ABSTRACT
Listeriae take up glucose and mannose predominantly through a mannose class phosphoenolpyruvate:carbohydrate phosphotransferase system (PTSMan), whose three components are encoded by the manLMN genes. The expression of these genes is controlled by ManR, a LevR-type transcription activator containing two PTS regulation domains (PRDs) and two PTS-like domains (enzyme IIAMan [EIIAMan]- and EIIBGat-like). We demonstrate here that in Listeria monocytogenes, ManR is activated via the phosphorylation of His585 in the EIIAMan-like domain by the general PTS components enzyme I and HPr. We also show that ManR is regulated by the PTSMpo and that EIIBMpo plays a dual role in ManR regulation. First, yeast two-hybrid experiments revealed that unphosphorylated EIIBMpo interacts with the two C-terminal domains of ManR (EIIBGat-like and PRD2) and that this interaction is required for ManR activity. Second, in the absence of glucose/mannose, phosphorylated EIIBMpo (P∼EIIBMpo) inhibits ManR activity by phosphorylating His871 in PRD2. The presence of glucose/mannose causes the dephosphorylation of P∼EIIBMpo and P∼PRD2 of ManR, which together lead to the induction of the manLMN operon. Complementation of a ΔmanR mutant with various manR alleles confirmed the antagonistic effects of PTS-catalyzed phosphorylation at the two different histidine residues of ManR. Deletion of manR prevented not only the expression of the manLMN operon but also glucose-mediated repression of virulence gene expression; however, repression by other carbohydrates was unaffected. Interestingly, the expression of manLMN in Listeria innocua was reported to require not only ManR but also the Crp-like transcription activator Lin0142. Unlike Lin0142, the L. monocytogenes homologue, Lmo0095, is not required for manLMN expression; its absence rather stimulates man expression.
IMPORTANCE Listeria monocytogenes is a human pathogen causing the foodborne disease listeriosis. The expression of most virulence genes is controlled by the transcription activator PrfA. Its activity is strongly repressed by carbohydrates, including glucose, which is transported into L. monocytogenes mainly via a mannose/glucose-specific phosphotransferase system (PTSMan). Expression of the man operon is regulated by the transcription activator ManR, the activity of which is controlled by a second, low-efficiency PTS of the mannose family, which functions as glucose sensor. Here we demonstrate that the EIIBMpo component plays a dual role in ManR regulation: it inactivates ManR by phosphorylating its His871 residue and stimulates ManR by interacting with its two C-terminal domains.
INTRODUCTION
Glucose is recognized as a preferred carbon source for numerous bacteria. When multiple carbohydrates are present in a growth medium, glucose represses the synthesis of the enzymes required for the utilization of other carbon sources and is therefore the first to be transported and metabolized. This phenomenon is known as carbon catabolite repression. Glucose is taken up by bacterial cells either by ion-driven permeases, such as GlcP (1) and GlcU (2), or via the phosphoenolpyruvate (PEP):glycose phosphotransferase system (PTS), with the latter being the most common bacterial glucose transporter. The PTS is composed of four proteins or protein domains, which form a phosphorylation cascade, and one or two membrane-spanning proteins (3). The phosphorylation cascade starts with the PEP-requiring autophosphorylation of a histidyl residue in enzyme I (EI) (4). Phosphorylated EI (P∼EI) then transfers the phosphoryl group to histidine 15 in HPr, the second general PTS protein (Fig. 1). In the next step, P∼His-HPr phosphorylates a histidyl residue in a carbohydrate-specific EIIA component; bacteria usually contain several EIIA components. P∼EIIA subsequently phosphorylates a cysteyl or histidyl residue in the EIIB component of the same carbohydrate specificity. In the last step, P∼EIIB passes the phosphoryl group to a carbohydrate molecule that is bound to the membrane protein EIIC or sometimes both EIIC and EIID. Phosphorylation of the substrate decreases its affinity for the membrane protein(s), and the phosphorylated carbohydrate is thus released into the cytoplasm (3). In total, there are seven different PTS families, distinguished on the basis of their amino acid sequences and the structures of the soluble, substrate-specific EII proteins or domains and membrane-spanning EIIC (5). Of these proteins, at least two are involved in the transport of glucose. The first is the glucose class PTS, in which the glucose transporter is usually called PtsG (6). The second is the mannose class PTS, which, despite its name, contains PTSs that play an important role in the transport of glucose. Glucose-specific transporters in this class are usually composed of a fused cytoplasmic EIIA/EIIB protein and the two membrane components EIIC and EIID; however, in some glucose-transporting mannose-type PTSs (referred to here as glucose/mannose class PTSs), EIIA and EIIB are distinct proteins. Members of the glucose/mannose class PTS family frequently transport more than one substrate. For example, the PTSMan transports mannose, glucose, and in some bacteria also several other carbohydrates albeit with low efficiency. This PTS family shows the greatest variability in transported substrates; its members also catalyze the uptake of less-common carbon sources such as gluconate (7), d-glucosaminate (8), sorbose (9), d-ribitol (10), lacto-N-biose, and galacto-N-biose (11).
Schematic presentation of the PTS phosphorylation cascade and PTS-mediated sugar transport and phosphorylation and model of regulation of B. subtilis LevR, a ManR homologue, via phosphorylation by the two PTS components HPr and EIIBLev. While phosphorylation by HPr at the EIIAMan domain stimulates LevR activity, phosphorylation by EIIBLev inhibits it.
Listeria monocytogenes, the causative agent of the foodborne disease listeriosis (12), is a Gram-positive soil bacterium belonging to the order Bacillales. It has the capacity to switch from a saprophytic to a virulent life-style (13). In order to exploit the numerous carbon sources in soil derived from decaying plants and other organisms, L. monocytogenes possesses a large number of carbohydrate transporters (14), many of which belong to the PTS (15). The utilization of an efficiently metabolized PTS substrate was suggested to indicate to the bacterium its presence in a saprophytic environment, because it represses the expression of most virulence genes (16). Cellobiose and glucose are the two carbohydrates that exhibit the strongest repressive effect on virulence gene expression (17). The main L. monocytogenes virulence genes are organized on a pathogenicity island (18). Their expression requires the transcription activator PrfA (positive regulatory factor A), a member of the Crp/Fnr family of transcription regulators (19). PrfA activity and prfA expression, among others, respond to changes in the oxidative state (20) and in temperature (21), respectively. The efficient utilization of PTS substrates was reported to inhibit the activity of PrfA without altering the expression of its gene, and PTS components were suspected to regulate PrfA activity (22). Repression of the L. monocytogenes virulence genes therefore does not follow the general catabolite control protein A- and P-Ser-HPr-mediated carbon catabolite repression mechanism operative in firmicutes (23–25).
