INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Two-component signal transduction proteins are commonly employed by bacteria to respond to changes in environmental conditions (11, 32, 38). In their simplest forms, two-component systems consist of histidine kinases and response regulators. Histidine kinases transduce environmental cues into intracellular signals by interacting with and modifying response regulator proteins. Signal processing involves the transfer of phosphate between a histidine residue within the kinase and an aspartate residue located within the response regulator.

A large family of response regulator proteins has been identified through genetic and genomic analyses of many bacteria (32, 42). These proteins generally consist of multiple domains and are characterized by a conserved receiver domain, which contains the site of aspartyl phosphorylation, and an output domain, which regulates transcription. Response regulators have been subdivided into families based on their output domains (31, 42). The pattern of conserved residues within the receiver domain defines this superfamily and strongly supports the idea that these domains have a common structure and potentially employ a common mechanism of activation. The three-dimensional structures of several receiver domains have been determined (CheY, NtrC, FixJ, SpoOF, NarL, CheB, and PhoB) (2, 3, 5, 7, 27, 35, 36, 41, 43). In each of these proteins, the receiver domain has a doubly wound / topology consisting of a central five-stranded parallel -sheet (1 to 5) surrounded by five -helices (1 to 5). A prominent feature of the receiver domain is an acidic pocket, which is found at the C-terminal edge of the -sheet. This pocket contains the phosphoaccepting aspartate residue. The structures of intact multidomain response regulators NarL and CheB have recently been determined (2, 5). Although the structures of all receiver domains are similar, these proteins do not have the same domain-packing arrangements.

The mechanism(s) by which the phosphorylation signal originating within the receiver domain is propagated to the output domain is not known. However, several recent studies of activated receiver domains have demonstrated a common structural change involving the repositioning of a conserved tyrosine or phenylalanine residue in 5 from a solvent-exposed position into a hydrophobic pocket (3, 4, 9, 15). This conserved change leads to slightly different structural alterations in each of the receiver domains studies.

A well-characterized adaptive response in Escherichia coli that employs a two-component signaling pathway is triggered by inorganic phosphate (P_i) limitation (44). The phosphate response permits cells to acquire P_i with high affinity and to utilize alternate phosphorus sources. The genes under phosphate control are positively regulated and are called the Pho regulon. When P_i becomes limiting, transcription is initiated from the promoters of the regulon; for example, the expression of alkaline phosphatase, the product of the phoA gene, is stimulated more than 150-fold (45).

The signaling proteins that operate on the cytoplasmic side of the inner membrane are two-component regulators PhoR and PhoB. PhoR is a histidine kinase that receives environmental input from the high-affinity phosphate transporter (20, 21). When phosphate levels are low, PhoR donates a phosphoryl group to a conserved aspartate residue within response regulator PhoB (18). PhoB is a soluble 229-amino-acid protein that consists of two domains: an ~125-amino-acid N-terminal receiver domain and an ~100-amino-acid C-terminal output domain that binds DNA and interacts with the ⁷⁰ subunit of RNA polymerase (16, 34). The output domain is a member of the winged-helix-turn-helix family of transcription factors, which is represented by OmpR from E. coli (22, 23). The three-dimensional structure of the output domain of PhoB was recently solved (30). Upon phosphorylation, PhoB forms a dimer and its affinity for target DNA sequences, called pho boxes, is increased, which leads to enhanced levels of transcription (8, 18, 24).

We have recently demonstrated that the receiver domain of PhoB negatively regulates its output domain (6). We have shown that the liberated output domain of PhoB binds to pho box DNA more tightly and activates transcription better than the intact unphosphorylated protein. In this paper, we extend those studies to show that the 5 helix of the receiver domain is involved in the interdomain interactions that negatively control the output domain of PhoB. We also provide data that suggest that the phosphorylation-generated activation signal requires the 5-5 loop and the 5 helix to be propagated to the output domain.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, growth conditions, and plasmids. The E. coli strains and plasmids used in this study are listed in Table 1. Cells were grown in either Luria-Bertani (LB) medium which was supplemented with ampicillin (100 µg/ml) or in modified glucose-morpholinepropanesulfonic acid (MOPS) minimal medium containing 5.0 mM KH₂PO₄ and ampicillin (100 µg/ml) (25, 28).

