The Pta–AckA pathway controlling acetyl phosphate levels and the phosphorylation state of the DegU orphan response regulator both play a role in regulating Listeria monocytogenes motility and chemotaxis
Biology of Gram-Positive Pathogens, Department of Microbiology, CNRS URA 2172, and
DegU is considered to be an orphan response regulator in Listeria monocytogenes since the gene encoding the cognate histidine kinase DegS is absent from the genome. We have previously shown that DegU is involved in motility, chemotaxis and biofilm formation and contributes to L. monocytogenes virulence. Here, we have investigated the role of DegU phosphorylation in Listeria and shown that DegS of Bacillus subtilis can phosphorylate DegU of L. monocytogenes in vitro. We introduced the B. subtilis degS gene into L. monocytogenes, and showed that this leads to highly increased expression of motility and chemotaxis genes, in a DegU-dependent fashion. We inactivated the predicted phosphorylation site of DegU by replacing aspartate residue 55 with asparagine and showed that this modified protein (DegUD55N) is no longer phosphorylated by DegS in vitro. We show that although the unphosphorylated form of DegU retains much of its activity in vivo, expression of motility and chemotaxis genes is lowered in the degUD55N mutant. We also show that the small-molecular-weight metabolite acetyl phosphate is an efficient phosphodonor for DegU in vitro and our evidence suggests this is also true in vivo. Indeed, a L. monocytogenesΔptaΔackA mutant that can no longer synthesize acetyl phosphate was found to be strongly affected in chemotaxis and motility gene expression and biofilm formation. Our findings suggest that phosphorylation by acetyl phosphate could play an important role in modulating DegU activity in vivo, linking its phosphorylation state to the metabolic status of L. monocytogenes.
Signal transduction in bacteria occurs essentially through two-component systems (TCSs), using phosphorylation as a mechanism to adapt genetic expression to meet the challenges of environmental changes. A prototypical TCS is composed of a histidine kinase, usually membrane-bound, which is autophosphorylated in response to external signals. This is followed in turn by phosphotransfer to a response regulator which then acts to control gene expression (Hoch and Silhavy, 1995). Phosphorylation of the response regulator is the end result of multiple enzymatic reactions: autophosphorylation of the kinase, autophosphatase activity of the kinase, phosphotransfer to the regulator, back-transfer to the kinase, autophosphatase activity of the regulator and phosphoprotein phosphatase activity of the kinase towards the regulator. In some systems, additional proteins also act as response regulator phosphatases such as CheZ of Escherichia coli or the Rap phosphatase family (Hess et al., 1988; Perego et al., 1996).
The unphosphorylated response regulator is usually in an inactive form, with phosphorylation being required to induce conformational changes within the active site that lead to functional activation (Stock et al., 2000). However, several regulators such as DegU of Bacillus subtilis or AlgB and AlgR of Pseudomonas aeruginosa are known to be active in their unphosphorylated form (Msadek et al., 1990; 1995; Dahl et al., 1992; Ma et al., 1998). Indeed, a rather unique feature of the B. subtilis DegS/DegU TCS is the fact that the DegU response regulator acts as a molecular switch, positively controlling two alternate sets of genes: the phosphorylated form is essential for the expression of a wide variety of genes, including those encoding degradative enzymes such as levansucrase (sacB) and the alkaline protease subtilisin (aprE), whereas the unphosphorylated form is needed for synthesis of the ComK regulator which in turn activates expression of late competence genes (Msadek et al., 1990; 1995; Dahl et al., 1992; van Sinderen et al., 1995). The DegS kinase is also remarkable in that it is one of a small number of histidine kinases whose localization is entirely cytoplasmic, with no transmembrane domains or extracytoplasmic loops, unlike the vast majority of these kinases (Kunst et al., 1988; Inouye and Dutta, 2003).
Partner recognition among TCS proteins is highly specific, with promiscuous interactions being the exception rather than the rule. Indeed, specificity determinants of the interaction surfaces between the response regulator and the kinase are organized in a pattern of anchor and recognition residues, and are crucial in preventing non-specific interactions, often referred to as ‘cross-talk’ (Hoch and Varughese, 2001; Mukhopadhyay and Varughese, 2005; Laub and Goulian, 2007; Skerker et al., 2008). Although cross-phosphorylation of a response regulator by a non-cognate histidine kinase has often been reported in vitro, there are very few examples where non-specific interactions occur in vivo in a physiologically relevant manner (Matsubara et al., 2000; Howell et al., 2006).
It is well known, on the other hand, that phosphorylation of response regulators can take place in a kinase-independent manner using small-molecule phosphodonors such as acetyl phosphate, carbamoyl phosphate or phosphoramidate (Lukat et al., 1992). The most well studied of these phosphodonors is acetyl phosphate, which has been shown to regulate several cell processes, including chemotaxis, nitrogen and phosphate assimilation, outer membrane porin expression, flagellar and capsule biosynthesis, stress response and biofilm formation, and shown to act through TCS phosphorylation in some cases (Wolfe, 2005).
Listeria monocytogenes is a Gram-positive facultative intracellular pathogen very closely related to B. subtilis, which can cause severe diseases in immunocompromised hosts, including meningitis, septicaemia and gastro-enteritis. Although an orthologue of the DegU response regulator is present in Listeria, the degS gene encoding its cognate histidine kinase is absent (Glaser et al., 2001). We and others have previously shown that DegU is required for motility and chemotaxis in L. monocytogenes and that it also plays a role in virulence (Williams et al., 2005a,b; Gueriri et al., 2008). Additionally, we demonstrated that DegU was required for biofilm formation by Listeria, and that it represses its own synthesis by binding directly to the degU promoter region (Gueriri et al., 2008).
