Correspondence: Boris Görke, Department of General Microbiology, Institute of Microbiology and Genetics, Georg-August-University, Grisebachstrasse 8, 37077 Göttingen, Germany. Tel.: +49 551 393796; fax: +49 551 393808; e-mail: firstname.lastname@example.org
Bacillus subtilis possesses carbon-flux regulating histidine protein (Crh), a paralog of the histidine protein (HPr) of the phosphotransferase system (PTS). Like HPr, Crh becomes (de)phosphorylated in vitro at residue Ser46 by the metabolite-controlled HPr kinase/phosphorylase HPrK/P. Depending on its phosphorylation state, Crh exerts regulatory functions in connection with carbohydrate metabolism. So far, knowledge on phosphorylation of Crh in vivo has been limited and derived from indirect evidence. Here, we studied the dynamics of Crh phosphorylation directly by non-denaturing gel electrophoresis followed by Western analysis. The results confirm that HPrK/P is the single kinase catalyzing phosphorylation of Crh in vivo. Accordingly, phosphorylation of Crh is triggered by the carbon source as observed previously for HPr, but with some differences. Phosphorylation of both proteins occurred during exponential growth and disappeared upon exhaustion of the carbon source. During exponential growth, ~ 80% of the Crh molecules were phosphorylated when cells utilized a preferred carbon source. The reverse distribution, i.e. around 20% of Crh molecules phosphorylated, was obtained upon utilization of less favorable substrates. This clear-cut classification of the substrates into two groups has not previously been observed for HPr(Ser)~P formation. The likely reason for this difference is the additional PTS-dependent phosphorylation of HPr at His15, which limits accumulation of HPr(Ser)~P.
The histidine protein (HPr) of the carbohydrate : phosphotransferase system (PTS) has a dual role in Firmicutes bacteria. In its transport function HPr delivers phosphoryl-groups from Enzyme I (EI) to the Enzyme II (EII) transport proteins, which phosphorylate their sugar substrates during uptake. During this phosphoryl-group transfer, HPr becomes transiently phosphorylated at residue His15. In addition, HPr also exerts important regulatory functions (Deutscher et al., 2006). It is the key player in carbon catabolite repression (CCR), which allows the bacteria to repress functions for the utilization of secondary carbon sources when a preferred substrate is simultaneously present (Deutscher, 2008; Görke & Stülke, 2008).
To be active in CCR, HPr must be phosphorylated at a different site, Ser46. HPr(Ser)~P binds the global transcriptional regulatory protein CcpA, which thereby gains DNA-binding activity (Fujita, 2009). Phosphorylation as well as de-phosphorylation of HPr at Ser46 is catalyzed by a single enzyme, the HPr kinase/phosphorylase (HPrK/P). The decision as to whether kinase or phosphorylase activity will prevail is controlled by the quality of the available carbon source. Preferred carbon sources such as glucose or fructose, which allow the fastest growth rates, activate the kinase function of HPrK/P and thereby trigger the formation of HPr(Ser)~P. In contrast, phosphorylase activity of HPrK/P predominates when cells grow on less favorable substrates such as succinate or ribose. The various substrates form a hierarchy in their ability to trigger phosphorylation of HPr at Ser46 (Singh et al., 2008). Fructose-1,6-bisphosphate (FBP) has been identified as the main factor that allosterically activates kinase activity of HPrK/P, but other metabolites may also play a role (Jault et al., 2000; Ramström et al., 2003).
