Cysteine-Scanning Analysis of the Dimerization Domain of EnvZ, an Osmosensing Histidine Kinase

Abstract

EnvZ and OmpR are a transmembrane sensor and its cognate response regulator, respectively, regulating the transcription of porin genes in response to medium osmolarity in Escherichia coli. The cytoplasmic domain of EnvZ (EnvZc) possesses both kinase and phosphatase activities and can be dissected into two functional domains, A and B. Here, we performed a cysteine-scanning analysis of domain A, a 67-residue central dimerization and phosphatase domain containing His-243 as the phosphorylation site, and we examined the effects of the cysteine substitution mutations on the enzymatic activities of domain A. The substitution mutations were made at 31 residues, from which 24 mutant domain A proteins were biochemically characterized. From the analysis of the phosphatase activity of purified mutant proteins, it was found that there are two regions in domain A which are important for this activity. Cysteine mutations in these regions dramatically reduce or completely abolish the phosphatase activity of domain A. The mutations that have the most-severe effects on domain A phosphatase activity also significantly reduce the phosphatase activity of EnvZc containing the same mutation. Using an in vitro complementation system with EnvZc(H243V), these cysteine mutants were further characterized for their autophosphorylation activities as well as their phosphotransfer activities. The results indicate that some mutations are specific either for the phosphatase activity or for the kinase activity.

The EnvZ-OmpR system in Escherichia coli belongs to the histidyl-aspartyl phosphorelay signal transduction system (also known as the two-component signal transduction system), which serves as the major signal transduction system in prokaryotes in response to various environmental stresses and growth conditions. The basic components of this system are a dimeric histidine kinase serving as a signal receptor and its cognate response regulator mediating the signal to specific gene expression or cellular locomotion (8, 17).

EnvZ is a transmembrane histidine kinase that monitors environmental osmolarity changes. It is autophosphorylated using ATP at the highly conserved His residue (His-243) in the cytoplasmic domain. This high-energy phosphoryl group is subsequently transferred to the highly conserved Asp-55 residue of OmpR, forming phosphorylated OmpR (OmpR-P). OmpR-P is a transcription factor which binds upstream promoter regions of outer membrane porin genes ompF and ompC and differentially modulates their expression according to the cellular OmpR-P level. Notably, EnvZ also acts as a phosphatase that dephosphorylates OmpR-P to regulate the cellular OmpR-P concentrations. Thus, the medium osmolarity mediates the porin composition by adjusting the kinase-to-phosphatase ratio of EnvZ (mainly by adjusting phosphatase activity [10]). Recently, the cellular contents of EnvZ and OmpR were estimated to be 100 and 3,500 molecules per cell, respectively (2). While it has been reported that phosphorylation of OmpR decreases its binding affinity to EnvZ (13), we have demonstrated in a number of different ways that the cytoplasmic region of EnvZ (EnvZc) interacts with OmpR-P almost as well as with OmpR and that EnvZc forms a stable stoichiometric complex with OmpR-P as well as with OmpR (20, 21).

The dual activities of EnvZ are located in its cytoplasmic region, which consists of a linker region, domain A, and domain B (14). The linker region was found to be important for transducing the osmolarity signal from the extracellular receptor domain through the transmembrane domain to the cytoplasmic enzymatic domain. However, the linker region is not directly involved in kinase or phosphatase activities. Domain A, containing the autophosphorylation site His-243, forms a stable dimer and can be phosphorylated in the presence of ATP by another molecule of EnvZc such as EnvZc(H243V), an autophosphorylation-defective mutant. The phosphorylated domain A subsequently transfers the high-energy phosphoryl group to OmpR (14). Domain B binds with ATP and phosphorylates the domain A of its partner in an EnvZ dimer. Because of their unique features, domains A and B are termed the DHp (mnemonic for dimerization and histidine phosphotransfer) domain and the CA (catalysis-assisting and ATP-binding) domain, respectively (3, 8). The structures of both domains have been solved by nuclear magnetic resonance (NMR), which revealed that a dimer of domain A exhibits a four-helix bundle structure (19) and that domain B has a signature protein kinase motif (α/β sandwich fold [18]). These structural features are highly conserved in other histidine kinases, including CheA (1) and PhoQ (12).

