NprR is a quorum sensor of the RNPP family found in bacteria of the Bacillus cereus group. In association with its cognate peptide NprX, NprR controls the expression of genes essential for survival and sporulation of Bacillus thuringiensis during its necrotrophic development in insects. Here, we report that the nprR–nprX genes are not autoregulated and are co-transcribed from a σA-dependent promoter (PA) located upstream from nprR. The transcription from PA starts at the onset of the stationary phase and is controlled by two transcriptional regulators: CodY and PlcR. The nutritional repressor CodY represses nprR–nprX transcription during the exponential growth phase and the quorum sensor PlcR activates nprR–nprX transcription at the onset of stationary phase. We show that nprX is also transcribed independently of nprR from two promoters, PH and PE, dependent on the sporulation-specific sigma factors, σH and σE respectively. Both promoters ensure nprX transcription during late stationary phase while transcription from PA has decreased. These results show that the activity of the NprR–NprX quorum sensing system is tightly co-ordinated to the physiological stage throughout the developmental process of the Bacillus.
Quorum sensing is mediated by the secretion and detection of diffusible biomolecules which accumulate in a cell density-dependent manner (Waters and Bassler, 2005). When the concentrations of these signalling molecules reach a threshold, bacteria respond by inducing the expression of genes involved in specific functions such as sporulation, genetic competence and expression of virulence factors. Many of the signalling molecules used by Gram-negative bacteria are acyl-homoserine lactones, whereas many of those in Gram-positive bacteria are oligopeptides. Signalling peptides elicit a response when they are recognized by their cognate membrane-bound two-component sensor histidine kinase or by binding directly to their intracellular receptor in the responder cell (Slamti and Lereclus, 2002; Pottathil and Lazazzera, 2003; Bassler and Losick, 2006). The RNPP family (named after the key members Rap/NprR/PlcR/PrgX) includes the quorum sensing systems in which the regulatory protein interacts directly with the signalling peptide (Declerck et al., 2007). Their signalling peptides are encoded as a precursor by a small gene located downstream from the gene encoding the regulatory protein. This precursor is cleaved to yield the active oligopeptide (Perego and Brannigan, 2001; Slamti and Lereclus, 2005; Bouillaut et al., 2008) which is then reimported by an oligopeptide permease to interact with its target (Perego et al., 1991; Rudner et al., 1991; Gominet et al., 2001).
In Bacillus subtilis, the Rap family consists of 11 members including phosphatases and proteins that inhibit the activity of other proteins by a protein–protein interaction. The Rap proteins control diverse processes involved in bacterial physiology: the initiation of sporulation, genetic competence and degradative enzyme production (Perego et al., 1996; Ogura et al., 2003; Auchtung et al., 2006). The rap genes are generally co-transcribed with their phr genes from a promoter located upstream from the rap gene (Perego and Brannigan, 2001). Transcription from this promoter is controlled by various regulators: transcription of rapA, rapC and rapE is positively regulated by the response regulator ComA (Mueller et al., 1992; Lazazzera et al., 1999; Jiang et al., 2000) and negatively affected by the transcriptional regulator CodY (Lazazzera et al., 1999; Molle et al., 2003), and rapD, rapH and rapG are negatively regulated by the transcriptional repressor RghR (Hayashi et al., 2006; Ogura and Fujita, 2007). In addition, the expression of rapA is negatively regulated by DegU (Ogura et al., 2001) and the expression of rapC, rapE, rapF and rapI is indirectly activated by Spo0A (Fawcett et al., 2000). Most of the phr genes have a second promoter which is controlled by the alternative σH factor (McQuade et al., 2001). The Rap proteins in the B. cereus group have not been described in as much detail, although six Rap proteins have been identified in Bacillus anthracis. Two of them are involved in the dephosphorylation of Spo0F∼P and affect sporulation initiation (Bongiorni et al., 2006).
In B. cereus and B. thuringiensis, PlcR is the major virulence regulator activating the expression of several extracellular factors, including degradative enzymes, cytotoxic proteins and cell surface proteins (Agaisse et al., 1999; Gohar et al., 2008). PlcR activity depends on the PapR signalling peptide encoded by a small gene located immediately downstream from plcR (Lereclus et al., 1996; Slamti and Lereclus, 2002). The PlcR–PapR complex binds to a DNA target site (the PlcR box) upstream from the σA promoter of the regulated genes (Agaisse et al., 1999; Gohar et al., 2008). A PlcR box has been identified in the promoter regions of the plcR and papR genes and it has been demonstrated that their transcription is auto-induced (Lereclus et al., 1996; Agaisse et al., 1999). Moreover, the transcription of plcR is repressed by the sporulation factor Spo0A (Lereclus et al., 2000). Spo0A binds to two DNA target sites flanking the PlcR box upstream from the plcR gene and represses its expression, probably by preventing the binding of the PlcR–PapR complex to its recognition site. Consequently, the expression of the PlcR regulon is abolished in a sporulation-specific medium, in which the level of Spo0A∼P increases sharply at the end of exponential growth. In addition, it has been recently shown that the expression of the PlcR regulon requires the presence of the global regulator CodY (Frenzel et al., 2012; Lindbäck et al., 2012).
NprR, another member of the RNPP family, has recently been described in bacteria of the B. cereus group (Perchat et al., 2011). This quorum sensor interacts with the signalling peptide NprX and the NprR–NprX complex activates the expression of nprA, a gene encoding a metalloprotease, during late stationary phase. A transcriptomic analysis revealed that NprR regulates at least 40 additional genes encoding proteins involved in nutrient supply, stress and antibiotic resistance, and in the synthesis of a lipopeptide (Dubois et al., 2012). In vivo experiments with insect larvae indicated that these genes are expressed after the death of the host allowing the bacteria of the B. cereus group to survive and to sporulate in the insect cadaver.
