Development of Transcriptional Fusions to Assess Leptospira interrogans Promoter Activity

Background Leptospirosis is a zoonotic infectious disease that affects both humans and animals. The existing genetic tools for Leptospira spp. have improved our understanding of the biology of this spirochete as well as the interaction of pathogenic leptospires with the mammalian host. However, new tools are necessary to provide novel and useful information to the field. Methodology and Principal Findings A series of promoter-probe vectors carrying a reporter gene encoding green fluorescent protein (GFP) were constructed for use in L. biflexa. They were tested by constructing transcriptional fusions between the lipL41, Leptospiral Immunoglobulin-like A (ligA) and Sphingomielynase 2 (sph2) promoters from L. interrogans and the reporter gene. ligA and sph2 promoters were the most active, in comparison to the lipL41 promoter and the non-induced controls. The results obtained are in agreement with LigA expression from the L. interrogans Fiocruz L1-130 strain. Conclusions The novel vectors facilitated the in vitro evaluation of L. interrogans promoter activity under defined growth conditions which simulate the mammalian host environment. The fluorescence and rt-PCR data obtained closely reflected transcriptional regulation of the promoters, thus demonstrating the suitability of these vectors for assessing promoter activity in L. biflexa.


Introduction
Leptospira interrogans is the main causative agent of leptospirosis, a zoonotic infectious disease with worldwide distribution. Chronically infected reservoir hosts, such as rats, do not exhibit overt disease but are colonized by leptospires in their renal tubules and shed bacteria in the urine. Humans become infected by exposure to contaminated water, soil, or urine [1]. Severe manifestations of the disease, as observed in Weil's disease, are frequent and associated with significant mortality, up to 15% [1,2]. In addition, leptospirosis may evolve to severe pulmonary haemorrhage syndrome (SPHS), for which case fatality is .50% [3,4].
The Leptospira genus is composed of 20 recognized species and includes strains that belong to the saprophyte, intermediate or pathogen groups [5]. Currently, almost 300 serovars are recognized, of which more than 200 are considered to be pathogenic [2,5,6,7]. The available genome sequences for pathogenic [8,9,10] and saprophytic [11] Leptospira spp. have been employed to search for new diagnostic reagents and vaccine candidates for leptospirosis.
Prompt diagnosis and early treatment of leptospirosis are essential to avoid severe outcomes [12]. The early phase leptospirosis is often misdiagnosed due to its presentation with nonspecific clinical signs [1], low sensitivity and frequent poor specificity of the results exhibited by the microscopic agglutination test (MAT) and the commercially available assays [7,13,14,15,16,17,18,19].
The use of vaccines as prevention measures appears to be a costeffective approach to prevent worldwide diseases. Commercially available whole-cell vaccines confer protection in a limited and incomplete manner, limiting their use among humans. E. g. whole-cell preparations produce only short-term immunity, requiring administration semi-annually; present low cross-protection and adverse reactions due to both residual media components and leptospiral lipopolysaccharide [2,6,7,20,21]. Efforts have been made for over a decade towards identifying immunoreactive [22,23,24,25,26,27,28] or protective [29,30,31,32,33,34,35,36,37,38,39,40,41] antigens via recombinant DNA technology. Despite these advances, limitations still remain to be overcome. The rational identification of novel candidate antigens is therefore necessary and the development of genetic tools to Leptospira spp. may be helpful for this purpose.
The mechanisms of Leptospira pathogenesis remain unclear, despite the efforts to identify virulence factors and their role in the pathogen-host interaction. To this aim, new genetic tools have been developed in the last years [11,42,43]. Functional characterization of outer membrane and cytoplasmic proteins [44,45,46,47,48,49], and more recently the generation and study of knock-out mutants [50,51,52,53,54,55,56] have provided important contribution.
Many efforts have been done to understand the influence of environmental signals on the leptospiral transcriptome and proteome, aiming to identify antigens involved in pathogenesis [44,46,47,48,57,58]. Temperature, physiological osmolarity, iron availability and the growth phase, in addition to the multitude of factors existing in the host serum and during pathogen-host cell contact, are known to affect the expression of several leptospiral proteins [57,58,59,60,61,62]. pH was also found to be responsible for protein regulation, such as LipL36 and P31LipL45 (Qlp42), in the kidney tubules of hamsters [61,63]. However, many leptospiral coding sequences (CDSs) still remain to be characterized [64]. The L. interrogans genome carries a considerably large number of genes supposedly involved in response regulation [9]. L. borgpetersenii, which possesses a reduced ability to survive outside the host, contains a lower number of regulatory genes [10]. However, the saprophyte L. biflexa contains a larger number of putative transcription factors than the other sequenced species. This suggests that L. biflexa can be used as a model to study the gene regulation of pathogenic Leptospira spp. [11,57]. Despite the sequence diversity between both species, these findings suggest that pathogens and saprophytes might share some similar mechanisms to respond to the environment.
