K+ stimulates specifically the autokinase activity of purified and reconstituted EnvZ of Escherichia coli.

The histidine kinase/response regulator system EnvZ/OmpR of Escherichia coli regulates transcription of the genes ompF and ompC, encoding two porins of the outer membrane. Although the total amount of OmpF and OmpC remains constant, the relative levels of the two proteins fluctuate in a reciprocal manner depending on medium osmolality. The membrane-anchored sensor EnvZ somehow monitors changes in environmental osmolality. To characterize the nature of the stimulus perceived by EnvZ, this protein was overproduced, purified, and reconstituted into proteoliposomes. Autokinase activity of purified and reconstituted EnvZ was stimulated by an increase of the K(+) concentration. Rb(+), Na(+), and NH4(+) also stimulated the activity but to a smaller extent, whereas an osmotic upshift imposed by various sugars or increasing concentrations of glycine betaine, proline, or Tris/MES were without influence. Neither the transfer of the phosphoryl group from EnvZ approximately P to OmpR nor the EnvZ-mediated OmpR approximately P dephosphorylation were affected by one of the tested solutes. Experiments with the reconstructed signal transduction cascade including DNA fragments demonstrated a substantial increase of the amount of phosphorylated OmpR in the presence of K(+) and to a lower extent in the presence of Na(+), Rb(+), and NH4(+). Various K(+) salts were tested indicating that the determined effects were K(+)-specific and not dependent on the anion. In a further in vitro test system, which utilizes right-side-out membrane vesicles, the K(+)-specific activation of EnvZ autokinase from the luminal side was confirmed. These results clearly indicate a regulation of EnvZ autokinase activity by monovalent ions, specifically K(+). Whether K(+) accumulation, which is one of the first responses of E. coli after an osmotic upshift, is related to the stimulation of the EnvZ autokinase activity in vivo is discussed.

The adaptation of Escherichia coli to varying osmolalities is studied thoroughly (1). However, how changes in external osmolality are detected is still under investigation. We follow this question by studying the stimulus for EnvZ, a postulated osmosensor of E. coli. The membrane-bound histidine kinase EnvZ modulates in interplay with the soluble response regula-tor OmpR expression of ompC and ompF, encoding porins of the outer membrane. OmpF and OmpC differ in their pore diameter, which is 1.16 nm in the case of OmpF and 1.08 nm in the case of OmpC, leading to a diffusion rate 10 times higher for OmpF compared with OmpC (2,3). Although the sum of OmpF and OmpC remains constant, the relative levels change in a reciprocal manner depending on the osmolality of the medium, whereby low osmolality favors the synthesis of OmpF, and high osmolality leads to preferential synthesis of OmpC (4).
EnvZ is autophosphorylated by the use of ATP at a highly conserved His residue (His-243), and subsequently, the phosphoryl group is transferred to a highly conserved Asp residue of the response regulator OmpR (Asp-55). EnvZ also has phosphatase activity to dephosphorylate OmpRϳP (see Ref. 5 for review). Phosphorylated OmpR is a transcriptional factor that binds upstream of the promoter regions of ompC and ompF to modulate their expression (6 -8). In vitro, the equilibrium between phosphorylated and nonphosphorylated OmpR is drastically shifted toward the phosphorylated form in the presence of DNA fragments comprising OmpR binding sites (9,10).
Mutational analyses of EnvZ have identified amino acid regions that are critical for signaling. The most critical domains are centered in the linker region, the region between the second transmembrane and the cytoplasmic transmitter domain (14), and around the site of autophosphorylation including the X-region (15). Periplasmic deletions result in a constitutive OmpF Ϫ OmpC ϩ phenotype (16) or have no effect at all (17). A truncation of 38 amino acids prevents proper localization of EnvZ into the cytoplasmic membrane; the protein can be phosphorylated, but the response to osmolality is lost (16), indicating the necessity of membrane binding. Amino acid replacements in the transmembrane domains result in various phenotypes (15,18). Amino acid replacements of the conserved asparagine residue Asn-347 lead to an OmpF ϩ OmpC Ϫ phenotype probably as a result of the shift in the enzymatic activities of EnvZ (19). Thr-237 plays a critical role for EnvZ phosphatase activity (20). A number of kinase Ϫ phosphatase ϩ or kinase ϩ phosphatase Ϫ mutants were identified; most of them are EnvZ derivatives with amino acid replacements in a structural motif, called the X-region, following His-234 (15).