These various observations highlight the prominent role of the RNPP family regulators in the developmental process of bacteria of the B. cereus group in their host: PlcR participates in virulence, NprR is involved in necrotrophic development in the host cadaver, and several Rap proteins regulate the commitment to sporulation by controlling the phosphorylation level of Spo0A. The regulation of expression of plcR–papR and rap–phr has been relatively well studied, but the regulation of nprR–nprX expression was unknown. Here, we show that the expression of nprR and nprX is controlled through the developmental process by at least two transcriptional regulators and by two sporulation-specific sigma factors. This complex regulation allows co-ordination of the quorum sensing system to the physiological stage of the bacteria.
Transcriptional analysis of the nprR–nprX locus
Like other regulators of the RNPP family, nprR is upstream from a small open reading frame, designated nprX. The two genes are separated by 2 bp, suggesting that they are co-transcribed (Fig. 1A). To verify this, transcription of the nprR–nprX locus in B. thuringiensis strain 407 was investigated (Fig. 1B). RNA was extracted from cultures of B. thuringiensis strain 407 2 h after initiation of sporulation (t2) and subjected to RT-PCR with primers overlapping the bthur0002_5070 and nprR genes; no amplification product was detected on 1% agarose gel. However, the same experiment with primers overlapping the nprR and nprX genes produced a PCR product of the expected size. These results indicate that nprR and nprX were co-transcribed from a promoter upstream from nprR.
We investigated whether the inverted repeated sequence downstream from the nprX gene is a transcriptional terminator. DNA fragments including or excluding this putative terminator were inserted into the low-copy-number plasmid pHT304-18Z upstream from the lacZ reporter gene (Fig. 1A and C). The β-galactosidase activity 4 h after entry into the stationary phase (t4) was about 11000 Miller units in the absence of the inverted repeat sequence, whereas it was less than 200 Miller units in the presence of the inverted repeat sequence. This indicates that the inverted repeat sequence acts as a rho-independent terminator.
To determine the nprR transcription start site, primer extension analysis was performed using oligonucleotide Npr-P and RNA extracted from B. cereus cells grown in LB medium from t−1 to t1 (Fig. 2A and B). A transcript was detected at t1 and the start site was identified at nucleotide position −27 with reference to the nprR translation start codon. Putative −35 and −10 regions (TTTATA and TATTAT with a 17 bp spacer), resembling those recognized by σA-RNA polymerase, are present upstream from the transcription start site. The nprR promoter region also contains a 16 bp palindromic sequence (between positions −105 and −89) matching the PlcR consensus binding sequence (TATGnAwwwwTnCATA, where w is A or T and n is any base) (Gohar et al., 2008): the sole difference is an A in the first position instead of a T. However, a mutagenesis analysis indicated that the first base of the palindromic sequence may be changed without abolishing activity (Gohar et al., 2008). These results suggest that nprR and nprX are co-transcribed from a σA promoter and that transcription is controlled by the pleiotropic regulator PlcR. Moreover, a putative CodY target sequence presenting two mismatches with the proposed consensus sequence AATTTTCwGAAAATT (Belitsky and Sonenshein, 2008) is located between positions −90 and −75, thus suggesting control of nprR–nprX transcription by this global regulator of transition phase.
PlcR specifically binds to the nprR promoter
To determine whether the PlcR–PapR complex specifically binds to the putative PlcR box of the nprR promoter, we performed Isothermal Titration Calorimetry (ITC) experiments using a 25 bp dsDNA fragment containing the putative PlcR box of the nprR promoter (nprR-25) (Fig. 3A). Calorimetric titrations were performed in which graduated amounts of DNA probes were sequentially injected into the sample cell containing the PlcR–PapR7 complex. This result was compared with a non-specific signal obtained with a 22 bp dsDNA fragment centred on the −35 box of the nprA promoter devoid of PlcR box (nprA-22) (Fig. 3B). No interaction was detected using the non-specific nprA-22 probe while nice saturation curve with a Kd value of 1 μM was obtained with the nprR-25 probe. This experiment demonstrates that the PlcR–PapR complex specifically binds to the promoter region of nprR and may thereby directly control nprR–nprX transcription from PA.
The nprR–nprX transcription is activated by PlcR
We constructed a chromosomal fusion between the nprR promoter region (PnprR) and the lacZ gene in the B. thuringiensis strain 407 (strain nprR′Z) to study nprR–nprX expression. This strain harbours a promoterless lacZ gene in place of the nprR–nprX coding sequence, and is deficient for NprR and NprX. The kinetics of PnprR′–lacZ expression was determined in a rich medium (LB) delaying the initiation of sporulation and in a sporulation-specific medium (HCT) (Fig. 3C and D). In LB medium, the transcription started at the end of exponential growth and increased during the stationary phase. In HCT medium, PnprR′–lacZ expression started earlier, increased slightly during the first stages of sporulation and stopped around t3. Genetic complementation of the nprR′Z strain with pHT304-RX did not affect transcription, indicating that the transcription of nprR–nprX is not auto-regulated (Fig. S2A). To investigate the involvement of PlcR in the expression of nprR–nprX, the plcR gene in the nprR′Z strain was disrupted by homologous recombination. In LB medium, the transcription of nprR′Z in the ΔplcR strain was half that in the parental strain (Fig. 3C). In agreement with the presence of a PlcR consensus binding site upstream from the nprR promoter, this result indicates that the transcription of the nprR–nprX genes is activated by PlcR. However, in HCT medium, PnprR′–lacZ expression in the ΔplcR stain was indistinguishable from that in the parental strain. This absence of difference may result from the repression of plcR transcription by Spo0A∼P, the phosphorylated form of Spo0A. Indeed, plcR transcription is abolished in sporulation-specific medium because of Spo0A∼P (Lereclus et al., 2000). We therefore measured PnprR′–lacZ expression in Δspo0A and Δspo0A ΔplcR mutant strains grown in HCT medium (Fig. 3D). β-Galactosidase activity of the Δspo0A mutant was higher than that of the parental strain in this medium, suggesting that Spo0A∼P acts as a negative regulator of nprR–nprX expression. However, β-galactosidase activity of the Δspo0A ΔplcR mutant strain was similar to that in the parental strain. These results demonstrate an indirect negative control of Spo0A∼P via PlcR on nprR–nprX transcription.
