Reverse phosphotransfer from OmpR to EnvZ in a kinase-/phosphatase+ mutant of EnvZ (EnvZ.N347D), a bifunctional signal transducer of Escherichia coli.

EnvZ of Escherichia coli is a transmembrane histidine kinase belonging to the family of two-component signal transducing systems prevalent in prokaryotes and recently discovered in eukaryotes. In response to changes in medium osmolarity EnvZ regulates the level of phosphorylated OmpR, its conjugate response-regulating transcription factor for ompF and ompC genes. EnvZ has dual opposing enzymatic activities; OmpR-phosphorylase (kinase) and phospho-OmpR-dephosphorylase (phosphatase). The osmotic signal is proposed to regulate the ratio of the kinase to the phosphatase activities of EnvZ to modulate the level of OmpR phosphorylation. In this work we used a COOH-terminal fragment of a previously identified kinase/phosphatase EnvZ mutant (EnvZ•N347D) to demonstrate that the phosphoryl group on phospho-OmpR is transferred back to EnvZ to the same histidine residue (His) that is utilized for the autokinase reaction by the wild type protein. Phospho-EnvZ•N347D thus formed could also transfer its phosphoryl group back to OmpR. The phosphotransfer reaction from phospho-OmpR to EnvZ•N347D was inhibited by ADP while Mg ions stimulated the dephosphorylation reaction, resulting in release of inorganic phosphate. These results indicate that the energy levels of phosphoryl groups on OmpR and EnvZ are very similar and that the phosphatase reaction in the EnvZ•N347D mutant involves a reversal of the phosphotransfer reaction from EnvZ to OmpR using the identical His residue.

EnvZ of Escherichia coli is a transmembrane histidine kinase belonging to the family of two-component signal transducing systems prevalent in prokaryotes and recently discovered in eukaryotes. In response to changes in medium osmolarity EnvZ regulates the level of phosphorylated OmpR, its conjugate response-regulating transcription factor for ompF and ompC genes. EnvZ has dual opposing enzymatic activities; OmpR-phosphorylase (kinase) and phospho-OmpR-dephosphorylase (phosphatase). The osmotic signal is proposed to regulate the ratio of the kinase to the phosphatase activities of EnvZ to modulate the level of OmpR phosphorylation. In this work we used a COOH-terminal fragment of a previously identified kinase ؊ /phosphatase ؉ EnvZ mutant (EnvZ⅐N347D) to demonstrate that the phosphoryl group on phospho-OmpR is transferred back to EnvZ to the same histidine residue (His 243 ) that is utilized for the autokinase reaction by the wild type protein. Phospho-EnvZ⅐N347D thus formed could also transfer its phosphoryl group back to OmpR. The phosphotransfer reaction from phospho-OmpR to EnvZ⅐N347D was inhibited by ADP while Mg 2؉ ions stimulated the dephosphorylation reaction, resulting in release of inorganic phosphate. These results indicate that the energy levels of phosphoryl groups on OmpR and EnvZ are very similar and that the phosphatase reaction in the EnvZ⅐N347D mutant involves a reversal of the phosphotransfer reaction from EnvZ to OmpR using the identical His 243 residue.
Phosphorylation and dephosphorylation of cellular proteins plays a critical role in signal transduction to regulate numerous cellular functions in both prokaryotes and eukaryotes. Protein histidine kinases and their response regulators, so-called "two-component" systems, constitute a large family (Ͼ50) of signal transducing systems that enable bacteria to adapt to the changing environment (reviewed by Parkinson and Kofoid (1992)). Classical eukaryotic protein kinases are tyrosine, serine, and threonine kinases. Recently, however, histidine kinases have been found not only in yeast (Maeda et al., 1994;Ota and Varshavsky, 1993) but also in plants (Chang et al., 1993) (reviewed by Alex and Simon (1994) and Swanson et al. (1994)) and in mammalian cells (Crovello et al., 1995). This represents a novel paradigm for eukaryotic cell signaling. Prokaryotic histidine kinases exhibit considerable homology in their COOH-terminal domains, and it is believed that they share a common mechanism of action.