Bacillus subtilis and other Bacilli possess carbon-flux-regulating histidine protein (Crh), which shares over 40% sequence identity with HPr (Galinier et al., 1997). Crh lacks His15, but contains Ser46 and accordingly it becomes (de)phosphorylated by HPrK/P in vitro (Lavergne et al., 2002). Crh~P can likewise form a complex with CcpA and contributes to CCR, but to a weaker extent than HPr(Ser)~P. Hence, Crh~P can only partially replace HPr(Ser)~P in CCR (Galinier et al., 1997; Singh et al., 2008). This weaker contribution of Crh~P to CCR can be ascribed to its much lower levels in the cell and its lower binding affinity for CcpA as compared with HPr(Ser~P) (Görke et al., 2004; Seidel et al., 2005). Therefore, Crh was regarded for a long time as back-up factor for CCR. However, recently a distinct role for Crh has been identified. It was found that non-phosphorylated Crh binds to and inhibits activity of the metabolic enzyme methylglyoxal synthase, MgsA, in B. subtilis (Landmann et al., 2011). MgsA catalyzes the formation of methylglyoxal from dihydroxyacetone-phosphate, initiating a glycolytic bypass. This pathway may relieve cells from sugar-phosphate stress, when carbohydrate uptake rates exceed the capacity of the lower branch of the Embden–Meyerhof–Parnas (EMP) pathway (Weber et al., 2005).
To understand the physiological conditions under which Crh exerts its regulatory functions, it is crucial to know its phosphorylation state in vivo. Indirect evidence from studies on CCR suggested that the phosphorylation of Crh and HPr at their Ser46 sites has similar dynamics (Galinier et al., 1997; Singh et al., 2008). Direct proof of this hypothesis has so far been hindered technically by the low cellular abundance of Crh (Görke et al., 2004). This might explain why Crh~P was detected in only one of several phosphoproteome studies (Eymann et al., 2007).
In the present work, we overcame these limitations and analyzed phosphorylation of Crh in vivo in response to different nutritional conditions. A direct method was used that involves separation of Crh and Crh~P in cell extracts by non-denaturing gel electrophoresis. Crh was detected using a tailor-made sensitive antiserum specifically directed against a C-terminal peptide of Crh.
Materials and methods
Bacterial strains and growth conditions
Bacillus subtilis strains were grown at 37 °C in CSE minimal medium (C-medium supplemented with sodium succinate and potassium glutamate; Martin-Verstraete et al., 1995) supplemented with 50 mg L−1 tryptophan and 0.5% of the indicated carbon source. The strains used were B. subtilis 168 (trpC2; wild-type), QB7097 (trpC2 Δcrh::spec; Singh et al., 2008) and GP82 (trpC2 ΔhprK::cat; Landmann et al., 2011).
Determination of the phosphorylation states and of the total amounts of HPr and Crh in vivo
Bacteria were grown in CSE minimal medium supplemented with 0.5% of the indicated carbon source. At the time of harvest, the pH of the culture medium was lowered to pH 4.5 using 12 M HCl to stop HPrK/P activity, as described previously (Singh et al., 2008). Aliquots of 45 mL culture were pelleted by centrifugation, washed with 500 mM NaCl and 20 mM sodium acetate pH 4.5, and finally re-suspended in 1 mL of the same buffer. Cells were disrupted by three passages through a French pressure cell. All protein extracts were separated by non-denaturing gel electrophoresis (100 V, 75 min) using gels prepared from 12% acrylamide in 375 mM Tris/HCl pH 8.8. Crh and HPr were subsequently detected by Western blotting. For the determination of total Crh and HPr amounts, SDS sample buffer [62.5 mM Tris/HCl pH 6.8, 5% (v/v) β-mercaptoethanol, 2% (w/v) SDS, 10% (v/v) glycerol, 0.05% (w/v) bromophenol-blue] was added to protein extracts. The samples were heated for 10 min at 70 °C and subsequently separated by SDS-PAGE using 15% (Fig. 1) or 12% polyacrylamide gels (Figs 2 and 3) prepared in 375 mM Tris/HCl pH 8.7 and 0.1% SDS. Gels were analyzed by Western blotting.
Generation of an antiserum directed specifically against Crh
To identify an epitope in Crh hat was suitable for its specific detection, a sequence alignment of various Crh proteins and their cognate homologs was performed using clustalw software. This analysis showed that the primary sequences of the Crh proteins deviate considerably from the HPr sequences in three regions. Subsequent analysis of the secondary structures of these regions (Janin, 1979; Parker et al., 1986) suggested the 23 C-terminal amino acids of Crh as an epitope suitable for the generation of an antiserum that allows specific detection of Crh and does not cross-react with HPr. Consequently, this peptide, which additionally carried a cysteine at the N-terminus for its coupling to a protein carrier, was synthesized and used for the immunization of rabbits (performed by Pineda, Antikörper Services, Berlin, Germany).
