Identification of σE-Dependent Promoter Upstream of clpB from the Pathogenic Spirochaete Leptospira interrogans by Applying an E. coli Two-Plasmid System

There is limited information on gene expression in the pathogenic spirochaete Leptospira interrogans and genetic mechanisms controlling its virulence. Transcription is the first step in gene expression that is often determined by environmental effects, including infection-induced stresses. Alterations in the environment result in significant changes in the transcription of many genes, allowing effective adaptation of Leptospira to mammalian hosts. Thus, promoter and transcriptional start site identification are crucial for determining gene expression regulation and for the understanding of genetic regulatory mechanisms existing in Leptospira. Here, we characterized the promoter region of the L. interrogans clpB gene (clpBLi) encoding an AAA+ molecular chaperone ClpB essential for the survival of this spirochaete under thermal and oxidative stresses, and also during infection of the host. Primer extension analysis demonstrated that transcription of clpB in L. interrogans initiates at a cytidine located 41 bp upstream of the ATG initiation codon, and, to a lesser extent, at an adenine located 2 bp downstream of the identified site. Transcription of both transcripts was heat-inducible. Determination of clpBLi transcription start site, combined with promoter transcriptional activity assays using a modified two-plasmid system in E. coli, revealed that clpBLi transcription is controlled by the ECF σE factor. Of the ten L. interrogans ECF σ factors, the factor encoded by LIC_12757 (LA0876) is most likely to be the key regulator of clpB gene expression in Leptospira cells, especially under thermal stress. Furthermore, clpB expression may be mediated by ppGpp in Leptospira.


Introduction
Leptospira interrogans is one of the many pathogenic species of the genus Leptospira and is composed of several strains that can cause leptospirosis in mammals, including humans. It has been reported that over 1 million human cases of severe leptospirosis occur worldwide each year, with approximately 60,000 deaths from this disease [1]. It has to be noted that leptospirosis also generates huge economic losses in a number of countries due to reproductive disorders in cattle, sheep, pigs, and horses [2,3]. Furthermore, recent serological and microbiological studies have indicated a high rate of leptospiral infections in domestic animals [2,[4][5][6]. Despite a high risk of leptospirosis, especially in the tropical and subtropical countries, and its global importance, molecular mechanisms of both the leptospiral virulence and the disease pathogenesis currently remain largely unknown [7,8], mainly due to the historical lack of standard genetic tools for use in work with the pathogenic Leptospira species. Recent advances in genetic manipulation of these species have made it possible to identify several leptospiral virulence factors. However, many of them have turned out not to be required for virulence in animal models [7]. Unfortunately, the limitations of modern genetic tools available for pathogenic Leptospira spp. still have an enormous impact on the understanding of the molecular and cellular mechanisms involved in the pathogenesis of leptospirosis.
Like other pathogenic bacteria, L. interrogans is exposed to environmental stresses during infection of mammalian hosts. Bacteria fight various environmental stressors by altering the expression of genes involved in host adaptation and promoting their survival. Transcriptional regulation, and especially sigma factors controlling the promoter selectivity of bacterial RNA polymerase, play a crucial role in this stress-induced gene expression response. Unfortunately, there is limited information on gene regulation in Leptospira spp. Comparative genomics and genome-wide in silico analyses had shown that the Leptospira genome contains one basic sigma factor-encoding gene-rpoD (σ 70 ), and alternative sigma factor genes: rpoF (σ 28 ), rpoN (σ 54 ), and several (5)(6)(7)(8)(9)(10)(11) genes encoding extracytoplasmic function (ECF) sigma factors, referred to as rpoE (σ E ) in the literature [9,10] (Table 1). Table 1. Sigma factors identified in E. coli and predicted in L. interrogans (based on [10,11]).

Species.
σ Factor Function/Controlled Genes L. interrogans σ 70 housekeeping genes (>1000 genes) σ 28 genes encoding components of the endoflagellum and the flagellin-specific chaperone FliS σ 24 /σ E (ECF σ factor) extracytoplasmic function (469 putative binding sites in the Leptospira genome), stress response and virulence (clpB, this study) σ 54 genes encoding putative lipoproteins and the ammonium transporter AmtB E. coli σ 70 housekeeping genes σ S /σ 38 stationary phase gene expression σ 32 /σ H heat shock response σ 28 /σ F flagellar genes/motility genes σ 24 /σ E (ECF σ factor) extracytoplasmic function, cell surface stress response, resistance to heat shock and other environmental stresses σ 19 /σ FecI (ECF σ factor) extracytoplasmic function σ 54 /σ N nitrogen metabolism genes According to these analyses, the housekeeping sigma factor, σ 70 , coordinates transcription of most genes in the leptospiral cells, while alternative sigma factors play key roles in environmental adaptation of bacteria and their virulence. In that light, it is intriguing that the Leptospira species differ in the number of ECF σ factors. For example, pathogenic L. interrogans encodes ten different ECF σ factors (Table 2), while saprophytic species encode only five [9].
On the other hand, the well-known bacterium E. coli possesses only two ECF σ factors [14] ( Table 1). It is postulated that the ECF σ factors existing only in the pathogenic Leptospira species are important for the Leptospira life cycle within the host [9]. Unraveling regulation of virulence genes' expression is a particularly important challenge because it is necessary for understanding the molecular basis of the disease caused by the pathogen and its underpinning. Molecular chaperone ClpB, a member of the Hsp100/Clp subfamily of the AAA + ATPases (ATPases associated with a variety of cellular activities), is among the known leptospiral virulence factors. It has been demonstrated that ClpB deficiency in L. interrogans resulted in bacterial growth defects under oxidative and heat stresses, and also in its loss of virulence [15]. Furthermore, in previous studies, we had shown that L. interrogans ClpB (ClpB Li ) is not only synthesized but is also immunogenic during the infection process, further supporting its involvement in Leptospira pathogenicity [16]. Our recent studies suggest a possible role of ClpB Li , i.e., its aggregate-reactivation activity is necessary for maintaining the energy-generating metabolism of the Leptospira cell [17], again strongly supporting ClpB's importance in leptospiral virulence and implying the importance of clpB gene expression regulation. Here, primer extension analysis combined with promoter activity assays using a genetic strategy revealed that clpB Li transcription is σ E dependent. We postulate that factor encoded by LIC_12757 (LA0876) is a key ECF σ E factor that promotes clpB transcription in pathogenic Leptospira, especially under stressful conditions. In addition, the ppGpp alarmone may act as a regulator of clpB expression in those cells. Our findings provide the first insight into Leptospira clpB Li transcription regulation. Still, further studies are needed to discover all regulatory elements affecting clpB expression in Leptospira cells.