L. monocytogenes transports glucose mainly via a glucose/mannose class PTS that is composed of the three proteins EIIABMan (ManL), EIICMan (ManM), and EIIDMan (ManN). When the PTSMan is inactivated, L. monocytogenes still utilizes glucose at ∼30% of the efficiency of the wild-type strain (26). This remaining activity is due mostly to a second PTS of the mannose class (27), known as the PTSMpo (Mpo stands for “mannose permease one”). This PTS is composed of EIIAMpo (MpoA), EIIBMpo (MpoB), EIICMpo (MpoC), and EIIDMpo (MpoD) (15, 26). In L. monocytogenes, several other PTS and non-PTS transporters are able to take up glucose with low affinity, most notably a PtsG-like permease of the glucose class, which is encoded by lmo0027. However, the glucose transport activities of these transporters can be detected only in mutants that lack the two major glucose PTSs (15, 26) or, for non-PTS transporters, the general PTS protein EI (26).
While the expression of the manLMN operon is inducible by glucose and mannose, the mpo operon is constitutively expressed (15, 26) but is σB dependent (28, 29). Interestingly, the gene upstream of the mpo operon encodes the LevR-type transcription activator ManR, which has been reported to be controlled by the general PTS proteins and by the soluble components of the PTSMpo (26, 27). However, the detailed regulatory mechanisms remain obscure. LevR-type regulators usually function as enhancer-binding proteins for σ54-requiring transcription units expressed from −12 and −24 promoters (30, 31). Although the man and mpo operons contain nearly identical −12 and −24 promoters, ManR controls the expression of only the manLMN operon (26). LevR-type regulators are composed of an N-terminal helix-turn-helix motif, which is followed by a domain resembling the central domain of NifA/NtrC-type transcription regulators and four regulatory domains with potential phosphorylation sites for PTS proteins: two PTS phosphorylation domains (PTS regulation domain 1 [PRD1] and PRD2), an EIIAMan-like domain, and an EIIBGat-like domain (Fig. 1) (for a review, see reference 32). Based on genetic data, the regulatory domains of ManR have been predicted to be the targets of both phosphorylation by (33) and interactions with (26) PTS components.
In Listeria innocua, the lin0142 gene, which is located upstream of the manLMN operon, encodes a Crp- and PrfA-like transcription regulator that was identified as a second activator of the manLMN operon in this species. A knockout mutant carrying an insertion of the Tn917 transposon in this gene was isolated based on its resistance to class IIa bacteriocins (34). Similar resistance, together with the inhibition of glucose transport, has also been reported for listeriae and some other firmicutes as a result of mutations that affect the membrane components of the PTSMan (31, 35). In this case, however, the Tn917 insertion occurred not in the man operon but in the region between the presumed −35 and −10 promoter sites upstream of lin0142. Nevertheless, quantitative reverse transcription-PCR (qRT-PCR) experiments revealed that the Tn917 insertion mutant exhibits a very low manLMN expression level (34), which is probably responsible for its bacteriocin resistance. Normal manLMN expression and bacteriocin sensitivity were restored when the mutant was complemented with the lin0142 gene. In L. monocytogenes, the gene upstream of manLMN, lmo0095, encodes a protein that resembles Lin0142, but whether this gene plays a role in man expression was not known.
In order to better understand the utilization of glucose by L. monocytogenes and its repressive effect on virulence gene expression, we carried out an in-depth study of ManR- and Lmo0095-mediated regulation of manLMN expression. We report here that, depending on its phosphorylation state, the glucose/mannose-specific PTS component MpoB either stimulates ManR via a direct interaction with its two C-terminal domains or inhibits the transcription activator by phosphorylating a conserved histidine in the regulatory domain PRD2. We also found that, unlike L. innocua Lin0142, L. monocytogenes Lmo0095 is not required for manLMN expression.
MATERIALS AND METHODS
Bacterial strains, plasmids, and growth conditions.The L. monocytogenes and Escherichia coli strains and plasmids used in this study are listed in Table 1. For growth studies, β-d-glucuronidase assays, and qRT-PCR experiments, L. monocytogenes cells were grown in chemically defined minimal medium (MM) (36), which contains leucine, isoleucine, arginine, methionine, valine, cysteine (each at 100 mg/liter), glutamine (600 mg/liter), riboflavin and biotin (each at 0.5 mg/liter), thiamine (1.0 mg/liter), thioctic acid (0.005 mg/liter), and ferric citrate (88 mg/liter) as a source of Fe3+. MM was complemented with glucose or glycerol at the indicated concentrations and contained 5 μg/ml erythromycin when appropriate. In solid MM (37), the glucose, cellobiose, and glycerol concentrations ranged from 2.5 to 50 mM, and erythromycin (5 μg/ml) was added when appropriate. For all other purposes, L. monocytogenes cells were grown in brain heart infusion (BHI) broth, and E. coli cells were grown in Luria-Bertani (LB) medium, at 37°C with shaking.
Bacterial strains and plasmids used in this study
The concentrations of antibiotics used for the selection of E. coli transformants were 100 μg/ml ampicillin and 20 μg/ml kanamycin; 5 μg/ml erythromycin was used for the selection of L. monocytogenes integrants. X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) and X-GlcA (5-bromo-4-chloro-3-indolyl-β-d-glucuronide) were added to solid medium at a concentration of 100 μg/ml to screen for L. monocytogenes clones that express the lacZ gene during mutant construction and to monitor PrfA-controlled expression of a Φ(Phly-gus) fusion in strain AML73 (38) and mutants derived from it, respectively.
Plasmid pMAD (39) confers resistance to erythromycin and ampicillin and contains both a thermosensitive pE194 replication origin (40) and the promoterless Geobacillus stearothermophilus bgaB gene, which encodes a thermostable β-galactosidase (under the control of the constitutive Staphylococcus aureus clpP promoter). The pAT18 plasmid confers resistance to erythromycin and contains the multiple-cloning site and replication origin of pUC18 as well as the pAMβ1 replication origin (41). For the purification of His-tagged proteins, the corresponding genes were cloned into the expression vector pQE30 (Qiagen). In order to increase the amount of soluble wild-type and mutant ManR proteins, the strain used for the production of the His-tagged proteins was also transformed with plasmid pREP4(GroES/GroEL) (42).
DNA techniques.General molecular biological and genetic experiments were carried out as previously described (43). Chromosomal and plasmid DNAs were extracted from L. monocytogenes strains with the DNeasy blood and tissue kit and the Qiaprep Spin Miniprep kit, respectively (Qiagen), and purified with the Wizard kit (Promega). Restriction enzymes and T4 DNA ligase were obtained from New England BioLabs and were used as recommended by the manufacturer. DNA sequencing was carried out by Cogenics (Essex, United Kingdom). E. coli and L. monocytogenes strains were transformed by electroporation with a Gene Pulser apparatus (Bio-Rad Laboratories) (44). PCRs were performed with a Mastercycler apparatus (Eppendorf, Hamburg, Germany) programmed for 30 cycles, each composed of three steps, 30 s at 95°C, 30 s at 55°C, and 1 to 5 min at 72°C, followed by a final extension step at 72°C for 10 min.