The extent to which the manipulation of in vitro conditions can be used to reproduce the full spectrum of mammalian host signals, which trigger differential gene expression in pathogenic Leptospira spp. remains uncertain. In this study, we sought to develop a new genetic tool to help elucidating the biology of Leptospira spp. We found that our promoter-probe methodology is useful for assessing promoter activity under defined conditions. The comparative analysis of the fluorescence produced by a specific L. biflexa reporter strain, PAG (carrying a copy of the ligA promoter fused to the gfp gene), and LigA expression from L. interrogans Fiocruz L1-130 strain, grown in vitro under the same systematic conditions, validated the use of L. biflexa as a model to assess L. interrogans promoter activity. The reporter strain containing a copy of the sphingomyelinase 2 promoter, Psph2 (P2G) was also strongly induced by the conditions tested, above an established cut-off, whereas the non-virulence factor promoter control, from the lipL41 lipoprotein gene, was not. We believe this promoter-probe methodology may support the existing methodologies to the identification of novel virulence factors of pathogenic Leptospira spp.

Bacterial strains and growth conditions
Bacterial strains and constructs are listed in Table 1. Strains were obtained from the collection of the Faculdade de Medicina Veterinária e Zootecnia, Universidade de São Paulo, São Paulo, Brazil; Laboratoire de Biologie des Spirochetes, Institut Pasteur, Paris, France and the ATCC. The virulence of the low-passage L. interrogans serovar Copenhageni Fiocruz L1-130 strain was maintained by passage through Golden Syrian hamsters. Lowpassage refers to strains that were sub-cultured in EMJH liquid medium up to 10 times. All strains were cultured at 30uC in liquid EMJH medium supplemented with 1% rabbit serum [65,66]. Chemically competent E. coli TOP10F cells were used as host for genetic manipulation of plasmids. E. coli transformants were typically selected on LB agar plates containing spectinomycin (50 mg/ml) or ampicillin (100 mg/ml). Electrocompetent L. biflexa sorovar Patoc strain Patoc 1 was prepared as previously described [55], and transformed with replicating shuttle vectors containing the promoter-probe cassettes (Table 1). EMJH plates were prepared using 1% agar and supplemented with spectinomycin at 50 mg/ml.
Liquid EMJH pH 6.7 was prepared by adding concentrated HCl until the pH of choice was reached. Spermine (Sigma) was dissolved according to the manufacturer's instructions and supplemented in cultures at 200 mM. Physiologic osmolarity was induced by supplementation with 120 mM NaCl [44]. Before induction, cultures were grown at 30uC in EMJH until the lateexponential phase was reached (culture density of 10 8 to 10 9 /ml). Growth was monitored by measuring the OD 420 using an Ultrospec 2100 pro spectrophotometer (GE Healthcare). Cells were harvested (OD 420 0.5) and immediately stored at 220uC.

Bioinformatics analyses
The sequences used for this study were obtained from the complete genome sequence of L. interrogans serovar Copenhageni strain Fiocruz L1-130 by using the SpiroScope (http://www. genoscope.cns.fr/agc/mage) database [67] and Leptospira Genome Project (http://aeg.lbi.ic.unicamp.br) database [9]. The Neural Network Promoter Prediction v.2.2 [68], TRES -Transcription Regulatory Element Search, which performs searches within the TRANSFAC database [69] and the PromScan program, that generates an alignment of known sequences and matrix frequency [70], were used to scan DNA sequences for potential binding sites. Repeats were identified by the EMBOSS programs available at http://bioweb.pasteur.fr/nucleic/intro-en.html#repeat.

Construction of reporter vectors
The gfp gene [43] was amplified by PCR with Gfp5/Gfp3 primers pair, which introduced the SmaI, SacI, XhoI, BamHI, SmaI restriction sites. The resulting fragment was cloned into pGEM-T Easy (Promega) yielding the pGEM-T Easy/gfp construct. The PlipL41, PligA and Psph2 promoters from L. interrogans serovar Copenhageni strain Fiocruz L1-130 were amplified by PCR using primers PlipL41300F/PlipL41300R, PligA300F/PligA300R and Psph2600F/Psph2600R, which introduced the XhoI and BamHI restriction sites, and cloned into pGEM-T Easy/gfp via the same sites. Promoter-probe vectors containing the L. interrogans promoters-gfp cassettes were constructed in the E. coli-L. biflexa pSLe94 shuttle-vector [71]. The cassettes were removed from pGEM-T Easy (Promega) by SmaI digestion and cloned in via PvuII restriction site to give the new vectors, pSLe94/PlipL41/gfp, pSLe94/PligA/gfp and pSLe94/Psph2/gfp ( Table 1). Electroporation of leptospires was performed as previously described [55,71]. After 1 to 2 weeks of incubation, spectinomycin resistant transformants were used to inoculate liquid medium.