Identification of the clpB Transcriptional Start Sites by Primer Extension
In order to determine clpB transcriptional start sites and identify functional promoters responsible for its expression in Leptospira, primer extension assays were performed. Three synthetic oligonucleotides, p 51-72 , p 290-321 , p 565-591 (see Table 3) were hybridized to RNA extracted from both, Leptospira interrogans and E. coli MC4100∆clpB cells carrying the L. interrogans clpB gene (clpB Li ) cloned into a low-copy pGB2 plasmid together with a 500-bp DNA fragment containing sequence located upstream of the gene (pClpB Li ). A BLAST analysis did not show a potential cross-hybridization of the designed primers with regions of the E. coli genome, suggesting high specificity of the designed primers. To prepare RNA, the L. interrogans strain was grown at 28 • C for optimal in vitro growth and also under thermal stresses at 37 and 42 • C for 4 and 2 h, respectively, whereas E. coli cells were grown at 30 or 42 • C for 30 min (heat shock). Following reverse transcription and analysis on a sequencing gel, one major product was observed when the p 51-72 primer was used in reactions with RNA isolated from Leptospira cells grown at all tested temperatures ( Figure 1). This product corresponds to a transcriptional start site at a cytidine residue located 41 bp upstream of the first clpB ATG codon (−41 relative to ATG). In addition, a lesser band is observed at an adenosine residue located at the −39 position relative to ATG. It is interesting to note that this alternative start point of transcription (A-39) coincides with the primary transcriptional start site (tctAaac) previously predicted by [10] using differential RNA-sequencing (dRNA-seq). It is also worth mentioning that the secondary transcriptional start site (G-526 of the GenBank clpB gene sequence, accession number M28364) was found to be close to the primary transcriptional start site (A-525) in primer extension analysis of the E. coli clpB transcripts [18]. It is also known that Mycoplasma pneumoniae reference strain M129 uses two different transcriptional start sites (G-10 and A-7 relative to the ATG initiation codon) to produce ClpB [19]. Thus, the presence of the primary and secondary transcriptional start site in the clpB Li promoter region is not surprising. Further, the amount of both Leptospira clpB transcripts is increased after exposure of cells to a temperature of 42 • C for 2 h. This result indicates that clpB Li expression in Leptospira is induced by temperature stress, which is consistent with the findings of [15], who demonstrated that clpB Li transcription in Leptospira had increased by 3-folds after 2 h exposure to heat stress. One possible candidate regulating this expression could be RpoH (σ 32 ; heat shock sigma factor); however, comparative genomics and genome-wide in silico analyses had shown no presence of a rpoH gene in the Leptospira genome [9,10]. Thus, Leptospira spp. does not possess σ 32 that is required for heat shock gene expression and heat shock response in E. coli. On the other hand, as shown in Figure 1, for RNA extracted from the E. coli cells, no extension product was detectable with the p 51-72 primer, indicating that E. coli either does not possess an appropriate factor for clpB Li expression or this factor is not induced in E. coli by temperature stress. Additional analysis has been performed with two other primers. Primer extension assay utilizing the p 290-321 primer did not detect any transcripts either with Leptospira or with E. coli RNA (not shown), while the use of the p 565-591 primer revealed a few weak heat-inducible transcripts for Leptospira and one transcript for E. coli RNA that was not induced by heat shock stress and did not coincide with any leptospiral transcription start sites ( Figure 1). All of these transcripts initiate within the region that is 20-96 nt upstream of a potential internal translation initiation site of clpB Li with a good Shine-Dalgarno sequence (AGGGAA) and the ATG codon (see Figure 1). It is known that E. coli uses two translational start sites to produce two isoforms of ClpB, the full-length ClpB95 and shorter ClpB80, which does not contain the substrate-interacting N-terminal domain [18,20,21]. It has been demonstrated that the two isoforms of ClpB from E. coli cooperate in reactivation of aggregated proteins to form a highly efficient chaperone system [21]. It is very likely that the functional cooperation between ClpB isoforms arises from interactions between them because it was found that ClpB95 and ClpB80 associate into hetero-oligomers, which boost the aggregate-reactivation potential of the ClpB chaperone [21]. Thus, it was postulated that E. coli produces two isoforms of ClpB to optimize its disagregase activity. The presence of two forms of ClpB, a 96-kDa and an 80-kDa protein, was also demonstrated in L. interrogans [15]. Our results point to the occurrence of an additional transcriptional start site located within the clpB Li coding region in both E. coli and Leptospira. Still, it is intriguing that even though the internal clpB Li transcriptional start sites were detected in both instances, they differ depending on the host. It is also interesting to note the similarity between one of the internal transcriptional start sites (see Figure 1, gagtTtt) that we mapped by primer extension analysis and one of the sixteen internal start points (gagTttt) previously predicted by [10]. Thus, it is highly possible that clpB Li has an additional promoter. On the other hand, it cannot be excluded that the mapped minor transcripts result from the processing of the major transcript produced. Still, further studies are needed to examine the nature of those minor transcripts in Leptospira.