Deletion mutagenesis. Inverse PCR was performed on plasmid pDE1 to generate all deletions used in this study. All synthetic oligonucleotides were purchased from Life Technologies (Rockville, Md.) and are listed in Table 2. The primers were designed to flank the region to be deleted, and each contained a BamHI site so that the linear PCR product could be ligated after restriction digestion with BamHI. The digested fragments were joined using T4 ligase to create each of the pMP plasmids. The design of each of the deletions was based on the predicted secondary structure of PhoB from the crystal structures of the homologous proteins CheY and OmpR (22, 37). Mutagenized plasmid DNA was transformed into E. coli DH5 for plasmid maintenance and into E. coli DWE1 for phenotypic evaluation. Deletion mutagenesis was verified by DNA sequencing using a LI-COR (Lincoln, Nebr.) 4000L automated sequencer. The sequence information was compared to the phoB sequence previously published (19).

Alkaline phosphatase assays. Alkaline phosphatase assays were performed as described previously (48).

Construction of chimeric genes. The cheY/phoB chimeric genes were constructed using a "gene SOEing" process previously described (12). Gene fragments were generated from cheY and phoB, which contained an overlapping 15-bp sequence. For each construct, four primers were used (Table 2), an A::C primer pair for CheY and a D::B primer pair for PhoB. The 15-bp complementary region was created by using a 30-mer for the D primer that contained at its 5' end 15 bases complementary to the C primer for CheY and 15 residues complementary to the PhoB coding sequence at its 3' end. The C and D primers specify the location of the splice site between CheY and PhoB. The A primer contains an internal NdeI site that provides the start codon for cheY, and the B primer contains an internal BamHI site downstream of the phoB gene termination codon. The A::C and D::B amplification products were separated from primers by agarose gel electrophoresis and were purified using the Qiaex II resin from Qiagen Inc. (Valencia, Calif.). They were then combined, denatured, and reannealed under PCR conditions. The overlaps were extended with Taq polymerase, and the new chimeric gene was further amplified using the A::B primer pair. This amplification product was purified following agarose gel electrophoresis using Qiaex II resin, was digested with NdeI and BamHI, and was cloned into expression vector pJES307 (26) to give plasmids pT7-Ch1 and pT7-Ch3.

Overexpression, purification, and analysis of chimeric proteins. The expression and purification of insoluble proteins were performed as previously described (26). The phosphotransfer assays were conducted by incubating [³²P]phospho-CheA in the presence of various phosphoacceptors. CheA was phosphorylated at room temperature for 15 min in a 20-µl reaction mixture containing 10 µM CheA, 10 mM MgCl₂, 25 mM Tris-HCl, pH 7.2, and 0.2 µM [-³²P]ATP. Phosphotransfer reactions were initiated by adding 2.5 µl of the CheA phosphorylation reaction mixture to a tube containing 6 µl of a phosphoacceptor. These reaction mixtures were incubated for 2 min at room temperature, and the reactions were stopped by the addition of sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample loading buffer containing 25 mM EDTA. The final concentrations of the phosphoacceptors in the phosphotransfer reaction mixtures were as follows: CheY, 12 µM; PhoB, 15 µM; Ch1, 2 µM; Ch3, 12 µM. Samples were separated on an SDS-15% polyacrylamide gel, after which the gel was dried and exposed to X-ray film for autoradiography.