The primary signal to which EnvZ responds is still unclear.
In contrast to other sensor processes where ligand binding is involved (e.g. nitrate binds to the sensor kinase NarX (21), and citrate binds to the sensor kinase CitA (22)), sensing of changes in osmolality is difficult to attribute to a specific ligand. Because EnvZ responds both to polar and nonpolar solutes, it is proposed that EnvZ is activated by changes in the cytoplasmic, periplasmic, or extracellular water activity (a W ) (23). However, the picture is even more complex. It is known that glycine betaine antagonizes the osmotic repression of ompF (24) and that procaine represses ompF and induces ompC at low osmolality (25,26). Most of the in vitro analyses for EnvZ were done with a soluble truncated form (EnvZc), because of the lack of a method for solubilization and reconstitution of EnvZ. Here we describe the influence of various solutes on the activities of purified and reconstituted full-length EnvZ. The results obtained with two different in vitro test systems reveal a stimulation of the autokinase activity of EnvZ by monovalent cations, especially K ϩ .

EXPERIMENTAL PROCEDURES
Materials-[␥-32 P]ATP was purchased from Amersham Pharmacia Biotech. Goat anti-rabbit IgG-alkaline phosphatase conjugate was obtained from Biomol. Ni 2ϩ -NTA-agarose 1 and His 5 antibody were from Qiagen, and Bio-Beads were from Bio-Rad. Detergents were from Calbiochem. Purified E. coli lipids were purchased from Avanti Polar Lipids. Synthetic oligonucleotide primers were from Life Technologies, Inc., and the pET16b vector was from Novagen. All other reagents were reagent grade and obtained from commercial sources.
Oligonucleotide-directed Site-specific Mutagenesis-Construction of plasmid pFR29 -6His was achieved by insertion of six codons for His after the triplet corresponding to amino acid 450 of EnvZ by the overlap extension polymerase chain reaction method (33) using synthetic primers containing six triplets for His.
Preparation of Everted Membrane Vesicles-E. coli strain WH56 transformed with plasmid pFR29 -6His was grown aerobically at 37°C in KML complex medium (1% tryptone, 0.5% yeast extract, and 1% KCl) supplemented with ampicillin (100 g/ml) and kanamycin (50 g/ml). Overexpression of envZ-6His under control of the lac promoter was achieved by addition of 0.3 mM isopropyl-1-thio-␤-D-galactopyranoside when the culture reached an absorbance at 600 nm of ϳ0.8. 3 h later the cells were harvested. The everted membrane vesicles were prepared by passage of cells through a Ribi press and washed twice in EDTAcontaining buffer of low ionic strength (34). The vesicles were resuspended in 50 mM Tris/HCl, pH 8.0, containing 10% (v/v) glycerol, frozen in liquid nitrogen, and stored until use at Ϫ80°C.