In the present article, we investigated the role of DegU phosphorylation in Listeria by three different approaches. We introduced the B. subtilis degS gene into L. monocytogenes, and showed that this leads to highly increased expression of motility and chemotaxis genes, in a DegU-dependent fashion. We also inactivated the predicted phosphorylation site of DegU and showed that although the unphosphorylated form of DegU retains much of its activity in vivo, expression of motility and chemotaxis genes is lowered in the degUD55N mutant as well as biofilm formation. We present evidence indicating that acetyl phosphate is an efficient phosphodonor for DegU in vitro and probably in vivo as well, based on the phenotypes of a L. monocytogenesΔptaΔackA mutant. Our findings indicate that phosphorylation enhances DegU activity, and that acetyl phosphate levels in the cell may play a part in controlling the activity of this orphan response regulator in the absence of the DegS kinase.
Phosphorylation of DegU
In order to study the role of DegU phosphorylation in Listeria, we first determined whether DegU of L. monocytogenes could be phosphorylated by the B. subtilis DegS histidine kinase even though the cognate kinase is not present in L. monocytogenes. The corresponding coding sequences were cloned in plasmid pET28/16, creating translational fusions adding a carboxy-terminal extension containing six histidine residues and placing the genes under the control of an inducible T7 bacteriophage promoter, and the overproduced proteins were purified by immobilized metal affinity chromatography (IMAC) in a single step using Ni-NTA agarose columns (see Experimental procedures). As shown in Fig. 1A, the Bacillus DegS and Listeria DegU proteins were obtained with a purity greater than 95% and displayed the expected apparent molecular masses of approximately 45.9 and 26.8 kDa respectively (Fig. 1A, lanes 2 and 3).
As shown in Fig. 1B, autophosphorylation of the DegS histidine kinase was demonstrated by incubating the protein with [γ-32P]-ATP, followed by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and autoradiography (Fig. 1B, lane 1; see Experimental procedures). When the B. subtilis DegS and L. monocytogenes DegU purified proteins were incubated together, phosphotransfer to DegU was observed upon addition of [γ-32P]-ATP followed by SDS-PAGE and autoradiography (Fig. 1B, lanes 2–4). Phosphorylation of DegU by DegS reached equilibrium in approximately 30 min (Fig. 1B, lane 4). No difference in phosphorylation efficiency was observed whether the two proteins were pre-incubated before the addition of [γ-32P]-ATP or if DegS was first autophosphorylated and then added to DegU (data not shown). This demonstrates that even though the DegS kinase is not present in L. monocytogenes, the heterologous B. subtilis kinase is able to efficiently phosphorylate DegU, and also indicates that the purified L. monocytogenes DegU protein is correctly folded and active as a phosphoacceptor.
B. subtilis DegS overproduction in L. monocytogenes leads to high-level expression of motility and chemotaxis genes
In order to investigate whether in vivo phosphorylation of DegU influences motility and chemotaxis gene expression, we introduced the B. subtilis degS gene into L. monocytogenes and overexpressed it from the multicopy plasmid pMK4. Overexpression of the plasmid-borne degS gene from its native promoter was verified by primer extension (data not shown). Total RNA was isolated from L. monocytogenes LM1000 (EGDe/pMK4), LM1002 (ΔdegU/pMK4) and LM1004 (EGDe/pMK4degS) cells grown in rich medium [brain–heart infusion (BHI)] to an OD600 of 0.6 at 25°C or 1.2 at 37°C.
Listeria monocytogenes chemotaxis and motility genes are expressed at 25°C but not at 37°C, in a DegU-dependent fashion. Quantitative real-time PCR (qRT-PCR) analysis was used to examine motility and chemotaxis gene expression in the presence of DegS. As shown in Fig. 2A, overexpressing B. subtilis DegS in Listeria strongly increased the relative transcript levels of motility and chemotaxis genes (flaA, cheA and motB), between three- and fivefold. The effect of DegS is strictly DegU-dependent, as shown in Fig. 2B, since increased flaA expression due to DegS is lost when the degU gene is deleted (strain LM1005). Transcription of the flaA gene, encoding flagellin, was also examined by primer extension, following RNA extraction after growth at 25°C or 37°C. As shown in Fig. 2CflaA was expressed during growth in BHI at 25°C in the parental strain EGDe but not in the LM1001 ΔdegU mutant, and was strongly increased in strain LM1004 when DegS was overproduced in vivo (Fig. 2C); however, no expression of flaA was observed at 37°C (Fig. 2C), as previously described (Peel et al., 1988), consistent with the reported repression of motility gene expression by MogR at 37°C (Shen and Higgins, 2006).
In order to test whether the DegS-dependent increase in motility gene expression was correlated with enhanced motility, we compared the LM1000 (EGDe parental strain with the control vector pMK4), LM1004 (EGDe/pMK4degS) and LM1005 (ΔdegU/pMK4degS) strains by motility plate assays. When DegS was present in the EGDe strain, migration was slightly faster at 25°C, giving colony halo diameters on soft agar motility assay plates that were consistently 13% greater than those of the control strain (Fig. 2D, Student's t-test P-value 1 × 10−4). The effect of DegS specifically requires the presence of DegU, since increased motility due to DegS is lost when the degU gene is deleted (strain LM1005) (Fig. 2D).