Western immunoblotting. E. coli DWE1 cells containing various plasmids were grown overnight in LB media supplemented with ampicillin (100 µg/ml). Equivalent amounts of cellular protein, adjusted according to the optical densities of the overnight cultures, were separated on an SDS-15% PAGE gel and transferred onto a nitrocellulose membrane using the Mini Trans-Blot transfer cell (Bio-Rad) according to the manufacturer's instructions. The membrane was blocked overnight at 25°C in TBS (20 mM Tris-HCl [pH 7.5], 500 mM NaCl)-3% gelatin-5% (wt/vol) nonfat dried milk and then incubated with anti-PhoB rabbit polyclonal antiserum for 2 h. Proteins were detected using the Bio-Rad Immun-Star chemiluminescent protein detection system as indicated by the manufacturer. The membranes were then wrapped in plastic wrap and exposed to X-ray film and then developed in an automated film processor.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

To better understand the mechanism of activation of the PhoB protein, several in-frame deletion mutations were introduced into the phoB gene using inverse PCR and the phenotypes of these mutants were examined. Each mutant was tested for the ability to activate transcription of the phoA gene. Initially, four deletion mutations were created (Fig. 1A). The corresponding proteins are designated PhoB4-122, PhoB130-227, PhoB4-110, and PhoB125-131 and are encoded on plasmids pMP40, pMP7, pMP8, and pMP17, respectively. The nomenclature for each of the mutant proteins indicates which residues of PhoB have been deleted. For example, PhoB4-122 consists of the first three amino acids of PhoB, followed by Gly-Ser (from the introduction of a BamHI site at the point in the plasmid corresponding to the site of the deletion; see Materials and Methods), followed by residues 123 to 229. PhoB4-122 and PhoB130-227 lack the receiver and output domains, respectively. PhoB4-110 lacks 80% of the receiver domain but retains the 5 helix of the receiver domain plus the entire output domain. PhoB125-131 retains both domains but is missing the predicted interdomain linker.

View larger version (48K): [in this window] [in a new window]   FIG. 1.   Structures of the deletion and chimeric proteins used in this study. (A) The domain structure of PhoB is represented as two white rectangles separated by a black linker region. The amino acid numbers are shown above the map of the secondary structures of PhoB (arrows, -strands; ovals, -helices) (30, 35). For the 10 deletion proteins the white bars represent the protein segments that remain whereas the lines correspond to the deleted segments. The name of each protein designates which amino acid residues have been deleted from PhoB. For example, PhoB4-122 contains residues 1 to 3 of PhoB, followed by Gly-Ser (from an inserted BamHI site in the coding sequence), followed by residues 123 to 229. (B) Schematic representation of the chimeric proteins used in this study. Ch1 joins the N-terminal 108 residues of CheY to the C-terminal 125 residues of PhoB. Ch3 joins the N-terminal 127 amino acid residues of CheY to the C-terminal 106 residues of PhoB.

The induction of alkaline phosphatase by mutant proteins. Plasmids expressing PhoB or the four deletion derivatives were introduced into the phoB mutant strain, DWE1, and the levels of alkaline phosphatase produced by these strains were determined following growth in phosphate-sufficient media (Fig. 2). DWE1 contains the phoB23 allele, which has a transition mutation in the ninth codon of the phoB gene that results in the conversion of a glutamate residue to a lysine (46). Under these growth conditions the Pho regulon remains uninduced in wild-type cells and, as expected, the strain harboring pDE1 (which contains the full-length phoB gene) produced very low levels of alkaline phosphatase. In contrast, the strain producing the PhoB4-122 protein, consisting of only the output domain, induced alkaline phosphatase to high levels. These results confirm our previous findings that the unphosphorylated receiver domain of PhoB inhibits the activity of the output domain (6). As anticipated, PhoB130-227, consisting of only the receiver domain, was unable to activate transcription. Expression of the PhoB125-131 protein also induced the synthesis of alkaline phosphatase, but to slightly lower levels than did expression of PhoB4-122. This observation suggests that in PhoB a functional interdomain linker is required for the receiver domain to inhibit the output domain. This interdomain linker is probably required to correctly position the two domains relative to each other.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   The transcriptional activation activities of PhoB deletion proteins were determined by measuring the amounts of alkaline phosphatase (AP) synthesized in phosphate-sufficient medium. E. coli DWE1 cells were transformed with plasmids encoding PhoB deletion proteins. The genes for PhoB, PhoB4-122, PhoB130-227, PhoB4-110, and PhoB125-131 were contained on plasmids pDE1, pMP40, pMP7, pMP8, and pMP40, respectively. The cells were grown overnight in LB medium containing ampicillin, and alkaline phosphatase assays were performed.