CodY binds to the nprR promoter
nprR–nprX expression is only partially activated by PlcR suggesting that additional factors are required to trigger nprR–nprX expression. There is a putative CodY box upstream from PA (Fig. 2B); moreover, in B. anthracis, a transcriptome analysis and an affinity purification assay suggest that nprR expression is repressed by CodY (van Schaik et al., 2009; Chateau et al., 2013). To determine whether CodY binds to this putative CodY box, we performed ITC experiments. Calorimetric titrations were performed in which graduated amounts of DNA probes were sequentially injected into the sample cell containing CodY and its cofactors (GTP and isoleucine). We used two dsDNA probes. The first was a 23 bp dsDNA fragment containing the putative CodY box of the nprR promoter (codY-23) (Fig. 4A). The second, used as a control, was a 29 bp dsDNA fragment of the nprA promoter devoid of CodY box (nprA-29) (Fig. 4B). No specific interaction was detected using the non-specific nprA-29 probe while a saturation curve with a Kd value of 18 μM was obtained with the codY-23 probe. Specificity of the interaction was further investigated by an electrophoretic mobility shift assay. A 114 bp DNA fragment encompassing the nprR promoter region and centred on the putative CodY binding box showed a lowered mobility when incubated with increasing concentration of CodY protein (Fig. 4C, lanes 1–5). As control, a higher mobility of the labelled probe was restored when competitor DNA was added in the reaction (Fig. 4C, lane 6). The specificity of the CodY binding box (5′-ATTTTTTAGAAATT-3′) was assessed by introducing three mutated bp (underlined) in the motif, giving the sequence 5′-ATTTTTTATGGAATT-3′. The mobility of the 114 bp DNA fragment carrying these mutations is not affected in the presence of increasing amounts of CodY (Fig. 4C, lanes 7–10). Together, these experiments demonstrate that CodY specifically binds to the promoter region of nprR and may also directly control nprR–nprX transcription from PA.
The nprR–nprX transcription is repressed by CodY
We failed to inactivate codY in the B. thuringiensis strain 407. Therefore, we investigated the involvement of CodY in the expression of nprR–nprX using B. cereus ATCC14579 and B. cereus ATCC14579 ΔcodY strains (Lindbäck et al., 2012) carrying a PnrpR′–lacZ transcriptional fusion cloned into the plasmid pHT304-18Z. The β-galactosidase activity during growth of these strains in LB medium was determined (Fig. 4D). In the wild-type strain, transcription of the PnrpR′–lacZ fusion started at the end of exponential growth and increased during stationary phase. Identical results were obtained with a B. thuringiensis strain 407 harbouring the same plasmid transcriptional fusion (not shown). The use of plasmid transcriptional fusions (Fig. 4D) instead of chromosomal transcriptional fusions (Fig. 3C) resulted in higher levels of β-galactosidase production. The plasmid copy number (four copies per chromosome) may explain the fourfold difference of β-galactosidase production (3000 Miller units versus 12000 Miller units for the 407 wild-type strain) in these experiments. In the codY mutant, the transcriptional activity was high during exponential growth and increased during stationary phase (Fig. 4D). The genetic complementation of the codY mutant strain by a plasmid carrying codY restored the repression of nprR transcription during exponential growth. These results demonstrate that CodY has a direct negative control on nprR–nprX expression. The activation of PnrpR′–lacZ expression during stationary phase of the codY mutant strain is not due to PlcR since this regulator is not active in a codY mutant (Frenzel et al., 2012; Lindbäck et al., 2012). Therefore, these results suggest that, in addition to PlcR, nprR–nprX transcription is activated at the onset of stationary phase by another uncharacterized regulation factor.
