B. burgdorferi B31MI grown at 37°C. The resulting amplicons were analyzed in 1.8% agarose gels. Molecular mass markers (in base pairs; New England Biolabs) are indicated. Controls for these analyses included flaB and reactions either lacking template (NT) or in which RT was omitted (N-RT).

TABLE 3.

Features of bdrF₂ and its upstream genes

Gene name	Gene family (no. of paralogs)	Intergenic spacer	Mass (kDa)	Protein characteristics
BBG29	PGF62 (9)	G29-30 (10 bp)	56.7	Hypothetical conserved
BBG30	None	G30-31 (80 bp)	16.3	Hypothetical conserved
BBG31	PGF50 (23)	G31-32 (31 bp)	21.0	Hypothetical conserved
BBG32	PGF46 (2)	G32-33 (94 bp)	30.9	Putative DNA helicase
bdrF2 (BBG33)	PGF80 (18)	Not applicable	30.5	Inner membrane protein; central repeats; putative Ser-Thr phosphorylation motifs

Analysis of the influence of cultivation conditions and the host environment on the expression of the bdrF₂ operon.

To determine if the bdrF₂ operon is temperature regulated, bacteria were cultivated at either 23, 33, or 37°C and harvested, and RNA was isolated and real-time RT-PCR was performed using SYBR Green with primers specific for each gene in the operon (Fig. 2). In all real-time RT-PCR analyses, reactions lacking reverse transcriptase served as the negative controls. To allow comparison of the relative expression levels of the genes, a correction for the different efficiency of SYBR Green binding by each amplicon was performed. To accomplish this, the dye incorporation levels that result upon PCR with each primer set and a standardized amount of template DNA were determined. These values were then used to normalize the results. The flaB gene, which is constitutively expressed, served as a control in these analyses, and as expected, its expression was not affected by temperature (data not shown). Although BBG29, BBG30, BBG31, BBG32, and bdrF₂ can be expressed as a polycistronic mRNA, the data revealed a significantly higher level of bdrF₂ mRNA than those of BBG29, BBG30, BBG31, and BBG32 under all conditions tested. The differing levels of mRNA for each gene suggest that bdrF₂ can also be expressed from an independent internal promoter.

FIG. 2.

Real-time RT-PCR of the expression of genes in the B. burgdorferi bdrF₂ locus cultivated at different temperatures and during residence in the tissues of infected mice. Real-time RT PCR was conducted as described in the text. The RNA for these analyses was isolated from B. burgdorferi B31MI cultivated at different temperatures (23, 33, and 37°C) and from ear punch biopsies from infected mice (Tissue). The ear punch biopsy specimens were collected from mice infected for 2 weeks. The constitutively expressed flaB gene served as a control in these analyses (data not shown). Note that expression levels were normalized based on the values obtained for the flaB gene. All reactions were done in triplicate, and each set was run twice on two different occasions. The standard deviations for these experiments are indicated by the error bars.

The expression of the bdrF₂ operon and its individual genes in spirochetes residing in mouse tissue was also assessed. RNA was extracted from ear punch biopsy specimens collected from three different C3H/HeJ mice that had been infected with B. burgdorferi B31MI by needle inoculation. The biopsy specimens were collected at 2 (early infection) and 4 (late infection) weeks. As described above, to allow comparison of the expression levels of transcript in mouse tissue with that observed during in vitro cultivation, the flaB transcript levels were used for normalization. All genes of the bdrF₂ locus were slightly up-regulated in the host at the 2-week infection time point relative to the levels seen in bacteria cultivated at 37°C (Fig. 2), and the relative level of each transcript did not change in spirochetes in tissue during the early stages of infection. However, an important observation was that transcripts for BBG29, BBG30, BB31, BBG32, and BdrF₂ were not detected in any of the infected mice at week 4. In contrast, there was only a slight decrease in the amount of flaB transcript detected at 4 weeks compared with that seen at 2 weeks. The slight decrease in the amount of flaB mRNA detected at 4 weeks is completely consistent with the decrease in spirochete numbers that occurs in tissue at this point in infection. The data indicate that genes of the bdrF₂ locus are down-regulated as infection progresses, suggesting that these proteins play a role specifically during the early stage of infection.

Identification of transcriptional start sites and demonstration of internal promoters within the bdrF₂ operon.