View larger version (73K): [in this window] [in a new window]   FIG. 3.   Western immunoblot analysis of PhoB, PhoB4-122, PhoB130-227, PhoB4-110, and PhoB125-131. E. coli DWE1 cells were transformed with plasmids encoding PhoB and the four deletion mutants. The cells were grown overnight in LB medium containing ampicillin, were collected by centrifugation, and were lysed in SDS-PAGE sample buffer. Equal amounts of cellular extracts were separated by SDS-PAGE, transferred onto a nitrocellulose membrane, and detected with a chemiluminescence detection system using rabbit anti-PhoB polyclonal sera. Purified PhoB was run as a standard (48).

Deletion analysis of the 5 helix. An additional series of deletions was created to better define the amount of the receiver domain that was required to silence the output domain. Several of these deletions removed residues from the 5 helix, whereas other deletions extended the amount of PhoB from the 5 helix. PhoB4-113 and -4-116 lacked 3 and 6 amino acid residues, respectively, whereas PhoB4-104, -4-98, -4-92, and -4-89 contained an additional 6, 12, 18, and 21 residues, respectively (Fig. 1A). Plasmids encoding these deletion mutations were introduced into DWE1 cells, and their phenotypes were determined. There was no alkaline phosphatase induction in DWE1 cells expressing mutant proteins that extended the 5 helix (Fig. 4). In contrast, as the 5 helix was deleted, inhibition of the output domain was decreased and the proteins behaved similarly to PhoB4-122. These results demonstrate that the minimum amount of the receiver domain that is required to silence the output domain is the entire 5 helix. These observations raise the possibility that in full-length PhoB the 5 helix of the receiver domain interacts with the output domain in a specific manner. It is the modulation of this interaction through conformational changes that is triggered by phosphorylation of Asp53, which controls the activity of the output domain.

View larger version (22K): [in this window] [in a new window]   FIG. 4.   The transcriptional activation activities of a series of PhoB deletion proteins localize the inhibitory region of the receiver domain to the 5 helix. E. coli DWE1 cells were transformed with plasmids encoding PhoB deletion proteins. The genes for PhoB, PhoB4-89, PhoB4-92, PhoB4-98, PhoB4-104, PhoB4-110, PhoB4-113, PhoB4-116, and PhoB4-122 were contained on plasmids pDE1, pMP41, pMP42, pMP44, pMP46, pMP8, pMP48, pMP49, and pMP40, respectively. The cells were grown overnight in glucose-MOPS minimal medium containing 5.0 mM KH2PO4, and alkaline phosphatase (AP) assays were performed.

Design and construction of chimeric proteins. To more fully understand the role of the 5 helix in controlling the activity of the output domain of PhoB, two chimeric proteins in which portions of CheY were swapped for homologous regions of PhoB were constructed. The first chimera, Ch1, has a splice site at the end of the 5 -strand at the conserved Lys-Pro-Phe triplet (residues 105 to 107 of PhoB) and maintains the 5-5 loop and the entire 5 helix from PhoB (Fig. 1B). The design of this construct is based on the idea that regions of amino acid identity between CheY and PhoB may result from structural or functional constraints and that it may be necessary to maintain these identities to generate a functional protein. The second chimera, Ch3, has a splice site downstream of the 5 helix and substitutes the entire response regulator domain of CheY for that of PhoB. The chimeric genes were created by extending engineered overlaps in PCR products that contain the cheY and phoB gene fragments (12). The analysis of these proteins was designed to focus on the role of the 5 helix in propagating an input signal into an appropriate output response. Phosphorylation of the receiver domain by the CheA protein provided the input, whereas the regulated production of alkaline phosphatase was the output.