Exposure of bacteria to high osmolarity leads to dehydration, collapse of ion gradients over the cytoplasmic membrane, and decrease in cell viability. Therefore, the first response of bacteria to osmotic stress consists of changes in the activities of enzymes and transport systems so that the turgor pressure is restored and the cytoplasmic environment is optimized. Somewhat later, changes in gene expression provide additional flexibility in adapting cells to osmotic shock (for reviews see Csonka (1989), and Csonka and Hanson (1991). One of the means that Escherichia coli adopts to handle osmotic stress is to modulate the type of diffusion pores that exist in the outer membrane (Csonka and Hanson, 1991). These pores are formed by homotrimeric association of the porin proteins, OmpF and OmpC. These proteins are highly expressed (approximately 10 5 molecules/cell), and the rate of diffusion through OmpF has been measured to be 10 times faster than through OmpC (Nikaido and Vaara, 1985). The expression of these two porins is differentially regulated through the members of the ompB operon, ompR and envZ (Hall and Silhavy, 1981;Forst and Inouye, 1988). OmpF is preferentially produced at low osmolarity and OmpC at high osmolarity.
Taz1 is a hybrid receptor consisting of the NH 2 -terminal (256-residue) ligand-binding domain of Tar (a chemoreceptor for aspartate) and the COOH-terminal (228-residue) signaling domain of EnvZ (Utsumi et al., 1989). This chimeric receptor induces ompC-lacZ expression in response to aspartate, thus enabling one to study EnvZ function in response to a well defined ligand (the natural ligand for EnvZ is unknown). It has been shown that all substitution mutants at the conserved Asn residue (381 in Taz1) eliminated the kinase activity of Taz1, resulting in the elimination of ompC-lacZ expression in either the absence or presence of aspartate (Yang et al., 1993). However, all of them retained the phosphatase activity at a similar level to that of the wild type Taz1, regardless of the amino acid residues used for substitution (D, E, Q, H, A, T, and C). The Asn 347 substitution mutants could complement the H277V (the autophosphorylation site) mutant of Taz1 in vivo to restore wild type levels of osmoregulation. For a finer analysis of the role of the conserved Asn in EnvZ, the 271-residue (Arg 180 -Gly 450 ) cytoplasmic domain of the mutant EnvZ⅐N347D(C) was purified, and its enzymatic activity was characterized. We now report that the phosphatase reaction of EnvZ⅐N347D(C) involves the reverse phosphotransfer of the phosphoryl group from OmpR to EnvZ to the identical His 243 residue which is the site of autophosphorylation. The dephosphorylation of phospho-OmpR by the mutant protein was found to be inhibited by ADP but stimulated by Mg 2ϩ ions. These results indicate that the phosphate groups on EnvZ and OmpR are energetically similar and that the phosphatase reaction of the mutant EnvZ⅐N347D(C) involves the reverse phosphotransfer of the phosphoryl group from OmpR to EnvZ using the identical His 243 , the autophosphorylation site.