Domain A of EnvZ has been shown to be responsible for the phosphatase activity (24). A corresponding region in NtrB has also been demonstrated to show phosphatase activity toward phosphorylated NtrC (9, 11), indicating that the conserved DHp domain plays the major role in both kinase and phosphatase activities of the bifunctional histidine kinases. Since the phosphatase activity of domain A can be significantly stimulated by a covalently linked domain B in EnvZ, it is proposed that the osmolarity signal transduced through the membrane alters the relative spatial arrangement between domains A and B to modulate the phosphatase activity (24).

A number of phosphatase mutants of EnvZ have been identified, and the majority of them are located in the conserved H-box region containing His-243, indicating the importance of the H-box in the phosphatase reaction. These mutants include G240E, V241G, S242D, H243R, -V, -N, -Q, and -D, and T247A, -E, -K, -R, -C, -Y, -Q, and -N (4, 6). However, most of these studies were done with either full-length EnvZ or the cytoplasmic region of EnvZ containing domain B, which by itself does not have the phosphatase activity but is able to stimulate the phosphatase activity of domain A. Therefore, some mutations at the residues that are important for the interaction between domains A and B could indirectly affect the phosphatase activity of EnvZ. For example, an X region (a weakly conserved motif located at the middle and at the C terminus of helix II of domain A) mutant, L288P, that has been shown to severely affect the phosphatase activity of full-length EnvZ (7) has no effect on domain A's phosphatase activity in the absence of domain B (24). This result suggests that the major role of this region is to correctly position domain B toward domain A to modulate the phosphatase activity of domain A. To date, only His-243 and Thr-247 have been studied for their roles in the phosphatase activity of domain A. Mutations at either residue completely abolish domain A's phosphatase activity, indicating that these two residues may be directly involved in the phosphatase activity of domain A (4, 24).

Here, we performed cysteine-scanning mutagenesis on domain A in order to further dissect the functional regions of domain A. We chose cysteine-scanning rather than alanine-scanning mutagenesis for functional analysis because cysteine residues are more versatile, since they can be used for fluorescent probe attachment and for cross-linking between domain A (or EnvZc) and OmpR. Phosphatase analysis was performed with each mutant to identify residues and regions in domain A that are important for the phosphatase activity. Using an in vitro complementation system with EnvZc(H243V), these cysteine mutants were further characterized for their autophosphorylation activities as well as their phosphotransfer activities. These results indicate that some mutations are specific either for the phosphatase activity or for the kinase activity.

MATERIALS AND METHODS

Plasmids.

PET11a-EnvZc(223-289) encoding domain A and pET11a-EnvZc encoding EnvZc were constructed previously (14). All cysteine-scanning mutants of domain A were constructed by use of a site-directed mutagenesis kit (Stratagene, La Jolla, Calif.) using pET11a-EnvZc(223-289) as a template. EnvZc(T247R) and EnvZc(H243V) were created previously (4, 14). EnvZc(R246C), EnvZc(L254C), and EnvZc(K272C) were constructed by use of a site-directed mutagenesis kit (Stratagene) using pET11a-EnvZc as a template.

Protein purification.

All cysteine-scanning mutants of domain A were purified in the same way as described previously for wild-type domain A protein (24). All cysteine-scanning mutants of EnvZc were purified in the same way as described previously for wild-type EnvZc (15). It is important to note that 5 mM β-mercaptoethanol was present in all purification solutions. EnvZc(H243V) was purified as described previously (14).

Preparation of OmpR-P, in vitro phosphatase assay, and calculation of rate constant for phosphatase reaction.