Purification of EnvZ-6His-Proteins of everted membrane vesicles (140 mg) in buffer containing 50 mM Tris/HCl, pH 8.0, 10% glycerol, and 10 mM ␤-mercaptoethanol were solubilized with 1% (w/v) n-decyl-␤-Dmaltopyranoside (final protein concentration, 10 mg protein/ml). While stirring on ice, the detergent was added stepwise within 5 min, and incubation was extended for additional 25 min. This solubilization mixture was centrifuged at 264,000 ϫ g for 45 min. After centrifugation the supernatant containing the solubilized proteins was adjusted with NaCl and imidazole to reach the concentration of the purification buffer. Meanwhile 1.1 ml of Ni 2ϩ -NTA resin was preequilibrated with purification buffer (50 mM Tris/HCl, pH 8.0, 10% glycerol, 10 mM ␤-mercaptoethanol, 0.08% (w/v) n-decyl-␤-D-maltopyranoside, 0.5 M NaCl, and 30 mM imidazole). Solubilized proteins and preequilibrated resin were mixed in a centrifuge tube and incubated at 4°C for 30 min. The resin was allowed to settle down, unbound proteins were removed, and the resin was washed three times with purification buffer. EnvZ-6His was eluted by adding 2.5 ml of purification buffer containing 130 mM imidazole to the resin. The purified protein was either reconstituted into proteoliposomes or for experiments using the solubilized protein dialyzed against 50 mM Tris/HCl pH 8.0 buffer containing 10% glycerol.
Reconstitution of EnvZ-6His-Purified EnvZ-6His was reconstituted into E. coli phospholipids essentially as described (35). Briefly, E. coli liposomes (10 mg/ml) were solubilized with Triton X-100 (final concentration, 0.47% v/v). Then, EnvZ-6His in elution buffer was added (ratio of lipid to protein, 80:1 w/w), and the mixture was stirred for 10 min at room temperature. Pretreated Bio-Beads (36) at a bead to detergent ratio of 5 (w/w) were added, and the mixture was incubated under gentle movement at room temperature for 1 h. Fresh Bio-Beads were added, and the mixture was kept for another hour at room temperature. Finally, twice the amount of Bio-Beads was added, and the mixture was kept under gentle movement at 4°C overnight. The proteoliposomes solution was pipetted off into centrifugation tubes, and proteoliposomes were collected by centrifugation for 1 h at 372,000 ϫ g. The pellet was resuspended in 50 mM Tris/HCl, pH 8.0, and 10% glycerol. Proteoliposomes were either used instantly or stored in liquid nitrogen. The efficiency of reconstitution was calculated from the amount of protein obtained after ultracentrifugation.
Preparation of Right-side-out Membrane Vesicles-RSO-MV of E. coli strain MHA1160 transformed with plasmid pFR29 were prepared according to the protocol described recently (37).
Phosphorylation Assays with RSO-MV-The lumen of RSO-MV can be made accessible for ATP in the presence of Mg 2ϩ because of a permeabilizing effect of this cation (38). Therefore, both buffers (inside and outside) contained 20 mM MgCl 2 . Autokinase activity of EnvZ in RSO-MV (3 mg protein/ml) (isoosmolar buffers outside and inside) was initiated by addition of 100 M [␥-32 P]ATP (0.476 Ci/mmol). Autokinase of EnvZ in RSO-MV was found to be linear within the first 2 min. To obtain sufficient amounts of phosphorylated EnvZ in all experiments, the reaction was stopped after 2 min by the addition of an equal volume of 2ϫ concentrated SDS sample buffer (39). When the osmolality outside of the RSO-MV was varied, the vesicles were incubated in lysis buffer for 1 min with ATP and centrifuged (14.000 ϫ g, 0.5 min), and then the pellet was resuspended in the higher osmolal buffer (50 mM Tris/HCl, pH 8.0, plus osmolytes) lacking ATP and Mg 2ϩ . After 1 min of incubation, the reaction was stopped as described above.
All samples were immediately subjected to SDS-polyacrylamide gel electrophoresis (PAGE) (39). Shortly before stopping SDS-PAGE, an [␥-32 P]ATP standard was loaded on the gels. The gels were dried, and phosphorylation of the proteins was detected by exposure of the gels to a Storage Phosphor Screen. Phosphorylated proteins were quantified by image analysis using the PhosphorImager SI of Molecular Dynamics.