Construction of a L. monocytogenes Δlmo0095 mutant.A mutant that carries an in-frame deletion of the lmo0095 gene, which is located upstream of the manLMN operon, was constructed by homologous recombination, using the pMAD vector (39) and strain AML73 (38). For this purpose, we amplified two slightly overlapping 500-bp DNA fragments corresponding to the upstream and downstream regions of lmo0095 by using genomic DNA as a template and appropriate primer pairs (Table 2). In a second PCR, we used the two distal primers and the two PCR products as a template, which allowed the fusion of the two DNA fragments. The resulting amplicon was cloned via the BamHI and NcoI restriction sites added by PCR to the 5′ and 3′ ends, respectively, into the pMAD vector cut with the same enzymes. The plasmid thus obtained, pMAD-Δlmo0095, was electroporated into cells of L. monocytogenes strain AML73, and blue colonies able to grow at 42°C owing to the integration of the plasmid into the chromosome were selected on solid BHI medium containing X-Gal and erythromycin. One integration mutant was isolated and grown for several days at 30°C in liquid BHI medium to allow for the excision of the plasmid. Subsequently, several erythromycin-sensitive white colonies that had been obtained on solid BHI medium containing X-Gal were screened by PCR in order to identify clones in which the second recombination step had resulted in the deletion of the gene of interest. The correct in-frame deletion was confirmed by DNA sequencing.
List of oligonucleotides used in this study
Construction of various manR alleles.Several manR alleles were constructed, which encode a regulator in which one or two of the presumed phosphorylatable histidines were replaced with alanine: manR encoding a His-to-Ala change at position 506 [manR(His506Ala)], manR(His585Ala), manR(His871Ala), and manR(His585Ala/His871Ala). In the first step, the wild-type manR gene was amplified by PCR using appropriate oligonucleotides and genomic DNA of L. monocytogenes strain EGD-e as a template. The amplicon was cut with the restriction enzymes BamHI and SalI and cloned into the His tag expression vector pQE30, which had been cut with the same enzymes. For one clone, the correct sequence of the insert was verified by DNA sequencing, and the plasmid was designated pQE30-manR(PstI). In order to construct the various manR alleles, we first removed the PstI site that was located between SalI and HindIII in plasmid pQE30-manR(PstI) by cutting it with PstI. The manR gene also contains a PstI site, and therefore, digestion with this enzyme enabled the removal of a 635-bp fragment from the 3′ end of manR, which was replaced with the 635-bp fragment obtained by PCR amplification with oligonucleotides ManR_pUC18_PstI871 and ManR_pUC18_HindIII871 and chromosomal DNA as a template. The resulting plasmid was designated pQE30-manR. The use of oligonucleotide ManR_pUC18_HindIII871 enabled the conservation of the SalI site (important for the insertion of the manR alleles into plasmid pAT18 [see below]) but removed the PstI site of pQE30. Together with appropriate mutagenic primers (Table 2), oligonucleotides ManR_pUC18_PstI871 and ManR_pUC18_HindIII871 were also used for the construction of the manR(His871Ala) allele by first amplifying two DNA fragments that overlapped at the mutated codon; these fragments were subsequently used as the template for a second PCR amplification with only the distal oligonucleotides ManR_pUC18_PstI871 and ManR_pUC18_HindIII871. The resulting amplicon was cloned into pUC18, and following sequence verification, the DNA fragment was used to replace the 635-bp 3′ end of manR in plasmid pQE30-manR, thus producing pQE30-manR(His871Ala). Essentially the same technique was used for the construction of the manR(His506Ala) and manR(His585Ala) alleles. DNA fragments of 853 bp containing the manR(His506Ala) or manR(His585Ala) mutations were amplified by using primers ManR_pBC_EcoRV506 and ManR_pBC_PstI506 together with the corresponding mutagenic primers (Table 2). The amplicons were cut with EcoRV and PstI and cloned into plasmid pBC KS. The presence of the mutated codons was verified by DNA sequencing, and the EcoRV/PstI fragments were subsequently cloned into pQE30-manR(PstI) that had been cut with the same enzymes. In doing so, we removed a 1,482-bp fragment from the 3′ end of the manR gene and replaced it with one of the 847-bp amplicons. In the last step, the two manR alleles were completed by adding the 635-bp PstI/HindIII DNA fragment obtained from pQE30-manR. The doubly mutated manR(His585Ala/His871Ala) allele was obtained by replacing the 3′ part of the manR(His585Ala) allele with the PstI/HindIII fragment derived from pQE30-manR(His871Ala).
Complementation of the ΔmanR mutant.To construct the plasmids for complementation of the ΔmanR mutant, we first amplified the upstream region of the manR gene, including the entire intergenic region between manR (lmo0785) and lmo0786, using appropriate primers (Table 2) and genomic DNA of L. monocytogenes strain EGD-e as a template. The resulting 164-bp amplicon therefore contained the manR promoter and the Shine-Dalgarno sequence. It was cut with EcoRI and BamHI and inserted into vector pAT18, cut with the same enzymes. The resulting plasmid, pAT18-PmanR, was used to transform E. coli NM522, and one clone with the correct insert was isolated. In the second step, the various manR alleles present in pQE30 (see above) were excised with BamHI and SalI and inserted into pAT18-PmanR that had been cut with the same enzymes, thus allowing for the expression of the manR alleles from the natural promoter. Following the transformation of E. coli strain NM522 with the resulting plasmids and verification of the correct inserts by DNA sequencing, they were used to transform L. monocytogenes ΔmanR mutant strain EM1002 (26). Control strains were constructed by transforming strain AML73 and ΔmanR mutant strain EM1002 with the empty vector pAT18.
β-d-Glucuronidase assays.AML73 and the various mutants derived from it were grown in MM buffered with 100 mM morpholinepropanesulfonic acid (MOPS) (pH 7.0) and supplemented with 20 mM glucose or glycerol. To carry out β-d-glucuronidase assays, bacteria were grown in 4 ml liquid medium until they reached the exponential phase, at which point cells were harvested, washed, and resuspended in 0.6 ml “gus” buffer (50 mM sodium phosphate [pH 7.0], 1 mM EDTA). Cells were lysed by shaking with glass beads, cell debris was removed by centrifugation, and 100- or 200-μl aliquots of the supernatant were used for the β-d-glucuronidase assay. β-d-Glucuronidase activity was measured by using p-nitrophenyl-β-d-glucuronide as the substrate. The reaction was stopped by the addition of 800 μl of a 0.4 M Na2CO3 solution to the mixture.