Cell Strain and Construction of Plasmids
E. coli B strain BL21-DE3 (F Ϫ ompT r B m B ) was used for the expression of the envZ(C) and envZ . N347D(C) gene fragments cloned into the pEt11a vector. The 1.5-kilobase pair EcoRII/EcoRI DNA fragment corresponding to the 271-amino acid residue (Arg 180 -Gly 450 ) cytoplasmic domain of EnvZ (EnvZ(C)) was recloned from the ompC:envZ plasmid previously constructed in the laboratory ) into a pEt11a T7 expression system to construct pEt11aE-Z(C). The EnvZ(C) protein thus expressed contains four extra amino acids (Met, Ala, Gly, Ile) at the amino-terminal end. The mutagenesis of envZ was performed by site-directed mutagenesis using M13 as described previously . The envZ⅐N347D(C) mutant was then subcloned into pEt11aE in the same manner as the wild type gene. 1

Purification of EnvZ(C) and EnvZ⅐N347D(C) Protein Fragments
EnvZ(C) was purified from E. coli B strain BL21(DE3) transformed with pEt11aE-Z(C) plasmid by a modification of the procedure described previously . The protein was precipitated using 30% ammonium sulfate saturation instead of 40%. The Affi-Gel blue affinity column was substituted with a Green A affinity column from which the protein was eluted with a linear gradient 0 -1.5 M KCl in buffer A (20 mM Tris-HCl (pH 7.8) containing 5% glycerol, 10 mM ␤-mercaptoethanol, and 1 mM EDTA). Protein fractions were pooled and centricon-concentrated.

Preparation of Membrane Fractions
PDR 200 cells harboring wild type EnvZ was grown in LB medium. Mid-logarithmic-phase cells were harvested by centrifugation and washed once and sonicated in buffer B (0.1 M sodium phosphate (pH 7.2) containing 10% (v/v) glycerol, 1 mM phenylmethylsulfonyl fluoride (PMSF), 2 5 mM EDTA, 5 mM 1,10-phenanthroline, 0.3 mM p-hydroxymercuribenzoate, and 1 M leupeptin). Cell debris was removed by centrifugation at 10,000 ϫ g for 15 min. The supernatant was then ultracentrifuged at 180,000 ϫ g for 90 min at 4°C. The pellet was washed with buffer C (20 mM Tris-HCl (pH 7.8), 5% glycerol, 1 mM EDTA, 10 mM ␤-mercaptoethanol) containing 2 M KCl and sonicated in the same buffer and then centrifuged at 393,000 ϫ g for 14 min. This washing step was repeated. The resulting membrane fraction (En-vZ(M)) was resuspended in buffer C. The concentration of total protein in the membrane preparation was estimated by the Bio-Rad assay.

In Vitro Autophosphorylation and pH Stability Assays
EnvZ membrane preparations or soluble purified cytoplasmic proteins were incubated in 0.1 M Tris-HCl buffer (pH 8.0) containing 50 mM KCl, 5 mM CaCl 2 , 1 mM PMSF, and 10% glycerol with 0.4 mM[␥-32 P]ATP (1,100 cpm/pmol) at 25°C for 10 min. The reaction was stopped by adding 5ϫ SDS gel loading buffer containing 10% (w/v) SDS, 3 mM ␤-mercaptoethanol, and 40% glycerol. The reaction mixture was then subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE). After electrophoresis protein bands were transferred to a BA85 nitrocellulose membrane (0.45 m). The membrane was exposed for autoradiography. Protein bands were visualized with anti-EnvZ polyclonal antiserum and goat anti-rabbit IgG alkaline phosphatase conjugate followed by a color reaction with nitro blue tetrazolium and 5-bromo-4chloro-3-indolyl phosphate p-toluidine salt. The pH stability of of the phosphate linked to EnvZ⅐N347D(C) and to EnvZ(C) was checked by the method described previously .

In Vitro Phosphorylation of OmpR Protein by EnvZ Membrane or Cytoplasmic Proteins
EnvZ proteins were first autophosphorylated under the same conditions as described above for 15 min. Purified OmpR (2 l, 0.3 g/l, obtained from S. Harlocker (Delgado et al., 1993)) was then added to the reaction mixture and incubated at 25°C for 20 min. The reaction was stopped by adding 5 ϫ SDS gel loading buffer, and the reaction mixture was analyzed by SDS-PAGE as described above.

Dephosphorylation of Phosphorylated OmpR by EnvZ Cytoplasmic Proteins
The phosphatase activity of EnvZ-soluble cytoplasmic proteins was measured as follows.