Presence of a Promoter Dependent on σ E Upstream of the ClpBLi Gene and its Activity in E. coli Cells
After the mapping of the two clpBLi transcriptional start sites (both probably originating from the same promoter) in Leptospira (see Figure 1), the nucleotide sequence upstream of clpBLi was carefully investigated. A putative σ E promoter element was successfully identified on the basis of sequence similarity to the well-characterized E. coli σ E core promoters. As shown in Figure 2, sequences of the −10 and −35 motifs of the potential clpB promoter (clpBLiP1) exhibit a considerable homology to the known promoters controlled by E. coli σ E . The temperature at which bacteria were grown prior to RNA isolation is indicated (28, 30, 37, or 42 • C). (B) clpB Li promoter and coding region with the clpB Li P1 promoter region underlined; transcriptional start sites are indicated by arrows as in (A), primer annealing sites and putative rbs sequences are also annotated; clpBLi-FL: start codon yielding ClpB95; clpBLi-short: internal ATG site producing ClpB80.

Presence of a Promoter Dependent on σ E Upstream of the ClpB Li Gene and its Activity in E. coli Cells
After the mapping of the two clpB Li transcriptional start sites (both probably originating from the same promoter) in Leptospira (see Figure 1), the nucleotide sequence upstream of clpB Li was carefully investigated. A putative σ E promoter element was successfully identified on the basis of sequence similarity to the well-characterized E. coli σ E core promoters. As shown in Figure 2, sequences of the −10 and −35 motifs of the potential clpB promoter (clpB Li P1) exhibit a considerable homology to the known promoters controlled by E. coli σ E .