In order to detect β-d-glucuronidase activity on solid medium, a preculture was grown in BHI broth. When it reached stationary phase, cells were washed with MM, and 5-μl aliquots were spotted onto buffered MM agar that contained different concentrations of glucose, cellobiose, or glycerol (2.5, 5, 10, 20, and 50 mM) and 100 μg/ml X-GlcA. The plates were kept at 37°C for 72 h in order to allow for the hydrolysis of X-GlcA and the resulting formation of blue color by the strains that produced β-d-glucuronidase (38).
Overproduction and purification of His-tagged proteins.His-tagged Bacillus subtilis EI and HPr were overproduced and purified as previously described (45). For the purification of EIIAMpo and EIIBMpo, the corresponding genes mpoA and mpoB were amplified by PCR using appropriate primers (Table 2) and genomic DNA of L. monocytogenes strain EGD-e. The PCR products were cut with BamHI and SalI and cloned into the His tag expression vector pQE30, cut with the same enzymes. The resulting plasmids were used to transform E. coli strain NM522, and the correct DNA sequence was verified. For the overproduction and purification of His-tagged EIIAMpo and EIIBMpo, we followed a protocol described previously (45). Because ManR was found to form inclusion bodies, the pQE30-derived plasmids, which contained either wild-type manR or one of the four manR mutant alleles described above, were used to transform E. coli strain M15[pREP4(GroES/GroEL)]. The presence of the chaperones GroES/GroEL and the addition of 200 mM NaCl to the sample during the preparation of cell extracts and dialysis improved the solubility of ManR and its mutated derivatives. In addition, the strains containing the manR alleles were grown at 28°C. Once the culture had reached an optical density at 600 nm (OD600) of between 0.5 and 0.6, expression of the manR alleles was induced with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) during 3 h of incubation at 28°C.
Synthesis of [32P]PEP.[32P]PEP was synthesized from [γ-32P]ATP by using the pyruvate kinase isotope exchange reaction method at equilibrium, as described previously (46).
Protein phosphorylation assays.Phosphorylation of ManR was carried out with purified PTS proteins in 30-μl assay mixtures that contained 50 mM Tris-HCl (pH 7.4), 5 mM MgCl2, and 16 μM [32P]PEP (0.5 μCi; specific radioactivity, 3 Ci/mmol). The samples contained the following amounts of the different PTS components: 0.1 μg of EI, 0.2 μg of HPr, 0.1 μg of EIIAMpo, 0.3 μg of EIIBMpo, and either 0.2 μg of wild-type ManR, 0.9 μg of ManR(His506Ala), 0.6 μg of ManR(His585Ala), or 1 μg of ManR(His585Ala/His871Ala). The assay mixtures were incubated for 30 min at 37°C before the reactions were stopped by the addition of SDS sample buffer to the mixture. The samples were not heated in order to avoid hydrolysis of the heat-labile phosphoamidate bonds. The proteins were separated by electrophoresis on a denaturing 1% SDS–15% polyacrylamide gel. After drying, the gels were exposed to a storage phosphor screen (Storm).
Yeast two-hybrid experiments.We PCR amplified the mpoB and ptsH genes along with DNA fragments that encoded (i) the entire L. monocytogenes ManR protein, (ii) its four regulatory domains fused together, (iii) each separate regulatory domain, or (iv) each regulatory domain fused to its neighboring domain. For each reaction, we used appropriate primers (Table 2) and genomic DNA of L. monocytogenes strain EGD-e as a template. All PCR products thus obtained contained a BamHI restriction site at the 5′ end and a SalI site at the 3′ end. They were cloned into both the bait vector pGBDU (URA3) and the prey vector pGAD (LEU2), which had been cut with the same enzymes. The resulting plasmids were used to transform E. coli strain NM522, and the correct DNA sequence of all cloned fragments was confirmed. The bait constructs were subsequently introduced into Saccharomyces cerevisiae strain PJ69-4a, and the prey constructs were introduced into strain PJ69-4α (47). The bait- and prey-containing “a” and “α” strains, respectively, were mixed in rich yeast extract-peptone-dextrose (YEPD) medium in order to allow mating. The resulting diploids were first selected on solid synthetic complete (SC) medium that lacked leucine and uracyl and subsequently tested for their ability to express interaction phenotypes during growth in SC medium that lacked the appropriate amino acids (Leu and His) or nucleotides (Ade and Ura) (48). An interaction of the two proteins that are encoded by the inserts of the pGBDU and pGAD vectors usually allows growth of diploid S. cerevisiae strains on solid SC medium lacking Leu and His and on solid SC medium lacking Ade and Ura. False-positive interactions generated by the yeast two-hybrid system were eliminated experimentally, as previously described (49).
RNA isolation and qRT-PCR experiments.L. monocytogenes strains were grown to exponential phase in MOPS-buffered MM (pH 7.0) supplemented with either 25 mM glucose or glycerol. Total RNA was isolated from Listeria cells as previously described (26). qRT-PCR experiments were carried out by using the LightCycler instrument (Roche) with previously described manL-specific primers (26) and SuperScript III reverse transcriptase (Invitrogen), according to the protocol for the LightCycler Faststart DNA Master SYBR green I kit (Roche). Data were processed with Roche Molecular Biochemicals LightCycler software. In order to normalize transcript levels of the studied genes, the constitutively expressed rpoB gene was used as a reference (50).
RESULTS
Conserved histidines in PRD1 and the EIIAMan-like domain are essential for ManR activity.In a previous study, a L. monocytogenes mutant that lacked EI exhibited very low expression levels of the manLMN operon (26). This effect was proposed to be due to the absence of PEP-dependent, EI- and HPr-catalyzed phosphorylation, which allows binding of the transcription activator ManR to a presumed operator and expression of the manLMN operon (26, 33). The predicted operator can form an imperfect palindrome (TTTCTGTATCAATAGACGATACACTAA [imperfect palindrome shown in boldface type]) probably serving as an upstream activating sequence, because it resembles the LevR-binding site in B. subtilis (51). In L. innocua, whose ManR protein is 98% identical to that of L. monocytogenes, phosphorylation by P∼His-HPr was assumed to occur at His506, the first conserved histidine in PRD1. This prediction was based on the observation that the replacement of His506 in PRD1 with Ala caused a loss of ManR activity (33). However, this conclusion was contradictory to data from studies of the LevR proteins from B. subtilis (52) and Lactobacillus casei (53), which strongly resemble listerial ManR. Previously reported in vitro experiments with the LevR proteins had established that stimulation of phosphorylation by EI and HPr occurs at the conserved histidine in the EIIAMan-like domain (Fig. 1). In addition, in B. subtilis, the replacement of the first conserved histidine in PRD1 of LevR was found to cause increased transcription activation under both inducing and noninducing conditions (52).