To identify the TSS for the bdrF₂ operon and to determine if internal promoters allow expression of bdrF₂ independent of the bdrF₂ operon, the 5′ RACE approach was employed. This approach allows the identification of the TSS and is particularly well suited for the analysis of low-level transcripts, such as bdr mRNAs. Using this approach, the TSS for the full bdrF₂ operon was identified (Fig. 3). The TSS mapped to an A residue located 2 bases upstream of the ribsomal binding site (RBS) sequence, GAGGA, and 20 bases upstream from the translational start codon of BBG29. Ten bases upstream of the TSS is an AT-rich region that presumably serves as the −10 site. An identifiable consensus −35 promoter element was not evident by sequence scanning.

FIG. 3.

Identification of the transcriptional start site for the bdrF₂ operon and demonstration of the existence of internal promoters for BBG31 and bdrF₂. To identify transcriptional start sites, the 5′-RACE approach was employed. All methods were as described in the text. The resulting amplicons were analyzed by electrophoresis in a 1.8% agarose gel in TAE buffer. The amplicons were then cloned and sequenced to identify the transcriptional start site, which is the first residue of the amplicon. Size standards (in base pairs) are indicated on the left.

These analyses also identified at least two putative promoters within the bdrF₂ operon that could drive the expression of one or more genes of the operon. TSSs that mapped just upstream of BBG31 and bdrF₂ were identified. The TSS for the internal promoter for the BBG31 gene mapped to an A residue located 117 bases upstream from the start codon of BBG31 and within BBG30. This would place the apparent promoter for this transcript within the coding sequence of the BBG30 gene. However, a strong RBS or −10 sequence was not evident. The TSS for the independently transcribed bdrF₂ was identified as a T residue located 5 bases upstream of the RBS (GGUGAG) and 17 bases upstream from the translation start codon. An appropriately spaced −35 promoter element with the sequence TTGTAAA was evident. This putative promoter is homologous to the repeated promoter elements of the ospC gene (TTGAAAA) (17). It was not possible to determine if BBG29, the first gene in the operon, is transcribed independently of other downstream genes, since transcript derived from expression of the operon cannot be distinguished from transcript derived from a BBG29-specific promoter using the approach taken here.

Analysis of the subcellular localization of BBG29, BBG30, BBG31, and BBG32.

Earlier analyses demonstrated that BdrF₂ and other Bdr proteins are inner membrane proteins (26). To determine the cellular localization of the proteins encoded by other genes of the bdrF₂ locus, bacteria were subjected to Triton X-114 extraction and phase partitioning. The proteins in each of the resulting fractions were separated by SDS-PAGE and immunoblotted (Fig. 4). The controls for these analyses included immunoblot detection of DbpA (decoring binding protein A, a known lipidated outer membrane protein) (11), FlaB (a protein associated with the protoplasmic cylinder), and Bdr (inner-membrane-localized proteins) (26). All of the control proteins partitioned into the appropriate phases. To facilitate the identification of the subcellular locations of BBG29, BBG30, BBG31, and BBG32, antiserum to each was generated using either paralog-specific peptides in the form of MAP constructs or recombinant proteins. Immunoblot analyses confirmed the specificities of all antisera (data not shown). However, only BBG29 and BBG31 (and the control, BdrF₂), which partitioned exclusively into the aqueous phase after Triton X-114 extraction and phase partitioning, could be detected in these analyses. In addition, these proteins could be detected only after the subcellular fractions were concentrated 50-fold by trichloroacetic acid precipitation. The fact that significant concentration is required to detect these proteins indicates that they are of low abundance, a finding that is completely consistent with the amount of mRNA detected for each in the transcriptional analyses described above. The inability to detect these proteins in the whole-cell lysates is not surprising in light of their low abundance. Note that it was not possible to concentrate and analyze the whole-cell lysates in the same manner as the subcellular fractions. The partitioning of BBG29 and BBG31 into the aqueous phase is consistent with the inferred properties of these proteins, which are predicted to be hydrophilic and to lack export signals, transmembrane domains, and lipidation signals. These properties and the demonstrated partitioning of BBG29 and BBG31 into the aqueous phase support a cytoplasmic localization. While BBG30 and BBG32 could not be detected in these analyses, they have properties analogous to those of BBG29 and BBG31 and hence are likely to reside in the cytoplasm as well.