Reporter gene assays
Cells for spectrofluorometry measurements were resuspended in 300 ml of deionized water and distributed 100 ml per well in a black 96-well OptiPlate-96F microplate (PerkinElmer). GFP fluorescence was measured using a 650-10 spectrofluorometer (PerkinElmer) at an excitation wavelength of 485 nm and an emission wavelength of 538 nm. The fluorescence intensity from samples was expressed as arbitrary fluorescence units, obtained at a wavelength of maximum emission. The mean specific activity from at least three independent assays is indicated in the results.

Fluorescence microscopy
Qualitative expression of GFP was examined under UV light for detection of fluorescence. The cultures were induced and samples were collected at representative time-points, spun down and the resulting pellets were washed twice with 16 PBS to remove the culture medium. The pellets were then resuspended in 16 PBS, 20% Glycerol. Aliquots of 20 ml of the bacterial suspensions were applied to slides and sealed with a concentrated formaldehyde resin. Images of leptospires on cover slips were acquired using an IX81 inverted microscope (Olympus) equipped with 206and 406 objectives and the CellR software (version 3.1). The UV filter sets used were DAPI (excitation: 350 nm -range is 330 to 380 nm, Emission: 460 nm) and DIC (transmitted light).

RNA isolation and rt-PCR
Cells were harvested (OD 420 0.25) and RNA was isolated from bacteria with TRIzol reagent (Invitrogen), as described by the manufacturer. Then, isolated RNA (2 mg) was treated with DNaseI (Invitrogen), following manufacturer's instructions. cDNA was synthesized using SuperScript III First-strand (Invitrogen) according to the manufacturer's protocol. Samples were quantified and checked for purity using an Ultrospec 2100 pro spectrophotometer (GE healthcare) and agarose gel electrophoresis. The cDNA (representing 100 ng of RNA per reaction) was amplified with Taq DNA polymerase (Invitrogen) using primers pairs fD1/ rP2 (control) [72], Gfp5/Gfp3 for reporter strains [43], or NligA5/ NligA3, for L. interrogans serovar Copenhageni strain Fiocruz L1-130.

Whole-cell ELISA
Whole-cell ELISA experiments were performed using a modification of previously described methods [45,73]. Flat-bottom polystyrene high-binding microtitre plates (Corning) were coated overnight at 4uC with 100 ml per well of 10 8 ml 21 whole L. interrogans serovar Copenhageni strain Fiocruz L1-130, which were previously centrifuged at 4,0006g and resuspended in 0.05 M

Statistical analysis
Differences between average values were tested for significance by performing an unpaired, two-sided Student's t-test [74]. Differences were considered statistically significant when the resulting p values were #0.05.

In vitro induction of virulent Leptospira interrogans
Many efforts have been made to identify and quantify the leptospiral transcriptome and proteome during growth under different conditions. But, so far, little is known about the influence that mammalian host conditions exert over promoter activity. Therefore, we initiated our investigation by studying ligA gene transcription and protein expression profiles using L. interrogans serovar Copenhageni strain Fiocruz L1-130 grown, for up to 24 h, under a combination of conditions that include the mammalian physiological osmolarity (,300 mosmol-120 mM), temperature (37uC), urine pH (6.7) and the supplementation with spermine, a component belonging to the intracellular environment. We focused our study on the first 24 h of induction, to observe the changes that L. interrogans promoters undergo during the early stages of host invasion.
Initially, we evaluated the influence of the different conditions by rt-PCR analysis ( Figure 1). It was performed on RNA extracted from both non-induced and induced cultures of the virulent, low-passage, L. interrogans serovar Copenhageni strain Fiocruz L1-130. Low level transcription (ligA-based) was detected before induction of the virulent strain, while a clear difference in the abundance of ligA transcript could be seen after induction, and between the time intervals. Three time-points were evaluated (1, 12 and 24 h post induction -p.i.) and, interestingly, ligA transcription was not constant after induction ( Figure 1). Under most conditions, the level of ligA transcript was higher than the non-induced control immediately one hour p.i. However, only spermine was able to upregulate ligA for an extended period, up to 12 h p.i., while the other treatments exhibited down-regulation (Figure 1). At 24 h p.i., only the temperature upshift and physiological osmolarity conditions stimulated up-regulation in comparison to the previous time-point ( Figure 1).