In order to study the role of His506 in L. monocytogenes ManR in more detail, we complemented the previously constructed manR deletion mutant strain EM1002 (26) with wild-type manR, the His506Ala allele, and the allele in which the conserved histidine in the EIIAMan-like domain (His585) was replaced with an alanine. Strain EM1002 and its parental strain AML73, both transformed with the empty vector pAT18, were used for control experiments. We measured the expression levels of the manLMN operon in the various complemented strains by qRT-PCR. While complementation with the wild-type manR allele fully restored glucose induction of the manLMN operon in EM1002, the level of activity of ManR remained very low in glucose-grown strains that were complemented with either the manR(His506Ala) or manR(His585Ala) mutant allele (Fig. 2A). When grown on glycerol, the manLMN expression levels in all four strains remained very low (Fig. 2B). It therefore appears that both the first conserved histidine in PRD1 and the conserved histidine in the EIIAMan-like domain are essential for ManR activity.
Expression of the manLMN operon in L. monocytogenes wild-type strain AML73 and ΔmanR mutant strain EM1002 (26) transformed with either the empty pAT18 vector or pAT18 that carried various manR alleles. The following strains were used in this study: AML73 and EM1002 transformed with the empty pAT18 vector and EM1002 transformed with pAT18 that contained either manR, manR(His506Ala), manR(His585Ala), manR(His871Ala), or manR(His585Ala/His871Ala). The manLMN expression level in the various strains was determined by qRT-PCR and compared to that of the wild-type (wt) strain carrying the empty plasmid, which was set to 1. The cells were grown in MM that contained 5 μg/ml erythromycin and either 25 mM glucose (A) or 25 mM glycerol (B). The mean values of data from at least three independent experiments are presented, with standard deviations not exceeding 10%.
Phosphorylation of L. monocytogenes ManR by EI and HPr occurs at His585.To test whether both histidines are modified, we carried out in vitro phosphorylation assays with wild-type and mutant ManR proteins. For this purpose, we purified His-tagged L. monocytogenes wild-type and His506Ala and His585Ala mutant ManR proteins as well as His-tagged EI and HPr from B. subtilis (45); we then carried out in vitro phosphorylation assays with [32P]PEP. As expected, wild-type ManR was not phosphorylated when the phosphorylation mixture contained only EI or HPr (Fig. 3A, lanes c and d). Phosphorylation of ManR occurred only when both EI and HPr were simultaneously present (Fig. 3A, lane e). In addition, we observed that ManR(His506Ala) still became phosphorylated by [32P]PEP, EI, and HPr (Fig. 3B, lane d) but that ManR(His585Ala) did not (Fig. 3B, lane e). In summary, these results confirm that the activation of listerial ManR occurs via EI- and HPr-catalyzed phosphorylation at the conserved histidine (His585) in the EIIAMan-like domain.
In vitro phosphorylation of wild-type ManR and several ManR mutant proteins by [32P]PEP and various mixtures of the general PTS proteins EI and HPr and the PTSMpo components EIIAMpo and EIIBMpo. The samples for the phosphorylation experiments were separated on 1% SDS–10% polyacrylamide gels, which were dried and exposed for autoradiography. The samples loaded onto the gels contained the following proteins. (A) Lane a, EI; lane b, ManR; lane c, ManR and EI; lane d, ManR and HPr; lane e, EI, HPr and ManR. (B) Lane a, EI; lane b, EI and HPr; lane c, EI, HPr, and wild-type ManR; lane d, EI, HPr, and ManR(His506Ala); lane e, EI, HPr, and ManR(His585Ala). (C) All samples contained EI, HPr, and the indicated additional proteins. Lane a, wild-type ManR; lane b, EIIAMpo and ManR(His585Ala); lane c, EIIAMpo, EIIBMpo, and ManR(His585Ala); lane d, EIIAMpo, EIIBMpo, and ManR(His585Ala/His871Ala). Only the top part of the gel is shown, including the migration positions of EI (64 kDa) and ManR (107 kDa) (A and B). In lanes b to d of panel C, a strong radioactive band, corresponding to the EIIAMpo dimer (32 kDa), is also visible. The radioactive bands that correspond to HPr (10 kDa), the EIIAMpo monomer (16 kDa), and EIIBMpo (18 kDa) are cut off in the three panels.
Phosphorylation of L. monocytogenes ManR by P∼EIIBMpo occurs at His871 in PRD2.In previous studies of L. monocytogenes, it was shown that ManR is also regulated by the EIIA and EIIB components of the PTSMpo (15, 26). Indeed, the deletion of L. monocytogenes mpoA caused constitutive expression of the manLMN operon (26), suggesting that this protein is required for the phosphorylation (and, thus, the inhibition) of ManR. In L. innocua, constitutive manLMN expression was also observed when His871, the conserved histidine in PRD2 of ManR, was replaced with an alanine (33), thus providing evidence that His871 might be the site of phosphorylation by the soluble EIIMpo components. In order to test this hypothesis, we purified doubly mutated L. monocytogenes ManR(His585Ala/His871Ala) protein and carried out in vitro phosphorylation assays.
As reported above, ManR(His585Ala) is not phosphorylated by [32P]PEP, EI, and HPr (Fig. 3B, lane e), and the addition of EIIAMpo to the phosphorylation mixture made no difference in this regard (Fig. 3C, lane b). Only when EIIAMpo and EIIBMpo were included in the reaction mixture did ManR(His585Ala) become phosphorylated (Fig. 3C, lane c). ManR was not phosphorylated by [32P]PEP, EI, HPr, EIIAMpo, and EIIBMpo when both His585 and His871 were replaced with alanine (Fig. 3C, lane d). These results confirm the prediction that His871 is the site of negative phosphorylation by EI, HPr, EIIAMpo, and EIIBMpo.
The inhibiting His585Ala mutation dominates over the stimulating His871Ala replacement.In order to confirm the presumed negative effect of the PTS-mediated phosphorylation of His871 in PRD2 on ManR activity, we complemented the manR deletion mutant with plasmids that contained the manR(His871Ala) and the manR(His585Ala/His871Ala) alleles under the control of the manR promoter. As previously observed for a L. innocua ΔmanR mutant, complementation with the manR(His871Ala) allele resulted in constitutive expression of the manLMN operon (33). When grown in either glucose- or glycerol-containing MM, the L. monocytogenes manR mutant complemented with the manR(His871Ala) allele showed 3.5-fold or 5.5-fold-higher manLMN expression levels, respectively, than did the glucose-induced wild-type strain (Fig. 1B and 2A). In contrast, complementation with the doubly mutated manR(His585Ala/His871Ala) allele did not restore manLMN expression in either glucose- or glycerol-grown cells (Fig. 1B and 2A), suggesting that the His585Ala mutation is dominant over the His871Ala replacement.