The preparation of ³²P-labeled OmpR-P and the in vitro phosphatase assay were carried out as described previously (24). OmpR-P (1.5 μM) was incubated with 2.5 μM concentrations of various domain A mutant proteins in phosphatase reaction buffer (50 mM Tris · HCl [pH 8.0], 50 mM KCl, 10 mM MgCl₂, 1 mM dithiothreitol, and 5% glycerol) at room temperature. Aliquots were taken at six different time points depending on the level of the phosphatase activity of domain A mutant proteins, analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and followed by phosphorimager quantification to estimate the amounts of OmpR-P. Accordingly, the half-life of OmpR-P (t_1/2) was calculated for each domain A mutant protein. The value of k (rate constant) was determined by the following formula: k = ln 2/t_1/2 − ln 2/t_1/2auto, where t_1/2auto is the half-life of OmpR-P alone and is experimentally determined to be 90 min.

Phosphorylation of domain A and phosphotransfer from phosphorylated domain A to OmpR.

Domain A protein fragments (2.5 μM) were mixed with EnvZc(H243V) (1 μM) in autophosphorylation buffer(50 mM Tris-HCl [pH 8.0], 50 mM KCl, 5 mM MgCl₂, 5 mM β-mercaptoethanol, and 5% glycerol) in the presence of 50 μM [γ-³²P]ATP at room temperature. Aliquots were taken at 5 and 30 min. After 30 min, OmpR (5 μM) was added to the reaction mixtures, and then aliquots were taken at 3 and 6 min after the addition of OmpR. All samples were analyzed by SDS-20% PAGE. The amounts of phosphorylated OmpR and phosphorylated domain A were estimated by a phosphorimager.

RESULTS

Cysteine scanning of domain A.

A total of 31 of 67 residues of domain A were replaced by Cys. The following 36 residues were not replaced with Cys: (i) the residues from Met-223 to Arg-234 at the N-terminal region and the residues Tyr-287, Leu-288, and Arg-289 at the C-terminal region, which are known to be unstructured according to the NMR structure of domain A (19); (ii) the residues L237, M238, V241, L245, P248, L249, I252, A255, L266, A267, I270, I274, C277, I280, I281, F284, and I285, forming the hydrophobic core of the four-helix bundle structure; (iii) the highly conserved His-243 residue essential for both kinase and phosphatase activities (24); (iv) Thr-247, since a number of substitution mutations at this residue have already been isolated and their effects on the EnvZ function have been well investigated (4); and (v) Ala-239 and Ala-279 because of their structural resemblance to Cys.

All the Cys mutants were cloned in pET11a and expressed in E. coli strain BL21(DE3) cells. Twenty-four mutant proteins were expressed at reasonable levels and subsequently purified. Note that 5 mM β-mercaptoethanol was added to all preparations. All these proteins were soluble, and their circular dichroism spectra were similar to that of the wild-type domain A, indicating that the four-helix bundle structure was maintained in these mutant proteins (data not shown). The remaining seven mutant proteins (D244C, R251C, R253C, M258C, M259C, D263C, and D273C) were produced at such low levels that we were unable to obtain sufficient amounts for their biochemical characterization. The residues mutated in the present study are shown in green in the domain A primary sequence (Fig. 1A) and in the helical wheel view (Fig. 1B). His-243 and Thr-247, whose mutants had been analyzed previously, are shown in purple in this figure.

FIG. 1.

Cysteine-scanning analysis of domain A of EnvZ. (A) Amino acid sequence of domain A. Every 10th amino acid in the sequence is marked with a triangle above it. The residues shown in green in panels A and B constitute the cysteine-scanning site, and mutants with these residues replaced were the subjects of the biochemical characterization in this report. His-243 and Thr-247, which were demonstrated previously to be important for the domain A's phosphatase activity, are shown in purple. The Cys mutants with mutations at seven residues (D244, R251, R253, M258, M259, D263, and D273) were produced at such low levels that we were unable to obtain sufficient amounts for their biochemical characterization. These seven residues along with other unmutated residues are shown in black. On the basis of the NMR structure of domain A (19), the regions for helix I, helix II, and the turn region between two helices are shown above the sequence. The regions (I and II) highly affected by mutations in the phosphatase activity are indicated under the sequence. (B) Helical wheel representation of domain A dimer in a projection down the bundle axis. N or C at the center of each wheel indicates the terminus that is closer to the reader. Subunits A and B are labeled in red and yellow, respectively. The hydrophobic residues are shown as filled yellow circles.