Taken together, these results suggest that in vivo phosphorylation of L. monocytogenes DegU by DegS of B. subtilis strongly upregulates motility and chemotaxis gene expression and also has a positive effect on bacterial motility.
Listeria DegUD55N is no longer phosphorylated by DegS
In order to determine whether DegU phosphorylation plays a role in vivo in Listeria despite the absence of the DegS kinase, we inactivated the DegU phosphorylation site by replacing the aspartate residue at position 55 by asparagine. The corresponding markerless mutation was introduced by a single-base-pair change in the Listeria chromosomal degU gene, using site-directed mutagenesis by strand overlap extension PCR (SOE-PCR) and the pMAD plasmid (Arnaud et al., 2004; see Experimental procedures). This residue corresponds to aspartate residue 56 in DegU of B. subtilis which we have previously established as its phosphorylation site (Dahl et al., 1992). In order to verify that Listeria DegUD55N is no longer phosphorylated, we overproduced and purified the protein as well as the B. subtilis DegU and DegUD56N response regulators for the control reactions.
The overproduced proteins were purified in a single step by IMAC using Ni-NTA agarose columns (see Experimental procedures) and analysed by SDS-PAGE, and displayed the expected apparent molecular mass of approximately 26.8 kDa (Fig. 3A).
As shown in Fig. 3B, DegU of L. monocytogenes was phosphorylated by DegS to the same extent as B. subtilis DegU, following incubation with [γ-32P]-ATP, SDS-PAGE and autoradiography, but the two mutationally modified proteins, Listeria DegUD55N and Bacillus DegUD56N, could not be phosphorylated by DegS, indicating that aspartate 55 is the likely phosphorylation site in Listeria.
Inactivation of the DegU phosphorylation site lowers expression of motility and chemotaxis genes
To investigate the in vivo role of DegU phosphorylation in Listeria, we studied the effect of inactivating the DegU phosphorylation site on motility and chemotaxis gene expression. As shown in Fig. 4A, expression of chemotaxis and motility genes as followed by qRT-PCR was lowered in strain LM1006 (degUD55N): flaA and motB expression were decreased 2.5-fold, whereas cheA expression was also reduced but only by approximately 20%. The decrease in flaA expression in strain LM1006 (degUD55N) was confirmed by primer extension analysis as shown in Fig. 4B. When we introduced the multicopy pMK4degS plasmid into strain LM1006 (degUD55N), flaA expression levels remained 2.5-fold lower than the parental EGDe strain (data not shown). This confirms that the DegS-dependent 4.5-fold increase in flaA expression shown in Fig. 2A requires phosphorylation of DegU on aspartate residue 55 by DegS and that inactivating this site downregulates motility and chemotaxis genes and prevents DegS-dependent activation.
In order to determine whether the decrease in motility and chemotaxis gene expression is correlated with altered cell motility, we compared strains EGDe, LM1006 (degUD55N) and LM1001 mutant (ΔdegU) on motility plate assays. As shown in Fig. 4C strain LM1006 clearly migrates more slowly than the parental EGDe strain, giving colony halo diameters on soft agar motility assay plates that were consistently 43% smaller than those of the control strain (Fig. 4C, Student's t-test P-value 1 × 10−9), whereas cells migrated faster when DegS was produced in Listeria.
These results indicate that although the DegUD55N protein still retains much of its activity as a transcriptional activator (Fig. 4A), inactivation of the DegU phosphorylation site decreases motility and chemotaxis gene expression as well as overall cell motility. This indicates that phosphorylation does play a role in controlling DegU activity in L. monocytogenes even though the cognate DegS kinase is absent. Alternative phosphorylation pathways could involve either cross-talk with a non-cognate kinase, or phosphorylation by a small chemical phosphodonor such as acetyl phosphate.
DegU is phosphorylated by acetyl phosphate in vitro
There are very few examples of physiologically relevant cross-phosphorylation of response regulators by a non-cognate histidine kinase, as evolutionary constraints have selected mechanisms for preserving high-specificity recognition determinants between cognate TCS pairs (Hoch and Varughese, 2001; Mukhopadhyay and Varughese, 2005; Laub and Goulian, 2007; Skerker et al., 2008). In contrast, acetyl phosphate has been proposed to play a role as a global signalling molecule in bacteria and shown to act as a high-energy phosphodonor for several, but not all TCS response regulators, with a larger ΔG° of hydrolysis than ATP (Lukat et al., 1992; McCleary et al., 1993; McCleary and Stock, 1994; Wolfe et al., 2003; Wolfe, 2005; Klein et al., 2007). Utilization of acetyl phosphate by DegU has not been previously reported. In order to test whether this can occur, DegU phosphorylation was analysed by high-performance liquid chromatography coupled to electrospray mass spectrometry (HPLC-MS) (Fig. 5). When DegU is incubated with ATP and DegS (Figs 1B, 3B and 5B), the mass spectrum shows the expected phosphorylated protein (peak marked with an asterisk in Fig. 5B) with a mass increase of 80 Da as compared with the control unphosphorylated protein (Fig. 5A). A second peak is also observed in Fig. 5B, with a mass decrease of 98 Da as compared with the phosphorylated product (marked with a triangle, Fig. 5B), due to spontaneous intramolecular dephosphorylation of the phosphorylated protein species, resulting in a succinimidyl derivative, as previously described (Napper et al., 2003). Aspartyl group derivative structures are indicated in each panel for the different forms of DegU, alongside the corresponding peaks.
