The transcriptional activator ManR of the Bacillus subtilis mannose utilization operon is composed of an N-terminal DNA-binding domain, two phosphotransferase system (PTS) regulation domains (PRDs), an EIIBBgl- and an EIIAFru-like domain. Site-specific mutagenesis of ManR revealed the role of conserved amino acids representing potential phosphorylation sites. This was investigated by β-galactosidase activity tests and by mobility shift assays after incubation with the PTS components HPr and EI. In analogy to other PRD-containing regulators we propose stimulation of ManR activity by phosphorylation. Mutations in PRD1 lowered ManR activity, whereas mutations in PRD2 abolished ManR activity completely. The Cys415Ala (EIIBBgl) and the His570Ala mutations (EIIAFru) provoked constitutive activities to different degrees, whereas the latter had the greater influence. Addition of EIIBAMan reduced the binding capability significantly in a wild-type and a Cys415Ala background, but had no effect on a His570Ala mutant. The different expression levels originating from the two promoters PmanR and PmanP could be ascribed to different 5′-untranslated mRNA regions. Sequences of 44 bp were identified and confirmed as the ManR binding sites by DNase I footprinting. The binding properties of ManR, in particular the equilibrium dissociation constant KD and the dissociation rate kdiss, were determined for both promoter regions.
The mannose utilization operon of Bacillus subtilis consists of manP, encoding the phosphoenolpyruvate (PEP):carbohydrate phosphotransferase system (PTS) mannose-specific transporter, manA, encoding a mannose 6-phosphate isomerase, and a distal gene, yjdF, which might play a role in formation of a structured, non-coding RNA of unknown function (Kunst et al., 1997; Weinberg et al., 2010). The manPA-yjdF operon is regulated by the transcriptional activator ManR (Sun and Altenbuchner, 2010). The corresponding manR gene is located immediately upstream and in the same orientation as the manPA-yjdF operon and is regulated by ManR itself (Reizer et al., 1999; Sun and Altenbuchner, 2010). Induction of the manPA-yjdF operon promoter, PmanP, with mannose revealed a fourfold higher expression rate compared with the manR promoter, PmanR (Sun and Altenbuchner, 2010).
ManR belongs to a class of regulators that are modulated in their activities by reversible phosphorylation events mediated by PTS components (Stülke et al., 1998; Greenberg et al., 2002; Sun and Altenbuchner, 2010). Those kinds of regulators contain two PTS regulation domains (PRDs) with conserved histidines as the target for phosphorylation and may act as antiterminators or as transcriptional activators. Antiterminators consist of an RNA-binding domain and two PRD domains. Some well-studied examples in B. subtilis are LicT, GlcT, SacT and SacY (Steinmetz et al., 1989; Débarbouillé et al., 1990; Krüger and Hecker, 1995; Stülke et al., 1997). The PRD-containing transcriptional activators are usually more complex. So far, four of them were identified and characterized in B. subtilis. These are LevR, LicR, MtlR and ManR (Débarbouillé et al., 1991; Tobisch et al., 1997; Joyet et al., 2010; Sun and Altenbuchner, 2010). ManR has a sequence length of 648 amino acids, a molecular weight of about 75 kDa and a theoretical pI of 8.15. It is structurally composed of several domains: an N-terminal DNA-binding region composed of two potential helix–turn–helix motifs, followed by duplicated regulatory PRDs, namely PRD1 and PRD2, and regulatory EIIBBgl-, EIIAFru-like domains.
The structural similarities of ManR with PRD-containing activators like MtlR of Geobacillus stearothermophilus (Henstra et al., 1999) or LicR and MtlR of B. subtilis (Tobisch et al., 1997; Joyet et al., 2010), indicate that its activity is most probably also modulated via phosphorylation of conserved amino acids in the PRDs, EIIB- and EIIA-like domains. Alignments of the PRDs and EIIBA-like sequences showed that ManR contains two conserved histidines in PRD1 and PRD2 each (His-222/His-281 and His-334/His-393 respectively), a conserved cysteine in EIIBBgl (Cys-415) and a conserved histidine in EIIAFru (His-570), which represent potential phosphorylation sites (Greenberg et al., 2002). Although there seems to be no general rule, most often PRD-containing antiterminators are active when PRD1 is dephosphorylated and PRD2 is phosphorylated, which is the case when substrate is available and glucose is absent. It could be shown for several examples that the corresponding PTS transporters and the general PTS protein HPr play major roles in phosphorylation and dephosphorylation of the PRDs in vivo (reviewed by Deutscher et al., 2006). For the PRD-containing activators, the situation is less clear since the EIIBA-like domains can also play regulatory roles. For example, activation of MtlR in B. subtilis seems to need the PRD2 domain phosphorylated and the EIIBA domains dephosphorylated, whereas the PRD1 has no obvious effect in regulation (Joyet et al., 2010).
Recently, it could be shown for MtlR of B. subtilis that also the EIIB domain of the cognate PTS transporter is essential for MtlR activation (Bouraoui et al., 2013). Deletion of the EIIBMtl domain prevented constitutive expression of the mtlAFD operon in a mtlA (EIIA) mutant. The authors postulated a membrane sequestration mechanism for MtlR in order to render it functional. As we reported previously, ManR is inactivated by the cognate EIIBCAMan transporter, ManP, when mannose is not available (Wenzel et al., 2011). In agreement with this finding, deletion of manP led to constitutive expression from both promoters, PmanR and PmanP, when glucose is absent (Sun and Altenbuchner, 2010). The negative regulatory effects of the transporter on ManR activity are most probably due to phosphorylation of the ManR EIIBA-like domains and a membrane sequestration mechanism for ManR seems not obvious.