The OmpR-P phosphatase activities of all mutant proteins were measured and compared to those of wild-type domain A under identical conditions and listed in Table 1. The half-lives of OmpR-P (t_1/2) were then converted into rate constants (k) that represent phosphatase activities. Since in the phosphatase reaction the hydrolysis of OmpR-P occurs not only because of the EnvZ phosphatase activity but also because of the autophosphatase activity of OmpR-P by itself, the rate constant of the autophosphatase activity (k_aut) was subtracted from the observed rate constant (k_obs) to obtain the actual rate constant (k =k_obs − k_aut) shown in Table 1. Note that the rate constant of the autophosphatase activity (k_aut) of OmpR-P is equal to ln 2/t_1/2 (90 min) for OmpR-P alone. The observed rate constants (k_obs) were calculated as ln 2 divided by the half-lives of OmpR-P for individual domain A mutants.

TABLE 1.

Phosphatase activities of domain A mutants^a

Cys mutant	t1/2 (min)	k (10−2 min−1)	Activity (%)	Region	Class
Wild type	12	5.0	100
T235C	14	4.2	84		B
L236C	25	2.0	40		D
G240C	10	6.2	123		D
S242C	39	1.0	20	I	A
H243S, -N, -K, -Y, -V	90	0.0	0	I
R246C	62	0.3	7	I	B
T247R	90	0.0	0	I
T250C	23	2.2	45	I	B
L254C	90	0.0	0	I	A
T256C	27	1.8	36	I	A
E257C	30	1.5	31	I	A
S260C	13	4.6	91		C
E261C	15	3.9	77		B
Q262C	11	5.5	111		B
G264C	12	5.0	100		B
Y265C	20	2.7	54		A
E268C	23	2.2	45		A
S269C	20	2.7	54		A
N271C	36	1.2	23	II	A
K272C	55	0.5	10	II	D
E275C	28	1.7	34	II	C
E276C	68	0.2	5	II	C
N278C	45	0.8	15	II	A
E282C	15	3.9	77		B
Q283C	13	4.6	91		C
D286C	16	3.6	71		C

^aThe relative activity of each domain A mutant is given relative to the rate constant of the wild-type domain A protein fragment. The half-lives of His-243S, -N, -K, -Y, and -V and T247R mutants are cited from previous reports (4, 24).

Interestingly, as shown in Table 1 and Fig. 1, residues affecting the phosphatase activity of domain A are mainly clustered into two regions (I and II). All mutations located in these regions significantly decreased the phosphatase activity to less than 50% of that of the wild-type domain A. Region I, from S242 to E257, contains His-243, the autophosphorylation site. It is important to note that in addition to previously known mutations (H243S, -N, -K, -Y, and -V and T247R) L254C in this region also completely abolished the OmpR-P phosphatase activity, since the half-life of OmpR-P in the presence of domain A(L254C) was 90 min, the same as that of OmpR-P alone. In the NMR structure of domain A, the hydrophobic side chain of Leu-254 is fully exposed to the solvent like that of His-243, suggesting that this residue may play a role in the interaction between OmpR-P and domain A. Moreover, cysteine mutations at two residues located in the H-box, Ser-242 and Arg-246, severely reduced the phosphatase activity to 20 and 7% of that of wild-type domain A, respectively (Table 1). Regions II containing Asn-271, Lys-272, Glu-275, Glu-276, and Asn-278 exist in the middle of helix II. Cys substitutions at these residues significantly lowered the phosphatase activity to 23, 10, 34, 5, and 15% of that of wild-type domain A, respectively (Table 1). It is interesting that Cys mutations at the turn structure, the N terminus of helix I, and the C terminus of helix II have little effect on domain A's phosphatase activity. In a three-dimensional structure, these residues are located either at the top or the bottom of the four-helix bundle.