Interaction with unphosphorylated EIIBMpo is required for ManR activation.Because EIIBMpo is required for the phosphorylation of ManR at His871 (and, thus, its inhibition) (Fig. 3C, lane c), it would be logical to expect that the deletion of the mpoB gene would also cause strong constitutive manLMN expression, similar to that which results from mpoA deletion. However, this is not the case: in a previous study, we reported that mutants that carried a deletion of mpoB or an inactivation of both mpoA and mpoB displayed very low levels of ManR activity (26). In that study, complementation of a ΔmpoB mutant with the wild-type mpoB allele restored the inducible manLMN expression of the wild-type strain, whereas complementation with the mpoB(His14Ala) allele, which produces an EIIBMpo protein that cannot be phosphorylated, caused constitutive ManR activity (26). It therefore appears that phosphorylated EIIBMpo inhibits ManR, whereas unphosphorylated EIIBMpo activates it. One possible explanation for EIIBMpo-mediated ManR activation was a direct interaction between the two proteins.
In order to test this hypothesis, we carried out yeast two-hybrid experiments. Growth of the diploid yeast strain that produced EIIBMpo and ManR on selective medium indeed confirmed that the two proteins interact with each other (Fig. 4A, circles). In addition, an interaction of ManR with itself was observed (Fig. 4A, box), indicating that ManR forms oligomers, most likely dimers. We then cloned DNA fragments that encode each of the four regulatory domains of ManR into the yeast two-hybrid vectors. Surprisingly, none of the domains interacted with EIIBMpo (Fig. 4A). However, we observed a strong interaction of the EIIAMan-like domain of ManR with itself (Fig. 4A, dotted circle) and a weaker interaction with the intact ManR (Fig. 4A, dotted boxes), suggesting that this domain contributes to ManR oligomerization. Indeed, several EIIA proteins of the mannose PTS family were reported to form dimers (54, 55). In Fig. 3C, a strong radioactive band migrating according to a molecular mass of 30 kDa, present in lanes b to d, shows that, if not heated prior to electrophoresis, the 16-kDa EIIAMpo migrates mainly as a dimer.
Yeast two-hybrid experiments with EIIBMpo (MpoB) and either intact ManR or its various domains. (A and B) Interactions of EIIBMpo were studied with intact ManR as well as each individual regulatory ManR domain (PRD1, PRD2, EIIBGat-like, and EIIAMan-like) (A) and with intact ManR as well as the four constructs of fused domains (B). (C) Schematic presentation of the domain fusions, including the four regulatory domains of ManR fused together (PRD1-PRD2), PRD1 fused to EIIAMan-like, EIIAMan-like fused to EIIBGat-like, and EIIBGat-like fused to PRD2. In panel B, we also tested the interaction of HPr with intact ManR and the various ManR fragments. Important protein-protein interactions are highlighted with boxes or circles. Control experiments with the empty pGBDU and pGAD vectors did not provide any positive interaction signals (data not shown).
To test whether EIIBMpo interacts with the N- or C-terminal half of ManR, we cloned DNA fragments that encompassed either the DNA-binding domain of ManR and its central domain with the ATP-binding site (amino acids 1 to 443) or the four C-terminal regulatory domains (PRD1-PRD2; amino acids 444 to 938) into the yeast two-hybrid vectors. An interaction of EIIBMpo was observed only with the four regulatory domains (Fig. 4B, boxes) and not with the N-terminal DNA-binding and central domains (data not shown). We therefore amplified DNA fragments that encoded each regulatory domain fused to either its preceding or following domain (Fig. 4C) and cloned them into the yeast two-hybrid vectors. We detected an interaction with EIIBMpo for one only combination, the EIIBGat-like domain fused to the C-terminal PRD2 (Fig. 4B, circles). In addition, all protein fragments that contained the EIIAMan-like domain of ManR interacted with each other, confirming that this domain plays a major role in the oligomerization of the transcription activator (Fig. 4B, dotted circles). EIIBMpo did not interact with ManR proteins that lacked the C-terminal PRD2 (data not shown), confirming that this domain is important for EIIBMpo binding.
Does phosphorylation of EIIBMpo prevent its interaction with ManR?As mentioned above, complementation of a ΔmpoB mutant with the mpoB(His14Ala) allele caused constitutive expression of the manLMN operon (26). This finding demonstrated that phosphorylation of EIIBMpo is not necessary for ManR activation. However, it is possible that the phosphorylation of EIIBMpo prevents its interaction with ManR. In order to test this possibility, we carried out yeast two-hybrid experiments with His14Asp mutant EIIBMpo and either the complete ManR protein, the four fused regulatory domains of ManR, or PRD2 fused to the EIIBGat-like domain. We found that the presumed phosphomimetic replacement of the phosphorylatable histidine in EIIBMpo indeed reduced its interaction with the four regulatory domains but had no effect on the interaction with the entire ManR protein (data not shown). We therefore decided to test whether the presumed phosphomimetic replacement would affect ManR activity. For this purpose, we complemented the ΔmpoB mutant with the mpoB(His14Asp) allele. Because of the His14Asp replacement, the complemented strain is not able to phosphorylate His871 in PRD2. There were thus two possible outcomes of this experiment: if mutant EIIBMpo interacted with ManR, manLMN expression should have been constitutive, while if the His14Asp replacement in EIIBMpo prevented this interaction, the activity level of the transcription activator should have been low. We found that the ΔmpoB mutant that was complemented with the mpoB(His14Asp) allele exhibited constitutive manLMN expression, because the operon was strongly expressed in cells grown on either glucose (Fig. 5) or glycerol (data not shown). The His14Asp replacement therefore does not prevent the interaction of EIIBMpo with ManR.
Expression of the manLMN operon in the ΔmpoB mutant (FA1017) (26) that was complemented with either wild-type mpoB or the mpoB(His14Ala) or mpoB(His14Asp) allele. The manLMN expression level was determined by qRT-PCR. Strain FA1017 transformed with the empty vector pAT18 was used as a negative control. The mean values of data from at least three independent experiments are presented, with standard deviations not exceeding 10%.