Step 1: Autophosphorylation of EnvZ(M) for 20 Min-EnvZ(M) (50 g) was autophosphorylated with [␥-32 P]ATP (100 Ci, specific activity: 3,000 Ci/mmol) in 200 l of 0.1 M Tris-HCl buffer (pH 8.0) (containing 50 mM KCl, 5 mM CaCl 2 , 1 mM PMSF, and 10% glycerol) at 25°C for 20 min. The reaction mixture was centrifuged at 393,000 ϫ g for 14 min at 4°C using a Beckman TL100 ultracentrifuge. The membrane pellet was washed three times with 0.1 M Tris-HCl buffer (pH 8.0) containing 50 mM KCl, 5 mM CaCl 2 , 1 mM PMSF, and 10% glycerol, sonicated, and then resuspended in the same buffer. The suspended membrane fraction was centrifuged and washed six more times. The removal of [␥-32 P]ATP and P i in the membrane fraction was monitored by thinlayer chromatography. The final membrane fraction containing phosphorylated EnvZ(M) was resuspended in 200 l of the same washing buffer.
Step 2: Isolation of Phosphorylated OmpR Protein-Purified OmpR protein (approximately 15 g) was incubated with the membrane fraction containing phosphorylated EnvZ protein in a total volume of 140 l for 20 min at 25°C. After incubation the reaction mixture was centrifuged at 393,000 ϫ g for 14 min to remove the EnvZ-containing membrane. The 120-l supernatant containing phosphorylated OmpR was collected and immediately used as substrate for the phosphatase assay.
Step 3: OmpR-P Phosphatase Assay-Phosphorylated OmpR protein (1.7 M) was mixed with purified cytoplasmic fragment of EnvZ (En-vZ(C) or EnvZ⅐N347D(C)) (8.7 M) and ADP (final concentration: 1 mM) in phosphatase buffer (0.1 M Tris-HCl buffer (pH 8.0) containing 50 mM KCl, 5 mM CaCl 2 , 1 mM PMSF, and 10% glycerol) for the required length of time at 25°C. The reaction was stopped by adding 5 ϫ SDS gel loading buffer and analyzed by SDS-PAGE, transferred to nitrocellulose, and then subjected to autoradiography as described above.

Reverse Transfer of Phoshoryl Group from EnvZ⅐N347D(C) to OmpR Protein
Phosphorylated OmpR protein was used as substrate for the phosphatase assay for EnvZ⅐N347D(C) as described above. For the study of the reverse transfer of the phosphoryl group from the mutant protein, purified non-phosphorylated NH 2 -terminal fragment of OmpR (1 g, obtained from S. Harlocker (Delgado et al., 1993)) was added to the reaction mixture containing phospho-OmpR and EnvZ⅐N347D(C) and incubated for an additional 10 min under the same conditions. The results were analyzed in the same manner as described above.
Endoproteinase Lys-C Digestion of Phospho-EnvZ(C) and  EnvZ(C) protein (8 g) was autophosphorylated as described above. EnvZ⅐N347D(C) protein was phosphorylated using phosphorylated OmpR following the protocol described above except that all the reactions were carried out in 25 mM Tris-HCl (pH 8.5). After the 20-min incubation the reaction was stopped by the addition of SDS loading buffer and an aliquot of the reaction mixture containing approximately 8 g of the mutant protein was subjected to endoproteinase Lys-C digestion along with the wild type protein. The phosphorylated proteins were incubated for 18 h at 37°C in the digestion buffer of 25 mM Tris-HCl (pH 8.5) containing 1 mM EDTA and 1 g of endoproteinase Lys-C. After incubation SDS loading buffer was added to the reaction mix and further incubated for 30 min at 40°C. The digestion products were separated on a Tricine-SDS-PAGE gel (Schä gger and von Jagow, 1987). The separating gel composition was 16.5% T and 6% C. the spacer gel was of 10% T and 3% C while the stacking gel was 4% T and 3% C. Insulin B chain was used as a marker. The protein bands were transferred to nitrocellulose and exposed for autoradiography.