Figure 2.
Nucleotide sequence alignment of the well-characterized E. coli σ E -dependent promoters and the potential clpBLi promoter (clpBLiP1). The two transcriptional start sites are marked in bold and underlined, and the major transcriptional start site is denoted by +1. The consensus sequence for the σ E -regulated promoter element is shown below the alignment.
To investigate whether the σ E factor may be really involved in the clpBLi transcriptional regulation, promoter activity assay was carried out under σ E -limiting (basal) and σ E excess (overproduction) conditions. It is worth to emphasize that under normal growth conditions, the expression level of rpoE in E. coli cells is low. Therefore, it was of great interest to compare the activity of the potential clpBLi promoter (clpBLiP1) under basal and increased levels of σ E . To our knowledge, there has been only one report to date, describing the construction and utility of the gfp reporter plasmid for assessing promoter activity in L. interrogans [22]. Therefore, due to the still existing limitations of genetic tools that could be easily utilized for examining gene expression in L. interrogans, experiments were performed in E. coli cells. A modified two-plasmid system has been employed that has been previously successfully used for the identification of many σ E -cognate promoters from E. coli and also from other bacteria in studies that demonstrated the suitability of this system for assessing promoter activity in E. coli [23,24]. Briefly, this system uses two compatible plasmids. One of them is pAC-rpoE4 carrying the E. coli rpoE gene under the control of an arabinose inducible pBAD promoter, and the second plasmid carries a reporter gene fused to a potential σ E promoter. We used the V. harveyi luxAB gene, encoding luciferase, as a reporter cloned into a low-copy pGB2 plasmid under control of the tested clpBLiP1 promoter (pclpBLiP1-luxAB). The method used assumes that shortly after arabinose addition σ E factor overproduced from pAC-rpoE4 interacts with the E. coli RNA polymerase core enzyme forming functional RNA polymerase holoenzyme that is capable of recognizing σ E -controlled promoters located upstream of a reporter gene carried on the second plasmid; in our study, it was the pclpBLiP1-luxAB plasmid carrying the luxAB reporter genes. As shown in Figure 3, the activity of luciferase in both, the exponential and stationary phase of bacterial growth was indeed significantly higher in the presence of the pAC-rpoE4 plasmid expressing the E. coli rpoE (i.e., under σ E excess conditions) than that in cells transformed only with pclpBLiP1-luxAB and producing a basal level of σ E . Figure 2. Nucleotide sequence alignment of the well-characterized E. coli σ E -dependent promoters and the potential clpB Li promoter (clpB Li P1). The two transcriptional start sites are marked in bold and underlined, and the major transcriptional start site is denoted by +1. The consensus sequence for the σ E -regulated promoter element is shown below the alignment.
To investigate whether the σ E factor may be really involved in the clpB Li transcriptional regulation, promoter activity assay was carried out under σ E -limiting (basal) and σ E excess (overproduction) conditions. It is worth to emphasize that under normal growth conditions, the expression level of rpoE in E. coli cells is low. Therefore, it was of great interest to compare the activity of the potential clpB Li promoter (clpB Li P1) under basal and increased levels of σ E . To our knowledge, there has been only one report to date, describing the construction and utility of the gfp reporter plasmid for assessing promoter activity in L. interrogans [22]. Therefore, due to the still existing limitations of genetic tools that could be easily utilized for examining gene expression in L. interrogans, experiments were performed in E. coli cells. A modified two-plasmid system has been employed that has been previously successfully used for the identification of many σ E -cognate promoters from E. coli and also from other bacteria in studies that demonstrated the suitability of this system for assessing promoter activity in E. coli [23,24]. Briefly, this system uses two compatible plasmids. One of them is pAC-rpoE4 carrying the E. coli rpoE gene under the control of an arabinose inducible p BAD promoter, and the second plasmid carries a reporter gene fused to a potential σ E promoter. We used the V. harveyi luxAB gene, encoding luciferase, as a reporter cloned into a low-copy pGB2 plasmid under control of the tested clpB Li P1 promoter (pclpB Li P1-luxAB). The method used assumes that shortly after arabinose addition σ E factor overproduced from pAC-rpoE4 interacts with the E. coli RNA polymerase core enzyme forming functional RNA polymerase holoenzyme that is capable of recognizing σ E -controlled promoters located upstream of a reporter gene carried on the second plasmid; in our study, it was the pclpB Li P1-luxAB plasmid carrying the luxAB reporter genes. As shown in Figure 3, the activity of luciferase in both, the exponential and stationary phase of bacterial growth was indeed significantly higher in the presence of the pAC-rpoE4 plasmid expressing the E. coli rpoE (i.e., under σ E excess conditions) than that in cells transformed only with pclpB Li P1-luxAB and producing a basal level of σ E . Figure 3. Activity of the clpBLiP1 promoter in E. coli cells during the exponential and stationary phase of growth. E. coli MG1655 cultures carrying the clpBLiP1-luxAB (luxAB) transcriptional fusion, or a two-plasmid system (pclpBLiP1-luxAB and pAC-rpoE4-E. coli rpoE) were grow in LB at 30 °C to OD600 nm of 0.45-0.5 and expression of rpoE was induced by addition of 0.02% arabinose for 1 h (for mid-exponential-phase culture) or 6 h (for early-stationary-phase culture), and the luciferase reporter assay was carried out. E. coli MG1655 cells carrying pSDLiluxAB (promoter-less luxAB) together with pAC-rpoE4 were used as a negative control. The luciferase activity (relative luminescence units/OD600 nm) in the E. coli strain without pAC-rpoE4 is taken as 100%; bars represent the percentage of luciferase activity normalized to that in the E. coli strain without rpoE overexpression; the results are presented as the average of three independent experiments, each performed with duplicate cultures, with standard deviations indicated. The paired t-test result: * p < 0.05 calculated with GraphPad Prism software. Importantly, no significant luminescent signal was detected for cells carrying a construct with promoter-less luxAB genes. These results strongly suggested that the activity of clpBLiP1 is dependent on σ E . Furthermore, we examined clpBLi expression driven by the clpBLiP1 promoter in E. coli cells under heat shock exposure at 42 °C and also under σ E excess conditions (in the presence of pAC-rpoE4). Thus, the main aim of this experiment was to compare the clpBLiP1 promoter behavior under thermal stress and σ E overproduction, by monitoring the amount of ClpBLi that was produced. Such an approach was impossible in case of experiments with pclpBLiP1-luxAB because luciferase is thermolabile and would not withstand the heat shock conditions. To this end, Western blotting analysis of whole-cell lysates of the E. coli MC4100ΔclpB strain carrying pClpBLi alone or together with pAC-rpoE4 was carried out using anti-ClpBLi158-334 serum [15]. Expression of E. coli rpoE was induced by the addition of 0.02% arabinose for 30-90 min, while the expression of pClpBLi was induced by exposure to 42 °C for 1 h. As shown in Figure 4a, thermal stress was not sufficient to induce the ClpBLi production (96-kDa protein) in the cells carrying the clpBLi gene under the control of the clpBLiP1 promoter (see lanes 5 and 6). This is in agreement with our results of primer extension analysis (see Figure 1a) that did not show the presence of the clpBLi transcripts in E. coli cells after heat shock stress. Consistent with our previous results [25], the use of pGB2-ClpBLi [25], which carries the heat-inducible promoter σ 32 from E. coli located upstream of clpBLi, resulted in increased synthesis of ClpBLi after heat treatment (see lanes 3 and 4). The most interesting observation pertains to the clpBLi expression under σ E excess conditions (Figure 4b), as the appearance of the 96-kDa ClpBLi in the E. coli cells 30-90 min after arabinose addition (see lanes [3][4][5] provides further evidence that clpBLiP1 is σ E dependent. As can be seen in Figure 4b, there was one more protein that occurred at a higher amount after σ E induction. It is likely that this protein corresponds to a shorter form of ClpBLi (80 kDa) that either represents transcription initiation from the internal promoter or exclusively translation initiation from the internal in-frame ATG codon. Figure 3. Activity of the clpB Li P1 promoter in E. coli cells during the exponential and stationary phase of growth. E. coli MG1655 cultures carrying the clpB Li P1-luxAB (luxAB) transcriptional fusion, or a two-plasmid system (pclpB Li P1-luxAB and pAC-rpoE4-E. coli rpoE) were grow in LB at 30 • C to OD 600 nm of 0.45-0.5 and expression of rpoE was induced by addition of 0.02% arabinose for 1 h (for mid-exponential-phase culture) or 6 h (for early-stationary-phase culture), and the luciferase reporter assay was carried out. E. coli MG1655 cells carrying pSD Li luxAB (promoter-less luxAB) together with pAC-rpoE4 were used as a negative control. The luciferase activity (relative luminescence units/OD 600 nm ) in the E. coli strain without pAC-rpoE4 is taken as 100%; bars represent the percentage of luciferase activity normalized to that in the E. coli strain without rpoE overexpression; the results are presented as the average of three independent experiments, each performed with duplicate cultures, with standard deviations indicated. The paired t-test result: * p < 0.05 calculated with GraphPad Prism software. Importantly, no significant luminescent signal was detected for cells carrying a construct with promoter-less luxAB genes. These results strongly suggested that the activity of clpB Li P1 is dependent on σ E . Furthermore, we examined clpB Li expression driven by the clpB Li P1 promoter in E. coli cells under heat shock exposure at 42 • C and also under σ E excess conditions (in the presence of pAC-rpoE4). Thus, the main aim of this experiment was to compare the clpB Li P1 promoter behavior under thermal stress and σ E overproduction, by monitoring the amount of ClpB Li that was produced. Such an approach was impossible in case of experiments with pclpB Li P1-luxAB because luciferase is thermolabile and would not withstand the heat shock conditions. To this end, Western blotting analysis of whole-cell lysates of the E. coli MC4100∆clpB strain carrying pClpB Li alone or together with pAC-rpoE4 was carried out using anti-ClpB Li158-334 serum [15]. Expression of E. coli rpoE was induced by the addition of 0.02% arabinose for 30-90 min, while the expression of pClpB Li was induced by exposure to 42 • C for 1 h. As shown in Figure 4A, thermal stress was not sufficient to induce the ClpB Li production (96-kDa protein) in the cells carrying the clpB Li gene under the control of the clpB Li P1 promoter (see lanes 5 and 6). This is in agreement with our results of primer extension analysis (see Figure 1A) that did not show the presence of the clpB Li transcripts in E. coli cells after heat shock stress. Consistent with our previous results [25], the use of pGB2-ClpB Li [25], which carries the heat-inducible promoter σ 32 from E. coli located upstream of clpB Li , resulted in increased synthesis of ClpB Li after heat treatment (see lanes 3 and 4). The most interesting observation pertains to the clpB Li expression under σ E excess conditions ( Figure 4B), as the appearance of the 96-kDa ClpB Li in the E. coli cells 30-90 min after arabinose addition (see lanes 3-5) provides further evidence that clpB Li P1 is σ E dependent. As can be seen in Figure 4B, there was one more protein that occurred at a higher amount after σ E induction. It is likely that this protein corresponds to a shorter form of ClpB Li (80 kDa) that either represents transcription initiation from the internal promoter or exclusively translation initiation from the internal in-frame ATG codon. Together, the results obtained with E. coli σ E support that ECF σ E factor may mediate the clpBLi transcription in L. interrogans cells.