So far, it was shown that ManR activates transcription from PmanR and PmanP (Sun and Altenbuchner, 2010); however, the mechanism of activation could not be clarified. Three general mechanisms have been described depending on the location of the regulator's binding site relative to the promoter elements: Class I activation, Class II activation and activation by conformational changes (reviewed by Browning and Busby, 2004). In Class I activation, regulator proteins bind to target sequences located upstream of the promoter −35 element. Activation occurs after recruitment of RNA polymerase, more precisely by interactions with the α-C-terminal domain. The best-known example is the cyclic AMP receptor protein, CRP, acting at the lac promoter (Ebright, 1993). In Class II activation, the target sequence of the regulator overlaps, or is adjacent to, the promoter −35 element. Activation occurs after recruitment of the RNA polymerase by interaction with domain 4 of the σ subunit. A well-studied example is the bacteriophage λ cI protein acting on PRM, which is one of the promoters located in the rightward regulatory region of bacteriophage λ (Nickels et al., 2002). In the third mechanism for activation, the regulator binds at, or near to, the promoter elements, and reorientates them, so that RNA polymerase can bind to the promoter. Members of the MerR-type regulator family act by this mechanism (Brown et al., 2003).
In this study, we compared the two promoter regions PmanR and PmanP and localized the operator sequences by DNase I footprinting experiments. The different expression levels obtained in transcriptional fusions to lacZ could be ascribed to the different 5′-untranslated mRNA sequences as shown by the exchange of promoter elements. We further studied the influence on ManR activity by the PTS components HPr, EI and EIIBAMan. Site-specific mutagenesis revealed the regulatory roles of the potential phosphorylation sites of ManR. Finally, the equilibrium dissociation constant KD and dissociation rate kdiss describing the binding properties of ManR to its operator sequence were determined.
Sequence-based analysis and modifications of the promoter regions
Mapping of the transcriptional start sites and subsequent sequence analysis of PmanR and PmanP by Sun and Altenbuchner (2010) revealed inverted repeats upstream of the −35 promoter elements of the PmanR and PmanP promoters. A sequence alignment of both promoter regions revealed a 44 bp region from position −75 to −32 relative to the transcription start sites with 59% sequence identity (Fig. 1). Through this comparison new inverted repeats were identified which extend beyond the previously identified ones and overlap with the potential −35 promoter elements. In the PmanR region an inverted repeat with an AAAAN6TTTT motif was identified. Similar inverted repeats with an AAAATN2CTTTWTT motif were found in the PmanP region. Interestingly, the 14 bp motif AAAATCGCTTTTTT could be found in the PmanR as well as in reverse orientation in the PmanP region. As both are located within the identified 44 bp region, we propose this region as the possible binding site of ManR and the 14 bp motifs building the boundaries thereof.
The low-copy lacZ-expression vector pSUN290 contains the manP promoter beginning with position −80 and shows mannose-dependent β-galactosidase activity (Sun and Altenbuchner, 2010). Deletion analysis starting from this vector revealed the importance of the putative operator sequence (Fig. 2A). First, 5 bp of the manP upstream sequence until position −75 was deleted. The resulting vector, designated as pMW712.2, was brought into B. subtilis 3NA. No significant effect of the deletion was observed in the subsequent β-galactosidase assay on the expression level compared with pSUN290 (Fig. 2B). Previously made deletions down to position −67 (pSUN380.1) reduced expression of lacZ to nearly 50% and deletions extending to position −53 (pSUN381.1) abolished it completely (Sun and Altenbuchner, 2010). These results are in accordance with the assumed operator sequence.
Location of the ManR binding sites
As we reported previously (Wenzel et al., 2011), formation of functional ManR is possible using the mannose transporter negative strain B. subtilis TQ356 under glucose limiting conditions. After putting manR under control of the strong PmanP in a high-copy expression vector, soluble ManR was obtained with B. subtilis TQ356 in LB liquid medium. The active state of ManR molecules in crude cell extracts of B. subtilis TQ356/pMW687.1 (overexpressed ManR) was shown by electrophoretic mobility shift assays (EMSAs) resulting in a clear band shift concerning the PmanR and PmanP regions (Fig. 3). As a negative control crude cell extracts from B. subtilis TQ356/pMW718.1 (no overexpressed ManR) were used resulting in no band shift. The intensity of the shifted band was strongly reduced after addition of non-labelled promoter fragment demonstrating the specificity of the binding.
For reliable location of the DNA binding sites of ManR by DNase I footprinting assays purified and active ManR molecules were needed. Therefore, ManR was overexpressed in soluble form with a C-terminal 6xHis- (pMW735.3) or Strep(II)-tag (pMW725.1) using B. subtilis TQ356 cells (Fig. S1). However, purification failed with both tags. Western blots did not show ManR-Strep(II)-tag bands (data not shown). Thus, B. subtilis ManR-Strep(II)-tag was overexpressed with Escherichia coli. Usage of purified protein in EMSAs showed that only a small fraction is active resulting in an incomplete band shift (Fig. 4, lane 2). The PTS components EI and HPr play major roles in phosphorylation and activation of PRD-containing regulators (reviewed by Deutscher et al., 2006). In order to render ManR into an active state, purified B. subtilis PTS components HPr, EI and PEP were added to ManR resulting in a complete band shift (Fig. 4, lane 10). Using HPr and EI, alone or in combination with and without PEP, no band shifts were observed (Fig. 4, lanes 3–6). Interestingly, when HPr and ManR were used without PEP, no or weaker band shifts were observed (Fig. 4, lanes 7 and 9). On the other hand, when EI and ManR were mixed together without PEP (Fig. 4, lane 8), the moderate binding activity was still observed. This indicates that HPr may also play a role in inactivation of ManR.