To examine the effect of the conditions over the expression patterns of native LigA, we cultivated L. interrogans serovar Copenhageni strain Fiocruz L1-130 until late-exponential phase before induction. The combination of both temperature upshift and physiological osmolarity promoted induction of protein synthesis ( Figure 2A). An up-regulation 4.98-fold higher than the non-induced control was observed at one hour p.i. (p,0.01). The influence of spermine was also evaluated in the kinetics assay and stimulated LigA expression above the control levels immediately after induction, one hour p.i. (5.95-fold, p,0.01), ( Figure 2B). The urine pH condition was simulated by reducing the pH from 7.2 to 6.7 and the highest up-regulation was obtained by this treatment, one hour p.i. (5.92-fold, p,0.01), ( Figure 2C). But it was the only condition to be followed by decreasing expression levels. The analysis of native LigA expression complemented the results obtained by rt-PCR, although correlation was not always observed between RNA and protein profiles (Figure 1 versus 2).

Prediction of regulatory sequences by Bioinformatics
Previous studies demonstrated that upon induction by physiological osmolarity and temperature shift from 20uC to 37uC and 30uC to 37uC the lipL41, ligA and sph2 genes are differentially expressed in L. interrogans serovar Copenhageni strain Fiocruz L1-130, L. kirschneri serovar Grippotyphosa strain RM52 and L. interrogans serovar Lai strain 56601 [44,46,47,48]. The rational to select the PlipL41, PligA and Psph2 promoters to compose and standardize our study was based on previous studies that characterized them extensively at the RNA and protein levels.
The PlipL41, PligA and Psph2 promoters were predicted by bioinformatics based on the analysis of the L. interrogans serovar Copenhageni strain Fiocruz L1-130 genome sequence (AE016823). Different regions were identified, which are composed of direct and inverted repeats (data not shown).
To examine whether the computationally determined sequences comprise leptospiral regulatory elements, we fused the DNA segments to a promoterless gfp reporter. Both elements were subsequently cloned into an E. coli-L. biflexa shuttle vector depicted in Table 1, yielding the plasmids pSLe94/PlipL41/gfp, pSLe94/ PligA/gfp and pSLe94/Psph2/gfp (Table 1). To ensure the accurate inclusion of the DNA sequences containing the regulatory elements in the study, we designed primers to amplify 300 bp upstream of the first ATG of lipL41 and ligA, and 600 bp upstream of sph2. The larger length of the amplified fragment upstream of sph2 was adopted to include possible distal elements that may play a role in the activity of the Psph2 promoter, since the intergenic space to the next ORF, LIC12630 (hypothetical protein), is 650 bp. In the case of the lipL41 and ligA genes the intergenic spaces are 114 bp and 456 bp, respectively. The plasmid constructs were used to transform L. biflexa serovar Patoc strain Patoc I and generated the reporter strains P41G (lipL41 promoter), PAG (ligA promoter) and P2G (sph2 promoter), which were induced for production of fluorescence by the green fluorescent protein (GFP).

Application of promoter-probe vectors in L. biflexa
A series of replicating promoter-probe vectors utilizing gfp were constructed to examine changes in promoter activity in response to a set of defined conditions ( Table 1). The constructs were based on the low-copy number shuttle-vector pSLe94 (Table 1), which contains the LE1 origin of replication that is capable of replication within saprophytic Leptospira spp. To validate the usefulness of these vectors as genetic tools to investigate promoter regulation, transcriptional fusions were constructed including the promoters from the well-characterized genes lipL41, ligA and sph2 of L. interrogans serovar Copenhageni strain L1-130. GFP reporter strains were established by electroporation of the promoterspecific reporter vectors into L. biflexa. The effectiveness of the reporter strains for assaying transcriptional activity was assessed quantitatively.
All three reporter strains of L. biflexa were initially cultivated on EMJH agar plates, at 30uC, and then transferred to liquid EMJH, both containing low-salt concentration (70 mosmol per liter NaCl -28 mM). When cultures reached the late-exponential phase they were aliquoted and either adjusted to 120 mM NaCl or temperature-upshfited from 30uC to 37uC. Quantitative assessment of fluorescence showed that, under physiological osmolarity conditions, the highest promoter activity was detected in the PAG reporter strain, compared to the control, p,0.01 (Figure 3). The PligA promoter was induced in the presence of 120 mM NaCl and the fluorescence reached the highest level 3 hours p.i. The average intensity at this time-point was 1.48-fold higher than the noninduced control ( Figure 3B). Neither the PlipL41 nor Psph2 promoters were significantly induced ( Figure 3A and 3C). Of note, there were no obvious differences between the growth curves of the wild-type and the reporter strains of L. biflexa, suggesting that the GFP expression did not cause any disadvantage to the host strain (data not shown).