Two additional promoters are located upstream from nprX
Our results indicate that nprR and nprX are co-transcribed from the promoter PA but did not exclude the possibility that there are other promoters upstream from nprX. The region upstream from nprX was mapped by constructing transcriptional fusions to the lacZ reporter gene in pHT304-18Z (Fig. 5A). Three different fragments (with the same 3′-end) amplified by PCR were inserted upstream from the lacZ gene to obtain pX1′Z, pX2′Z and pX3′Z. These plasmids were introduced into the B. thuringiensis strain 407 and β-galactosidase activity was assayed from t−1 to t4 in HCT medium at 37°C (Fig. 5B). The β-galactosidase appeared in bacteria harbouring pX1′Z and pX2′Z at t1 and increased sharply after t2; the β-galactosidase activity of bacteria harbouring pX3′Z was low throughout the culture. Therefore, the 120 bp region which differentiates pX2′Z from pX3′Z is necessary for promoter activity. A primer extension analysis was used to identify the transcription start site (Fig. 5C). Two transcripts were detected: one with a 5′-end at nucleotide position −101 and one, giving a very faint band, with a 5′-end at nucleotide position −114, with reference to the nprX start codon (Fig. 5D). The −35 and −10 regions upstream from the first 5′-end, GCAGGAAAT-11 bp-AAAGAAA, are similar to the consensus sequence for recognition by σH, RNAGGAWWW-(11–12 bp)-RNNGAAT (R is A or G, W is A or T and N is any base) as defined in B. subtilis (Helmann and Moran, 2002). The −35 and −10 regions upstream from the second 5′-end, GTATAAA-13 bp-AATAGAAT, resemble the consensus sequence for recognition by σE, DYMTRWW-(14 bp ± 1 pb)-CATAHAWT (D is A or G or T, H is A or C or T, M is A or C, R is A or G, W is A or T and Y is C or T) as defined in B. subtilis and B. thuringiensis (Zhang et al., 1998; Helmann and Moran, 2002).
To demonstrate the existence of these promoters, we introduced substitutions into the putative sigma factor binding sites and constructed transcriptional fusions between the modified promoter regions of nprX and the lacZ gene, carried on the pHT304-18Z (Fig. 6A and B). Substitutions in the putative σH binding site or σE binding site reduced β-galactosidase activity; substitutions in both lead to a total loss of activity suggesting a control of nprX transcription by these two alternative sigma factors. To confirm this result, the plasmid pX2′Z carrying the transcriptional fusion nprX′–lacZ was introduced into B. thuringiensis 407 ΔsigH, ΔsigE and ΔsigH ΔsigE strains (Fig. 6C). The β-galactosidase activities of the ΔsigH and ΔsigE mutant strains were about twofold lower than that of the wild-type strain, and β-galactosidase activity of the double ΔsigH ΔsigE mutant strain was completely abolished. This result is surprising since σH is essential for activation of σE in B. subtilis (Hilbert and Piggot, 2004). Thus, the expression of a σE-dependent gene should be abolished in a ΔsigH mutant as in a double ΔsigH ΔsigE mutant. To test this hypothesis we have measured the transcription of the spoIID promoter in the ΔsigH and ΔsigE mutant strains. This promoter was found to be exclusively transcribed by the σE-associated RNA polymerase in B. subtilis and B. thuringiensis (Rong et al., 1986; Bravo et al., 1996). The results show that PspoIID-directed transcription is drastically reduced but not completely abolished in the ΔsigH mutant strain (Fig. S3). This low σE-dependent transcription in a ΔsigH strain might account for the expression of nprX′–lacZ in B. thuringiensis 407 ΔsigH mutant strain (Fig. 6C).
These results indicate that, in addition to transcription from the nprR promoter, nprX was transcribed from two proximal promoters depending on σH (PH) and σE (PE). These two proximal promoters map within the nprR coding sequence. To determine whether nprX transcription is regulated by NprR, the pX2′Z was introduced into ΔRX mutant strains. The level of nprX transcription in the ΔRX strain was similar to that in the wild-type strain, indicating that the transcription of nprX from PH and PE is not auto-regulated by the NprR–NprX complex (Fig. S2B).
We show that the genes encoding the regulator and the signalling peptide of the NprR–NprX quorum sensing system of B. thuringiensis strain 407 are co-transcribed from a promoter upstream from nprR. The analysis of the promoter region strongly suggests that the sigma factor involved in transcription is σA. Transcription from this promoter (PA) started at the end of exponential growth in a sporulation-specific medium (HCT) and continued for 2−3 h during sporulation. In bacterial cells grown in LB medium, nprR–nprX expression started at the onset of stationary phase and was continued longer than in cells grown in HCT medium. The level of nprR–nprX expression in the two media was similar. This is not consistent with our previous findings that nprA transcription was fivefold higher in sporulation medium than in rich medium (Perchat et al., 2011) suggesting that NprR activity is modulated post-transcriptionally. As discussed below, the higher activity of NprR in sporulation medium may be due to a boost of nprX expression from the sporulation-specific σH and σE promoters.
The transcriptional regulators CodY and PlcR exert temporal control of nprR–nprX transcription. CodY is a global regulator which monitors the general nutritional state of the bacterial cells by sensing the intracellular pool of GTP and branched-chain amino acids (BCAA) (Sonenshein, 2007). In rich growth conditions, when GTP and BCAA are abundant, CodY is functional and acts as a repressor of stationary-phase genes and as an activator of genes involved in carbon overflow metabolism. At the onset of the stationary phase, the intracellular drop of GTP and BCAA availability inactivates CodY and thereby induces expression of the CodY-repressed genes. We report that the PnprR′–lacZ transcriptional fusion was expressed during exponential growth in B. cereus codY mutant strains. This result indicates that CodY represses nprR expression. Moreover, electrophoretic mobility shift assays reveal the presence of a CodY binding site 12 bp upstream from the potential −35 box of PA, thus suggesting that the repression by CodY is direct. This location is consistent with the observation that many CodY binding sites are upstream from the −35 box of CodY-repressed genes (Guedon et al., 2005; van Schaik et al., 2009). Binding of CodY close to the −35 box of the nprR promoter may cause steric hindrance that prevents binding of the RNA polymerase holoenzyme. Repression by CodY indicates that nprR–nprX transcription is regulated according to the nutritional state of the bacteria, and may explain why nprR expression started earlier in HCT than in LB medium. Indeed, CodY is expected to become inactive earlier in a sporulation medium than in a rich medium like LB.