As shown in Fig. 1B, among the residues strongly affecting the phosphatase activity (less than 50% of that of the wild-type domain A), some are located at the intrasubunit surfaces. These residues include Ser-242, Arg-246, Glu-268, Asn-271, Glu-275, and Asn-278. Some residues, including His-243, Thr-250, and Lys-272, are located at both intra- and intersubunit surfaces. The remaining residues (Thr-247, Leu-254, and Glu-276) are at the intersubunit surfaces.

Effects of domain A mutations on the phosphatase activity of EnvZc.

Next, we examined how the same mutations that remarkably reduced the phosphatase activity of domain A to less than 10% of that of the wild type affected the phosphatase activity of EnvZc. Those mutations are located at His-243, Arg-246, Thr-247, Leu-254, Lys-272, and Glu-276 (Table 1). Previous studies have already demonstrated that His-243, Thr-247, and Glu-276 play important roles in the phosphatase activity of EnvZc, showing that various substitutions at those residues in EnvZ or EnvZc either significantly reduced or completely abolished EnvZ's phosphatase activity (4, 6, 16, 23). Therefore, in the present study, we only examined how R246C, L254C, and K272C affected the phosphatase activity when they were introduced in the EnvZc construct. The phosphatase activity of EnvZc can be greatly stimulated by cofactor ADP, which binds to domain B, as the half-life of OmpR-P was less than 0.5 min in the presence of 1 mM ADP, compared with 8 min in the absence of ADP (Table 2). As shown in Table 2, in the absence of ADP mutation R246C completely abolished the phosphatase activity of EnvZc. The phosphatase activities with L254C and K272C mutations were severely reduced, as the half-lives of OmpR-P were 67 and 60 min, respectively. On the other hand, in the presence of ADP, the half-lives of OmpR-P were significantly shortened to 25, 16, and 36 min for R246C, L254C, and K272C, respectively. Nevertheless, these half-lives were still much longer than that observed with the wild-type EnvZc (less than 0.5 min). We also presented data about EnvZc(H243V) in Table 2 because in our study the phosphatase activity of EnvZc(H243V) was not detectable either in the presence or in the absence of ADP; but it has been reported that EnvZc containing H243V significantly reduced but did not completely abolish the phosphatase activity (16). The reason for this discrepancy is not known at present. Note that the H243V mutation in domain A completely abolished the phosphatase activity (24).

TABLE 2.

Phosphatase activities of domain A mutants of EnvZc

Mutants	t1/2 (min)
	No ADP	1 mM ADP
EnvZc(H243V)	90	90
EnvZc(R246C)	90	25
EnvZc(L254C)	67	16
EnvZc(K272C)	60	36
EnvZc	8	<0.5

In summary, our present results and previous reports demonstrated that these residues severely affecting domain A's phosphatase activity are also involved in the phosphatase reaction of EnvZc, implying that these residues may be directly involved in the phosphatase reaction. These results further confirm our previous conclusion that domain A is the phosphatase domain (24).

Effects of Cys mutations on the domain A phosphorylation by EnvZc(H243V) and on the phosphoryl transfer from domain A to OmpR.