The influence of the temperature upshift, from 30uC to 37uC, on promoter activity was also assessed. The levels of fluorescence produced by the P41G, PAG and P2G reporter strains were significantly higher than the non-induced controls (p,0.01). The PlipL41 and Psph2 promoters reached their maximum activity 3 hours p.i. (Figure 4A and 4C). At this time point there was significant (p,0.01) up-regulation of 1.46-fold and 1.26-fold, respectively. The P41G reporter strain exhibited the highest promoter activity, under this condition, followed by P2G. PligA promoter activity was most influenced, with expression 1.50-fold higher than the non-induced control, p,0.01. Interestingly, PligA and Psph2 activities remained constant during the assay conditions ( Figure 4B and 4C). The induction of PlipL41 and Psph2 by the temperature upshift led to production of fluorescence levels significantly higher than those induced under conditions of physiological osmolarity, p,0.05 ( Figure 3A versus 4A and 3C versus 4C), while the PligA activity induced by the temperature upshift did not differ from that observed by salt supplementation, at most of the time-points investigated ( Figure 3B versus 4B).
Effect of the combination of environmental signals on L. interrogans promoter activity To identify promoters that are differentially induced in response to a combination of conditions, we performed systematic experiments with different, overlapping, stimuli. Subcultures were initially induced under conditions of physiological osmolarity and normal human body core temperature (37uC). The average fluorescence intensity produced by the reporter strains was highly significant when compared to the non-induced controls, p,0.01. For P41G, the highest promoter activity was observed 3 hours p.i. (1.54-fold), and remained constant throughout the assay ( Figure 5A). Both PligA and Psph2 were induced to the highest levels 1 hour p.i., 2.59-fold and 1.62, respectively (p,0.01). However, they did not maintain a constant activity during the assay ( Figure 5B and 5C). Additionally, PlipL41 was induced to a similar extent by the temperature upshift and the combination of physiological osmolarity and temperature ( Figure 4A versus 5A). PligA and Psph2 were significantly induced, 1.74-fold and 1.57-fold, respectively ( Figure 3B versus 5B and 3C versus 5C), and 1.73-fold and 1.28-fold, respectively ( Figure 4B versus 5B and 4C versus 5C), compared to the single-condition treatments (physiological osmolarity or temperature upshift versus their combination).
The results indicate these promoters were preferentially active, in vitro, when L. biflexa was cultured under conditions emulating the mammalian host environment. Thus, we decided to evaluate another combination of conditions, the effect of the urine pH plus the physiological osmolarity and the temperature upshift. The reporter strains were sub-cultured into EMJH medium supplemented with 120 mM NaCl, pH 6.7, and shifted from 30uC to 37uC. This resulted in the activation of all promoters as seen by increasing levels of fluorescence. The results from three independent experiments are shown in Figure 6. The PligA promoter was the most affected, reaching an activity 2.08-fold higher than the non-induced control (3 hours p.i.), followed by Psph2 (1.96-fold) and PLipL41 (1.58-fold), both at 24 hours p.i, p,0.01 ( Figure 6). Curiously, we observed that both PligA and PlipL41 were slightly less stimulated by the combination that included pH 6.7 than that including salt and temperature only (p,0.05) ( Figure 5A versus 6A and 5B versus 6B).
Previous studies have demonstrated that several bacteria undergo translational regulation in response to spermine [75]. This polyamine is only produced by eukaryotic cells, it can reach millimolar levels within them [76], and it has significant effects on various cellular processes in liver, kidney and brain cells and in lymphocytes [76,77,78,79,80,81,82]. Here, we investigated whether spermine contributed to L. interrogans promoter regulation. The L. biflexa reporter strains were grown at 30uC until late-exponential phase and then supplemented with 200 mM spermine, 120 mM NaCl and upshifted to 37uC. The ligA promoter was the most affected, the PAG reporter strain produced a fluorescence level 2.74-fold higher than the non-induced control (3 hours p.i.), followed by P2G (2.12-fold) and P41G (1.58-fold), both 24 hours p.i, p,0.01 ( Figure 7). However, ligA promoter activity was not maintained, the average fluorescence decreased to 1.72-fold (24 hours p.i.), demonstrating that PligA is transiently induced by spermine ( Figure 7B). The spermine-induced activity of both PligA and Psph2 promoters was the highest observed in this study ( Figure 7B and 7C).
In agreement with the native LigA expression pattern, the fluorescence produced by the PAG reporter strain reached the highest (4.98-fold versus 2.59-fold) and lowest (2.19-fold versus 2.13-fold) levels at one and 12 hours p.i., respectively, when induced by the combination of physiological osmolarity and temperature upshift (Figure 2A versus 5B and Table 2), while the LigA expression pattern induced by spermine was different from the fluorescence intensity produced by the PAG reporter strain ( Figure 2B versus 7B). Finally, induction by pH 6.7 led to LigA expression that was consistent with the fluorescence levels produced by the PAG reporter strain ( Figure 2C versus 6B and Table 2). Protein and fluorescence levels were highest one hour p.i., but the amount of native LigA decreased significantly (from 5.95-fold to 4.04-fold), while the fluorescence level produced by the PAG reporter strain remained constant (from 2.07-fold to 2.03-fold). The results obtained in this study, by the use of the promoter-probe methodology described, allowed us to establish a cut-off based on the determined fluorescence levels ( Table 2). This information will be useful to support the future prediction of virulence factors as we are currently developing a follow-up study employing several other promoter sequences to construct a library of knock-in mutants.