RESULTS
The COOH-terminal Fragment of EnvZ⅐N347D Is Kinase Ϫ / Phosphatase ϩ -The purified EnvZ⅐N347D(C) was tested for its ability to autophosphorylate in the presence of [␥-32 P]ATP and thence to transfer the phosphoryl group to OmpR (kinase activity). Fig. 1A shows that EnvZ(C) was autophosphorylated (lane 1) and could concomitantly transfer the phosphoryl group to OmpR (lane 2). EnvZ⅐N347D(C), however, was deficient in autophosphorylation (lane 3), and therefore subsequent transfer of the phosphoryl group to OmpR was also not detected (lane 4). Equivalent quantities of purified COOH-terminal pro-teins were used for the reactions as determined by immunoblot (Fig. 1B).
Phosphorylation of EnvZ⅐N347D(C) by Phospho-OmpR-We next examined whether EnvZ⅐N347D(C) was capable of dephosphorylating phospho-OmpR. The phosphatase assay was performed as described under "Experimental Procedures"; 32 Plabeled EnvZ (Fig. 2, lane 1) was incubated with purified OmpR, and the reaction mixture was incubated at 25°C. [ 32 P]phosphoryl group was transferred to OmpR from phospho-EnvZ(M) as shown in Fig. 2, lane 2. 32 P-Labeled phospho-OmpR was then isolated by ultracentrifugation of the reaction mixture to remove the EnvZ(M). Phospho-OmpR, recovered in the supernatant, was then incubated with the COOH-terminal fragments of EnvZ for 1 and 20 min. With wild type EnvZ(C) approximately 90% of phospho-OmpR was dephosphorylated within 1 min (lane 7), and in 20 min phospho-OmpR was completely hydrolyzed (lane 8). There was no significant dephosphorylation of phospho-OmpR in the absence of EnvZ(C) during the same time period (lanes 3 and 4). In contrast to the wild type EnvZ(C), when EnvZ⅐N347D(C) was used for the reaction, dephosphorylation of phospho-OmpR was accompanied with the concomitant appearance of phospho-EnvZ⅐N347D(C), as shown in lanes 5 and 6. At 20 min there was an almost quantitative transfer of phosphoryl group from OmpR to EnvZ⅐N347D(C) (lane 6).
Reversibility of the Phosphoryl Group Transfer from EnvZ⅐N347D(C) to OmpR-The kinase activity of EnvZ proceeds in two steps; autophosphorylation at His243 with ATP, followed by phosphoryl group transfer from phospho-EnvZ to OmpR. Although EnvZ⅐N347D(C) was deficient in autophosphorylation with ATP (see Fig. 1), the experiment described above clearly indicated that it could become phosphorylated by phospho-OmpR. If the phosphorylated EnvZ⅐N347D(C) thus obtained is identical to the autophosphorylated wild type En-vZ(C), it is expected that the phosphoryl group on EnvZ⅐N347D(C) should be transferred back to OmpR if an excess of non-phospho-OmpR is added. To test this, purified NH 2 -terminal domain of OmpR (residues 1-134) was added after 20 min incubation of EnvZ⅐N347D(C) with phospho-OmpR (Fig. 2, lane 6) and was further incubated for an additional 10 min. OmpR(N) contains the Asp 55 residue and has been shown to be efficiently phosphorylated by EnvZ(C) (Delgado et al., 1993). As shown in Fig. 3, lane 2, the phosphoryl group was almost quantitatively transferred from OmpR-P to EnvZ⅐N347D(C) after a 20-min incubation. The molar ratio of OmpR to EnvZ⅐N347D(C) in the reaction was 1:5. In a duplicate reaction, after the initial 20-min incubation of EnvZ⅐N347D(C) with OmpR, OmpR(N) was added to the reaction mixture  2 and 4) purified OmpR protein (0.6 g) was added to the reaction mixture at the end of 5 min and further incubated for 15 min before stopping the reaction with 5 ϫ SDS sample buffer. The reaction mixtures were then subjected to 17.5% SDS-PAGE analysis. After electrophoresis protein bands were transferred to a nitrocellulose membrane and exposed for autoradiography. Protein bands were visualized with anti-EnvZ polyclonal antiserum as described under "Experimental Procedures." A, SDS-PAGE autoradiogram. B, immunoblot of the same gel as in A.