The same two peaks which correspond to the expected masses of phosphorylated and dephosphorylated products are detected following incubation of DegU in the presence of acetyl phosphate as the phosphoryl group donor (Fig. 5C), indicating that DegU is phosphorylated in vitro using acetyl phosphate. To determine whether this involves the same site as kinase-mediated phosphorylation, DegUD55N was also incubated with acetyl phosphate. Under these conditions, the protein remained unphosphorylated (Fig. 5C), showing that aspartate residue 55 is required for phosphorylation by acetyl phosphate as well as by phospho-DegS.
Molecular mass analyses indicate that all the proteins lack the N-terminal methionine residue, as they are in agreement with the theoretical masses of the different proteins without methionine at the N-terminus, as previously observed (Stock et al., 1985; Sanders et al., 1989). Previous studies in E. coli have suggested that excision of the N-terminal methionine is catalysed by aminopeptidase and depends on the length of the side-chain of the second amino acid residue in the polypeptide chain (Hirel et al., 1989).
These data are the first reported evidence that acetyl phosphate can serve as a phosphodonor for DegU, using the same site as ATP-dependent kinase-mediated phosphorylation, aspartate residue 55.
Acetyl phosphate levels play an important role in controlling motility, chemotaxis and biofilm formation
Synthesis of acetyl phosphate occurs through the acetate dissimilation pathway, generated as a high-energy intermediate during conversion of acetyl-CoA to acetate through the reversible activity of two enzymes, phosphotransacetylase (Pta) and acetate kinase (AckA) (Wolfe, 2005). Acetyl phosphate levels in E. coli can reach 3–5 mM under certain growth conditions, a concentration that is considered to be more than sufficient for efficient phosphotransfer to response regulators (Klein et al., 2007). Reasoning that in the absence of DegS, which as we have previously shown can also act as a DegU phosphoprotein phosphatase (Dahl et al., 1992), acetyl phosphate levels may contribute to phosphorylation-dependent control of DegU activity in L. monocytogenes, we set out to examine the effects of acetyl phosphate depletion on the DegU regulon.
A markerless L. monocytogenesΔptaΔackA mutant (LM1009) was constructed in two steps using the pMAD plasmid (see Experimental procedures) in order to completely block acetyl phosphate synthesis. We then examined DegU-dependent phenotypes in the resulting mutant strain. As shown in Fig. 6A by qRT-PCR analysis on RNA from cells grown at 25°C, motility and chemotaxis gene expression is lowered in the ΔptaΔackA mutant, to levels that are comparable to those in strain LM1006 (degUD55N), in agreement with the hypothesis that acetyl phosphate could act as a phosphodonor for DegU in vivo as well. Motility was also examined, showing that strain LM1009 (ΔptaΔackA) migrated only slightly slower than the parental EGDe strain (Fig. 6B) giving colony halo diameters on soft agar motility assay plates that were consistently 14% smaller than those of the parental strain (Fig. 6B, Student's t-test P-value 8 × 10−7), but remained larger than those of strain LM1006 (degUD55N) (Fig. 4C). Among other phenotypes, we noted that the LM1009 (ΔptaΔackA) mutant was able to grow efficiently at high temperatures, up to 44°C, whereas the EGDe parental strain cannot, and also that the mutant exhibited a slight growth defect in BHI medium, but not in modified Welshimer's broth (MWB) defined medium (data not shown).
We have previously shown that DegU is required for efficient biofilm formation by L. monocytogenes (Gueriri et al., 2008). In order to study the effect of acetyl phosphate and DegU phosphorylation on Listeria biofilm formation, we tested the capacity of strains EGDe, LM1004 (EGDe/pMK4degS), LM1006 (degUD55N) and LM1009 (ΔptaΔackA) to adhere to plastic surfaces. Cells were grown in microtitre plates in MWB medium at 25°C (see Experimental procedures) and biofilm formation was assessed by measuring OD595 following biomass staining with 0.1% crystal violet. Accumulated biomass was measured over time, after 0, 10, 24, 36, 40 and 48 h, and the results after 40 h are shown in Fig. 7.
As shown in Fig. 7, there was no significant difference in biofilm formation between strains EGDe and LM1004 (EGDe/pMK4degS), suggesting that increased DegS-dependent phosphorylation of DegU does not affect biofilm formation. However, strains LM1006 (degUD55N) and LM1009 (ΔptaΔackA) were both less efficient in biofilm formation, with reproducible adherent biomass decreases of 2.2- and 3.3-fold respectively (Fig. 7). As observed for chemotaxis and motility gene expression, this reduced biofilm formation is another common phenotype observed either when inactivating the DegU phosphorylation site or when lowering acetyl phosphate levels in the cell, strongly suggesting the two processes are connected. This indicates that DegU phosphorylation may play a role in biofilm formation and that this may be in part linked to acetyl phosphate levels in the cell.
Our results indicate a correlation between lowered intracellular acetyl phosphate levels and DegU-dependent phenotypes, suggesting that phosphorylation by acetyl phosphate could play an important role in controlling DegU activity in L. monocytogenes.