The ability of the promoter-probe system to allow for the differentiation between both non-induced and induced reporter strains suggests that the gfp regulation can be employed to accurately mimic the behaviour of the native genes. This is corroborated by similarities observed between the expression patterns of native LigA and the fluorescence levels produced by the PAG reporter strain, in comparison to the non-induced controls ( Table 2). Qualitative assessment of these reporter strains clearly demonstrated the presence of fluorescent leptospires upon induction in vitro (Figure 8).
To further substantiate the data obtained by the use of the promoter-probe vectors, rt-PCR analysis was performed on RNA extracted from non-induced and induced cultures of the L. biflexa reporter strains ( Figure 9). As expected, low level of transcripts (gfpbased) were detected before induction of the reporter strains. Three time-points were evaluated, 1, 3 or 6 and 24 h p.i., but only a slight difference was observed between the time-points for any given in vitro treatment, per reporter strain (Figure 9). Despite this, the amount of gfp mRNA varied considerably per treatment. Correlation between the fluorescence and transcription levels was not always evident, similarly to the lack of correlation observed during the analysis of the native strain (see Results). In both cases (native and reporter strains) the discrepancy observed between the mRNA and protein levels probably resulted from intrinsic regulatory mechanisms. Although we have no information on the signalling pathways involved, the L. interrogans promoters evaluated appear to behave similarly in both species, suggesting conservation of the regulatory mechanisms higher than expected.
The transcription data complement those obtained by fluorescence measurement from the reporter strains, reinforcing the utility of the promoter-probe vectors as a genetic tool to test promoter activity in L. biflexa.
Validation of the L. biflexa promoter-probe constructs was performed by comparison against the in vitro-induced pathogenic strain. In addition to the protein content, we compared the mRNA profiles produced by the wild-type virulent L. interrogans and the PAG reporter strain. Side-by-side analysis of two time-points (1 and 24 h p.i.) showed that all treatments produced similar patterns of both LigA and GFP expression, the latter being controlled by the native L. interrogans PligA promoter (Figure 1 versus 9). However, a noticeable difference was observed with respect to the transcript levels produced by the native ligA and the PligAcontrolled gfp (Figure 1 versus 9). Of note, up-regulation of the PligA was noted despite the treatment (Table 2), therefore validating the use of this promoter-probe methodology as a genetic tool to assess the ability of pathogens promoters to respond host stimuli.

Discussion
In this study we developed and characterized in detail a L. biflexa promoter-probe methodology for the analysis of L. interrogans promoters. The three promoters regions used in this study resemble regulatory regions of other bacterial species, based on bioinformatics characterization only (data not shown). Although we did not support in silico findings with in vitro data we observed that PlipL41 promoter sequence is similar to the E. coli s70 promoter, while PligA and Psph2 organization is more distant to it. The identification of different repeats into the promoter regions studied may be suggestive of specific response regulator binding sites. It is possible that the elements through which each promoter senses the changes on environmental cues may be different.
The lipL41, ligA and sph2 genes are known to be regulated in vitro by physiological conditions [44,46,47,48]. In addition, the L. interrogans transcriptome was evaluated in response to multifactorial conditions, upon exposure to serum [58]. To further substantiate the knowledge of Leptospira genetics, P41G, PAG and P2G reporter strains were constructed to contain gfp under the conditional control of the aforementioned L. interrogans promoters. Analysis of cell extracts harvested from exponentially growing cultures revealed low GFP activity under routine growth conditions, while variable activity was observed post-induction. In agreement to previous findings obtained by transcriptomics, our data demonstrate the sensitivity of leptospiral promoters under the various conditions evaluated.
ligA is a paralog of ligB, and originated by duplication events of the first ten immunoglobulin-like domains [83], consequently both are expected to respond to host stimuli through similar mechanisms. Both proteins seem to contribute to the pathogenesis of Leptospira spp. as important adhesins [84,85,86]. In addition, spectral counts of both proteins by a mass-spectrometry-based strategy revealed that both LigA and LigB are present in 553 and 914 copies per cell, respectively [87]. Recently, Lo and colleagues were not able to detect LigB among L. interrogans serovar Lai samples shifted from environmental to human body temperature (37uC), demonstrating that temperature upshift alone may be insufficient for induction of expression of Lig proteins [88]. Likewise, in our study, the PligA promoter activity was only weakly detected during non-induced and single-condition induced PAG cells. However, fluorescence was more than doubled after induction by combinations of stimuli ( Figure 5B-7B). The same was observed for the expression of native LigA by L. interrogans serovar Copenhageni strain Fiocruz L1-130 ( Figure 2).
