The saprophyte L. biflexa can be transformed at high rates with plasmids based on the LE1 replication origin, using kanamycin, spectinomycin, or gentamicin resistance as the selectable marker 8 27 28 . We chose the spectinomycin-resistant plasmid, pSLe94, as the backbone for our system: this shuttle plasmid containing the LE1 partition genes is stably maintained in L. biflexa in the absence of antibiotic selection 27 . Flagellin-encoding genes are usually both constitutively and strongly expressed. In addition, it has been reported that a kanamycin resistant cassette driven by the Borrelia burgdorferi flgB promoter is strongly expressed in B. burgdorferi 29 and in L. biflexa 4 . We therefore used the flgB promoter from B. burgdorferi to allow strong and stable expression of LigA and LigB proteins in L. biflexa serovar Patoc (further indicated as Patoc). The genes encoding LigA and LigB under the control of the flgB promoter were inserted into the L. biflexa replicative plasmid (Figure 1A). The Patoc wild-type (wt) strain was then electrotransformed by pSLePFligA and pSLePFligB, and the spectinomycin-resistant transformants were further analyzed. Lig expression by the lig-transformed Patoc strains was verified by Western blot analysis, which showed levels of protein comparable to the production by a low in vitro-passaged L. interrogans virulent strain (i.e. less than 10 in vitro passages). However, blots of the ligB transformant showed partial degradation of LigB (Figure 1B). The Patoc wt, ligA, and ligB strains had similar cell growth kinetics in EMJH liquid medium, indicating that the expression of the heterologous proteins did not affect cell growth (data not shown).

LigA and LigB proteins have been shown to be surface-exposed proteins in pathogenic Leptospira strains 11 . This was confirmed in this study with antibodies against LigA and LigB (see additional file 1: surface immunofluorescence assays in L. interrogans). Immunofluorescence studies found that antisera to LigA and LigB did not label the surface of the Patoc wt strain but did label the surface of the ligA- and ligB-transformed Patoc (Figure 2). The surface immunofluorescence binding assay specifically detected surface-exposed components because antiserum to whole bacteria labelled intact Patoc wt, Patoc ligA, and Patoc ligB whereas antisera to cytoplasmic heat-shock protein GroEL did not label live leptospires but was able to bind to permeabilized leptospires. LigA and LigB therefore appear to be surface-exposed when expressed in Patoc transformants carrying plasmid constructs pSLePFligA and pSLePFligB, respectively (Figure 2).

Interactions of Patoc wt, Patoc ligA, and Patoc ligB strains with mammalian host cells were assayed by examining adherence of leptospires to MDCK cells and translocation of leptospires across polarized MDCK cell monolayers. Adherence of L. interrogans strain Fiocruz L1-130 and Patoc ligA, but not Patoc wt and Patoc ligB, to MDCK cells was found to significantly increase in a time-dependent manner in two experiments (Figure 3). After a 240 min incubation period, approximately four times more Patoc ligA adhered to MDCK cells than Patoc wt and Patoc ligB. The number of adherent Patoc ligA leptospires per cell at 240 min incubation point was comparable (0.23 and 1.02 in experiments 1 and 2, respectively) to that observed for the pathogenic L. interrogans strain Fiocruz L1-130 (0.16 and 0.73 in experiments 1 and 2, respectively).

Patoc ligA and ligB strains did not demonstrate enhanced ability to translocate across MDCK monolayers in comparison with Patoc wt in three experiments (representative experiment in Figure 4). As reported previously 30 , we found that a small proportion (< 1%) of Patoc wt was able to translocate across MDCK monolayers after a 240 min incubation period. Proportions of translocating leptospires recovered from the lower transwell chamber were not significantly different between Patoc wt and Patoc ligA and ligB during the assay's time course (Figure 4). In contrast, > 6% of the inoculum of pathogenic L. interrogans strain Fiocruz L1-130 was recovered in the lower chamber after 240 min of incubation (Figure 4). As previously reported 30 , recovery of L. interrogans strain Fiocruz L1-130 was not associated with significant alterations in the TER (Figure 4), indicating that disruption of tight junctions of the monolayers did not occur during the translocation process. Together these findings indicate that whereas expression of LigA in the saprophyte Patoc was associated with an enhanced host cell adherence phenotype similar to that observed with pathogenic leptospires, it did not impart the ability to efficiently invade and translocate across polarized host cell monolayers.