Role of Leptospiral ECFσ Factors in ClpBLi Gene Expression
It has been demonstrated that σ E contributes to heat shock and oxidative stress response in a number of bacteria and is also involved in the regulation of virulence genes and virulence-associated genes in many bacterial pathogens as well [26]. Since L. interrogans rpoE expression was reported to be up-regulated in Leptospira cells exposed to elevated temperatures, it is possible that leptospiral σ E factors also regulate expression of genes required for virulence or pathogenesis [13]. This correlates well with a specific function of ClpB in L. interrogans and ClpB's importance during infections [15,17,25]. Of the ten genes encoding ECF σ (σ E ) factors listed in Table 2, LIC_12757 (LA0876), and LIC_10559 (LA3652) were found to be 1.5-and 2-fold up-regulated at higher temperatures, respectively [13]. Therefore, we considered their products as potential candidates ECF (σ E ) that transcribe clpBLi and we evaluated their effect on the transcriptional activity of clpBLiP1. To this end, we replaced the E. coli rpoE gene in the previously employed two-plasmid system with either LIC_12757 or LIC_10559 genes. Additionally, LIC_10144, which also encodes ECF σ factor (see Table  2) but is not up-regulated under heat shock, was also cloned and used in our assay. As shown in Figure 5, products of the LIC_12757 and LIC_10144 genes elevated the transcriptional activity of clpBLiP1 in both the exponential and stationary phase of bacterial growth, supporting our previous observations obtained for σ E from E. coli and indicating that an ECF σ E factor is indeed involved in clpBLi regulation (see Figure 3). Of note, clpBLiP1 displays a much higher activity in the presence of LIC_12757. This result suggests that its product may play a key role in clpBLi expression, especially under stressful conditions like high temperature. To our surprise, LIC_10559, whose expression level in response to thermal stress was reported to be more increased than that of LIC_12757 [13], did not significantly affect the transcriptional activity of clpBLiP1. Together, the results obtained with E. coli σ E support that ECF σ E factor may mediate the clpB Li transcription in L. interrogans cells.