To examine whether OmpR(N) is a better substrate than OmpR for the phosphotransfer reaction from EnvZ⅐N347D(C), the phosphatase assay was carried out with phospho-OmpR and EnvZ⅐N347D(C) at a molar ratio of 1:2. Aliquots were taken at 0.5, 1, 10, 30, and 45 min and analyzed by SDS gel electrophoresis (Fig. 4, lanes 1-5). By 30 min (lane 4), the ratio of 32 P in OmpR to that of EnvZ⅐N347D(C) reaches a constant (by densitometry), indicating that the phosphotransfer reaction between the two proteins have reached an equilibrium. At 45 min, three aliquots were taken. To one, additional non-phospho-OmpR was added (lane 6), to the second non-phospho-OmpR(N) was added (lane 7), and to the third both OmpR and OmpR(N) was added (lane 8). These reaction mixtures were incubated for another 10 min. From lanes 6 -8, it is clear that on the addition of OmpR or OmpR(N) the phosphoryl group is transferred to the added component. On the basis of the distribution of 32 P among these proteins, it appears that all three compete for the phosphoryl group. Also, OmpR(N) was not a better substrate than OmpR for the phosphotransfer from EnvZ⅐N347D(C). It is to be noted that when phospho-OmpR(N) was used instead of phospho-OmpR for the phosphorylation of EnvZ⅐N347D(C), it was at least 95% less efficient than that with phospho-OmpR (data not shown). This may suggest that the COOH-terminal domain of OmpR has some role in the reverse transfer of phosphoryl group from OmpR to EnvZ⅐N347D(C).
Effects of ADP and Metal Ions on the Reverse Phosphotransfer-Since ADP, ATP, and non-hydrolyzable analogues of ATP are known co-factors that enhance the rate of the phosphatase reaction (Aiba et al., 1989a;Igo et al., 1989), we investigated the effect of ADP on the phosphatase reaction with EnvZ⅐N347D(C). Interestingly, as the ADP concentrations increased from 0 to 10 mM, a progressive inhibition in the phosphoryl group transfer from phospho-OmpR to EnvZ⅐N347D(C) was observed (Fig. 3, lanes 2-5). At 10 mM ADP the phospho-transfer was completely blocked.
Next, we examined the effect of divalent cations on the phosphatase reaction. We carried out the reaction in 25 mM Tris-HCl (pH 8.5) in the absence of ADP and in the presence of 5 mM EDTA, 5 mM EGTA, 5 mM MgCl 2 , and 5 mM CaCl 2 . As shown in Fig. 5, lane 3, 5 mM EDTA inhibits the reaction. The EDTA inhibition can be reversed by addition of 5 mM MgCl 2 (data not shown). In contrast, EGTA does not inhibit the reaction (Fig. 5, lane 4). These results indicate that Mg 2ϩ is a necessary co-factor of the phosphatase reaction. Since very little phosphoprotein is detected in the presence of Mg 2ϩ , it clearly enhances the phosphatase reaction. Ca 2ϩ on the other hand stabilizes phospho-OmpR and retards the phosphotransfer (lane 6).