The PlcR regulator controls most of the virulence factors of B. thuringiensis and B. cereus during the early stationary phase of growth (Agaisse et al., 1999; Gohar et al., 2008). The PlcR–PapR complex binds directly to DNA and activates the expression of genes encoding enterotoxins, haemolysins, phospholipases and proteases. Experiments with transcriptional fusions showed that transcription from PA in the ΔplcR mutant strain is half that in the B. thuringiensis 407 wild-type strain indicating a positive control by PlcR on nprR–nprX transcription. The direct control by PlcR is supported by ITC experiments showing that the PlcR–PapR complex binds to a PlcR box located upstream from PA. This PlcR box partially overlaps the CodY box, suggesting that there may be competition between PlcR and CodY to bind to the nprR promoter. Thus, it is likely that the increase of transcription from PA between t0 and t2 was due to a gradual substitution of CodY by PlcR. However, PlcR is not absolutely necessary for strong expression of nprA in cells grown in sporulation-specific media, in which plcR is not expressed (Perchat et al., 2011). Similarly, large amounts of the protease NprA are found in the secretome of sporulating B. anthracis cells that are naturally deficient in PlcR (Chitlaru et al., 2006). Moreover, we show (Fig. 4D) that nprR transcription is still activated at the onset of stationary phase in a B. cereus codY mutant strain in which PlcR is known to be inactive (Frenzel et al., 2012; Lindbäck et al., 2012). In view of these data, we hypothesize that, in addition to CodY and PlcR, other regulation factor(s) positively affect nprR transcription during stationary phase and sporulation.
The nprX gene is co-transcribed with nprR, but it is also independently transcribed from two promoters located in the nprR coding sequence and recognized by the alternative σH factor (PH) and the early mother-cell-specific σE factor (PE). In B. subtilis, σE is encoded by the spoIIGB gene which is expressed at the initiation of sporulation from a promoter recognized by RNA polymerase containing the housekeeping σA factor in conjunction with a high concentration of Spo0A∼P (Satola et al., 1992; Baldus et al., 1994; Schyns et al., 1997; Fujita and Losick, 2005). σH is encoded by spo0H and is expressed at the transition from exponential growth to stationary phase. spo0H is directly controlled by the transcriptional repressor AbrB and indirectly by Spo0A∼P which represses abrB. σH activity is also modulated post-transcriptionally by many factors including pH, carbon source and the stringent response (Frisby and Zuber, 1994; Cosby and Zuber, 1997; Dixon et al., 2001; Eymann et al., 2001). In B. subtilis, σH controls genes involved in many physiological processes associated with the transition to the stationary phase including sporulation, competence, cytochrome biogenesis, generation of potential nutrient sources, transport and cell wall metabolism (Britton et al., 2002). Especially, many phr genes of B. subtilis are controlled by σH in a manner analogous to nprX (McQuade et al., 2001). In sporulation medium, nprX transcription from PH and PE starts 2 h after nprR–nprX transcription from PA and allows nprX to be transcribed belatedly while nprR transcription is abolished (Figs 3D and 5B). These observations raise the issue of the physiological function of these promoters. It is possible that the stability of NprR is greater than the stability of NprX: NprR is intracellular and NprX is secreted and subjected to greater proteolysis. The σH- and σE-dependent transcription of nprX could maintain a concentration of NprX sufficient to ensure expression of NprR-regulated genes during sporulation.
In contrast to the transcription of the plcR and papR genes, nprR–nprX transcription is not under control of a positive feedback loop. A recent review (Hense et al., 2007) helps to explain the role of this positive feedback loop on quorum sensing regulation in complex environments, such as in a host, when the spatial distribution of bacteria is not homogenous. A mathematical model, using a constant number of bacteria in a finite volume and varying the spatial arrangement of bacteria and the presence or not of a positive feedback loop, shows that a quorum sensing system can be activated at low bacterial density when the bacteria are clustered if there is a positive feedback loop. In contrast, at low bacterial density, a non-auto regulated quorum sensing system will always be shut off (Hense et al., 2007). This model is consistent with our knowledge of the physiological function and the regulation of the PlcR–PapR and NprR–NprX quorum sensing systems in in vivo conditions, as determined in insect larvae as an infection model. PlcR and the PlcR-regulated genes are activated during the first steps of the infection when the overall bacterial population in the insect larva is low (Salamitou et al., 2000; Dubois et al., 2012). This early expression of the PlcR regulon requires that an appropriate threshold of bacterial density be reached to allow the activation of PlcR by its cognate signalling peptide PapR. Ingestion of B. thuringiensis spores and Cry toxins by insect larvae results in midgut paralysis (Heimpel and Angus, 1959). This toxemic effect may prevent the loss of the bacteria and allow colonization and subsequent destruction of the midgut (Raymond et al., 2010). The bacteria in contact with the epithelial cells multiply to form clusters. In this microenvironment, the bacterial density is higher and the positive feedback loop acting on papR expression results in the massive production of virulence factors involved in the destruction of the midgut tissues. Thus, the PlcR auto-activation loop may be central to the success of the infection helping bacteria to sense their environment. In contrast, NprR and the NprR-regulated genes are activated after the death of the insect larva when the overall bacterial population is high (Dubois et al., 2012). During this necrotrophic phase, the bacteria use the contents of the cadaver as a substrate to survive and to sporulate. In this context, a positive feedback loop affecting nprR–nprX expression is not necessary to activate the NprR regulon.