Role of Leptospiral ECFσ Factors in ClpB Li Gene Expression
It has been demonstrated that σ E contributes to heat shock and oxidative stress response in a number of bacteria and is also involved in the regulation of virulence genes and virulence-associated genes in many bacterial pathogens as well [26]. Since L. interrogans rpoE expression was reported to be up-regulated in Leptospira cells exposed to elevated temperatures, it is possible that leptospiral σ E factors also regulate expression of genes required for virulence or pathogenesis [13]. This correlates well with a specific function of ClpB in L. interrogans and ClpB's importance during infections [15,17,25]. Of the ten genes encoding ECF σ (σ E ) factors listed in Table 2, LIC_12757 (LA0876), and LIC_10559 (LA3652) were found to be 1.5-and 2-fold up-regulated at higher temperatures, respectively [13]. Therefore, we considered their products as potential candidates ECF (σ E ) that transcribe clpB Li and we evaluated their effect on the transcriptional activity of clpB Li P1. To this end, we replaced the E. coli rpoE gene in the previously employed two-plasmid system with either LIC_12757 or LIC_10559 genes. Additionally, LIC_10144, which also encodes ECF σ factor (see Table 2) but is not up-regulated under heat shock, was also cloned and used in our assay. As shown in Figure 5, products of the LIC_12757 and LIC_10144 genes elevated the transcriptional activity of clpB Li P1 in both the exponential and stationary phase of bacterial growth, supporting our previous observations obtained for σ E from E. coli and indicating that an ECF σ E factor is indeed involved in clpB Li regulation (see Figure 3). Of note, clpB Li P1 displays a much higher activity in the presence of LIC_12757. This result suggests that its product may play a key role in clpB Li expression, especially under stressful conditions like high temperature. To our surprise, LIC_10559, whose expression level in response to thermal stress was reported to be more increased than that of LIC_12757 [13], did not significantly affect the transcriptional activity of clpB Li P1.