It has been shown that domain A can be easily phosphorylated by EnvZc(H243V), an autophosphorylation mutant, to form phosphorylated domain A (A-P). Subsequently the phosphoryl group on domain A can be efficiently transferred to OmpR, forming OmpR-P (14). Therefore, we next studied the effects of Cys mutations on the phosphorylation of domain A in a complementing system with EnvZc(H243V) (autophosphorylation activity) and the ability of the resultant phosphorylated domain A to transfer the phosphoryl group to OmpR (phosphotransfer activity). Figure 2A shows some examples of these reactions using the wild-type domain A (lanes 1 to 4), domain A(G264C) (lanes 5 to 8), and domain A(Y265C) (lanes 9 to 12). Similar experiments were performed with all other mutant domain A proteins. The efficiencies of the phosphorylation by EnvZc(H243V) were estimated by the relative amounts of mutant A-P compared to that of the wild-type A-P at 30 min (Fig. 2B). The efficiencies of the phosphotransfer reaction from A-P to OmpR were estimated as the ratios of the amounts of OmpR-P at 3 min after addition of OmpR (Fig. 2A, lanes 3, 7, and 11) to the amounts of A-P just before addition of OmpR (Fig. 2A, lanes 2, 6, and 10). The results are shown in Fig. 2C.

FIG. 2.

Phosphorylation of mutant domain A proteins by EnvZc(H243V) and the phosphotransfer from phosphorylated mutant domain A proteins to OmpR. (A) A typical gel showing the phosphorylation of domain A protein by EnvZc(H243V) and the phosphotransfer from phosphorylated domain A (A-P) to OmpR. Wild-type domain A or domain A cysteine mutants at a concentration of 2.5 μM were mixed with 1 μM EnvZc(H243V) in the presence of 50 μM [γ-³²P]ATP at room temperature. Aliquots were taken at 5 min (lanes 1, 5, and 9) and 30 min (lanes 2, 6, and 10). After 30 min, 5 μM OmpR was added into the reaction mixtures. Then, aliquots were taken at 3 min (lanes 3, 7, and 11) and 6 min (lanes 4, 8, and 12). Samples were analyzed by SDS-20% PAGE. The gels were exposed to and quantified by a phosphoimager. (B) Effects of domain A mutation on the phosphorylation by EnvZc(H243V). The amount of phosphorylated wild-type domain A at 30 min is set as 100%. (C) Phosphotransfer from phosphorylated mutant domain A to OmpR. The ratio of OmpR-P to A-P on the y axis represents the ratio of the amount of OmpR-P at 3 min after addition of OmpR to the amount of phosphorylated mutant domain A at 30 min before addition of OmpR. OmpR-P/A-P of wild-type domain A was calculated to be 0.26 and set as 100%.

On the basis of the results shown in Fig. 2B and C, domain A mutants may be classified into four classes, designated A to D (Table 1). Class A includes those for which no strong negative effects on both the phosphorylation of domain A and the phosphotransfer activity are observed. S242C, L254C, T256C, E257C, Y265C, E268C, S269C, N271C, and N278C are in this class. Class B includes those for which no serious effects on the phosphorylation reaction but poor phosphotransfer activities are observed. Mutations T235C, R246C, T250C, E261C, Q262C, G264C, and E282C are in this class. It is particularly noteworthy that Glu-261, Gln-262, and Gly-264 are clustered in the turn structure between the two helices (Fig. 1A). This indicates that this turn region is not directly involved in the autophosphorylation reaction of EnvZ but plays an important role in the phosphotransfer reaction from EnvZ to OmpR. Since these mutations did not affect the phosphorylation reaction of domain A and the circular dichroism spectra of mutant proteins were similar to that of the wild type, it seems that they did not disrupt the four-helix bundle structure. Therefore, one may speculate that these residues at the turn region are directly involved in the interaction with OmpR. The third class, class C, comprises those for which the phosphorylation activities are approximately 30 to 60% of that of the wild-type domain A but the phosphotransfer activities are as high as or substantially higher than that of the wild-type domain A. S260C, E275C, E276C, Q283C, and D286C are in this class. Except for S260C, which is at the turn structure, all these mutants are located in the middle to the upper part of helix II. The last class, class D, comprises those in which both activities are severely affected by the substitution mutations. L236C, G240C, and K272C belong to this class.