Increasing amounts of purified and activated ManR from E. coli were used to reveal the regions protected in the coding and the non-coding strands by ManR of each promoter region by DNase I footprinting experiments (Fig. 5). Additionally, crude cell extracts of B. subtilis TQ356/pMW687.1 were used to confirm the protection pattern (Fig. S3). The results of the DNase I footprinting concerning the protected regions and hypersensitive sites to DNase I restriction are summarized in Table 1.
Table 1. Results of the DNase I footprinting experiments concerning the protected region and hypersensitive sites to DNase I restriction on both strands of each promoter region
Results with purified and activated ManR are compared with ManR protection patterns observed with B. subtilis TQ356/pMW687.1 crude cell extracts comprising active ManR molecules.
With purified and activated ManR
from −79 to −49
From −77 to −26
from −79 to −31
From −78 to −29
With B. subtilis crude cell extracts comprising active ManR
from −80 to −50
From −77 to −29
from −78 to −30
From −78 to −30
Conserved histidines in PRD2 are mandatory for ManR activation
It is not known so far, which of the potential phosphorylation sites of ManR, His-222 and His-281 in PRD1, His-334 and His-393 in PRD2, Cys-415 in the EIIB-like domain, and His-570 in the EIIA-like domain, are important for its activation. By separate replacement of each potential phosphorylation site of ManR by alanine, the importance of these sites was examined. Wild-type and mutant ManR proteins were incubated with EI, HPr and PEP and checked for their DNA-binding capability in EMSA experiments (Fig. 6). Whereas the ManR-C415A and ManR-H570A mutants behaved like the wild-type protein, reduced binding could be observed with the PRD1 mutants H222A and H281A. Replacement of the putative PRD2 phosphorylation sites H334 and H393 with Ala resulted in a complete loss of the binding capability. These findings indicate that the PRD2 is mandatory for ManR activation by the PTS components HPr, EI and PEP. The conserved amino acids in the EIIBA-like domains seem not to play a role in ManR activation.
In vivo analysis of ManR mutants
Quantitative β-galactosidase activity assays were performed for further evaluation of the potential phosphorylation sites. For this purpose, ManR double mutants H222A/H281A, H334A/H393A and C415A/H570A were also analysed. B. subtilis strains carrying wild-type and mutant manR genes were analysed under non-inducing (no sugar), inducing (mannose) and catabolite repression conditions (mannose and glucose) (Fig. 7A). The basal β-galactosidase activity obtained with ManR-C415A reached about 29% of the value under inducing conditions, whereas the wild-type protein reached only 16%. Also carbon catabolite repression by glucose had a stronger effect compared with the wild-type. The basal activity of the ManR-H570A mutant was much more increased under non-inducing conditions and reached about 64% of the value under inducing conditions. The ManR-C415A/H570A double mutant had lost most of its activity but exhibited a constitutive phenotype, indicated by a basal level that reached 82% of the induced activity. Regarding the PRD mutants, only ManR-H222A showed β-galactosidase activity, albeit very low.
Since ManR also stimulates its own expression (Sun and Altenbuchner, 2010), a disadvantageous mutation may also reduce its intracellular level especially having only one chromosomal copy of the gene. To improve this situation manR alleles were alternatively delivered on low-copy expression vectors. Again, the β-galactosidase activity was analysed under non-inducing, inducing and repressing conditions (Fig. 7B). Whereas the activities of the wild-type ManR and the ManR-C415A mutant were comparable with the findings achieved with single copies of manR, the ManR PRD1 mutants showed much higher activities. Again the PRD2 mutants were not active at all. The ManR-H570A mutant and the ManR-C415A/H570A double mutant displayed constitutive behaviour. The basal β-galactosidase activities of these mutants were in the order of 70% and 97% of the activities obtained with induction by mannose respectively. The basal activity obtained with ManR-C415A was elevated to 37% compared with 19% obtained with the wild-type protein. Roughly, the results of the in vivo experiments concerning ManR activities were in accordance with the results of the EMSA experiments.
ManP-EIIB negatively stimulates ManR activity at His-570
Deletion of the mannose PTS transporter gene manP led to constitutive expression from PmanR and PmanP (Sun and Altenbuchner, 2010; Wenzel et al., 2011). In this study, it was shown that replacement of ManR-His-570 by alanine led to a nearly constitutive activity. In contrast, replacement of Cys-415 in the ManR EIIB-like domain by alanine only resulted in weak constitutive ManR activity. The ManR-C415A/H570A double mutant showed low, but fully constitutive activity. Recently, it was shown that the equivalent Cys-419 of MtlR of B. subtilis plays a major role in MtlR activity (Joyet et al., 2010). To confirm if ManR-Cys-415 and/or ManR-His-570 also play regulatory roles concerning ManR activity, in vitro activation experiments and subsequent EMSAs were performed using wild-type (wt) and mutant ManR proteins. Additionally, the soluble domains of the PTS transporter ManP, EIIAMan and EIIBMan, were overexpressed, purified and used to show the influence on ManR activation. Figure 8 shows the results of the performed EMSAs. When EIIAMan and EIIBMan were present together in the assay, ManR-wt binding activity became significantly reduced (Fig. 8, lane 5). No influence on the ManR-wt DNA-binding activity was observed when EIIAMan or EIIBMan were used separately (Fig. 8A, lanes 3 and 4). The ManR-C415A mutant behaved in a similar way as ManR-wt. In contrast, using the ManR-H570A mutant no reduction concerning the DNA-binding activity could be observed (Fig. 8A, lane 15).