Sph2 is expressed under routine in vitro conditions and temperature upshift to 37uC [87,88,89]. At the mRNA level, it Figure 6. Kinetics of GFP production by the L. biflexa reporter strains (P41G, PAG and P2G), after exposure to a combination of physiological salt levels, temperature upshift to 376C and urine pH 6.7. Cultures of the reporter strains P41G (A), PAG (B) or P2G (C) were induced by a combination of physiological osmolarity (PO), temperature upshift (TS) and pH reduction from 7.2 to 6.7 (PH). The induced wild-type control (P-24) and the uninduced reporter strains (NI-24) were included as controls. Transcriptional activity was measured for 24 h and presented as the mean 6 standard error (bars). Fluorescence from triplicate samples of each culture were standardized according to an OD 420 0.5 and are expressed as arbitrary fluorescent units. Data from a representative significant study are shown. doi:10.1371/journal.pone.0017409.g006 is up-regulated by physiological osmolarity [47] and temperature [46]. In our study, we corroborated these findings with the observation of low levels of fluorescence in the non-induced P2G reporter strain and elevated levels of Psph2 promoter activity after induction by the combination of different mammalian host conditions ( Figures 5C, 6, 7C). These findings suggest the mammalian host conditions may have contributed to this effect and reinforces the hypothesis that the sph2 gene is likely to be regulated at post-transcriptional level.
Interestingly, we observed LigA down-regulation when L. interrogans serovar Copenhageni was cultivated at pH 6.7 ( Figure 2C). This is in agreement with previous findings that leptospiral antigens are down-regulated when leptospires are excreted in the urine [90]. Of note, this was not observed during fluorescence analysis ( Figure 6). This suggests that specific response regulators may be triggered in the parental pathogenic strain, during renal colonization and/or urine shedding. In general, the activity of the PlipL41 promoter was not, or only weakly, influenced under the various conditions studied ( Figure 3A-7A), and this is consistent with previous observations [59,88,91,92]. In contrast to the study of Nally and colleagues [93] that observed LipL41 downregulation upon experimental infection, PlipL41 promoter activity was not reduced when the P41G reporter strain was cultivated under the majority of the conditions analysed ( Figure 3A-7A). This suggests that the reduction of LipL41 expression in guinea pig-recovered L. interrogans may be due to post translational regulation, and may involve complex regulation systems such as host cell contact, protein degradation, and energy metabolism.
Of note, the PlipL41 promoter was initially selected as a constitutive promoter. However, we noted that this promoter was also up-regulated at some of the selected time-points. An ideal constitutive gene should exhibit constant activity independently to the cell state or environmental conditions. However, constitutive expression may not be related to constant promoter activity. Moreover, expression levels of constitutive genes such as the flagellin or the ribosomal RNA synthesis genes, have been shown to be altered by temperature-shift or ciprofloxacin supplementation [87,88]. PlipL41 promoter activity was not always induced under the in vitro conditions. In addition, its activity was below the cut-off level (2-fold) established in this work for virulence factor promoters ( Table 2). This lead us to believe that the P41G reporter strain may serve as a negative control in the identification of novel virulence factors.
Although L. interrogans is a pathogen known to infect the mammalian host through active penetration, there is uncertainty over the existence of an intracellular phase during this process. In this study, we demonstrated that spermine, a polyamine found in abundance in eukaryotic cells [76] alters L. interrogans promoters activity. Previous studies showed leptospiral survival and replication within human macrophages, a spermine-rich eukaryotic cell [94]. We observed that supplementation of reporter strain cultures with spermine was able to stimulate promoter activity. In addition, we present evidence that similar transcriptional and translational changes occurred during induction of the native proteins, supporting the hypothesis that pathogenic Leptospira spp. can recognize polyamines as a signal of the intracellular environment. Polyamine recognition is likely a component of a sophisticated system that integrates multiple environmental signals and regulates gene expression in intracellular bacteria, i.e. Francisella [95,96,97], and biofilm formation in Vibrio spp. and Yersinia spp. [77,78]. As polyamines in the intracellular environment are likely to reach elevated levels, our findings on promoter regulation by spermine may be relevant bacterial pathogenesis in general.
It is not known what controls ligA or sph2 transcription initiation. However, previous studies have shown that both genes are upregulated by physiological osmolarity [44,47,48] and temperature [46]. Also, recent work presented evidence of a functional redundancy between LigA and LigB [52]. The finding that the PligA promoter is active under different in vitro conditions, which simulate the mammalian host environment, raises the hypothesis that ligA absence from several pathogenic serovars is likely to be due to spontaneous deleterious events that may have occurred inside or outside the mammalian host environment [83,98], without any influence on virulence or pathogenicity. Since most of the conditions studied induced constant activity of PligA in vitro ( Figure 3B-6B), it is possible that LigA may contribute to the early stages in the course of leptospirosis. The kinetics of native LigA expression also supports this hypothesis, although protein levels did vary (Figure 2).