Lig recombinant proteins have been shown to recognize in vitro host extracellular matrix proteins 13 14 . The introduction of the ligA or ligB gene from pathogenic L. interrogans into the nonpathogenic saprophyte L. biflexa enhanced the adhesion of the latter to the mammalian host protein fibronectin (Figure 5A). The lig transformants bound to both plasma and cellular fibronectin approximately two-fold better than the Patoc wild-type strain (2.0-fold average for 1.7- to 2.3-fold range in four independent determinations for the ligA cells; 2.2-fold average from 1.5- to 3.1-fold in five measurements with ligB). The wild-type cells showed non-Lig-mediated adherence to fibronectin, which may reflect the ability of the saprophyte to interact with related proteins in decaying material that it encounters in the environment. Transformation with the lig genes also increased laminin binding 1.2-fold in comparison to the Patoc wild-type strain (Figure 5B). However, the ligA or ligB cells did not appear to bind elastin better than wild-type cells, and all three strains interacted weakly with type I and type IV collagen (Figure 5B).

Leptospires were cultivated in liquid Ellinghausen-McCullough-Johnson-Harris (EMJH) medium 47 48 or on 1% agar plates at 30°C and counted in a Petroff-Hausser counting chamber (Fisher Scientific). The saprophyte Leptospira biflexa serovar Patoc strain Patoc I and the pathogen L. interrogans serovar Copenhageni strain Fiocruz L1-130 were used in this study. E. coli was grown in Luria-Bertani (LB) medium. When appropriate, spectinomycin or kanamycin was added to culture medium at the final concentration of 40 μg/ml.

The Borrelia burgdorferi flgB promoter was amplified with PflgA (5'-TAATACCCGAGCTTCAAGGAAG-3') and PflgB (5'-AA

CATATG

CATATG

CTCGAG

CATATG

CTCGAG

L. biflexa was prepared for transformation as previously described 4 . In brief, L. biflexa was grown at 30°C until the optical density reached 0.4 at 420 nm. Bacteria were collected by centrifugation at room temperature and washed by resuspension in deionized water followed by centrifugation. After removing the supernatant fluid, the bacteria were resuspended with deionized water to a final concentration of around 5 × 10¹⁰ cells/ml (100× concentration). 100 μl of the suspended bacteria were added to the plasmid DNA, and the DNA-bacteria mixture was added to chilled electroporation cuvettes with a 0.2 cm gap. The cuvette was placed in the electroporation unit (Bio-Rad Gene Pulser II) and subjected to electroporation at a setting of 1.8 kV, 25 μF, and 200 Ω. After adding 1 ml of EMJH, the bacteria were transferred to a 15 ml Falcon tube and incubated for 24 hours at 30°C with shaking. The culture (0.2 ml) was plated onto EMJH plates containing 40 μg/ml of spectinomycin and incubated at 30° for 10 days. Colonies were inoculated into liquid EMJH containing 40 μg/ml spectinomycin. L. biflexa transformants were maintained by serial passage in the liquid medium.

Exponential phase cultures of L. biflexa Patoc wild-type, Patoc ligA, Patoc ligB, and L. interrogans Fiocruz strains were washed, resuspended in PBS and solubilized in 62.5 mM Tris hydrochloride (pH 6.8)-10% glycerol-5% 2-mercaptoethanol-2% sodium dodecyl sulfate. A 20 μl volume of crude extracts containing 2 × 10⁸ bacteria/per well was resolved by 8% sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis using a discontinuous buffer system. After transfer to nitrocellulose membranes, immunoblots were blocked in 0.05 M Tris-buffered saline (pH 7.4)-0.05% (vol/vol) Tween 20 with 5% (wt/vol) nonfat dry milk. The blots were washed, incubated for 1 h at room temperature with a 1,000-fold dilution of mouse ascites containing MAb to the LigB identical repeat region (LigA/B) 6 and probed with goat anti-mouse conjugated to alkaline phosphatase (Sigma). Immunoblots were developed in a nitroblue tetrazolium--5-bromo-4-chloro-3-indolylphosphate (BCIP) solution (Bio-Rad).

We evaluated the localization of LigA and LigB by performing immunofluorescence labeling according to a modified protocol of Cullen et al. 50 . Suspensions of 10⁷ live leptospires in 10 μl of PBS were placed onto poly-L-lysine-coated slides (Sigma-Aldrich) for 1 h in a humidified chamber for adherence of the leptospires. In experiments in which the bacteria were permeabilized prior to incubation with antibody, slides were incubated with cold methanol for 10 min at -20°C, followed by two washes with PBS. Blocking with 1% bovine serum albumin (Sigma-Aldrich) (PBS-BSA) for 20 min was performed before incubation for 1 h at 37°C with normal rabbit serum, rabbit hyperimmune antisera to whole extracts of L. interrogans serovar Icterohaemorrhagiae strain RGA, LigB non-identical region (LigBNI), and LigA non-identical region (LigANI) 6 , and rat antiserum to GroEL, which were diluted 1:100 in PBS-BSA. The slides were washed gently with PBS-BSA and incubated with goat anti-rabbit IgG antibodies conjugated to Alexa dye (Molecular Probes) or goat anti-rat IgG antibodies conjugated to fluorescein isothiocyanate (Jackson ImmunoResearch Laboratories) for 1 h at 37°C. The slides were washed twice with PBS-BSA and incubated with 1 μg/ml DAPI (Molecular Probes) for 1 h at room temperature. Slides were then washed, then mounted in anti-fading solution (Prolong-Molecular Probes) and visualized by fluorescence microscopy (Olympus BX51).