Phosphorylation of TCS response regulators is widely accepted as the sine qua non requirement for their activation. Indeed, there are few examples of truly orphan response regulators, and fewer still instances where phosphorylation by a non-cognate histidine kinase occurs in vivo in a physiologically relevant manner, as these systems have instead evolved to ensure stringent interaction specificity among the cognate protein pairs (Matsubara et al., 2000; Hoch and Varughese, 2001; Mukhopadhyay and Varughese, 2005; Howell et al., 2006; Laub and Goulian, 2007; Skerker et al., 2008). Yet rather than behaving as simple ON/OFF binary switches, accessory signal inputs and additional regulatory factors such as phosphatases often act to confer analogue rheostat-type behaviour to otherwise digital relays (Msadek, 1999). The B. subtilis DegS/DegU TCS is one of the rare exceptions to the response regulator phosphorylation dogma, since DegU acts as a molecular switch, active in both its unphosphorylated and phosphorylated forms, each one controlling distinct sets of genes (Msadek et al., 1990; 1995; Dahl et al., 1992; Msadek, 1999).
In L. monocytogenes, a facultative intracellular pathogen closely related to B. subtilis, DegU is an orphan response regulator, as the degS gene is missing from the genome (Gueriri et al., 2008). We have previously shown that DegU of L. monocytogenes binds directly upstream from its own promoter region to negatively autoregulate its own synthesis, and that it also positively regulates motility and flagellar synthesis by binding upstream from the motB operon, which includes the gene encoding the GmaR antirepressor of flagellar synthesis (Gueriri et al., 2008). DegU is a highly pleiotropic regulator in B. subtilis, controlling competence gene expression, degradative enzyme and antibiotic synthesis, as well as motility and chemotaxis (Msadek et al., 1995; Msadek, 1999). We have shown that DegU is equally pleiotropic in Listeria, as it is required not only for flagellar synthesis, chemotaxis and motility, but also for growth at high temperature and in synthetic medium RPMI-1640, as well as biofilm formation (Gueriri et al., 2008). Although the DegU response regulator of Listeria has attracted a considerable amount of interest in the past few years (Knudsen et al., 2004; Williams et al., 2005a,b; Shen and Higgins, 2006; Shen et al., 2006; Gueriri et al., 2008), little attention has been paid to the role of phosphorylation and signalling in the control of DegU activity.
In this study we have used three different approaches to examine the role of DegU phosphorylation in Listeria. We demonstrated that B. subtilis DegS, the cognate histidine kinase of DegU, can phosphorylate the heterologous DegU protein of L. monocytogenes to the same extent as that of Bacillus in vitro. By then introducing and expressing the B. subtilis degS gene into Listeria, we showed that transcription of DegU regulon genes was strongly increased, and that this effect required DegU. We inactivated the DegU phosphorylation site by replacing aspartate residue 55 with asparagine, and showed that DegUD55N is no longer phosphorylated in vitro, while the strain carrying the chromosomal degUD55N point mutation showed lowered expression of DegU-dependent genes and reduced biofilm formation. Reasoning that in the absence of DegS and any orphan histidine kinases in Listeria, DegU might be using a kinase-independent mode of phosphorylation, we set out to examine whether acetyl phosphate was being used as a phosphodonor. Indeed, it has been known for some time that acetyl phosphate can act as a high-energy phosphodonor for response regulators, and in vivo levels of this metabolite have been linked to flagellar synthesis, biofilm formation and capsule synthesis (Lukat et al., 1992; Mccleary et al., 1993; Mccleary and Stock, 1994; Wolfe et al., 2003; Fredericks et al., 2006; Klein et al., 2007). We therefore constructed a ΔptaΔackA mutant of L. monocytogenes, effectively blocking in vivo acetyl phosphate synthesis, and showed that the mutant shared many common phenotypes with the Listeria degUD55N mutant, including lowered expression of flagellar synthesis and motility genes and diminished biofilm formation. Taken together, these results strongly suggest that DegU activity is enhanced by phosphorylation and that this response regulator may use acetyl phosphate as a phosphodonor in vivo, linking its activation to the metabolic status of the cell. However, despite several attempts we were unable to inactivate the ackA gene when an intact copy of the pta gene was present, whereas the mutation was readily obtained in the Δpta background. These results strongly suggest that the ackA gene is essential in L. monocytogenes in the presence of an intact pta gene, presumably due to the accumulation of large amounts of acetyl phosphate which are toxic to the cell. Therefore, although the evidence linking acetyl phosphate levels with DegU phosphorylation in vivo is compelling, we cannot exclude the alternative possibility that phenotypes shared between the ΔptaΔackA and degUD55N mutants of Listeria might be due to disruption of some general property of the Pta–AckA pathway, such as excretion of acetate, production of ATP or turnover of acetyl-CoA.
By comparing transcription levels in different genetic backgrounds of one of the major DegU-dependent genes in Listeria, flaA, which encodes flagellin, it is quite clear that its expression is directly linked to intracellular amounts of phosphorylated DegU. Indeed, as shown in Fig. 8, in the absence of DegU, flaA is no longer expressed. When either the phosphorylation site of DegU is inactivated (degUD55N) or acetyl phosphate synthesis is blocked (ΔptaΔackA), flaA is expressed at a low level, between a third and one-half of the levels seen in the parental EGDe wild-type strain, where both the DegU and acetyl phosphate synthesis pathways are intact (Fig. 8). Finally, when the heterologous DegS kinase is produced in Listeria, flaA transcription is increased approximately fivefold, and this effect requires DegU with an intact phosphorylation site. This strongly suggests that DegU activity in the cell functions as a rheostat, with gradually increasing levels of the phosphorylated form of DegU linked to progressively higher transcription levels of target genes (Fig. 8). This is very similar to the recently proposed model for DegU activation in B. subtilis, with the unphosphorylated form positively controlling genetic competence, low levels of DegS-independent phosphorylated DegU activating swarming, DegS-phosphorylated DegU being required for complex colony architecture and biofilm formation and a high level of phosphorylated DegU activating extracellular enzyme synthesis while inhibiting complex colony architecture, swarming motility and genetic competence (Kobayashi, 2007; Verhamme et al., 2007), with the caveat that acetyl phosphate levels would be playing a role in DegU phosphorylation in Listeria rather than the DegS kinase which is missing.