After gel electrophoresis, proteins were transferred to a polyvinylidene difluoride membrane (Bio-Rad) by electro-blotting (60 min at 0.8 mA cm−2). Polyclonal rabbit antisera directed against HPr (Monedero et al., 2001) or the C-terminus of Crh (this work) were used at dilutions of 1 : 10 000 to detect these proteins. The primary antibodies were visualized with goat anti-rabbit IgG secondary antibodies conjugated to alkaline phosphatase diluted 1 : 100 000 (Promega) and the CDP* detection system (Roche Diagnostics). Quantification of signal intensities was achieved using software imagej version 1.42 (Abramoff et al., 2004).
HPrK/P is an essential kinase for the phosphorylation of Crh in vivo
A useful technique that allows the investigation of the phosphorylation state of a protein in vivo involves the separation of the differently modified protein species by non-denaturing polyacrylamide gel electrophoresis (PAGE) followed by their sensitive detection by Western analysis. For Crh to be detected, an antiserum directed against the peptide comprising the 23 C-terminal amino acids of Crh was generated. Analysis of the primary and secondary structures of Crh suggested this epitope as being suitable for the sensitive and specific detection of Crh. Indeed, when protein extracts were separated by SDS-PAGE and subjected to Western analysis, a strong signal at the position expected for Crh (molecular weight 9.3 kDa) became visible in the wild-type, but not in the Δcrh mutant (Fig. 1). Thus, no cross-reactivity with HPr occurred. Next, we prepared protein extracts from the wild-type strain and its isogenic ΔhprK mutant, which were grown to exponential phase in minimal glucose medium.
The extracts were resolved by non-denaturing PAGE and the gel was analyzed by Western blotting using the Crh-specific antiserum. Two signals became detectable in the wild-type strain (Fig. 2a, lane 12). Quantification of the signal intensities revealed a threefold stronger signal for the faster migrating band, indicating that Crh is predominantly phosphorylated under these conditions. In contrast, only the slower migrating band corresponding to non-phosphorylated Crh was detectable in the hprK mutant (Fig. 2a, lane 13). Thus HPrK/P is essential for phosphorylation of Crh in vivo.
Carbon source quality triggers phosphorylation of Crh
The phosphorylation of HPr by HPrK/P is modulated by the carbon source. To determine whether this also holds true for Crh, we investigated the phosphorylation state of Crh in wild-type cells that were grown to exponential phase in minimal medium supplemented with various carbon sources. The degree of phosphorylation of Crh varied drastically with the carbon source utilized by the bacteria (Fig. 2a, top panel). In contrast, the total amount of Crh, as estimated from denaturing SDS gel electrophoresis, was only slightly affected by the carbon source and appeared to be somewhat higher when cells utilized unfavorable carbon sources such as succinate or ribose (Fig. 2a, bottom panel).
The relative proportions (in percent) of phosphorylated and non-phosphorylated Crh were determined by quantification of data obtained from at least three independent experiments (Fig. 2b). Crh was found predominantly in its non-phosphorylated form when bacteria utilized succinate, ribose or gluconate, all of which are unfavorable substrates. These substrates trigger no or only weak CCR and yield slower growth rates (with the exception of gluconate) in comparison with the other substrates (Singh et al., 2008). Under these conditions, 25% or less of all Crh molecules were phosphorylated. In contrast, the opposite distribution was observed with the other tested substrates. Those sugars triggered phosphorylation of ~80% of the Crh molecules.
Crh and HPr are predominantly non-phosphorylated in the stationary growth phase
We were keen to trace putative changes in the phosphorylation state of Crh when carbohydrates become exhausted and bacteria enter the stationary growth phase. To this end, we grew the wild-type strain in minimal medium containing succinate, ribose or glucose as carbon source. Succinate is an unfavorable carbon source and allows only slow growth. With ribose as substrate, growth rates are considerably improved, but still not as high as with glucose, which is the preferred carbon source of B. subtilis (Fig. 3a; Singh et al., 2008).