Phosphorylation and Dephosphorylation Assays-Solubilized EnvZ or EnvZ in proteoliposomes (1 M) was incubated with 100 M [␥-32 P]ATP (0.476 Ci/mmol) in phosphorylation buffer (50 mM Tris/HCl, pH 8.0, 10% glycerol, 2 mM dithiothreitol, and 110 M MgCl 2 ) containing varying concentrations of test solutes (assay volume, 21 l). At the times indicated, the reaction was stopped by the addition of an equal volume of 2ϫ concentrated SDS sample buffer.
To test EnvZ phosphotransfer and phosphatase activities, EnvZ in proteoliposomes (1 M) was phosphorylated as described above, whereby the buffer contained 50 mM KCl. After 5 min, proteoliposomes were collected by ultracentrifugation at 372,000 ϫ g for 45 min. The pellet was washed with 50 mM Tris/HCl buffer, pH 8.0, 10% glycerol, centrifuged, and resuspended in the same buffer, which contained in addition 5 mM MgCl 2 and various solutes (mostly at a concentration of 300 mM). Subsequently, equimolar amounts of purified 10His-OmpR were added, samples were taken, and the reaction was stopped as described above. After 5.5 min, ADP (1 mM) was added to maximize phosphatase activity, and further samples were taken.
To test the complete signal transduction cascade in vitro, EnvZ in proteoliposomes (1 M), purified OmpR (4 M), C1-C2-C3 DNA fragment (5 M) (9) were mixed in phosphorylation buffer containing vari-ous solutes as indicated. The reaction was started by the addition of 100 M [␥-32 P]ATP (0.476 Ci/mmol) and 8 M ADP, and samples were taken at the times indicated. The double-stranded C1-C2-C3 DNA fragment was obtained by annealing of two complementary oligonucleotides. The upper strand sequence (from 5Ј to 3Ј end) has the following composition: 5Ј-GGGGTTTACATTTTGAAACATCTATAGCGATAAATGAAACATC-TTAAAAGTTTTAGTATCATATTGGGG-3Ј.
In each case, the samples were immediately subjected to SDS-PAGE (39). The gels were dried, the radiolabeled proteins were detected by exposure of the gels to a phosphor screen, and the images were analyzed with a PhosphorImager system (Molecular Dynamics) using [␥-32 P] ATP as standard.
Analytical Procedures-Protein was assayed by the method described in Ref. 40 using bovine serum albumin as standard. The proteins were separated by SDS-PAGE (39) using 9 or 12% acrylamide gels and when indicated stained with silver (41). Immunodetection of EnvZ-6His or 10His-OmpR proteins with polyclonal antibodies against EnvZ or the His tag was performed as described (34).

Influence of the His Tag of EnvZ on the Expression
Pattern of ompF and ompC-To allow easy purification of EnvZ, six consecutive histidine residues were attached at the C-terminal end of the protein. The functionality of this protein was comparable with the untagged EnvZ because the ompC and ompF expression pattern was comparable with that of wild-type EnvZ as tested with E. coli strains WH56 and WH57, each transformed with plasmids, pFR29 or pFR29 -6His, respectively (data not shown).
Purification of EnvZ-6His-EnvZ-6His was purified by means of affinity chromatography. The following detergents were tested to be efficient in the solubilization of EnvZ-6His: n-octylglucoside, n-decylmaltoside, n-dodecylmaltoside, lauryldimethylamine oxide, zwittergent 3-12, and zwittergent 3-14. The autokinase activity of EnvZ-6His in these detergents was tested, and the highest activities were found with EnvZ-6His in decylmaltoside (data not shown). Therefore, for all subsequent steps of the purification, decylmaltoside was used as detergent. The highest purification results were achieved when binding of the protein to Ni 2ϩ -NTA-agarose was performed in the presence of 30 mM imidazole and 0.5 M NaCl. In a typical experiment 300 g of EnvZ-6His was obtained from 140 mg of membrane proteins. As judged from a silver-stained gel, the purity of EnvZ-6His was greater than 95% (Fig. 1). The purified protein was either dialyzed to remove imidazole or reconstituted into E. coli phospholipids.