In this study, we characterized the transcriptional regulation of the nprR–nprX genes. We establish a model of nprR–nprX transcriptional regulation according to the developmental state of the bacteria (Fig. 7). At the beginning of the infection, Cry toxins cause the destruction of host epithelial cells (Bravo et al., 2007). Cell lysis presumably releases nutrients causing spore germination and bacterial multiplication (Raymond et al., 2010). During this step, in some ways analogous to the exponential growth phase of bacteria in laboratory conditions, transcription of the nprR–nprX genes is repressed by CodY. When bacterial density is locally high, the nutrients are depleted, the negative control of CodY on PA is released and PlcR becomes active. During this stage, corresponding to the onset of the stationary phase, the PlcR-regulated genes are expressed inducing virulence factor production and nprR–nprX transcription is initiated. This virulence stage results in the death of the host. Expression of the PlcR-regulated genes becomes inappropriate for the necrotrophic lifestyle of the bacteria. As demonstrated in infected insect larvae, the bacterial density in the host cadaver is high (Dubois et al., 2012). Thus, the bacteria are rapidly faced with nutrient stress inducing a significant increase of Spo0A∼P. As a consequence, plcR transcription is repressed and the transcriptional activity of PA decreases. Simultaneously, the repression of spo0H transcription by AbrB is turned off inducing transcription of nprX from PH. At the end of the necrotrophic stage, which might be regarded as a late stationary phase, the high level of Spo0A∼P activates transcription of spoIIGB encoding σE. Transcription from PA is stopped and nprX is then mainly transcribed from PE.
This model highlights the key role played by CodY, PlcR and Spo0A∼P in the regulation of nprR–nprX transcription. These regulators allow the integration of multiple and diverse environmental signals such as nutritional and metabolic stress, DNA status, cell density and cell cycle signals (Phillips and Strauch, 2002; Slamti and Lereclus, 2002; Sonenshein, 2005; 2007). They contribute, with the quorum sensor NprR, to trigger the transition between the pathogenic and the necrotrophic state of B. thuringiensis. Our study provides a better understanding of the integration of environmental signals by quorum sensors and of strategies used by bacterial cells to survive and develop within their host environment.
Bacterial strains and growth conditions
The B. thuringiensis strain 407 Cry− (hereafter referred to as B. thuringiensis strain 407 or B. thuringiensis 407) is an acrystalliferous strain cured of its cry plasmid (Lereclus et al., 1989). This strain is phylogenically similar to B. cereus (Kolstø et al., 2009) and was used throughout this study as a model for the B. cereus group. The Bacillus strains used in this study are described in Table 1. Escherichia coli K-12 strain TG1 (Gibson et al., 1984) was used as host for the construction of plasmids and cloning experiments. Plasmid DNA for Bacillus electroporation was prepared from the Dam− Dcm−E. coli strain ET12567 (Stratagene, La Jolla, CA, USA). E. coli and Bacillus cells were transformed by electroporation as described previously (Dower et al., 1988; Lereclus et al., 1989). E. coli strains were grown in Luria Broth (LB) and Bacillus strains were grown in LB or in HCT, a sporulation-specific medium (Lecadet et al., 1980). The following antibiotic concentrations were used for bacterial selection: 100 μg ml−1 ampicillin and 50 μg ml−1 spectinomycin (for E. coli), 200 μg ml−1 kanamycin, 200 μg ml−1 spectinomycin and 10 μg ml−1 erythromycin (for Bacillus strains). Bacteria with the Lac+ phenotype were identified on LB plates containing 100 μg ml−1 X-gal. The xylA promoter in Bacillus was induced by adding xylose (20 mM final concentration) to the culture medium.
Table 1. Bacillus strains used in this study
Trait or relevant genotype
B. thuringiensis 407
Bacillus thuringiensis Cry− strain cured of its cry plasmid
CodY-deficient strain complemented by the codY gene of B. thruringiensis strain 407 (see Table S2)
Chromosomal DNA was extracted from Bacillus cells using the Puregene DNA Purification Kit (QIAgen, France). Plasmid DNA was extracted from E. coli by a standard alkaline lysis procedure using QIAprep spin columns (QIAgen, France). Restriction enzymes and T4 DNA ligase (New England Biolabs, USA) were used in accordance with the manufacturer's recommendations. Oligonucleotide primers (Table S1) were synthesized by Sigma-Proligo (Paris, France). PCRs were performed in an Applied Biosystem 2720 Thermak cycler (Applied Biosystem, USA). Amplified fragments were purified using the QIAquick PCR purification Kit (QIAgen, France). Digested DNA fragments were separated on 1% agarose gels after digestion and extracted from gels using the QIAquick gel extraction Kit (QIAgen, France). Nucleotide sequences were determined by Beckman Coulter Genomics (Takeley, UK)
The plasmid pRN5101 (Lereclus et al., 1992) was used for homologous recombination in the B. thuringiensis strain 407. The plasmid pHT304 (Arantes and Lereclus, 1991) and the plasmid pHT1618KΩPxyl (Perchat et al., 2011) were used for genetic complementation experiments. Transcriptional fusions were constructed in pHT304-18Z (Agaisse and Lereclus, 1994). All the recombinant plasmids used in this study are described in Table S2.
Point mutations were introduced into the σH and σE binding sites of the nprX promoter region by PCR amplification using degenerate primers. Plasmids carrying these mutated promoters (pX2H′Z, pX2E′Z and pX2HE′Z) were described in Table S2.