FIG. 4.

Analysis of the subcellular localization of proteins encoded by the bdrF₂ locus. To determine the cellular locations of BBG29, BBG30, BBG31, and BBG32, Triton X-114 extraction and partitioning were performed as described in the text. Note that due to the low abundance of these proteins, aliquots of the fractions obtained after extraction (excluding the whole-cell lysates) were concentrated 50-fold by precipitation with trichloroacetic acid. The fractions were then subjected to SDS-PAGE, immunoblotted, and screened with antiserum to each protein, FlaB, and DbpA. While markers are not indicated, all proteins detected were of the correct size. Note that BBG30 and BBG32 were not detected upon immunoblotting and hence are not shown. Abbreviations used are as follows: WC, whole-cell lysate; DI, detergent-insoluble phase; DS, detergent-soluble phase; AQ, aqueous phase.

Assessment of the potential Ab response to proteins encoded by the bdrF₂ operon.

To determine if the proteins encoded by the bdrF₂ locus are antigenic during infection in mice, recombinant S tag fusion proteins were generated for each protein and used as test antigens in immunoblot analyses. To confirm the expression of each protein, induced E. coli lysates were screened with S protein-horseradish peroxidase conjugate (Fig. 5). S tag fusion proteins of the appropriate size were detected for all proteins. However, BBG29 proved to be moderately toxic to E. coli and less stable than other proteins encoded by the operon, as a significant portion of the expressed protein was truncated and the growth rate of the cells was attenuated. Nonetheless, sufficient full-length BBG29 was produced to allow its use as one of the test antigens. Immunoblot analyses using sera from infected mice revealed that none of the recombinant proteins were detected by infection-induced antibodies. BBO39, an OspF paralog that has been demonstrated to be expressed during infection and to elicit a strong IgG response, served as the positive control (19). Antibodies to this protein were readily detectable 8 weeks postinfection. The lack of antibodies to the proteins encoded by the bdrF₂ locus is not surprising and is consistent with the low expression levels and subcellular-localization studies which indicate that these proteins are cytoplasmic and thus unlikely to elicit an Ab response during infection.

FIG. 5.

Analysis of Ab response to BBG29, BBG30, BBG31, BBG32, and BdrF₂ during infection in mice. Recombinant proteins were expressed in E. coli as S tag fusions, fractionated by SDS-PAGE, and immunoblotted. Four identical blots were generated and screened with S tag protein or infection sera collected at week 4, 8, or 12. The infection sera were recovered from mice infected with B. burgdorferi B31MI. Molecular mass standards (in kilodaltons) are indicated. BBO39, an OspF paralog expressed during infection, served as a positive control. Note that the S tag fusion present at the N terminus of each recombinant protein adds 17 kDa to the molecular masses of the proteins.

DISCUSSION

The bdr gene family of B. burgdorferi B31MI is comprised of 18 members that are plasmid encoded (6, 24, 25, 32). The overall biological rationale for the maintenance of large paralogous gene families like the bdr family is unclear, but it may provide a mechanism to rapidly adapt to changing environmental pressures through the differential expression of paralogs. This could be of particular importance to Borrelia spp., as they cycle between ticks and mammals. It is interesting that most of the paralogous gene families investigated to date exhibit differential expression of paralogs (9, 12, 13, 19, 24, 29). The lack of coordinate regulation of paralogous gene family members suggests that the proteins encoded by these genes are not redundant in function and that individual paralogs may carry out paralog-specific functions. Roberts et al. demonstrated that the production of some Bdr paralogs is influenced by environmental parameters (24). Production of BdrF₂ (at the protein level) was found to be up-regulated in host-adapted spirochetes generated using the dialysis membrane chamber implant model, while the production of selected other paralogs was down-regulated (24). However, little is known concerning the mechanisms of regulation of Bdr production. In this report, we have analyzed the transcriptional expression of bdrF₂ and its upstream genes under different environmental conditions and in infected animals.