Reconstitution of EnvZ-6His into Proteoliposomes-Reconstitution was carried out using the detergent-mediated method as described (35). E. coli phospholipids were solubilized with detergent and mixed with purified EnvZ-6His, and the detergents were removed by BioBeads. Detergent-dependent solubilization of liposomes can be followed by turbidity measurements, whereby three characteristic stages can be distinguished: onset, partial, and total solubilization. Rigaud et al. (35) found that the nature of the detergent used and the stage of solubilization of the liposomes influence the activity of the reconstituted protein. The following detergents were used for the solubilization of liposomes: Triton X-100, n-octylglucoside, n-decylmaltoside, and n-dodecylmaltoside. Purified EnvZ-6His was added at different solubilization stages. The highest activities of EnvZ-6His autokinase activity were determined when partially solubilized liposomes with Triton X-100 as detergent were used (data not shown). The efficiency of EnvZ-6His reconstitution was determined to be 53%.
Activities of Purified EnvZ-6His in Proteoliposomes and 10His-OmpR-In the presence of [␥-32 P]ATP reconstituted EnvZ was rapidly autophosphorylated. The autokinase activity was found to be linear for 2 min (Fig. 2A). Although autokinase activity of the solubilized protein was also detectable, activities of the reconstituted protein were about four times higher (data not shown). To test phosphotransfer and phosphatase activities, the following experimental approach was chosen. EnvZ-6His was phosphorylated under standard conditions for 5 min. Subsequently, proteoliposomes were collected by ultracentrifu- gation, washed, and resuspended in buffers containing 50 mM KCl and 5 mM MgCl 2 . 10His-OmpR was added, and at different time points samples were taken and analyzed (Fig. 2B). Although this experiment was started with the same amounts of EnvZ used in the autophosphorylation experiment ( Fig. 2A), the calculated sum of the phosphorylated proteins presented in Fig. 2B is much lower compared with the results presented in Fig. 2A. This discrepancy is on one hand due to the incomplete collection of proteoliposomes during ultracentrifugation and on the other hand due to smaller amounts of protein loaded onto the gels. Transfer of the phosphoryl group was detectable; however, the rate was slow, and the transfer was not completed within 5 min. The same setup was used to test EnvZ phosphatase activity. Thus, after 5.5 min ADP as a cofactor was added, and further samples were taken. As shown in Fig. 2B, after the addition of ADP the amounts of phosphorylated EnvZ and OmpR were rapidly declining, and the half-life of OmpRϳP was determined to be 3 min. Thus, purified EnvZ-6His and 10His-OmpR catalyzed all known enzymatic activities.
The Influence of Various Solutes on the Autokinase Activity of EnvZ-6His-EnvZ autokinase activity was tested in the presence of various solutes. It was found that this activity of EnvZ-6His was significantly stimulated in the presence of monovalent ions in a concentration-dependent manner (Fig. 3). The highest activities were detectable when KCl was added. NaCl, RbCl (Fig. 3), and NH 4 Cl (data not presented) had also stimulatory effects but to a lower extent. In contrast, no stimulation of the autokinase activity was found in the presence of Tris/ MES, sucrose, or trehalose, although these compounds were tested at equal osmolalities. Glycine betaine (Fig. 3) or proline (data not shown), which are accumulated in cells exposed to an osmotic upshift under certain conditions, did not influence EnvZ autokinase activity. For clarity, the results shown are representative for the whole concentration range tested for each compound, which was 25, 50, 100, 200, 300, and 500 mM in case of salts. In the case of the uncharged compounds, higher concentrations were used to achieve comparable osmolalities. For KCl and the other salts, a concentration-dependent stimulation was observed that reached the maximum at a concentration of 300 mM. Because monovalent ions, especially K ϩ , might be necessary for normal functioning of EnvZ autokinase activity, we also tested the influence of increasing concentrations of sucrose or trehalose in the presence of 50 mM KCl. Although, as expected, under these conditions higher amounts of EnvZϳP were detectable, the presence of sucrose or trehalose did not further increase activity.