Samples were taken periodically and the phosphorylation state of Crh was analyzed (Fig. 3b). For comparison, the phosphorylation state of HPr was also determined (Fig. 3c). To discriminate between HPr(Ser~P) and HPr(His~P), which migrate at the same position on the gel, a second aliquot of each sample was heated prior its loading onto the gel (Fig. 3c, even-numbered lanes). This leads to loss of the thermo-labile phospho-histidine bonds, whereas the serine-phosphate bonds are stable and remain intact.
The comparison of both aliquots allows an estimation of the degree of phosphorylation of each site. During growth on the various substrates, the phosphorylation patterns of both Crh and HPr changed in a similar manner. Both proteins were detectable in their non-phosphorylated as well as serine-phosphorylated forms during the exponential growth phase.
As observed before (Fig. 2 and Singh et al., 2008), the ratio of the two forms depended on the carbon source (Fig. 3b, compare lanes 1, 4, 8; Fig. 3c, compare lanes 2, 8, 16). However, upon transition to the early stationary phase, the amount of Ser-phosphorylated Crh and HPr decreased drastically. When glucose was the carbon source, Crh as well as HPr was completely non-phosphorylated at Ser46 when cells entered the stationary growth phase (Fig. 3b, lane 3; Fig. 3c, lane 6). When succinate or ribose was the carbon source, the extent of phosphorylation at Ser46 also decreased but a small amount of HPr(Ser)~P and Crh~P was detectable even upon entry into the stationary growth phase (Fig. 3b, lanes 7, 10; Fig. 3c, lanes 14, 20). The majority of phosphorylated HPr species detectable in this growth phase were phosphorylated at the His15 residue (Fig. 3c, compare lanes 5 and 6, lanes 13 and 14, lanes 19 and 20). There were no major changes in the total amounts of Crh or HPr under the various conditions (Fig. 3b and c, bottom panels).
Carbon source exhaustion inhibits phosphorylation of Crh and HPr by HPrK/P in the stationary growth phase
Finally, we wanted to confirm that scarcity of the carbon source prevents phosphorylation of Crh and HPr at their Ser46 sites when cells enter the stationary growth phase. To this end, the wild-type strain was grown once again in minimal medium supplemented with glucose. After 7 h growth, i.e. the time of transition to the stationary growth phase, the culture was split and glucose was added to one of the two resulting cultures. The culture treated with additional glucose resumed growth and reached a final OD600 nm of 8.4, whereas the untreated culture entered the stationary growth phase, yielding a final OD600 nm of 3.7 (Fig. 4a), demonstrating that scarcity of the carbon source is growth-limiting under these conditions.
Subsequently, the phosphorylation states of Crh and HPr were analyzed in samples that were taken periodically during growth (Fig. 4b). As observed before (Fig. 3), phosphorylation of Crh and HPr at Ser46 was strongly inhibited in the untreated cells (no additional glucose added) when growth ceased, i.e. after 9 h incubation (Fig. 4b, top panels). In contrast, much higher amounts of Crh~P and HPr(Ser)~P were detectable at that time (9 h) in the cells that were supplemented with additional glucose (Fig. 4b, compare lanes 3 and 10 in the top and bottom panels). This result unequivocally shows that exhaustion of the carbon source glucose prevents phosphorylation of Crh and HPr by HPrK/P when cells enter the stationary growth phase.
In this work, we analyzed the dynamics of phosphorylation of Crh in response to different nutritional conditions in vivo. Previous in vitro studies suggested that Crh becomes (de)-phosphorylated by HPrK/P at residue Ser46 like its homolog HPr, but whether this also applied to in vivo conditions was not clear. Our data confirm that HPrK/P is actually the kinase responsible for phosphorylation of Crh in vivo (Fig. 2). Thus, one might expect a similar dynamics of phosphorylation of Crh and HPr at their Ser46-sites. Overall, this was indeed the case, but with some remarkable deviations.