Involvement of the Conserved Autophosphorylation Site, His 243 , in the Phosphatase Reaction of EnvZ⅐N347D(C)-To identify the residue phosphorylated in EnvZ⅐N347D(C) by phospho-OmpR, we first examined its pH stability. Phospho-EnvZ⅐N347D(C) was alkali stable (pH 13) but acid-labile (pH 1) (data not shown), which is characteristic of histidyl phosphate. To determine which histidine residue(s), of the six histidine residues in EnvZ⅐N347D(C) was phosphorylated, both wildtype and mutant phosphoproteins were subjected to endoproteinase Lys-C digestion. Non-phosphorylated EnvZ(C) was also digested. Note that EnvZ(C) was autophosphorylated with [␥-32 P]ATP, while EnvZ⅐N347D(C) was phosphorylated using [ 32 P]phospho-OmpR. The products of digestion of nonphospho-EnvZ(C), autophosphorylated EnvZ(C), and phospho-EnvZ⅐N347D(C) were separated on a Tricine-SDS-PAGE gel (Fig. 6, lanes 2-4, respectively). Lanes 1 and 2 of the gel were stained with Serva blue dye. Lane 2 shows the seven predicted EnvZ(C) Lys-C digestion bands. Lanes 3 and 4 of the gel were subjected to autoradiography. For both phosphoproteins, a single band, migrating at the identical position corresponding to the fourth band in the stained gel, was phosphorylated. From its size the fragment was judged to be the peptide Gln 229 to Asn 272 , containing His 243 , the only histidine residue present in that fragment. This result clearly demonstrates that the same residue (His 243 ) which was autophosphorylated by ATP in the wild type EnvZ(C) was phosphorylated in EnvZ⅐N347D(C) by phospho-OmpR. DISCUSSION Previous work from this laboratory demonstrated that substitutions at the conserved asparagine residue blocks the kinase but maintains the phosphatase activity (Yang and Inouye, 1991;Yang et al., 1993). The present data show that although the N347D mutation (corresponds to the N381D mutation in FIG. 3. Reverse transfer of the phoshoryl group from EnvZ⅐N347D(C) to OmpR and the inhibition of the phosphatase reaction of EnvZ⅐N347D(C) by ADP. Lane 2, phosphorylated OmpR protein was isolated and used as substrate for the phosphatase assay for EnvZ⅐N347D(C) as described in the legend to Fig. 2, except that the reaction was performed in the absence of ADP. Incubation time was 20 min. Lane 1, in a duplicate reaction purified NH 2 -terminal OmpR protein (OmpR:OmpR(N)::1:5) was added to the reaction mixture after 20-min incubation and further incubated for 10 min under the same conditions. Lanes 3, 4, and 5, the phosphatase assay was carried out in the presence of 1, 5, and 10 mM ADP, respectively. Incubation time was 20 min. The results were analyzed in the same manner as described in the legend to Fig. 2.   FIG. 4. Reverse transfer of the phoshoryl group from EnvZ⅐N347D(C) to OmpR and/or to OmpR(N). The phosphatase assay was carried out as described in the legend to Fig. 2, except that ADP was omitted from the reaction buffer. The time course of the phosphotransfer from phospho-OmpR to EnvZ⅐N347D(C) was examined. Aliquots from the reaction mixture were withdrawn at 0.5-, 1-, 10-, 30-, and 45-min intervals (lanes 1-5, respectively). At the 45-min time point three aliquots were taken, to one was added OmpR (lane 6), to the second was added OmpR(N) (lane 7), and to the third was added both in equimolar amounts (lane 8), and the reaction mixes were incubated for a further 10 min. The results were analyzed in the same manner as described in the legend to Fig. 2. Taz) severely impaired the ATP-dependent autophosphorylation, EnvZ⅐N347D(C) can still receive the phosphoryl group readily from phospho-OmpR. Moreover, once phosphorylated it also retains its ability to transfer the phosphoryl group back to OmpR or OmpR(N). The phosphatase assay yielded unexpected results with the N347D mutant protein. The mutant could efficiently dephosphorylate phospho-OmpR by removing the phosphoryl group from it onto its own histidine (His 243 ) which is also the autophosphorylation site. Similar reverse phosphotransfer from phospho-OmpR to EnvZ was observed with purified EnvZ⅐N347H(C). In the presence of 5 mM Mg 2ϩ and absence of ADP very little phospho-protein is detected for both EnvZ⅐N347D and OmpR. The phosphoryl group is released as inorganic phosphate under these conditions (not shown). Therefore the "phosphatase" reaction by EnvZ⅐N347D(C) involves a phosphotransfer reaction from phospho-OmpR to the mutant protein. It is important to point out that the in vivo complementation experiments (Yang et al., 1993) clearly demonstrated that the asparagine substituent mutants can complement the H277V mutant (confers null phenotype) to restore functional signal transduction, indicating that the phosphatase activity of the conserved asparagine substituent mutants is functionally equivalent to the wild type phosphatase activity.