Construction of the B. thuringiensis 407 recombinant strains
The plasmid pRN5101ΩnprR::lacZ (Table S2) was introduced into B. thuringiensis 407 wild-type and Δspo0A mutant strains by electroporation. The chromosomal wild-type copy of nprR–nprX genes were replaced with the disrupted copy by homologous recombination as described previously (Lereclus et al., 1992). In the resulting B. thuringiensis 407 recombinant strains, the lacZ gene is transcribed from the nprR promoter: they were designated B. thuringiensis 407 nprR′lacZ and B. thuringiensis 407 Δspo0A nprR′lacZ, and are Lac+ and sensitive to erythromycin.
To disrupt the sigH and plcR genes, the plasmids pRNΩsigH::spc and pRNΩplcR::spc were constructed (Table S2). pRNΩsigH::spc was used to transform the strains B. thuringiensis 407 and B. thuringiensis 407 ΔsigE to obtain strains B. thuringiensis 407 ΔsigE and B. thuringiensis 407 ΔsigE ΔsigH respectively. pRNΩplcR::spc was used to transform strains B. thuringiensis 407 nprR′lacZ and B. thuringiensis 407 Δspo0A nprR′lacZ to obtain the strains B. thuringiensis 407 ΔplcR nprR′lacZ and B. thuringiensis 407 ΔplcR Δspo0A nprR′lacZ.
For β-galactosidase activity measurements, B. thuringiensis 407 cells containing lacZ transcriptional fusions were cultured in LB medium or HCT medium at 37°C. β-Galactosidase activities were measured as described previously (Perchat et al., 2011), and specific activities are expressed in units of β-galactosidase per milligram of protein (Miller units). Each assay was repeated at least three times.
Overproduction and purification of PlcR
The recombinant plasmid pET28.16ΩPlcR was used for producing PlcR (Slamti and Lereclus, 2002). The recombinant plasmid pET28.16ΩPlcR was introduced into E. coli strain C41 (DE3) (Invitrogen) and grown at 30°C in LB medium supplemented with 100 μg ml−1 ampicillin to an OD600 of 0.6. Expression of plcR was then induced by addition of isopropyl-d-thiogalactopyranoside (IPTG) to 0.5 mM and the cells were grown for a further 4 h at 30°C. Cultures were then harvested by centrifugation at 8750 g for 15 min (JLA-9.1000, Beckman), flash frozen in liquid nitrogen and kept at −80°C until further processing. The pellet was lysed by sonication (Misonix) in 25 mM Tris-HCl pH 8.0, 0.3 M NaCl, 20 mM imidazole, 5 mM β-mercaptoethanol, 30% glycerol after addition of a protease inhibitor cocktail (SIGMA). Benzonase (SIGMA) was added to the lysate. Cell debris were removed by centrifugation at 35 000 g for 30 min (JA-25.50, Beckman) and the cleared lysate was filtered through a 0.45 mm device (Sartorius). The supernatant was then loaded onto a 5 ml HiTrap chelating nickel column (Akta FPLC system, GE Healthcare) previously equilibrated with buffer A (25 mM Tris-HCl pH 8.0, 0.3 M NaCl, 20 mM imidazole, 5 mM β-mercaptoethanol, 10% glycerol). The protein was extensively washed with buffer A, and remaining contaminants were removed with buffer A supplemented with 10% of buffer B (25 mM Tris-HCl pH 8.0, 0.3 M NaCl, 0.3 M imidazole, 5 mM β-mercaptoethanol, 10% glycerol). PlcR was eluted in 100% of buffer B and loaded onto a HiLoad 26/60 Superdex 200 prep-grade gel-filtration column (GE Healthcare) pre-equilibrated in buffer E (20 mM Tris-HCl pH 8.0, 150 mM NaCl and 5 mM β-mercaptoethanol). The purified protein (> 98%) was typically concentrated to ∼ 2 mg ml−1 in elution buffer using Vivaspin ultrafiltration systems with a cut-off of 30 kDa. The concentrated sample was aliquoted by 1 ml, flash-frozen in liquid nitrogen and stored at −80°C. Storage of PlcR at higher concentrations promoted protein precipitation.
Overproduction and purification of CodY
The B. thuringiensis 407 codY gene was PCR-amplified from genomic DNA using primers codY-1/codY-2 and cloned in the plasmid pET28a between NdeI and XhoI restriction sites generating a fusion with an N-terminal hexa-histidine tag. The recombinant plasmid pET28-codY was introduced into E. coli strain BL21 (DE3) (Invitrogen) and grown at 37°C in LB medium supplemented with 25 μg ml−1 kanamycin to an OD600 of 0.6. Expression of codY was then induced by addition of IPTG to 1 mM and growth was continued for 20 h at 20°C. Culture was then harvested by centrifugation at 5000 g for 10 min (JA-14, Beckman), washed in cold lysis buffer (CodY-buffer A: 50 mM Tris-HCl pH 8.0, 250 mM NaCl, 5 mM imidazole, 7.5% glycerol, 1 mM PMSF) and kept at −20°C until further processing. The pellet was resuspended in 5 ml of CodY-buffer A supplemented with 100 μl of lysozyme (50 mg ml−1) and incubated on ice for 1 h. Cells were then sonicated and DNase I (SIGMA) was added to the lysate. Cell debris were removed by centrifugation at 20 000 g for 30 min (JA-17, Beckman). The supernatant was then mixed with 1 ml of Ni-NTA agarose resin (Qiagen) previously equilibrated with CodY buffer A and incubated for 2 h at 4°C. The resin with bound protein was poured into a column, and successively washed with 4 column volumes of CodY-buffer A containing 25 mM and 50 mM imidazole. CodY was eluted with the same buffer containing 250 mM and 500 mM imidazole. Purity was checked by SDS/PAGE followed by Coomassie Blue staining. Fractions containing CodY were pooled and loaded onto a HiLoad 26/60 Superdex 200 prep-grade gel-filtration column (GE Healthcare) pre-equilibrated in CodY-buffer B (50 mM Tris–HCl pH 8.0, 250 mM NaCl and 5% glycerol). The purified protein was typically concentrated to ∼ 2.9 mg ml−1 in buffer using Vivaspin ultrafiltration systems with a cut-off of 10 kDa. The concentrated sample was aliquoted by 1 ml, flash-frozen in liquid nitrogen and stored at −20°C.