Madin Darby canine kidney (MDCK) cells were grown in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum (Cultilab), 2% sodium bicarbonate, 25 mM HEPES, and 5 mM L-glutamine (Sigma) at 37°C in an atmosphere of 5% CO₂. MDCK cells were harvested by treating cell cultures with 0.05% trypsin and 0.02% EDTA in PBS. For adhesion assays, MDCK cells were plated onto 24-well plates in DMEM, containing 13-mm-diameter glass coverslips at 37°C in an atmosphere of 5% CO₂ until they were confluent. The number of MDCK cells in wells was determined by lysing cells with 0.1 M citric acid containing 0.05% crystal violet (Sigma-Aldrich) and 1% Cetrimide (Sigma) 51 , then the nuclei were counted in a hemacytometer. The cells were incubated with a suspension of Patoc wild-type, Patoc ligA, Patoc ligB and Fiocruz L1-130 strains in cell culture medium at the final bacteria: cell ratio of 10:1. Incubations were performed for periods of 30 to 240 min. Prior to staining, the cells were washed three times in PBS to remove nonadherent bacteria and then fixed with cold methanol for 10 min. An immunofluorescence assay was performed to detect adherent leptospires for which rabbit polyclonal antisera against whole extracts of L. interrogans strain RGA and goat anti-rabbit antibodies conjugated with Alexa488 (Molecular Probes) were used as first and second antibodies, respectively. DAPI and Alcian Blue were used to stain the nucleus and cytoplasm, respectively. The number of leptospires and MDCK cells was determined by examining ten high-magnification (1000×) fields during fluorescence microscopy. All incubation points were performed in triplicate. The ANOVA test was used to determine statistically significant (p < 0.05) differences between numbers of adherent leptospires/cell. We performed a translocation assay according to a protocol modified from that described by Barocchi et al 30 . MDCK cells at a concentration of 2 × 10⁵ cells in 500 μl of DMEM were seeded onto 12-mm-diameter Transwell filter units with 3- μm pores. Monolayers were incubated at 37°C in 5% CO₂ for 3 to 4 days with daily changes in media until the transmonolayer electrical resistance (TER) reached a range of 200 and 300 Ω/cm², as measured with an epithelial voltohmmeter (EVOM, World Precision Instruments, Sarasota, Fla.). The TER for polycarbonate filters without cells was approximately 100 Ω/cm². The upper chamber of the transwell apparatus was inoculated with leptospires at a multiplicity of infection (MOI) of 100 by adding 500 μL of bacteria which were resuspended in 1:2 v/v ratio of DMEM and EMJH media. Duplicate transwell chamber assays were performed for each leptospiral strain which were tested. Aliquots were removed from lower chamber (100 μl) at 30, 120 and 240 min and the number of leptospires was counted in triplicate by using the Petroff-Hausser chamber. The ability of leptospires to translocate MDCK polarized monolayers was determined by calculating the proportion of leptospires in the lower chamber in comparison to the initial inoculum for duplicate assays at each time point. The ANOVA test was used to determine significant differences in the proportions of translocating leptospires and TER values obtained during incubations with different leptospiral strains.

The adhesion of live L. biflexa strains to immobilized fibronectin was measured with an ELISA. Two to three × 10⁸ cells in serum-free EMJH or medium alone was incubated at 30°C for 1 h in a microtiter well pre-coated with 1 μg of fibronectin (from human plasma or foreskin fibroblasts, Sigma-Aldrich), collagen type I (bovine skin, Sigma-Aldrich), collagen type IV (human placenta, Sigma-Aldrich), laminin (murine, Sigma-Aldrich), elastin (human skin, Elastin Products Company, Owensville, MO), or left in PBS, pH7.2, overnight at 4°C. Uncoated sites in the well were covered with Protein-Free Blocker (Thermo Scientific) before the addition of cells. Adherent cells were fixed with 4% formaldehyde (Thermo Scientific) at room temperature for 1 h, tagged with a rabbit polyclonal antibody for intact L. biflexa (MyBioSource), and detected by spectrometry at 450 nm to measure the activity of horseradish peroxidase conjugated to a donkey antibody for rabbit IgG (GE Healthcare). Backgrounds from uncoated wells (PBS) and medium only were subtracted. Triplicate assays were done and statistically significant differences in adhesion were determined with one-way ANOVA compared to the wild-type cells.