Phosphorylation with phospho-CheA. To determine whether the addition of an output domain to CheY would prevent proper interactions with CheA, we conducted phosphotransfer assays. The two chimeric genes were cloned into a T7 expression vector from which high levels of protein expression were obtained. Since most of the overexpressed protein was insoluble, the chimeric proteins were purified from inclusion bodies. Phosphotransfer reactions were initiated by adding an aliquot of each protein to a sample of [³²P]phospho-CheA. The reactions were analyzed by SDS-PAGE and autoradiography (Fig. 5). The reaction mixture containing only CheA showed, in addition to the full-length phosphoprotein, two low-abundance, faster-migrating bands. These bands were most likely proteolytic fragments of CheA that were either capable of autophosphorylation or substrates for transphosphorylation from CheA and could also serve as phosphodonors since they and the full-length band disappeared upon incubation with CheY. Dephosphorylation of phospho-CheA and the subsequent phosphotransfer to the CheY moiety were observed with both of the chimeric proteins. However, no phosphotransfer or phospho-CheA dephosphorylation was observed when phospho-CheA was incubated with PhoB. These results show that both chimeric proteins are functionally active in receiving input from CheA. Note that the relative intensities of the phospho-Ch1 and phospho-Ch3 bands in Fig. 5 reflect the amounts of each protein in the phosphotransfer reactions and not the stability of the phosphorylated proteins.

View larger version (54K): [in this window] [in a new window]   FIG. 5.   Phosphotransfer reactions between CheA and chimeric proteins Ch1 and Ch3. CheA was phosphorylated with [-32P]ATP. Aliquots of [-32P]phospho-CheA were mixed with the indicated phosphoacceptors, incubated for 2 min at room temperature, and analyzed by SDS-PAGE and autoradiography. The final concentrations of the phosphoacceptors in the phosphotransfer reactions were as follows: CheY, 12 µM; PhoB, 15 µM; Ch1, 2 µM; Ch3, 12 µM.

Activation by the chemotaxis signaling pathway. To test whether the phosphorylation signal within the receiver domains of Ch1 and Ch3 could be transmitted to their output domains, we examined the ability of Ch1 and Ch3 to induce the expression of the chromosome-located alkaline phosphatase gene when presented with a chemotactic stimulus. The two chimeric constructs were subcloned into a regulatable expression vector and transformed into the appropriate tester strains. These strains were selected to assay the activation of CheY by the methylated chemotaxis protein-CheA pathway (1). E. coli PS2001 constitutively activates CheA and produces high levels of phospho-CheY when CheY is encoded on a plasmid, whereas E. coli PS2002 has a deletion of most chemotaxis genes and cannot activate CheY when it is encoded on a plasmid. By comparing the levels of alkaline phosphatase produced in the strains expressing either Ch1 or Ch3 it is possible to determine their levels of phosphorylation-dependent activation.

View larger version (34K): [in this window] [in a new window]   FIG. 6.   Alkaline phosphatase (AP) assay to measure the output activities of chimeric proteins Ch1 and Ch3. The genes for Ch1, Ch3, and PhoB were cloned into vector pMLB1120.215 and transformed into either PS2001 or PS2002. The data show means and standard deviations of assays performed in triplicate.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This study reports experiments on the mechanism by which the receiver domain of response regulator PhoB controls the activity of its output domain. We propose that the 5 helix of the receiver domain participates in interdomain interactions that control the activity of the output domain and that these interactions are modulated through phosphorylation of the receiver domain. This proposal is based on two different lines of experimental evidence. The first line is based on results from a deletion analysis of the PhoB protein. A deletion of the entire receiver domain generated a constitutively active protein (PhoB4-122). The addition of the 5 helix from the receiver domain to the output domain (found in PhoB4-110) silenced the activity of the output domain. Our experiments cannot distinguish whether this silencing results from blocked DNA-binding and/or RNA polymerase interactions or by locking the output domain in an inactive conformation. If PhoB is like the NarL and CheB response regulators, then inhibition is achieved by blocking the active site of its output domain (2, 5).