Preliminary phylogenetic analyses of the Firmicute phylum suggest that the degS gene was lost in Listeriaceae, rather than acquired in Bacillaceae, and a detailed investigation of this evolutionary event is in progress (S. Gribaldo, Unité de Biologie Moleculaire du Gene chez les Extrêmophiles, Institut Pasteur, Paris, pers. comm.). Because the major lifestyle difference between Listeria and Bacillus resides in its ability to function as an intracellular pathogen, it is conceivable that the activity of the DegS kinase might be incompatible with some specific step of the infectious process and as a result has been counterselected, resulting in the ‘use it or lose it’ loss of the degS gene. This is reminiscent of the incompatibility of an active PlcR regulator with the sporulation process, which is an essential aspect of pathogenesis in Bacillus anthracis, leading to inactivation of the plcR gene in B. anthracis strains carrying the pXO1 virulence plasmid (Mignot et al., 2001).
DegU is somewhat unique among response regulators as it activates transcription of competence genes in its unphosphorylated form (Dahl et al., 1992; Msadek et al., 1995; Msadek, 1999). As we show here, although phosphorylation enhances Listeria DegU activity, the unphosphorylated protein retains much of its activity and is able to activate flagellar synthesis and motility gene expression. We have also shown through in vitro DNA/protein interaction experiments that DegU binds upstream from its own promoter to negatively autoregulate its own synthesis (Gueriri et al., 2008) but no difference in binding efficiency was observed when DegU was first phosphorylated by DegS using unlabelled ATP (data not shown). In B. subtilis, DegU acts together with ComK as an antirepressor of Rok and CodY, allowing comK expression to be activated, creating a positive feedback loop allowing expression of competence genes and establishing development of the competence bistability state (Hamoen et al., 2000; Smits et al., 2005; 2007). In Listeria, DegU is known to positively control expression of gmaR, encoding a bifunctional O-GlcNac transferase which acts as an antirepressor, antagonizing MogR repression activity to restore expression of flagellar synthesis and motility genes (Grundling et al., 2004; Shen and Higgins, 2006; Shen et al., 2006). We have shown that DegU binds directly to the promoter region of the motB-gmaR operon, indicating it likely controls gmaR expression directly (Gueriri et al., 2008). We have pointed out previously (Gueriri et al., 2008) the intriguing observation that the ComK binding site sequence, AAAA-N5-TTTT (Hamoen et al., 1998; 2000), is the identical sequence on the antiparallel strand that is bound by MogR, TTTT-N5-AAAA (Shen and Higgins, 2006; Shen et al., 2006). We therefore suggest that the reason DegU is active in its unphosphorylated form could be due to its unusual role as a response regulator, acting in an antirepressor mechanism (Smits et al., 2007) by aiding ComK in Bacillus in preventing repression by Rok and CodY, and assisting GmaR in Listeria by relieving MogR-dependent repression of flagellar synthesis genes.
Bacterial strains and growth media
Listeria monocytogenes and B. subtilis strains used in this study are shown in Table 1. E. coli K12 strain DH5α™[F- (ϕ80lacZΔM15) Δ(lacZYA-argF) U169 recA1 endA1 hsdR17 (rk−, mk+) phoA supE44 λ-thi-1 gyrA96 relA1] (Invitrogen) was used for cloning experiments, and E. coli strain BL21 λ DE3 (Studier and Moffatt, 1986) (Novagen) for protein overproduction and purification. E. coli strains were grown in Luria–Bertani medium and transformed by electroporation, with selection on plates supplemented with ampicillin (100 μg ml−1) and kanamycin (25 μg ml−1) when appropriate (Sambrook et al., 1989). L. monocytogenes EGDe and its derivatives were grown in BHI (Difco), tryptic soy broth (TSB) (Difco) or MWB minimal medium (Premaratne et al., 1991) and transformed by electroporation, with selection on BHI plates supplemented with chloramphenicol (10 μg ml−1) or erythromycin (1 μg ml−1) when necessary. B. subtilis strains were grown in LB medium. Bacterial motility assays were carried out by spotting a 3 μl inoculum for each strain from liquid cultures grown in TSB to an OD600 = 2 onto TSB plates containing 0.25% agar (TSA) and incubated at 25°C, and bacterial motility was measured as the diameter of the bacterial swarm over time, after 9, 24, 34, 48 and 53 h. Biofilm formation was assayed as previously described (Gueriri et al., 2008). Motility and biofilm assays were repeated between 6 and 10 times and were highly reproducible, systematically giving the same results.
Table 1. Bacterial strains and plasmids used in this study.
pET28/16 derivative for overproduction of L. monocytogenes DegUD55N
pET28/16 derivative for overproduction of B. subtilis DegU
pET28/16 derivative for overproduction of B. subtilis DegUD56N
pET28/16 derivative for overproduction of B. subtilis DegS
Plasmid and genomic DNA were prepared as previously described (Gueriri et al., 2008). T4 DNA ligase and restriction enzymes (New England Biolabs), PCR reagents and Pwo thermostable DNA polymerase (Roche) were used according to the manufacturer's recommendations. Nucleotide sequencing of plasmid constructs was carried out by Genome Express-Cogenics.