According to amino acid sequence alignments, Cys-9 of EIIBMan is putatively responsible for phosphoryl group transfer. With wild-type EIIBMan a significant decrease in the binding activity of ManR was observed (Fig. 8B, lane 6), whereas ManR did not show any decrease in its binding activity when EIIBMan-C9A was used (Fig. 8B, lane 7).
DNA-binding properties of ManR
Mobility shift experiments were used to estimate the equilibrium dissociation constant KD and the dissociation rate kdiss of ManR to its DNA binding sites. The concentration of ManR that is necessary to shift 50% of the labelled DNA fragments is expressed by the KD value. A shift of 50% of the DNA fragment containing the manR promoter region occurred with 51.7 ± 6.2 ng of ManR (Fig. 9A); with the manP fragment, 56.7 ± 2.4 ng of ManR was needed (not shown). The corresponding KD values are 68.7 ± 8.2 nM for the manR promoter region and 75.3 ± 3.2 nM for the manP promoter region. The same experiment was conducted using crude cell extracts of B. subtilis TQ356/pMW687.1 leading to similar band shift patterns (data not shown).
ManR–DNA complexes were formed and equilibrated to determine the dissociation rates kdiss and the corresponding half-lives using a binding competition assay. For the binding site in the manR promoter fragment, a kdiss of (193 ± 9.5) × 10−6 s−1 and a half-life of 59.9 ± 3 min was obtained (Fig. 9B). Accordingly, for the binding site in the manP promoter fragment, a kdiss of (226 ± 8.2) × 10−6 s−1 and a half-life of 51.3 ± 1.9 min was obtained (not shown). The same experiment was conducted using crude cell extracts of B. subtilis TQ356/pMW687.1 leading to similar band shift patterns (data not shown).
Differences in the 5′-untranslated regions of manR and manP mRNAs are the reason for the different expression levels of the lacZ reporter gene
Another issue to be clarified was the lower expression level of PmanR compared with PmanP, reported by Sun and Altenbuchner (2010). Considering the nucleotide sequence (Fig. 1), PmanR exhibits the optimal σA consensus −10 promoter element with 5′-TATAAT-3′, whereas PmanP shows two mismatches with 5′-TACAGT-3′. The potential −35 promoter elements (‘−35’) are very similar in both promoters and do not fit with the σA consensus sequence 5′-TTGACA-3′ at all. The unexpected lower expression level from PmanR might be caused by either the operator sequence, the spacer region between −35 and −10 or the translational leader region from position +1 to the ATG start codon.
Since the low-copy PmanR–lacZ-expression vector pSUN291 was not suitable for the required cloning steps, pMW692.1 was constructed from pSUN291, which has additional restriction sites and lacks the cre site (catabolite repression element). This new vector was used to exchange the operator sequence (pMW720.1), the spacer region between the −35 and −10 elements (pMW711.2) and the translational leader sequence (pMW722.4) with that of the corresponding PmanP–lacZ expression vector pSUN290. The relevant sequences of the constructs are compared in Table 2. The β-galactosidase activities of B. subtilis 3NA carrying these vectors were measured under non-inducing (no additive), inducing (mannose) and catabolite repression conditions (mannose and glucose) (Fig. 10). Compared with pSUN291, vector pMW692.1 showed only a twofold higher expression level under catabolite repression conditions, presumably due to the missing cre site. The first replacement concerning the operator sequence (pMW720.1) had no effect. Replacement of the spacer region (pMW711.2) led to a slight decrease of the expression level. In contrast, replacement of the translational leader region (pMW722.4) increased induction of the modified PmanR to the same level as that found with PmanP (pSUN290). However, the basal activity (non-induced) was about threefold higher whereas expression under catabolite repression conditions was nearly at the same level. Obviously, the two promoters have about the same promoter strength.
Table 2. Relevant sequences starting from position −80 to ATG start codon of the lacZ expression vectors concerning the replacement experiments
Relevant sequence from −80 to +1 (5′ → 3′ direction)
The putative operator sequences, −10 promoter elements, transcriptional start sites, ribosomal binding sites, and ATG start codons are emphasized in bold face. Inserted restriction sites are indicated in italics (XbaI, BglII, EcoRV and XhoI).