The mechanism of the phosphatase reaction has not been elucidated for EnvZ. Also the role of the conserved asparagine is currently undefined. The results obtained with the EnvZ⅐N347D(C) mutant protein offers some insights into the roles of the conserved Asn 347 and the phosphatase function of EnvZ. Since EnvZ⅐N347D(C) is deficient in autophosphorylation (Fig. 1), it is possible that the conserved asparagine is involved in ATP binding and/or in the subsequent transfer of the phosphoryl group to histidine. It has been postulated that EnvZ may have a modulator binding site which may or may not overlap with the putative ATP binding site. Contrary to what is observed with the wild type protein ADP serves as a negative effector for the phosphatase reaction of the N347D mutant protein (Fig. 3), suggesting that the putative modulator binding site has been modified in the EnvZ⅐N347D(C) mutant. Therefore the conserved asparagine in EnvZ may be involved directly or indirectly in defining the putative ATP and/or modulator binding sites in EnvZ.
EnvZ⅐N347D(C) is phosphorylated during the dephosphorylation of phospho-OmpR and in the presence of Mg 2ϩ and the absence of ADP the phosphoryl group is released very efficiently as inorganic phosphate. This raises the possibility that the phosphatase reaction of the mutant EnvZ⅐N347D(C) proceeds in two steps. In the first step phosphoryl group transfer occurs from phospho-OmpR to His 243 on EnvZ, which is the autophosphorylation site of EnvZ. In the second step this phosphorylated intermediate is hydrolyzed to inorganic phosphate which is a Mg 2ϩ -dependent reaction. Therefore the phosphorylation of the mutant EnvZ during the dephosphorylation of phospho-OmpR just represents the reverse phosphotransferase reaction of the the phosphorylation of OmpR by EnvZ. If His 243 is involved in the phosphatase reaction of wild type EnvZ (as follows from the first possibility), an important implication would be that non-phosphorylatable substitutions at His 243 should inactivate both the kinase and phosphatase functions of EnvZ. Known mutations at this position in EnvZ include H243V (Yang and Inouye, 1991;Yang et al., 1993) and H243R (Tokishita and Mizuno, 1994), all of which confer a null phenotype. This is in agreement with the prediction. It is of interest to note here that in the nitrogen regulatory system, mutating the conserved histidine at position 319 to asparagine (H319N) does not eliminate the phosphatase activity of NtrB (Atkinson and Ninfa, 1993;Kamberov et al., 1994). Although the Ntr system closely parallels the osmoregulatory system, they differ in many respects. EnvZ is a membrane-localized sensor kinase, while NtrB is a soluble cytoplasmic protein. Unlike EnvZ, NtrB requires an accessory protein, P II, for its phosphatase activity (Ninfa and Magasanik, 1986;Keener and Kustu, 1988). Thus it is possible that the mechanism of the phosphatase reaction of NtrB may be different from that of EnvZ.
The present results suggest that the energy levels of the phosphoryl group is similar between the sensor, EnvZ, and the response regulator, OmpR, so that the phosphotransfer reaction is readily reversible. We are currently investigating whether the wild type EnvZ also dephosphorylates phospho-OmpR by the reverse reaction of the OmpR-phosphorylation reaction.