The bdrF₂ gene is located on a linear plasmid of ∼28 kb designated lp28-2 or plasmid G (10). Open reading frames BBG29 through BBG32 reside just upstream of bdrF₂, and based on the relatively short intergenic spacers between these genes, we hypothesized that they may be cotranscribed. This hypothesis was confirmed through RT-PCR analyses. BBG29, BBG30, BBG31, and BBG32 were found to be cotranscribed with bdrF₂ to yield a transcript of at least 4,400 bases. To determine if environmental conditions influence the expression of the bdrF₂ operon, SYBR Green-based real-time RT-PCR was conducted using RNA isolated from spirochetes propagated under different environmental conditions. For these analyses, primer sets that target each individual gene of the operon were used. Temperature (37 versus 25°C) was found to increase transcription of the operon as a whole, and as expected, the increases for the individual genes of the operon were similar, ranging from ∼3- to 4-fold. It is important to note, however, that at any given temperature the amount of bdrF₂-derived transcript was at least three- to sevenfold greater than that of any other gene in the operon. Slightly higher BBG32 transcript levels were also noted. These data suggest that there may be some degree of independent transcription of BBG32 and bdrF₂ from internal promoters. In an earlier analysis, it was observed that the level of BdrF₂ protein did not significantly increase when bacteria were cultivated at 37°C but did increase in host-adapted bacteria generated by implantation of dialysis membrane chambers in the peritoneal cavities of rats (24). Transcript levels were not assessed in that report. While the data presented here indicate that a temperature shift to 37°C leads to an increase in transcription, immunoblot analyses of BdrF₂ using anti-Bdr antisera confirmed our earlier observation that this does not result in increased protein production in cultivated spirochetes (data not shown). This suggests that posttranscriptional control mechanisms are involved and that increased expression, at least during the early stages of infection, may be dependent on host factors yet to be identified.

The expression levels of genes of the bdrF₂ locus were also determined for spirochetes residing within mouse tissue. While expression of these genes was readily detected at an early infection time point (2 weeks), expression was completely turned off by week 4 of infection. It is important to note that while flaB transcript was readily detected in RNAs extracted from three different mice, transcript derived from the bdrF₂ locus was not. This suggests that the proteins derived from this locus are produced only during early infection and therefore may play a role in transmission or in the establishment of infection. It is not yet known if these genes are expressed during residence of the spirochetes in ticks. The implications of these expression data are discussed in more detail below.

In light of the differing levels of transcript detected for individual genes within the bdrF₂ operon, we sought to determine if internal promoters for bdrF₂ and BBG32 are involved in the expression of these genes. With this goal, the 5′ RACE approach was employed to identify internal TSSs that would serve as an indication of internal promoters. This approach is well suited for identifying the TSSs of low-abundance transcripts. The data presented here indicate that the amount of bdrF₂ operon transcript is ∼1,000-fold less than that of flaB. While Northern hybridization analyses could provide important information regarding the expression patterns of the bdrF₂ locus and could serve to directly demonstrate the sizes of different transcripts derived from this locus, due to the low expression levels of these genes in cultivated bacteria, we have not been successful in detecting bdr mRNA by this approach. Several earlier reports also indicated that bdr transcript levels are below the threshold level of detection by Northern hybridization (7, 24, 28). Using the 5′ RACE approach and subsequent DNA sequence analysis of the products, three TSSs were identified. One mapped upstream of BBG29 and served as the TSS for the full-length operon. The second occurred 5′ of BBG31 and mapped within the BBG30 gene. The third TSS mapped within the BBG32-bdrF₂ intergenic spacer. The putative bdrF2 promoter (TTGTAAA) exhibits strong homology with the promoters that have been demonstrated to drive expression of the ospC gene (TTGAAAA) (17). Other similarities between ospC and bdrF₂ are that both have 7-bp repeat motifs in their promoter regions and both genes initiate transcription with a T residue (17). It is possible that the repeat motifs of ospC and bdrF₂ influence the transcription of these genes. Analysis of the upstream sequences of other bdrF subfamily genes (bdrF₁ and bdrF₃) revealed that except for the RBS there is little or no conservation of sequence among them (10). This is perhaps not surprising, as earlier studies by Roberts et al. demonstrated that the production of members of the bdrF subfamily is not coordinately regulated (24). Sequence variation within the promoter regions for the bdrF genes may be the basis for their observed differential regulation. In fact, paralogs of most Borrelia gene families characterized to date have been found to be differentially expressed and not coordinately regulated (12, 19, 29). Of particular relevance to this study is the mlp gene family (29). These genes, a subset of which reside immediately downstream from some bdr genes, are not coordinately regulated, and in fact, some may be cotranscribed with the other bdr genes (29).