Sequence from −80 to +1
AGGGA AAAATGCCTTTATTACCGGAACCTATGGTAAAAAAAGCGATTTT AATGAGCTGATTTCGGTA TACAGT TGAGACA A
GTATA AAAATCGCTTTTTTCCGGAAGCTTCGGTAAAAAACGAAACTTTT GTCTCTATGATTTTGTTT TATAAT GTAAACG G
GGTTA AAAATCGCTTTTTTCCGGAAGCTTCGGTAAAAAACGAAACTTTT GTCTAGATCTTTTTGTTT TATAAT GTTATAG G
GGTGA AAAATGCCTTTATTACCGGAACCTATGGTAAAAAAAGCGATTTT AATGAGATCTTTTTGTTT TATAAT GTTATAG G
GGTTA AAAATCGCTTTTTTCCGGAAGCTTCGGTAAAAAACGAAACTTTT GATGAGCTGATTTCGGTA TATAAT GTTATAG G
GGTTA AAAATCGCTTTTTTACCGGAAGCTTCGGTAAAAAAAGAAATTTT GTCTAGATCTTTTTGTTT TATAAT GTTATAG A
The location of the operators confirmed by DNase I footprinting experiments seem to overlap the corresponding promoter −35 elements. This let us assume a class II activation mechanism for ManR, i.e. it may recruit the RNA polymerase by interaction with domain 4 of the σ subunit. ManR bound the PmanR and PmanP operators in a different way. In the coding and non-coding strand of the PmanP, as well as in the non-coding strand of PmanR, a region of about 50 bp was protected by ManR, whereas in the coding strand of PmanR the right half of this sequence was unprotected. This was surprising since the affinity and half-life of ManR bound to PmanR and PmanP were similar. Assuming that ManR binds as a dimer to the operators, the two monomers could have different affinities to the corresponding half-sites due to differences in the nucleotide sequence. A similar case for such a half protection of only one of the strands by a transcriptional regulator was reported for CsdR of Pseudoalteromonas piscicida (Miyamoto et al., 2007). Different affinities to the half-sites of an operator or asymmetric binding have often been observed with multimeric regulators, like the tetrameric CggR repressor of B. subtilis (Zorrilla et al., 2007) or the tetrameric LacI repressor of E. coli (Kalodimos et al., 2002). Additionally, on both sides of the ManR protected regions a small part of the strands seems to be exposed as can be seen by DNase I hypersensitive sites. It is possible that the right half of the coding sequence of the PmanP operator is more exposed and thus, the whole side becomes a substrate for DNase I.
According to the molar excess of the PTS components used in the activation mixtures and the prolonged incubation time we granted conditions to allow all of the ManR molecules to become activated most probably by phosphorylation. However, we cannot exclude that there were also some ManR molecules present in the assays, which were only partially or completely unphosphorylated. Therefore, the determined values for KD and kdiss rather represent the minimum levels. To our knowledge, no such values have been reported for other PRD-containing transcriptional activators of B. subtilis. However, the KD values of the PRD-containing antiterminators SacY and LicT have been determined as 3 μM and 10 nM respectively (Declerck et al., 1999; 2002). Comparable KD values of B. subtilis transcriptional regulators from other classes have been reported: TnrA, which controls gene expression in response to nitrogen availability together with GlnR, binds with a KD of 55 nM to the GlnR/TnrA site in the tnrA promoter region (Zalieckas et al., 2006). Recently, the KD value of 33.5 nM of the transcription factor AlsR binding to the alsSD operon promoter was reported (Frädrich et al., 2012). All these examples show that ManR has similar binding properties as other transcription factors of B. subtilis.
The weaker expression from PmanR, which could be attributed to an obviously less favourable 5′-untranslated region of the mRNA, is physiologically reasonable since the regulatory protein is not needed in such high concentrations as the corresponding metabolic enzymes. Nevertheless, compared with the similar B. subtilis mannitol system, the expression from PmanR is still very high. In the mtl system, expression from PmtlR was about 130-fold lower compared with the PmtlAFD operon promoter, whereas in the mannose system expression from PmanR is only about fourfold lower compared with PmanP (Sun and Altenbuchner, 2010; Heravi et al., 2011). The reason for this high expression is unknown. However, this has a positive effect since the usage of high-copy expression vectors with PmanP suffice the single chromosomal copy of manR to achieve high expression rates of the target genes.
Previously, it could be shown that induction with mannose using the PmanR and PmanP promoters is not possible if B. subtilis carries an HPr-His15Ala mutation (Sun and Altenbuchner, 2010). For several PRD-containing regulators it could be shown that specific phosphorylation via the PTS components PEP, EI and HPr is needed for their activation (reviewed by Deutscher et al., 2006). In analogy to the cognate MtlR regulators of G. stearothermophilus and B. subtilis (Henstra et al., 1999; Joyet et al., 2010) and other PRD-containing regulators, we propose that B. subtilis ManR also becomes activated via phosphoryl group transfer from PEP, EI and HPr. In principle, the system should respond to three different situations. Without mannose as a carbon source ManR should be inactive, and in the presence of mannose alone ManR should be active and in presence of mannose and a second, more favourable carbon source like glucose ManR should be inactivated again. Our findings let us propose a model of ManR regulation by phosphorylation and dephosphorylation of the various ManR domains in response to the three different situations (Fig. 11).
For activation of ManR by EI, HPr and PEP the PRD2 conserved histidines seem to play a mandatory role as replacements with alanines have shown. Mutations in the PRD1 had no such significant effect on ManR binding activity. His-570 in the EIIAFru-like domain of ManR was identified as the main target for negative regulation by the EIIBMan. The equivalents His-598 of G. stearothermophilus MtlR and His-599 of B. subtilis MtlR have been identified previously as targets for phosphorylation by the cognate EIIB domains of the PTS transporters (Henstra et al., 1999; Joyet et al., 2010). At least as important as His-599 is Cys-419 in B. subtilis MtlR, which seems to be the main target for negative regulation (Joyet et al., 2010). Whereas in MtlR single substitutions to alanine led to full constitutive expression, elevated basal activities were observed with the corresponding ManR mutants. Thereby, ManR-H570A showed a much stronger effect as the ManR-C415A mutant. Full constitutive behaviour was only observed with the ManR-C415A/H570A double mutant, proposing that both amino acids are targets for phosphorylation by the cognate PTS transporter ManP, albeit His-570 seems to play the more important role. As the ManR-H570A mutant could not be inactivated by EIIAMan or EIIBMan, we also propose that Cys-415 phosphorylation is not sufficient for significant reduction of ManR activity. However, it has to be considered that the ManR-C415A/H570A double mutation was accompanied with a strong reduction of overall activity which might lead to misinterpretations.