Effect of the manR mutations on virulence gene expression.Glucose is one of the carbohydrates that strongly represses virulence gene expression in L. monocytogenes (17, 26). In addition, components of the glucose-transporting PTSGlc in enterobacteriaceae (56) and the PTSMan in firmicutes (57) are involved in the general carbon catabolite repression mechanism. PTSMan components were therefore suggested to function as general virulence gene regulators in L. monocytogenes. However, the deletion of components of the L. monocytogenes PTSMan or PTSMpo prevented only glucose- and mannose-mediated virulence gene repression while leaving repression by other carbohydrates unaffected (26). We therefore tested whether ManR might play a general role in L. monocytogenes virulence gene repression. In the ΔmanR mutant complemented with the manR(His871Ala) allele, in which ManR is constitutively active, the Φ(Phly-gus) fusion was repressed already at low glucose concentrations (2.5 mM), similar to the wild-type strain. However, the ΔmanR mutant and its derivatives that were complemented with the manR(His506Ala), manR(His585Ala), or manR(His585Ala/His871Ala) allele demonstrated reduced repression of the Φ(Phly-gus) fusion by glucose at concentrations of up to 10 mM (Fig. 6A). These strains poorly expressed the manLMN operon. The relief from virulence gene repression, which was confirmed by measuring β-glucuronidase activity in crude extracts of the various strains (data not shown), is therefore probably due to slowed glucose utilization (26). At higher concentrations (20 and 50 mM), glucose inhibited PrfA activity in all strains tested (data not shown). In addition, the relief from repression was restricted to glucose, because cellobiose repressed virulence gene expression normally in all strains (data not shown). When grown in glycerol-containing MM, all tested strains expressed the Φ(Phly-gus) fusion, including the ΔmanR mutant and the strain that was complemented with the manR(His871Ala) allele, in which ManR is constitutively active (Fig. 6B). It therefore seems that ManR does not function as the general regulator of virulence gene expression in response to the presence of efficiently metabolizable carbon sources. Its deletion or inactivation causes only a specific relief from glucose repression.
Effect of manR deletion and complementation of the ΔmanR strain with various mutant manR alleles on PrfA-dependent expression of the Φ(Phly-gus) fusion. Cells were grown on solid MM that contained X-GlcA and various concentrations (2.5, 5, and 10 mM) of either glucose (A) or glycerol (B). The cells were incubated at 37°C for 72 h.
Deletion of lmo0095 stimulates manLMN expression.As mentioned above, the gene located upstream of the manLMN operon, lmo0095, encodes a protein that exhibits similarity to Crp- and PrfA-like transcription activators. In a previous study, insertion of the Tn917 transposon in the region between the presumed −35 and −10 promoter sites upstream of lin0142, the L. innocua homologue of the lmo0095 gene, was reported to strongly diminish manLMN expression (34). Similarly to ManR, Lin0142 was therefore thought to function as an activator of the manLMN operon. However, the signal that controls the activity of Lin0142 as well as the operator site to which it binds were unknown. We therefore decided to study the role of Lmo0095 in L. monocytogenes manLMN expression and virulence gene repression in more detail. We first tried to confirm that the Lin0142 homologue Lmo0095 is indeed required for L. monocytogenes manLMN expression by deleting the lmo0095 gene in strain AML73 and carrying out qRT-PCR experiments. Surprisingly, the manLMN expression level in the lmo0095 deletion mutant was 2-fold higher than that in the wild-type strain (Fig. 7). In addition, lmo0095 deletion did not relieve the repressive effect of glucose (2.5, 5, and 10 mM) on the expression of the Φ(Phly-gus) fusion (data not shown).
Expression of the manLMN operon in the Δlmo0095 mutant. The Δlmo0095 mutant was grown in MM that contained 25 mM glucose, and the manLMN expression level was determined by qRT-PCR and compared to those of wild-type strain AML73, ΔmpoB mutant strain FA1017 (26), and ΔmanR mutant strain EM1002 (26). The mean values of data from at least three independent experiments are presented, with standard deviations not exceeding 10%.
DISCUSSION
As previously reported in a study of ManR in L. innocua (34), we found that the activity of this transcription activator in L. monocytogenes was strongly inhibited by the replacement of His506, the second conserved histidine in PRD1, with an alanine. Because ManR needs to be phosphorylated by PEP, EI, and HPr in order to be active, this previous study proposed that His506 serves as the phosphorylation site for P∼His-HPr. However, our tests of this hypothesis instead identified His585 in the EIIAMan-like domain as the phosphorylation site for P∼His-HPr (Fig. 8). Nevertheless, in both listeriae, the second conserved histidine in PRD1 seems to be important for ManR activity. In the PRD-containing transcription activator MtlR from Geobacillus stearothermophilus (58) and B. subtilis (59), the second conserved histidine in PRD2 is necessary for EI- and HPr-catalyzed phosphorylation and therefore also for MtlR activity. This histidine was therefore proposed to play a catalytic role during the phosphoryl group transfer. However, our results suggest that His506 must perform a different function, because in ManR(His506Ala), His585 was still phosphorylated. It might play a role in the transduction of the activating signal from phosphorylated His585 to the DNA-binding or the ATP-hydrolyzing domain of ManR, or its replacement with alanine might simply induce structural changes leading to the inactivation of ManR, as was proposed previously for the equivalent mutation in B. subtilis LevR (52).
Schematic presentation of ManR regulation by PTS components in the presence and absence of glucose. In the absence of glucose or other efficiently metabolized carbohydrates (bottom), the PTS components are present mainly in a phosphorylated form. ManR is therefore phosphorylated by P∼His-HPr at His585 and by P∼EIIBMpo at His871. The inhibitory phosphorylation at His871 dominates over the activating phosphorylation at His585, and ManR is therefore inactive. In the presence of glucose, sufficient P∼His-HPr is formed, allowing the phosphorylation and activation of a major fraction of the ManR molecules. In contrast, EIIBMpo is present mostly in a dephosphorylated form, because its phosphoryl group is rapidly transferred to glucose bound to the EIIC and EIID components. His871 is therefore barely phosphorylated. In addition, probably only dephosphorylated EIIBMpo interacts with the two C-terminal domains of ManR and stimulates its transcription activator function. EIIBMpo and P∼EIIBMpo therefore carry out antagonistic functions in ManR regulation: P∼EIIBMpo phosphorylates ManR at His871 and thus inhibits its activity. In contrast, binding of unphosphorylated EIIBMpo to the two C-terminal domains renders ManR active.