The analyses presented here indicate that transcription of bdrF₂ from its internal promoter is responsive to increasing temperature and that upon entering the host it is expressed only during early infection. Similarly, ospC is also upregulated by increasing temperature and, like bdrF₂, is down-regulated after the establishment of infection (16, 27). It was hypothesized that the down-regulation of OspC during infection is driven by immune pressure or selection (16). Since proteins encoded by the bdrF₂ locus do not elicit an antibody response, immune pressure or selection is clearly not involved in down-regulating their transcription. The similarity between the transcriptional control elements of these genes and their similar expression patterns raises the possibility that immune pressure or selection may not be the key driving force in down-regulating OspC. The decrease in transcription of OspC upon the development of an anti-OspC antibody response may simply be coincidental. In any event, the analyses presented here suggest that bdrF₂ and ospC may be regulated through similar molecular mechanisms. However, one caveat to consider is that while ospC is regulated by the RpoS-RpoN regulatory network (15), the bdr genes are not (24). While the real-time RT-PCR data provide strong evidence that the internal promoter associated with transcription of bdrF₂ is active, the contribution of the internal promoter that resides within BBG30 to the transcription of BBG31 (and possibly BBG32 and bdrF₂) is less clear. It remains to be determined if this particular promoter is biologically relevant. It is possible that it may be more active under environmental conditions that have not been analyzed in this study.

Prior to this report, several studies employing microarray-based analyses investigated differential gene expression by Borrelia spp. (4, 20, 21, 23, 31). However, due to the inherent difficulty in assessing the expression of paralogous gene family members using global approaches such as microarrays, definitive conclusions regarding the expression pattern of each gene analyzed in this study could not be reached. Nonetheless, the information presented in those analyses pertaining to the genes analyzed here are for the most part in good agreement. Consistent with what is reported here, Zhong and Barbour demonstrated that the BBG30 homolog carried by Borrelia hermsii is significantly down-regulated during late infection (31). However, Ojami et al. analyzed the influence of temperature on BBG30 expression in B. burgdorferi and noted an ∼2-fold increase in expression at 23 versus 35°C (21). In this report, we noted the opposite effect. The basis for this discrepancy is unclear. Brooks et al. reported that BBG31 was down-regulated in host-adapted bacteria generated using the dialysis chamber implant model (4). This finding is consistent with the decreased expression of BBG31 that we observed 4 weeks into infection.

In conclusion, the analyses presented here demonstrate that a complex transcriptional regulation system is in place to control the expression of bdrF₂ and its flanking genes. As described above, with the exception of BBG30, all of these genes belong to paralogous gene families. BBG31, a hypothetical conserved protein, belongs to a family with 23 members. The ability to tightly regulate the transcriptional expression of these genes may allow the Lyme disease spirochetes to tailor the expression patterns of individual paralogs to specific environmental conditions. Regarding the function of the Bdr proteins, the low abundance of these proteins and the inherent difficulty associated with gene inactivation of members of extensive gene families has hampered the identification of their specific functional roles. However, the demonstration that bdrF₂ is coexpressed with other genes and identification of the environmental conditions that influence their expression may provide clues to their functions. While the biological roles of most genes of the bdrF₂ operon are not known, BBG32 represents a putative DNA helicase (10). We previously demonstrated that BdrF₂, as well as other Bdr proteins, are inner membrane anchored via a strong C-terminal transmembrane-spanning domain (TMpred value, >2,000) (24). The remainder of the protein, including the putative Ser-Thr phosphorylation domain, extends into the cytoplasm of the cell. It is our hypothesis that BdrF₂ (and perhaps other Bdr proteins) serves as the scaffolding for the formation of an inner membrane protein complex that may involve the cytoplasmic proteins BBG29, BBG30, BBG31, and BBG32. This complex may interact with DNA through BBG32, the putative DNA helicase. We further speculate that phosphorylation-dephosphorylation of the Bdrs may play a regulatory role and possibly influence the potential interaction of this putative inner membrane protein complex with DNA. Future analyses will focus on the transcriptional expression patterns of other bdr paralogs and will seek to test the hypothesis that a putative BdrF₂ inner membrane protein complex exists, interacts with DNA, and potentially plays a regulatory role in Borrelia.