The findings of this study and the results published previously (Wenzel et al., 2011) showed that EIIBMan is not essential for ManR activation and a membrane sequestration mechanism like it was proposed for MtlR of B. subtilis (Bouraoui et al., 2013) is not obvious for ManR.
The weak point in our model is that the replacement of the histidine residues in the two PRD domains by alanine led to low or complete loss of β-galactosidase activity. We were not able to get constantly active ManR regulators by replacing the histidine residues by aspartate or glutamate as it was the case with other B. subtilis PRD-containing regulators (Charrier et al., 1997; Tortosa et al., 2001; Joyet et al., 2010). This may be due to problems in precise protein folding. Another point to consider are the two possible isomeric forms of phosphohistidine, 1-phosphohistidine and 3-phosphohistidine, which both were found in vivo (reviewed by Attwood et al., 2007), and the phosphomimicry by aspartate or glutamate may only work, if at all, for one of these isomers. For all that reasons, we cannot exclude the possibility that the inactivation of ManR is related to misfolded protein. It will be interesting to further investigate each phosphorylation event, for example by mass spectroscopy and determine the consequences on the activity, the oligomeric state and conformation of ManR.
Plasmids, bacterial strains and growth conditions
Relevant bacterial strains and plasmids used in this study are summarized in Tables S1–S3. Cloning steps were performed with E. coli JM109 (Yanish-Perron et al., 1985) using standard recombinant DNA techniques (Sambrook et al., 1989). Transformation of E. coli JM109 with plasmid DNA occurred as described (Chung et al., 1989). B. subtilis strains were transformed with plasmid DNA according to the modified ‘Paris method’ (Harwood and Cutting, 1990). If not stated otherwise, all strains were grown at 37°C using LB medium (Bertani, 1951). Antibiotics were used in the following concentrations: ampicillin (amp), 100 μg ml−1; spectinomycin (spc), 100 μg ml−1; erythromycin (erm), 5 μg ml−1.
Chemicals were purchased from Sigma-Aldrich Corporation (Taufkrichen, Germany), Carl Roth GmbH & Co. KG (Karlsruhe, Germany) and Merck KGaA (Darmstadt, Germany). Restriction endonucleases and other DNA modifying enzymes were purchased from New England Biolabs GmbH (Frankfurt am Main, Germany). PCRs were run with High Fidelity PCR Enzyme Mix (Thermo Fisher Scientific GmbH, Schwerte, Germany) on a TPersonal Thermocycler (Biometra GmbH, Goettingen, Germany). Synthetic DNA oligonucleotides (Table S4) were purchased from Eurofins MWG Operon GmbH (Ebersberg, Germany) and DNA sequence analyses were performed by GATC Biotech AG (Constance, Germany).
Overproduction and purification of ManR with B. subtilis
The manR expression vector pMW687.1 for B. subtilis is a derivative of the expression and shuttle vector pMW168.1 (Wenzel et al., 2011). It comprises the replication regions for E. coli (pUC18) and B. subtilis (pUB110), one common spectinomycin resistance marker and the PmanP-manR expression cassette flanked by strong transcriptional terminators. The construction started with PCR amplification of the manR gene using oligonucleotides s6934/s6935. The resulting fragment was cloned into pMW168.1 via BamHI and BsrGI to give pMW677.6. A shortened PmanP region was transferred from pSUN290 via Acc65I and BglII into pMW677.6, resulting in pMW687.1. The negative control vector pMW718.1 was obtained by excision of the expression cassette with PmeI and religation of the vector fragment. For purification, both Strep(II)- and 6xHis-tags were added to the C-terminus of ManR. By inserting a SmaI/AgeI fragment from pSUN383.1 (laboratory stock) into pMW687.1, the PmanP-manR-Strep(II) expression vector pMW725.1 was obtained. Insertion of a SmaI/AgeI fragment from pSUN390.1 (laboratory stock) into EcoRV/AgeI-digested pMW725.1 gave the PmanP-manR-6xHis expression vector pMW735.3.
Cells of B. subtilis TQ356 were transformed with the constructed expression vectors. Selection occurred at 37°C on LB agar plates supplemented with spectinomycin and 1% (w/v) glucose to prevent premature expression of manR by carbon catabolite repression. Overnight cultures of the transformants were grown at 30°C in shaking flasks with LB medium supplemented with spectinomycin. After 16 h incubation time, approximately 1 × 1010 cells were harvested and centrifuged. The pelleted cells were resuspended with 1 ml of 100 mM sodium phosphate buffer (pH 7.5), and crude cell extracts were prepared using ultrasonic sound (Heat Systems-Ultrasonics, model W-385 sonicator, Farmingdale, New York, USA; 3 × 45 s, 50% duty cycle). The crude cell extracts were centrifuged and, thereby, separated into a soluble (supernatant) and insoluble protein fraction (pellet). After SDS-PAGE analysis (Fig. S1), the soluble fractions (containing 1.02 ± 0.24 mg ml−1 of total protein) were used for EMSAs and DNase I footprinting experiments.
For purification of 6xHis- and Strep(II)-tagged ManR proteins in the soluble fraction of the crude cell extracts Ni-NTA Agarose (QIAGEN, Hilden, Germany) and Strep-Tactin® Sepharose® resins (IBA, Goettingen, Germany), respectively, were used according to the manufacturer's instructions.