Figure 5.
Transcriptional activity of the clpBLiP1 promoter in the presence of leptospiral ECF σ (σ E ) factors during the exponential and stationary phase of growth. E. coli MG1655 cultures carrying the clpBLiP1-luxAB (luxAB) transcriptional fusion or a two-plasmid system (pclpBLiP1-luxAB and an appropriate LIC gene coding σ E factor) were grow in LB at 30 °C to OD600 nm of 0.45-0.5 and expression of LIC genes was induced by addition of 0.02% arabinose for 1-3 h (for exponential-phase culture) or 6 h (for early-stationary-phase culture), then the luciferase reporter assay was carried out. The luciferase activity (relative luminescence units/OD600 nm) in the E. coli strain without LIC genes and 1 h after arabinose addition is taken as 100% (*); bars represent the percentage of luciferase activity normalized to that in the E. coli strain without σ E overexpression (*); the results are presented as the average of three independent experiments, each performed with duplicate cultures, with standard deviations indicated. The paired t-test result: *** p < 0.001 calculated with GraphPad Prism software.
Interestingly, comparison of promoter transcriptional activity in cells harboring pclpBLiP1-luxAB alone or together with LIC_10144 or LIC_12757 genes, as well as with E. coli rpoE, revealed that clpBLiP1 is more active in the presence of excess σ E in the stationary growth phase than in the mid-exponential phase when compared to basal σ E level (see Figures 3 and 5). Similar effect has been previously observed and carefully investigated for the E. coli rpoH promoter [27,28] and also for other σ E -controlled promoters [29]. It has been proposed that an increase in activity of σ E promoters during stationary phase results from elevation in the ppGpp level in this growth phase. To examine whether ppGpp could have an impact on the clpBLi promoter's transcriptional activity and, therefore, on clpBLi expression, the activity of luciferase was determined in both the wild-type strain (MG1655) and its derivative, i.e., the ΔrelAΔspoT mutant strain lacking ppGpp. We observed ( Figure 6) that in the ΔrelAΔspoT mutant strain, the luciferase activity was significantly lower in the stationary phase than in the wild-type strain, while the opposite was observed in the logarithmic phase of growth. This observation suggests that the clpBLiP1 promoter may be differently regulated by ppGpp during growth-it is repressed in the exponential phase and activated in the stationary phase. It is known that ppGpp is a key factor of the stringent response, which is a widespread response to changing environmental conditions found in all bacterial species tested so far [30]. Importantly, ppGpp is necessary for bacterial virulence and pathogenicity [31]. Presence of a single Rel-like bifunctional protein with (p) ppGpp-hydrolase/synthase activity, RelLin, in L. interrogans points to the existence of the stringent response in this bacterium [32] and corroborates our findings. Still, although ppGpp regulation of clpBLiP1 in Leptospira is very likely, it cannot be excluded that the promoter transcriptional activity observed in E. coli does not reflect regulation in Leptospira. The contribution of the stringent response to Leptospira zoonotic lifecycle is yet to be examined. Figure 5. Transcriptional activity of the clpB Li P1 promoter in the presence of leptospiral ECF σ (σ E ) factors during the exponential and stationary phase of growth. E. coli MG1655 cultures carrying the clpB Li P1-luxAB (luxAB) transcriptional fusion or a two-plasmid system (pclpB Li P1-luxAB and an appropriate LIC gene coding σ E factor) were grow in LB at 30 • C to OD 600 nm of 0.45-0.5 and expression of LIC genes was induced by addition of 0.02% arabinose for 1-3 h (for exponential-phase culture) or 6 h (for early-stationary-phase culture), then the luciferase reporter assay was carried out. The luciferase activity (relative luminescence units/OD 600 nm ) in the E. coli strain without LIC genes and 1 h after arabinose addition is taken as 100% (*); bars represent the percentage of luciferase activity normalized to that in the E. coli strain without σ E overexpression (*); the results are presented as the average of three independent experiments, each performed with duplicate cultures, with standard deviations indicated. The paired t-test result: *** p < 0.001 calculated with GraphPad Prism software.
Interestingly, comparison of promoter transcriptional activity in cells harboring pclpB Li P1-luxAB alone or together with LIC_10144 or LIC_12757 genes, as well as with E. coli rpoE, revealed that clpB Li P1 is more active in the presence of excess σ E in the stationary growth phase than in the mid-exponential phase when compared to basal σ E level (see Figures 3 and 5). Similar effect has been previously observed and carefully investigated for the E. coli rpoH promoter [27,28] and also for other σ E -controlled promoters [29]. It has been proposed that an increase in activity of σ E promoters during stationary phase results from elevation in the ppGpp level in this growth phase. To examine whether ppGpp could have an impact on the clpB Li promoter's transcriptional activity and, therefore, on clpB Li expression, the activity of luciferase was determined in both the wild-type strain (MG1655) and its derivative, i.e., the ∆relA∆spoT mutant strain lacking ppGpp. We observed ( Figure 6) that in the ∆relA∆spoT mutant strain, the luciferase activity was significantly lower in the stationary phase than in the wild-type strain, while the opposite was observed in the logarithmic phase of growth. This observation suggests that the clpB Li P1 promoter may be differently regulated by ppGpp during growth-it is repressed in the exponential phase and activated in the stationary phase. It is known that ppGpp is a key factor of the stringent response, which is a widespread response to changing environmental conditions found in all bacterial species tested so far [30]. Importantly, ppGpp is necessary for bacterial virulence and pathogenicity [31]. Presence of a single Rel-like bifunctional protein with (p) ppGpp-hydrolase/synthase activity, Rel Lin , in L. interrogans points to the existence of the stringent response in this bacterium [32] and corroborates our findings. Still, although ppGpp regulation of clpB Li P1 in Leptospira is very likely, it cannot be excluded that the promoter transcriptional activity observed in E. coli does not reflect regulation in Leptospira. The contribution of the stringent response to Leptospira zoonotic lifecycle is yet to be examined. ΔrelAΔspoT mutant (ppGpp 0 ) cultures carrying pclpBLiP1-luxAB along with LIC_10144 or LIC_12757 genes were grow in LB at 30 °C to OD600 nm of 0.45-0.5 and expression of the LIC genes encoding ECF σ E was induced by addition of 0.02% arabinose for 1 h (for mid-exponential-phase culture) or 6 h (for early-stationary-phase culture), and then the luciferase reporter assay was carried out. The luciferase activity (relative luminescence units/OD600nm) in the E. coli ppGpp + cells during exponential phase is taken as 100%; bars represent the percentage of luciferase activity normalized to that of E. coli ppGpp + cells during exponential phase; results are presented as the average of three independent experiments, each performed with duplicate cultures, with standard deviations indicated. The paired t-test result: ** p < 0.01 and *** p < 0.001 calculated with GraphPad Prism software.

Cloning and PCR Methods
To generate a transcriptional fusion between the clpBLi promoter region (500 bp upstream of the first ATG codon) and luciferase reporter genes, the V. harveyi luxAB genes were amplified from pLucVh [21] by PCR using Pfu Turbo DNA polymerase (Agilent Technologies/Perlan Technologies, Warsaw, Poland) with appropriate primers (see Table 3). First, the PCR product was cloned into a pJET1.2 blunt vector, then digested with appropriate restriction enzymes and ligated with the linearized pGB2 NdeI-HindIII vector and the clpBLi promoter region to produce the clpBLiP1-luxAB fusion. The sequence of the resulting construct was confirmed by DNA sequencing (Genomed S.A., Warsaw, Poland). To construct a negative control reporter plasmid (pSDLiluxAB; promoter-less luxAB), V. harveyi luxAB genes were amplified together with the Shine-Dalgarno sequence of clpBLi Figure 6. Transcriptional activity of the clpB Li gene promoter is dependent on ppGpp during the entry of bacteria into the stationary phase. E. coli wild-type (MG1655; ppGpp + ) and ∆relA∆spoT mutant (ppGpp 0 ) cultures carrying pclpB Li P1-luxAB along with LIC_10144 or LIC_12757 genes were grow in LB at 30 • C to OD 600 nm of 0.45-0.5 and expression of the LIC genes encoding ECF σ E was induced by addition of 0.02% arabinose for 1 h (for mid-exponential-phase culture) or 6 h (for early-stationary-phase culture), and then the luciferase reporter assay was carried out. The luciferase activity (relative luminescence units/OD600nm) in the E. coli ppGpp + cells during exponential phase is taken as 100%; bars represent the percentage of luciferase activity normalized to that of E. coli ppGpp + cells during exponential phase; results are presented as the average of three independent experiments, each performed with duplicate cultures, with standard deviations indicated. The paired t-test result: ** p < 0.01 and *** p < 0.001 calculated with GraphPad Prism software.