As expected, both Crh~P and HPr(Ser)~P levels decreased drastically or even disappeared when cells entered the stationary growth phase (Fig. 3). Exhaustion of the carbon source is responsible for accumulation of the non-phosphorylated proteins in this growth phase (Fig. 4). Consequently, stationary cells are released from CCR and primed for the uptake and utilization of alternative carbon sources.
The degree to which Crh became phosphorylated during exponential growth depended on the quality of the carbon source. The various substrates could be classified into two distinct groups, triggering the formation of either low or very high levels of Crh~P (Fig. 2). Such a splitting of the carbon sources into two distinct groups has not been observed previously in the formation of HPr(Ser)~P. In this case, a more gradual transition between the various substrates was detected (Singh et al., 2008). Nonetheless, the carbon sources that trigger either very low or very high levels of phosphorylation are the same for both proteins. Only a little Crh~P and HPr(Ser)~P is formed (Fig. 2; Singh et al., 2008) when cells utilize succinate, ribose or gluconate. Consequently, these gluconeogenic carbon sources cause no or only weak CCR (Singh et al., 2008). Except for gluconate, these substrates also yield slower growth rates in comparison with the other tested substrates (Fig. 2a; Singh et al., 2008).
In contrast, high Crh~P as well as HPr(Ser~P) levels were detectable when a substrate of the PTS (glucose, fructose, mannitol, salicin, sucrose), sorbitol or glycerol was the carbon source (Fig. 2; Singh et al., 2008). Accordingly, all these sugars, which exert a strong CCR, enter the upper branch of the EMP pathway directly (Singh et al., 2008). These substrates trigger fast growth rates and generate high intracellular FBP concentrations that lead to activation of the kinase function of HPrK/P (Singh et al., 2008).
The remaining substrates arabinose and maltose caused the efficient phosphorylation of Crh~P (80%) but no comparable accumulation of HPr(Ser)~P (21% and 13% of total HPr, respectively; Singh et al., 2008). Therefore, CCR caused by these substrates is weak. How can this discrepancy be explained? When arabinose or maltose is utilized, more than 60% of all HPr molecules are phosphorylated either at His15 or at both sites (Singh et al., 2008). Neither of these forms, HPr(His)~P or doubly phosphorylated HPr, is active in CCR because phosphorylation at His15 impedes complex formation with CcpA (Schumacher et al., 2004). It would appear that the phosphorylation at His15 provides an additional level of control that allows integration of information about the phosphorylation status of the PTS into the global mechanism of CCR.
Evidence is accumulating that Crh has no dedicated role in CCR. However, it appears to regulate glycolytic flux through interaction with two metabolic enzymes, methylglyoxal synthase (MgsA) and glyceraldehyde-3-phosphate dehydrogenase (GapA). Non-phosphorylated Crh inhibits MgsA (Landmann et al., 2011), whereas phosphorylated Crh~P, in concert with HPr(Ser)~P, inhibits GapA activity (Pompeo et al., 2007). Non-phosphorylated Crh accumulates when bacteria grow on less favorable (gluconeogenic) carbon sources or when carbohydrates become exhausted and cells enter the stationary growth phase (Figs 2-4). Consequently, MgsA activity and concomitantly flux through the methylglyoxal pathway is expected to be inhibited by Crh under these famine conditions. Under feast conditions, Crh is predominantly phosphorylated. Thus MgsA gains activity, whereas GapA is repressed, leading to re-direction of flux from the EMP pathway towards the methylglyoxal pathway. This mechanism may prevent accumulation of sugar-phosphates when there is an excess of sugars and uptake rates exceed the capacity of EMP pathway.
We thank Sabine Lentes for excellent technical assistance. We are grateful to Gerald Seidel for providing information on the Crh antiserum and for insightful discussion. This work was supported by the Federal Ministry of Education (Research SYSMO network) to J.S. and W.H., and by grant GO1355/7-1 of the Deutsche Forschungsgemeinschaft to B.G. J.J.L. was supported by a stipend of the Fonds der Chemische Industrie.