DISCUSSION

In this study we performed cysteine-scanning analysis of EnvZ domain A, the phosphatase domain in the histidine kinase, and studied the effects of the Cys mutations on three enzymatic activities: OmpR-P phosphatase activity, autophosphorylation activity by EnvZc(H243V), and phosphotransfer activity.

From phosphatase analysis, we identified two regions in domain A that play important roles in the phosphatase activity. These regions are located at the middle of helices I and II and contain three residues (Arg-246, Leu-254, Lys-272) that are reported for the first time to be critical not only for the phosphatase activity of domain A but also for the phosphatase activity of EnvZc. Alignment of the H-box sequences of histidine kinases revealed that only basic residues, mainly Arg, appear at position 246 (5). These results taken together with our present results indicate that a basic residue located at one turn downstream of the phosphorylation site may be essential for the phosphatase reaction.

Recently the structure of the Spo0B and Spo0F complex has been determined (22). Spo0B is a phosphotransferase involved in the sporulation pathway of Bacillus subtilis. It consists of a dimeric four-helix bundle in the center that is flanked by two α/β sandwich folds and thus has a tertiary structural arrangement similar to that of EnvZ. Spo0F is the cognate response regulator for Spo0B. In the complex, Spo0F contacts Spo0B through a hydrophobic patch. Leu-38 in Spo0B, located two helix turns downstream of His-30 (the phosphorylation site in Spo0B), is important for this hydrophobic interaction because it interacts with Gly-14 and Ile-15 of Spo0F. In the present study, we found that Leu-254, located three helix turns downstream of His-243 and completely solvent exposed, plays an important role in the phosphatase reaction. A mutation at this position led to a complete loss of the phosphatase activity of domain A and severely affected the phosphatase activity of EnvZc (Tables 1 and 2). However, the phosphotransfer from phosphorylated domain A(L254C) to OmpR was not affected (Fig. 2C). In a similar role of Leu-38 in Spo0B, Leu-254 in EnvZ may be a crucial residue for the hydrophobic interaction with phosphorylated OmpR. By mutating this residue to Cys, domain A is no longer able to make contact with OmpR-P, thus losing its phosphatase activity.

When the classification of domain A mutants based on a comparison of the phosphorylation reaction by EnvZc(His243V) and the phosphotransfer reaction to OmpR with the phosphatase activities (Table 1) is examined, an interesting correlation emerges. All class A mutations failed to affect the phosphorylation of domain A or its phosphotransfer activity, yet they substantially reduced its phosphatase activity. In particular, no phosphatase activity was detected in domain A(L254C). The majority of class B mutations, which have no strong effect on phosphorylation but have poor phosphotransfer activity, maintain reasonably high phosphatase activities. This may indicate that domain A with those mutations may interact poorly with OmpR while these domain A mutants may be still able to interact reasonably well with OmpR-P, probably because the phosphoryl group facilitates OmpR interaction with domain A. Among class B mutations, R246C, which caused a significant reduction in the phosphatase activity, is an exception. Arg-246 may be important for the enzymatic reactions of domain A, as it is located close to His-243, the active center for both OmpR kinase and OmpR-P phosphatase activities. Both class C and class D mutations are divided into two subclasses: one maintains rather high phosphatase activities (G240C, S260C, Q283C, and D286C), and the other has low phosphatase activities (L236C, K272C, E275C, and E276C). The significance of these subclasses is not clear at present.

The cysteine-scanning experiments described in the present paper provide interesting clues to further characterization of OmpR(OmpR-P)-EnvZ interaction. Among the mutations studied here, those mutations in the turn region between helix I and helix II are the most interesting because residues in this region appear to directly interact with OmpR and OmpR-P. Cross-linking experiments between EnvZc with a cysteine substitution mutation in the turn region and OmpR with a cysteine substitution mutation at an appropriate site may provide insights into the question as to how OmpR and OmpR-P interact with EnvZ. Such experiments are currently in progress in our laboratory.