Overproduction and purification of B. subtilis ManR and PTS components with E. coli
ManR with a C-terminal Strep(II)-tag was overexpressed in E. coli JM109/pSUN246.1 as follows: cells were grown for 2 h at 37°C. After induction with 0.2% (w/v) rhamnose, growth continued at 30°C for 4 h. Approximately 3 × 1010 cells were harvested by centrifugation, and the pelleted cells were either stored at −20°C or resuspended with 1 ml of buffer W (100 mM Tris-HCl pH 8.0, 150 mM NaCl, 1 mM EDTA). Crude cell extracts were prepared using ultrasonic sound (see section Overproduction and purification of ManR with B. subtilis) and centrifuged. The supernatant containing the soluble protein fraction was used for affinity tag chromatography using Strep-Tactin® Sepharose® resin according to the manufacturer's instructions (IBA, Goettingen, Germany) and led to a maximum protein concentration of 0.15 mg ml−1, quantified using the Bradford method (Bradford, 1976). A typical SDS-PAGE analysis is shown in Fig. S2.
Construction of ptsH (HPr), ptsI (EI), manP1–120 (EIIB) and manP438–650 (EIIA) expression vectors started with PCR amplification of the genes using oligonucleotide primer pairs s7654/s7655, s7656/s7657, s7658/s7659 and s7660/s8043, respectively, and chromosomal DNA from B. subtilis 3NA as template. The PCR products were inserted into pSUN390, a rhamnose promoter-containing expression vector, using BamHI/NdeI giving pMW850.2 (HPr-6xHis), pMW851.2 (EI-6xHis), pMW852.1 (ManP-EIIB-6xHis) and pMW987.1 (ManP-EIIA-6xHis) respectively. Overproduction and cell disruption occurred as described above. The soluble protein fraction was used for affinity tag chromatography using Ni-NTA Agarose resin according to the manufacturer's instructions (QIAGEN, Hilden, Germany). Purified proteins were quantified using the Bradford method (Bradford, 1976).
In vitro ManR activation assay
In order to activate purified B. subtilis ManR from E. coli, PTS components were added in molar excess in 100 μl of assay mixtures containing 50 mM Tris-HCl (pH 7.4), 5 mM MgCl2, 5 mM PEP, 266 nM ManR, 317 nM HPr and 1.5 μM EI. The mixtures were incubated for 30 min at 37°C.
Construction of B. subtilis strain MW678
Initially, the PmanP–lacZ expression cassette from pSUN290 was inserted into the integration vector pHM31 via BamHI, resulting in pMW678.6. Using the pHM30/pHM31-system (Motejadded and Altenbuchner, 2007), the PmanP–lacZ expression cassette was integrated downstream of the his operon of B. subtilis TQ276 (spo0A3 ΔmanR::ermC), resulting in B. subtilis MW678 (spo0A3 hisI::PmanP–lacZ ΔmanR::ermC).
Construction of manR alleles and mutant strains
Potential phosphorylation sites of ManR were replaced by alanine residues using the overlap fusion PCR method (Ho et al., 1989). First, manR, including its own promoter region, was blunt end cloned from pSUN063.11 (Sun and Altenbuchner, 2010) into the positive selection vector pJOE4786.1, a derivative of pJOE773 (Altenbuchner et al., 1992), giving pMW376.8. In order to replace each potential phosphorylation site with an alanine residue, codon usage for B. subtilis ManR revealed 5′-GCA-3′ as the most used codon for alanine. Next, mutagenic reverse and forward primers (b and c) overlapping the region with the desired mutation were designed and used for PCR amplification with corresponding forward (a/b) and reverse primers (c/d) containing unique restriction sites at their 5′-ends. DNA of pMW376.8 was used as template. The fragments were mixed equally and fused together by PCR using the flanking forward and reverse primers (a/d). The fusion PCR products were inserted into pMW376.8 and the mutation confirmed by DNA sequencing. Table S5 summarizes the used primers, cloning sites and resulting pMW376.8 derivatives. The PRD1 double mutant H222A/H281A was constructed by inserting the fragment that was used for pMW401.1 via AgeI/MunI into pMW408.3, giving pMW436.1. The PRD2 double mutant H334A/H393A was constructed like pMW407.1 using pMW402.1 as template for PCR amplifications, giving pMW420.1. The EIIB(C415A)/EIIA(H570A) double mutant was constructed by inserting a MunI/BglII fragment from pMW404.1 into the vector pMW403.1, giving pMW467.1.
For chromosomal integration the manR alleles were inserted via BamHI into the amyE integration vector pDG1730 (Guérout-Fleury et al., 1996). These vectors, listed in Table S2, were brought into B. subtilis MW678. The first selection occurred on LB agar plates containing spectinomycin. Double cross-over events were confirmed on LB starch and counterselection on LB plates with erythromycin. Colonies that were sensitive to erythromycin and had lost the ability to degrade starch were steaked out to gain single colonies.
The manR alleles, including their own promoter, were also excised with BamHI from the corresponding pMW376.8 derivatives and inserted into BamHI-digested pSUN290, a pBS72-based low-copy expression vector (Titok et al., 2003). These vectors, listed in Table S3, were brought into B. subtilis MW678.
Electrophoretic mobility shift assays
Formation of ManR–DNA complexes was detected via electrophoretic mobility shift assays (EMSAs), which were performed essentially as described previously (Shewchuk et al., 1989; Rother et al., 1999). Fluorescent DNA probes (c. 150–200 bp) containing the putative ManR operator sites were prepared by PCR amplification using Cy5-labelled primers according to Table S6. The same fragments were PCR amplified using analogous non-labelled primer pairs for binding competition experiments (Table S6). Each EMSA comprised 2.5–5 nM purified, Cy5-labelled DNA fragment, and a varying amount of either crude cell extract or purified protein in a total volume of 10 μl. The binding reaction occurred for 30 min at 4°C in EMSA binding buffer consisting of 10 mM Tris-HCl (pH 7.5), 50 mM KCl, 50 μg ml−1 BSA, 5 μg ml−1 salmon sperm DNA, 5% glycerol and 2 mM DTT. Protein–DNA complexes were separated at 20 mA for 60 min on native polyacrylamide gels. The fluorescent bands were detected with the Storm 860 PhosphorImager System (GE Healthcare, Munich, Germany). All EMSA experiments were independently reproduced at least three times.