We found that native LigA was up-regulated upon induction by the combined conditions of physiological osmolarity, temperature upshift to 37uC and spermine, with a 5.95-fold increase. The corresponding reporter strain, PAG, was up-regulated by the same treatment (2.74-fold), demonstrating a good correlation between the promoter-probe methodology and native protein expression in L. interrogans, suggesting that LigA is post-transcriptionally regulated ( Table 2). This also suggests that L. interrogans protein expression may involve particular mechanisms or pathways to respond to specific host stimuli. The overall up-regulation of native LigA observed in comparison to the non-induced control was in agreement with the induction of the L. biflexa PAG reporter strain. Our results are in concordance with previous studies whereby the lig genes were expressed in vitro at both transcript and proteins levels [99], although a previous study using an L. interrogans serovar Pomona type Kennewicki strain was only able to detect the lig genes at the transcript level, [100]. We believe that these differences may be due to the different strains used. Moreover, ligB a paralogous gene to ligA, was up-regulated by an overnight temperature upshift from 30uC to 37uC (1.7-fold). Yet, physiological osmolarity stimulated ligA and ligB up-regulation (4.41 to 5.27fold -based on the transcriptional signals employing oligonucleotides to both the ligA/ligB identical region and the ligB unique region, respectively) [46,48]. Additionally, a recent study evaluated the influence of serum on ligB up-regulation (1.89-fold) [58]. Although we did not quantify the transcription of the ligA gene, a clear variation was observed in comparison to the non-induced controls, which is in agreement with previous findings (Figure 1). This reinforces the applicability of the promoter-probe vectors to assess promoter activity in L. biflexa. Based on these results, we conclude that (i) the combination of the in vitro conditions reliably simulated the host environment, (ii) L. biflexa can be used as a model to characterize L. interrogans promoters and (iii) this promoter-probe methodology may be helpful in the prediction of potential virulence factors of pathogenic Leptospira spp. Furthermore, previous studies showed ligA gene up-regulation shortly after temperature upshift [46,88], thus reinforcing the hypothesis that LigA may contribute to the early stages of infection and host adaptation. In addition, the Psph2 promoter was upregulated in the presence of spermine (2.12-fold). We observed some conflicting evidence in the correlation between native ligA mRNA and LigA abundance levels ( Figure 1 versus 2), similar to previous studies [88,101,102,103,104]. As previously suggested by Lo and colleagues [88], it is possible that the lack of correlation might be due to the longer half-life, greater stability or post transcriptional regulation of the mRNA transcript, i.e. as a result of the activity of small non-coding RNAs. Native LigA was detected at very low levels in the non-induced L. interrogans cells, while a significant up-regulation, at similar rates to recent studies [48], was observed when leptospires were grown under mammalian host conditions. In light of this information, we may conclude that the expression of both LigA and Sph2 is likely to be regulated at the post-transcriptional level.
Despite advances in the development of genetic tools for Leptospira spp. some basic questions remain unanswered. It is still unclear what regulators and pathways are associated with the expression of virulence factors. The level at which members of a given regulatory cascade exert induction/repression of transcription of LigA, Sph2 and other virulence factors is unknown. In this sudy we demonstrate that L. biflexa can serve as a model to study the genetics of L. interrogans. The pathogen-derived promoters exhibited activity in the L. biflexa, suggesting the existence of shared regulatory mechanisms among the saprophytes and pathogenic Leptospira spp. In addition, the establishment of a cut-off based on the promoter activity of virulence and non-virulence factors (Table 2), and the different fluorescence levels expressed by the reporter strains upon induction eliminated the possibility that the effects were simply due to stress responses. Furthermore, the suite of promoter-probe vectors developed in this study can be modified to make diverse translational fusions to investigate protein expression patterns Figure 8. Inducible expression of GFP in the L. biflexa promoter-probe constructs. Microscope images of the in vitro induced and uninduced P2G reporter strain transformed with pSLe94-sph2 promoter-gfp vector. Uninduced and physiological osmolarity/temperature upshift/ spermine exposed leptospires were fixed and visualized by phase contrast and by GFP fluorescence. Wild-type L. biflexa str. Patoc 1 was used as the negative control. doi:10.1371/journal.pone.0017409.g008 and complex regulatory networks involved in leptospiral protein regulation.
In conclusion, this study demonstrates the potential of a novel genetic tool for the identification and characterization of virulence factors of Leptospira spp. Our transcription and expression findings suggest that a combination of in vitro signals may be important to accurately simulate the host environment. We expect to provide further information towards understanding Leptospira spp. genetics and that will eventually serve as a basis for further studies. The understanding of the extent to which leptospiral promoters are regulated by mammalian host conditions, as well as the expression kinetics, may reveal useful information about the biology of Leptospira spp. A knock-in mutant library of different promoters is currently under construction with this in mind.