For screening recombinant bacteria (E. coli or L. monocytogenes), single colonies were toothpicked into microtubes and bacteria were disrupted using a microwave oven (1000 W for 5 min) and used directly as templates for colony PCR.
Plasmids and mutant construction
Oligonucleotides used in this study were synthesized by Sigma-Proligo and their sequences are listed in Table 2. Plasmid pMK4 (Sullivan et al., 1984) was used to express the degS gene of B. subtilis in Listeria. An EcoRI/BamHI PCR-generated DNA fragment (1485 bp) corresponding to the entire B. subtilis degS gene with its upstream promoter region was amplified from genomic DNA of B. subtilis strain 168 using oligonucleotides OSA1 and OSA9 and cloned between the corresponding restriction sites of plasmid pMK4 to give plasmid pMK4degS.
Table 2. Oligonucleotides used in this study.
flaA primer extension
flaA primer extension
pMAD DNA sequence
pMAD DNA sequence
DegU B. subtilis overproduction
DegU B. subtilis overproduction
degS expression in L. monocytogenes
degS expression in L. monocytogenes
DegU or DegUD55N L. monocytogenes overproduction
DegU or DegUD55N L. monocytogenes overproduction
DegUD56N B. subtilis overproduction
DegUD56N B. subtilis overproduction
ackA DNA sequence
ackA DNA sequence
pta DNA sequence
pta DNA sequence
16S rRNA qRT-PCR
16S rRNA qRT-PCR
The markerless L. monocytogenesΔptaΔackA mutant (LM1009) was constructed in two steps. Two DNA fragments, of 643 and 782 bp, were generated by PCR using oligonucleotide pairs HD97/HD98 and HD99/HD100, respectively, corresponding to the chromosomal DNA regions located directly upstream and downstream from the pta gene. These DNA fragments were digested with SalI and ligated together, PCR was performed on the ligation mixture to amplify the resulting 1413 bp fragment using oligonucleotides HD97/HD100, which was then cloned between the EcoRI and BamHI sites of the pMAD vector (Arnaud et al., 2004), resulting in plasmid pMAD Δpta. The plasmid was introduced by electroporation into L. monocytogenes strain EGDe, and transformants were selected at 30°C on BHI plates containing erythromycin and 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) (50 μg ml−1). Integration and excision of pMAD Δpta was performed as previously described (Arnaud et al., 2004), with a non-permissive growth temperature of 42°C, yielding strain LM1008 (Δpta) in which the entire pta coding sequence was removed. The gene deletion was confirmed by PCR amplification with oligonucleotides HD101 and HD102, which hybridize to chromosomal DNA sequences located upstream and downstream from the pta gene.
The same strategy was used to remove the coding sequence of the ackA gene. Plasmid Pmad ΔackA was constructed by cloning two PCR-generated DNA fragments corresponding to the chromosomal DNA regions located directly upstream and downstream from the ackA gene (662 and 843 bp, oligonucleotide pairs HD89/HD90 and HD93/HD92 respectively) between the EcoRI and BamHI sites of the pMAD vector (Arnaud et al., 2004). Plasmid pMADΔackA was introduced by electroporation into L. monocytogenes strain LM1008 (Δpta), yielding L. monocytogenes strain LM1009 (ΔptaΔackA) following integration and excision of the plasmid. The chromosomal ackA deletion was confirmed by PCR using oligonucleotides HD126/HD94.
Site-directed mutagenesis by SOE-PCR (Ho et al., 1989) was used to construct the degUD55N mutant (LM1006 strain). Two DNA fragments (292 and 302 bp) with overlapping 3′ and 5′ ends were generated by PCR with oligonucleotide pairs HD84/HD86 and HD85/HD88. Oligonucleotides HD85 and HD86 are complementary and carry a single nucleotide mismatch, changing codon 55 from an aspartate codon (GAT) to an asparagine codon (AAT) through a G→A transition (Table 2). The purified DNA fragments were mixed in equal amounts and used as a DNA matrix for SOE-PCRs with oligonucleotides HD84 and HD88. The full-length resulting 564 bp DNA fragment was then cloned between the EcoRI and BamHI sites of the pMAD vector (Arnaud et al., 2004) resulting in pMADdegUD55N. E. coli clones were screened by colony PCR using oligonucleotide pair OP74/OP75, which hybridize within the pMAD vector, to identify those harbouring the recombinant plasmid. Plasmid pMADdegUD55N was transformed into L. monocytogenes, generating strain LM1006 (degUD55N) following integration and excision. As a result of the point mutation changing the aspartate 55 codon to asparagine, a new restriction site was generated for SspI. The presence of this restriction site within the degU coding sequence was used to identify positive clones by PCR amplification of the region using oligonucleotide pair HD84/HD88 followed by restriction with SspI, and the DNA fragment corresponding to the mutagenized chromosomal degU coding region was sequenced in order to verify the presence of the intended single-base-pair change.