Determination of equilibrium dissociation constants and dissociation rates
Determination of the binding properties of ManR, in particular the equilibrium dissociation constant KD and the dissociation rates kdiss, occurred as described previously (Rother et al., 1999). KD was determined by EMSAs as follows: varying amounts (140, 70, 20, 14, 7, 2 and 1.4 ng corresponding to 186, 93, 26.5, 18.6, 9.3, 2.7 and 1.9 nM) of purified and activated ManR-Strep(II)-tag were incubated for 30 min at 4°C with 5 nM Cy5-labelled DNA fragment containing the operator/promoter region to form a stable complex. After electrophoresis, band intensities were analysed using an ImageQuant software package. The ratio of free and bound DNA molecules was determined and plotted over the amount of ManR used in the assay. The amount of ManR required to shift 50% of the DNA was taken from the plot, and the corresponding KD value was determined.
Determination of the dissociation rates kdiss and the corresponding half-lives of the ManR–DNA complexes were also determined by EMSAs, but with a total starting volume of 100 μl. The concentration of activated ManR was 186 nM, which was sufficient to shift all of the 5 nM Cy5-labelled DNA fragment after incubation for 30 min at 4°C. Dissociation of the ManR–DNA complex was initiated by addition of a 50-fold molar excess of non-labelled DNA fragment. Samples were taken at 10 min intervals and directly applied to a native polyacrylamide gel with the current switched on. The intensities of the bands representing the ManR–DNA complex were analysed with the ImageQuant software package and plotted logarithmically over the time. From the equation of the best fit straight line cDNA(t) = cDNA(t = 0) × e−kdiss×t, according to the common decay law, the kdiss values were determined. The term cDNA(t) represents the concentration of the ManR–DNA complex at a given point of time (t). The half-life t½ equals ln(2) divided by kdiss.
All experiments were independently reproduced at least three times.
DNase I footprinting assays
The Cy5-labelled PCR fragments listed in Table S6, were used in binding reactions with a total volume of 30 μl, i.e. 6 μl of × EMSA binding buffer [50 mM Tris-HCl (pH 7.5), 250 mM KCl, 250 μg ml−1 BSA, 25 μg ml−1 salmon sperm DNA, 25% glycerol and 10 mM DTT], 2 μl of end-labelled DNA probe (c. 10 ng), and 22 μl either of crude cell extracts of B. subtilis TQ356/pMW687.1 (ManR) and B. subtilis TQ356/pMW718.1 (negative control) or of in vitro activation reaction mixture with and without ManR. After 30 min incubation at 4°C, 13 μl of ddH2O, 5 μl of 10 × DNase I Reaction Buffer (NEB) and 2 μl of DNase I (2000 units ml−1, NEB) were added. After incubation for 1 min at room temperature, 50 μl of stop solution (50 mM EDTA, 15 μg ml−1 calf thymus DNA) was added and mixed by vortexing. The DNA was phenol-chloroform extracted, precipitated with ethanol, and resuspended in loading buffer from the AutoRead™ Sequencing Kit (formerly Amersham Pharmacia Biotech, Piscataway, NJ, USA). After heating for 3 min at 80°C, the mixture was stored at −20°C or directly loaded onto a 0.3-mm-thick sequencing gel (ReproGel High Resolution, GE Healthcare Company, Munich, Germany).
The sequencing reactions were carried out using the dideoxy method (Sanger et al., 1977) with the AutoRead™ Sequencing Kit according to the manufacturer's instructions. For each sequencing reaction, 10 μg of pSUN351.3, which comprises both PmanR and PmanP, were used as template DNA. Both strands were analysed using the Cy5-labelled oligonucleotide primers s7135 (manR region, coding strand), s5097 (manR region, non-coding strand), s7232 (manP region, coding strand) and s8001 (manP region, non-coding strand) respectively. Separation of the fragments with the ALFexpress II DNA sequencer (formerly Amersham Pharmacia Biotech, Piscataway, NJ, USA) was conducted according to the manufacturer's instructions. DNase I footprinting experiments were independently reproduced at least three times for each promoter region.
Measurement of β-galactosidase activity
The β-galactosidase activity was measured as described previously (Miller, 1972): after cells were grown in 5 ml of LB medium with appropriate antibiotics for 2 h at 37°C, either no sugar, 0.2% (w/v) mannose, or 0.2% (w/v) mannose and 0.2% (w/v) glucose were added. After further incubation for 1 h at 37°C, the OD at 600 nm was measured, and 100 μl of cell culture was treated with 900 μl of Z buffer (60 mM Na2HPO4, 40 mM NaH2PO4, 10 mM KCl, 1 mM MgSO4 and 40 mM 2-mercaptoethanol) and 10 μl of toluene for 30 min at 37°C. The β-galactosidase reaction in the toluene-treated cells was started by adding 200 μl of 20 mM o-nitrophenyl-β-galactopyranoside (in Z buffer) at 22°C. After the suspension turned yellowish, 500 μl of 1 M Na2CO3 was added to stop the reaction, for a maximum of 30 min. Miller Units (M.U.) were determined by measuring the OD at 420 and 550 nm.
We thank Lonza AG (Visp, Switzerland) for financial support of this study.