Cloning and PCR Methods
To generate a transcriptional fusion between the clpB Li promoter region (500 bp upstream of the first ATG codon) and luciferase reporter genes, the V. harveyi luxAB genes were amplified from pLucVh [21] by PCR using Pfu Turbo DNA polymerase (Agilent Technologies/Perlan Technologies, Warsaw, Poland) with appropriate primers (see Table 3). First, the PCR product was cloned into a pJET1.2 blunt vector, then digested with appropriate restriction enzymes and ligated with the linearized pGB2 NdeI-HindIII vector and the clpB Li promoter region to produce the clpB Li P1-luxAB fusion. The sequence of the resulting construct was confirmed by DNA sequencing (Genomed S.A., Warsaw, Poland). To construct a negative control reporter plasmid (pSD Li luxAB; promoter-less luxAB), V. harveyi luxAB genes were amplified together with the Shine-Dalgarno sequence of clpB Li from pclpB Li P1-luxAB by PCR using Pfu Turbo DNA polymerase (Agilent Technologies) with appropriate primers (see Table 3). Next, the same cloning procedure was used as described above.
To study the clpB Li expression in E. coli MC4100∆clpB cells, clpB Li (2583 bp) and its promoter region (500 bp) were amplified by PCR from genomic DNA of L. interrogans (extracted with a QIAamp DNA mini kit (Qiagen, Germantown, MD, USA), by using Pfu Turbo DNA polymerase (Agilent Technologies) with appropriate primers (see Table 3). Subsequently, the same cloning procedure was used as described above for the construction of transcriptional fusion.
To clone leptospiral genes encoding ECF σ factors, LIC_10144, LIC_10559, and LIC_12757, into pAC7 under control of a pBAD promoter, the appropriate fragments of genomic DNA isolated from L. interrogans were amplified by PCR using Pfu Turbo polymerase and appropriate primers (see Table 3). Then, the same cloning procedure was used as described above for the construction of transcriptional fusion.

RNA Isolation and Primer Extension Assays
Total RNA was isolated from L. interrogans and E. coli MC4100∆clpB cells carrying the clpB Li gene along with a region 500 bp upstream of the first ATG codon, using the Total RNA Mini Plus kit (A&A Biotechnology). L. interrogans strain was grown in liquid Ellinghausen McCollough Jonhson and Harris medium (EMJH), as described previously [17]. E. coli strain was cultured in the LB medium supplemented with 50 µg/mL spectinomycin and 30 µg/mL kanamycin at 30 • C to OD 600 nm of 0.35, then transferred to 42 • C for 30 min (heat shock), whereas the thermally non-induced bacteria were further grown at 30 • C. The primers used for primer extension analyses are described in Table 3. Briefly, 2 µg of total RNA were combined with 0.6 pmol of a primer labeled at the 5'-terminus with [γ-32 P]ATP (specific activity 6000 Ci/mmol; Hartmann) using polynucleotide kinase (Promega), denatured at 75 • C for 10 min, and annealed at 58 • C for 20 min. This was followed by the addition of extension buffer (1x AMV reverse transcriptase buffer (Promega, Medison, WI, USA), 1 mM dNTPs, 20 U RNase-In (Promega), and 10 U of the AMV reverse transcriptase (Promega). The final reaction volume was 25 µL. After incubation at 42 • C for 30 min, 2.5 µg of RNase H (Promega) were added, and incubation was continued at 37 • C for 15 min. cDNA was precipitated in ethanol with 0.35 M sodium acetate (pH 5.5). The samples were then resuspended in a loading buffer (95% formamide, 0.05% bromophenol blue, 0.05% xylene cyanol) and resolved on a 7 M urea, 8% polyacrylamide sequencing gel, run in parallel with a sequencing reaction performed with the same labeled primer and pClpB Li plasmid DNA (see Section 2.1), using a cycle sequencing kit (Jena Bioscience, Jena, Germany).

Luciferase Activity Assay
V. harveyi luciferase activity assay was performed as described previously [21], namely, 200 µL culture aliquots were withdrawn at the indicated times and mixed with 7 µL of 10% n-decanal (Sigma/Merck KGaA, Darmstadt, Germany) in ethanol for up to 1 min. Luminescence produced by the enzyme was monitored using a Berhold luminometer.

Conclusions
To our knowledge, this is the first study providing insight into the transcriptional regulation of clpB in the pathogen L. interrogans. Primer extension analysis, in combination with promoter transcriptional activity assays permitted us to identify a major heat-inducible promoter located upstream of the clpB encoding sequence. Our results clearly indicate that the clpB Li transcription is under control of one of the σ E -type factors, which have been also reported to regulate the biosynthesis of virulence factors in some bacterial pathogens. It has to be noted that ClpB Li is among the known leptospiral virulence factors. It is likely that the product of LIC_12757 plays a key role in clpB transcription in L. interrogans cells. Furthermore, we show that the ppGpp alarmone, which also allows expression of virulence genes in some bacterial pathogens, may mediate clpB expression in Leptospira cells. We believe that findings presented in this study help to improve our knowledge of regulatory mechanisms used by Leptospira and may also have an impact on the understanding of the leptospiral virulence mechanisms that allow this pathogen to adapt to its host organisms.