The second line of investigation involved domain swapping experiments using the CheY protein from Salmonella enterica serovar Typhimurium and the PhoB protein from E. coli. Two chimeric CheY/PhoB proteins in which either 85% (Ch1) or 98% (Ch3) of the receiver domain of PhoB was replaced by the corresponding regions of CheY were constructed. Ch1 maintains the 5-5 loop and the 5 helix from PhoB's receiver domain, whereas Ch3 does not. These proteins were used to test whether an input signal could be transduced into an appropriate output response. Of the two chimeras examined, only the Ch1 protein transduced the input signal into the appropriate response. This result supports the proposal that the 5-5 loop and the 5 helix from the receiver domain of PhoB are required to propagate the phosphorylation-triggered signal from the receiver domain to the output domain. The Ch3 protein was constitutively active, consistent with the idea that, in the unphosphorylated receiver, the 5 helix directly participates in interdomain interactions that silence the output domain. By the incorporation of the 5 helix from CheY into Ch3, the interdomain interactions are abrogated and the inhibition imposed on the output domain by this helix does not occur, which leaves the output domain active.

Expression of the PhoB125-131 protein in cells grown in high-phosphate media resulted in production of alkaline phosphatase. PhoB125-131 consists of the receiver and output domains but is missing the interdomain linker region. We propose that the interdomain linker of PhoB is important for the correct positioning of the 5 helix relative to the output domain. It has previously been shown that the linker region of OmpR is essential in relaying conformational changes between its two domains (14, 40).

The levels of alkaline phosphatase that were induced by the chimeric constructs were only one-fifth of those routinely observed in our laboratory when wild-type cells are grown in phosphate-limiting media (data not shown). Part of this difference in expression levels may be due to the lack of a positive regulatory circuit in the tester cells in which phospho-PhoB induces its own expression (19). In the experiments reported in this study, the expression of the chimeric constructs was under the control of a lac promoter and the levels of protein should remain constant upon induction. In addition, the increase of expression between uninduced cells, PS2002(pCh1), and induced cells, PS2001(pCh1), was only four- or fivefold and was much lower than that observed for wild-type cells grown in phosphate-sufficient and phosphate-limiting media (45). A potential explanation for these results is that the 5 helix of the receiver domain is not the only component involved in interdomain interactions and that other parts of PhoB's receiver domain (perhaps other helices and/or loops) are necessary for complete induction. Another difference could be in the half-lives of the phosphorylated proteins; phospho-PhoB has a half-life of approximately 15 min, whereas phospho-CheY has a half-life of ~15 s (10, 24). This difference may alter the relative amounts of activated proteins within the cells and influence the amount of induction.

Recent studies on the activation of the NtrC protein have suggested that the 4 helix of its receiver domain is involved in an interdomain interaction that propagates the phosphorylation-induced signal (13, 15, 17, 29). In FixJ, the propagation signal is transmitted through the 4-5 surface (3). The interdomain interface for the NarL protein, which must be modulated for activation to occur, involves the 2-3, 3-4, and 4-5 loops as well as the end of 5 (2); In CheB, it is the 4-5-5 surface that constitutes the interdomain interface (5). Taken together with the work presented in this study, these results show that different response regulators employ different molecular surfaces for their interdomain interactions and imply that slightly different signal propagation strategies may be used to control the activities of different output domains.

We propose that response regulator proteins are composed of three functional units: a phosphorylation-triggered switch, a relay, and an output domain. The switch receives input either from a cognate kinase or from a small-molecule phosphodonor (25). This information is transmitted to the relay structure through a conserved conformational change that involves the repositioning of the conserved tyrosine or phenylalanine residue in 5 from a solvent-exposed position into a hydrophobic pocket (3, 4, 9). We propose that this conformational change is at least somewhat conserved because the CheY moiety of Ch1 functioned with the relay unit from PhoB. The relay interprets the conformational change and propagates this information to the output domain.