Plasmid pET28/16 (Chastanet et al., 2003), a derivative of plasmid pET28a (Novagen), was used for production of the B. subtilis DegS, DegU and DegUD56N proteins, and the L. monocytogenes DegU and DegUD55N proteins, using plasmids pETDegS, pETDegUBsu, pETDegUBsuD56N, pETDegULmo and pETDegULmoD55N respectively. These were constructed by cloning PCR-generated NcoI/XhoI DNA fragments corresponding to the B. subtilis degS (1491 bp, oligonucleotides TM178/TM183), degU (720 bp, oligonucleotides TM180/TM182) or degUD56N (704 bp, oligonucleotides degUBsuF/degUBsuR) coding sequences or the L. monocytogenes degU or degUD55N coding sequence (698 bp, oligonucleotides degUgenF/degUgenR) between the NcoI and XhoI sites of plasmid pET28/16, replacing the stop codons with an XhoI restriction site. Genomic DNA of B. subtilis strain QB4414 (degU146D56N) (Dahl et al., 1991) or L. monocytogenes strain LM1006 (degUD55N) was used to amplify the mutated degU alleles.
Overproduction and purification of DegS and DegU
pET28/16 derivatives designed to overproduce DegS and DegU proteins were introduced into E. coli strain BL21 λ DE3, in which the T7 RNA polymerase gene is under the control of the inducible lacUV5 promoter, which also carries a pREP4 plasmid derivative allowing coproduction of the GroESL chaperonin in order to optimize recombinant protein solubility (Amrein et al., 1995). The resulting strains were grown in 2 l of LB medium at room temperature, expression was induced during the mid-exponential growth phase by addition of 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG), and incubation was pursued for 4 h. Purification of the DegS and DegU proteins by single-step IMAC was then performed as previously described (Dubrac and Msadek, 2004).
Phosphorylation reaction mixtures (9 μl) contained 1 μg of purified DegU and/or DegS protein in phosphorylation buffer [100 mM Tris-HCl pH 8.0, 200 mM KCl, 4 mM MgCl2, 0.5 mM dithiothreitol, 0.1 mM EDTA, 3.5% (v/v) glycerol]. The reaction was initiated by adding 1 μl of an ATP mixture containing 4 μCi of [γ-32P]-ATP (Perkin Elmer) diluted with unlabelled ATP to give a final concentration of 2.5 μM and incubated at 25°C. Reactions were stopped by adding 2 μl of SDS loading buffer and analysed by SDS-PAGE on 12% acrylamide gels, followed by autoradiography.
In vitro phosphorylation of Listeria DegU by acetyl phosphate
DegU (20 μg, 2.97 × 10−2 pM) was incubated with 1 mM ATP and DegS (10 μg, 8.9 × 10−3 pM) or 25 mM acetyl phosphate in phosphorylation buffer in 25 μl of reaction mixtures for 2 h at 37°C. The DegU modified protein (DegUD55N) was incubated with 25 mM acetyl phosphate in the same conditions as a negative control. The wild-type DegU used as a control was incubated in the same conditions except that the phosphoryl group donor was omitted.
The reaction was diluted twofold in water. Aliquots (3 μl, 0.4 mg ml−1) were withdrawn and analysed by HPLC (Alliance, Waters) coupled to a mass spectrometer (Q-Tof Micromass, Waters) equipped with an electrospray ionization interface and operated in a positive mode. The samples were injected on a C18 column (XBridge BEH300, 3.5 μ, 100 × 2.1 mm, Waters) and eluted with a linear gradient of 0–100% acetonitrile (0.05% formic acid) in water (0.1% formic acid) over 20 min, at 0.35 ml min−1. Mass analysis was performed on the protein peaks which were eluted at ∼55% acetonitrile.
Primer extension reactions
Total RNA was extracted from cells in mid-exponential phase or beginning of stationary phase, during growth at 25°C or 37°C, respectively (Gueriri et al., 2008), and used as a template for primer extension reactions using a radiolabelled flaA-specific oligonucleotide (HD72, see Table 2) as previously described (Chastanet et al., 2001). The corresponding dideoxy chain termination DNA sequencing reactions were carried out by using the same oligonucleotide primer and PCR-amplified fragments corresponding to the flaA (407 bp) upstream region (oligonucleotide pair HD73/HD72, respectively; Table 2) with the Sequenase PCR product sequencing kit (USB).
cDNA synthesis and qRT-PCRs
RNA samples for qRT-PCR reactions were treated with DNase I using the TURBO DNA-free reagent (Ambion, Austin, TX) in order to eliminate residual contaminating genomic DNA. cDNA synthesis and qRT-PCRs were then carried out as previously described (Dubrac et al., 2007), using the L. monocytogenes rpoB gene as an internal standard (Schmittgen and Zakrajsek, 2000) and specific oligonucleotide pairs for each gene (see Table 2). Experiments were performed twice on three independent biological replicates. For comparisons with the ΔptaΔackA mutant, the 16S rRNA gene was used as an internal standard since it was consistently more stably expressed than rpoB in strain LM1009 (ΔptaΔackA), under the experimental conditions tested (Tasara and Stephan, 2007).
This work was supported by research funds from the European Commission (Grant BACELL Health, LSHG-CT-2004-503468), the Centre National de la Recherche Scientifique (CNRS URA 2172) and the Institut Pasteur (Programme Transversal de Recherche N° 18 and Grand Programme Horizontal N° 9). We would like to thank Nadia Benaroudj for critical reading of the manuscript, Alan J. Wolfe for helpful discussion, Olivier Poupel for assistance with qRT-PCR experiments and statistical analysis and Simonetta Gribaldo for helpful discussion and phylogenetic analyses.