Preparation of the peptide PapR7 and DNA probes
High-grade purified heptapeptide PapR7 chemically synthesized and analysed by MS and HPLC (GenScript USA) was resuspended in protein buffer E to a final concentration of 50 mM. It was flash-frozen in liquid nitrogen and stored at −20°C. The nprR-25F, nprA-22F, codY-23F and nprA-29F oligonucleotides chemically synthesized and purified by HPLC (Eurofins MWG Operon) were resuspended in protein buffer E (nprR-25F and nprA-22F) or CodY-buffer B (codY-23F and nprA-29F) to a concentration of 200 μM (nprA-29F) or 500 μM (nprR-25F, nprA-22F and codY-23F). Complementary strands (nprR-25R, nprA-22R, nprA-29R and codY-23R) were then equimolarly mixed and subsequently hybridized at 95°C under constant shaking at 300 r.p.m. on a Thermomixer R (Eppendorf France S.A.S.) for 40 min. After overnight cooling at room temperature, the dsDNA samples were flash-frozen in liquid nitrogen and stored at −20°C.
ITC experiments were performed at 20°C with an ITC200 isothermal titration calorimeter (Microcal). To assay the interaction between PlcR and the nprR promoter region, an equimolar PlcR–PapR7 complex was prepared by mixing both samples in buffer E at a final concentration of 40 μM in the microcalorimeter cell. The interaction between the PlcR–PapR7 complex and the control probe (nprA-22) was tested in the same condition. To assay the interaction between CodY and the nprR promoter region, the CodY protein was prepared in the microcalorimeter cell at a final concentration of 50 μM in CodY-buffer B with GTP (2 mM) and isoleucine (2 mM). The interaction between CodY and the control probe (nprA-29) was tested in the same buffer with CodY at 20 μM, GTP at 1 mM and isoleucine at 1 mM. A total of 20 injections of 2 μl of dsDNA at 500 μM (nprR-25, nprA-22 and codY-23) or 200 μM (nprA-29) were made at intervals of 180 s while stirring at 1000 r.p.m. The data were integrated to generate curves in which the areas under the injection peaks were plotted against the ratio of injected sample to cell content. Analysis of the data was performed according to the one-binding-site model using the MicroCal Origin software provided by the manufacturer.
Gel mobility shift assay
A 114 bp fragment corresponding to the nprR regulatory region (positions −116 to −3) was amplified by PCR using 5′ biotinylated nprR-114F and nprR-114R, or nprR-114R* as primers pair, and purified from a 1% agarose gel. The primer nprR-114R* was designed in order to replace three bases of the putative CodY binding site giving 5′-ATTTTTTATGGAATT-3′ instead of 5′-ATTTTTTAGAAAATT-3′. Binding reactions were carried out with 5 fM labelled probe and increasing amounts of CodY in 20 μl of reaction buffer containing 20 mM Tris-HCl pH 8.0, 50 mM sodium glutamate, 5 mM EDTA, 10 mM MgCl2, 0.05% (v/v) NP40, 5% (v/v) glycerol, 250 ng of salmon sperm DNA, 2 mM GTP and 10 mM of a mix of the three branched-chain amino acids, at room temperature for 30 min. For the competition reaction, protein was first incubated in presence of 200-fold molar excess of competitor DNA (unlabelled specific DNA) followed by incubation with the labelled probe. The bound product was electrophoresed on a non-denaturing 8% Tris/Borate/EDTA polyacrylamide gel and run in a 35 mM HEPES–43 mM Imidazol buffer (pH 7.4) on ice. The gel was electroblotted onto a nylon membrane in the same buffer. DNA was then UV-cross-linked and the DNA migration point was detected using the LightShift Chemiluminescent EMSA kit (Pierce).
Bacillus thuringiensis strain 407 was grown with shaking in LB or HCT medium at 37°C until t2. RNA extraction and primer extension were performed as described previously (Cadot et al., 2010). To detect the transcription start site of the nprR and nprX genes, primer extension was performed using the oligonucleotides Npr-P and NprX-P2 complementary to the 5′-ends of these genes respectively.
RNA was reverse transcribed into cDNA with SuperScriptIII (Invitrogen, France) or with a Stratascript QPCR cDNA Synthesis kit (Stratagene, USA) as described in the manufacturer's instructions. PCR amplifications were performed with cDNA as the template and the primers listed in Table S1. The absence of contaminating DNA in RNA preparations and the specificity of primers were confirmed by PCR using RNA and DNA as the template respectively.
We thank Dr A.T. Kovács for the gift of the B. cereus ATCC14579 wild-type and ΔcodY mutant strains. We are grateful to Gilles Vergnaud and the Délégation Générale de l'Armement for providing a PhD grant to T. Dubois. This work was supported by the French Agence Nationale de la Recherche (Cell.com project; N°ANR-09-Blan-0253).