Based on genetic data, L. innocua ManR was suggested to be inactivated by phosphorylation at His871 in PRD2 (34). Here, we confirm that L. monocytogenes ManR is indeed phosphorylated at His871 and show that this modification requires not only EI and HPr but also EIIAMpo and EIIBMpo (Fig. 8) and that the prevention of this phosphorylation leads to constitutive ManR activity. This result corroborates a previously reported finding that the deletion of EIIAMpo also leads to constitutive manLMN expression (26). Our data are also in agreement with data from previous studies carried out on LevR proteins from B. subtilis (52) and L. casei (53), which are also phosphorylated by EI, HPr, and the levan-specific EIIA and EIIB components at the conserved histidine in truncated PRD2. Based on the regulatory model that has been proposed for the LevR proteins (52, 53), dephosphorylation of His871 in L. monocytogenes ManR probably serves as an induction mechanism for manLMN expression. In the absence of glucose or mannose, most ManR molecules in a wild-type strain are probably phosphorylated at His871. Dephosphorylation of His871 and, consequently, ManR activation are triggered when P∼EIIBMpo donates its phosphoryl group to glucose or mannose during their uptake via the PTSMpo (Fig. 8).
In contrast, the absence of EI- and HPr-catalyzed phosphorylation at His585 of ManR probably serves as an autorepression mechanism, which prevents excessive utilization of glucose. This hypothesis is supported by the observation that growth on glycerol leads to 5-fold-higher expression levels of the manLMN operon in the ΔmanR mutant that was complemented with the manR(His871Ala) allele than in the glucose-grown wild-type strain (compare Fig. 2A and B). This means that in a wild-type strain grown in glucose, ManR is not fully active because a significant amount of the ManR molecules might not be phosphorylated by P∼His-HPr. This could be a consequence of the small amount of P∼His-HPr present in glucose-grown firmicutes (60, 61). Interestingly, a mutant in which both phosphorylatable histidines of ManR had been replaced with alanine exhibited a very low expression level of the manLMN operon. This result indicates that the loss-of-function effect that is caused by the absence of EI- and HPr-catalyzed phosphorylation at His585 dominates over the positive effect on ManR activity that is caused by the absence of phosphorylation at His871.
EIIBMpo carries out a dual function in ManR regulation: P∼EIIBMpo inhibits ManR by transferring its phosphoryl group to His871 in PD2, while unphosphorylated EIIBMpo activates the regulator by interacting with its EIIBGat-like domain and PRD2 (Fig. 8). EIIB components have previously been reported to interact with the transcription regulator MtlR of B. subtilis (62) and the E. coli antiterminator BglG (63). The regulation of these two PRD-containing proteins occurs via membrane sequestration, because the corresponding EIIBs are fused to the membrane-integral EIICs, whereas ManR activation probably occurs in the cytoplasm. While the interaction with the cognate EIIB stimulates MtlR (62), the interaction with P∼EIIB inhibits BglG (63). Because both effects, ManR activation by interaction with EIIBMpo and ManR inhibition by P∼EIIBMpo-mediated phosphorylation, involve PRD2, the EIIBMpo and P∼EIIBMpo interaction sites might be similar. It is possible that the EIIBGat-like domain of ManR contributes only to the binding of unphosphorylated EIIBMpo by specifically increasing the affinity for the dephosphorylated protein. Similarly, the binding of the EIIBMtl domain of MtlA to B. subtilis MtlR also required the EIIBGat-like domain (62). Interestingly, when yeast two-hybrid experiments were carried out with HPr, this domain interacted with neither the intact ManR protein nor any of the fused regulatory domains (Fig. 4B), although P∼His-HPr phosphorylates the EIIAMan-like domain of the regulator. It is possible that we did not observe any binding between these proteins because the transient interactions during phosphoryl group transfer are usually not detected in yeast two-hybrid experiments (64).
The activation of ManR by EIIBMpo does not require phosphorylation of the PTS component, because complementation with the mpoB(His14Ala) allele led to constitutive ManR activity (26). Surprisingly, the presumed phosphomimetic His14Asp replacement in EIIBMpo also caused constitutive manLMN expression. One explanation for this result might be that the His14Asp replacement in EIIBMpo does not truly mimic its phosphorylated form, as has been observed for other phosphoproteins. An alternative explanation could be that only the phosphorylation of both proteins, His14 in EIIBMpo and His871 in PRD2 of ManR, causes sufficient electrostatic repulsion to prevent the EIIBMpo/ManR interaction. This concept is supported by the observation that the His871Ala mutant ManR protein is 3.5-fold more active than wild-type ManR. The elevated ManR activity might partly be due to an enhanced interaction of EIIBMpo with the mutated transcription activator, which cannot be phosphorylated at position 871. Similarly, in the mpoB(His14Asp) mutant, ManR is also not phosphorylated at His871, and His14Asp mutant EIIBMpo might therefore still interact with ManR. In the wild-type strain, phosphorylation at His871 occurs when no inducer is present, and the interaction of unphosphorylated EIIBMpo with ManR probably serves as a second induction mechanism.
The L. innocua Crp/Fnr-like transcription regulator Lin0142 has previously been identified as a second activator of manLMN expression (34). In contrast, deletion of the equivalent L. monocytogenes protein, Lmo0095, increased ManR activity 2.5-fold. We have no explanation for these contradictory results. Nevertheless, we noticed that Lmo0095 and Lin0142 share a much lower degree of sequence identity (only 59%) than other homologous protein pairs of the two organisms, which typically exhibit between 93% and 100% sequence identity. For example, the three components of the PTSMan, as well as ManR, exhibit between 98% and 100% sequence identity. Moreover, the manLMN genes also share ∼97% sequence identity. The conserved region also includes the −12 and −24 promoters that precede manL, but the remaining intergenic regions between manL and the upstream gene lmo0095 (in L. monocytogenes) or lin0142 (in L. innocua) exhibit only 61% sequence identity. We therefore suspect that the poor conservation of these sequences in the two species is responsible for the functional differences of the two proteins Lmo0095 and Lin0142.
In summary, we tested the prediction that ManR activation requires phosphorylation by P∼His-HPr at His506 in PRD1 and found that this was not the case. Instead, we identified His585 in the EIIAMan-like domain as the phosphorylation site. We established that EIIBMpo plays a dual role in ManR regulation that depends on its phosphorylation state. P∼EIIBMpo inhibits ManR by phosphorylating His871 in PRD2, while unphosphorylated EIIBMpo stimulates ManR activity by interacting with its EIIBGat-like domain and PRD2. Finally, while the L. innocua Crp/Fnr-like transcription regulator Lin0142 enhances manLMN expression, its L. monocytogenes equivalent, Lmo0095, inhibits it.
ACKNOWLEDGMENTS
We thank Nancy Freitag for providing us with L. monocytogenes strain AML73 and Nguyen Cao for technical assistance with the β-glucuronidase assays.
This research was supported by Agence Nationale de la Recherche (ANR) project no. ANR-09-BLAN-0273. A.C.Z. received a fellowship from Campus France and the Ministère de l'Enseignement Supérieur et de la Recherche Scientifique de la Côte d'Ivoire.
FOOTNOTES
- Received 8 December 2014.
- Accepted 6 February 2015.
- Accepted manuscript posted online 17 February 2015.
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.