The Influence of Various Solutes on the Phosphotransfer and OmpRϳP Phosphatase Activity-To test the influence of various solutes on the further activities of EnvZ-6His the experimental approach described above and in Fig. 2B was used. Centrifugation of proteoliposomes containing phosphorylated EnvZ allowed easy changes of buffers. Activities were tested in the presence of the following solutes: KCl, K ϩ glutamate, NaCl, RbCl, Tris/MES, glycine betaine, sucrose, and trehalose (each 0.3 M). None of them affected the transfer of the phosphoryl group to OmpR nor the dephosphorylation of OmpRϳP significantly (data not shown). The results obtained were identical to those shown in Fig. 2B.
Reconstruction of the Whole Signal Transduction Cascade in Vitro-Because in whole cells a stepwise addition of components of the signal transduction cascade does not exist, our next aim was to establish the whole signal transduction cascade in vitro. Recent experiments of Inouye and co-workers (9) demonstrated a dramatic shift toward the phosphorylated form of OmpR in the presence of DNA comprising OmpR binding sites. Thus, purified EnvZ-6His in proteoliposomes and 10His-OmpR in a ratio of 1:4 were mixed with DNA comprising the C1-C2-C3 binding sites of OmpR (9), and the reaction was started by the addition of a mixture of ADP and ATP (ratio of 1:12.5) and [␥-32 P] ATP. Samples were taken after 1 and 30 min. The autoradiograph is shown in Fig. 4A. As shown before, the amount of phosphorylated OmpR was increased in the presence of DNA by a factor of about 6, whereas the amount of phosphorylated EnvZ remained at a basal level in each case. This experiment was performed at a KCl concentration of 50 mM. In the next experiment the time-dependent signal transduction in the presence or absence of KCl was tested (Fig. 4, B and C). Under both conditions phosphorylated EnvZ was detectable, however at a very low level. The amount of phosphorylated OmpR increased over time. In the presence of 50 mM KCl the initial rate was 17 times higher than in the absence of KCl. Whereas in the presence of KCl half-maximal phosphorylation of OmpR was already reached after 5 min, in the absence of KCl the amount of phosphorylated OmpR was steadily rising within the tested time range.
The Influence of Different Solutes on the EnvZ/OmpR Signal Transduction Cascade in Vitro-The next experiments were undertaken to test the influence of various solutes on the EnvZ/OmpR signal transduction cascade. To calculate initial rates, the reaction was stopped after 2.5 min. As already shown before (Fig. 4), under all conditions phosphorylated EnvZ was detectable at a basal level (Fig. 5). The presence of NaCl, KCl, RbCl, K ϩ glutamate, and NH 4 Cl significantly increased the amount of phosphorylated OmpR, whereby the highest values were reached in the following order KCl Ͼ K ϩ glutamate Ͼ RbCl Ͼ NH 4 Cl Ͼ NaCl (Fig. 5). Maximal stimulation was observed in the presence of KCl. The stimulatory effect of these salts was found to be concentration-dependent, and maximal values were detected at a concentration of 100 mM. The values for K ϩ glutamate were lower compared with KCl. Because the presence of K ϩ glutamate affected the resolution of proteins in SDS gels, the determined differences might be due to the quantitative analysis of the phosphorimages. According to our experience, sharp distinct bands gave higher values than broad diffuse bands. An increase of the Tris/MES buffer concentration or the addition of sucrose, trehalose, or betaine did not affect the amounts of EnvZϳP or OmpRϳP.
At the highest concentration of KCl (500 mM), the amount of phosphorylated OmpR was reduced, whereas the amount of phosphorylated EnvZ was increased. Similar effects were observed for RbCl. However, the values represent initial rates. Analyses of the time courses of the signal transduction cascade at low and high K ϩ concentrations indicated that the phosphorylation of OmpR was delayed at higher K ϩ concentrations; however, OmpR reached the maximal phosphorylation level at later time points (data not shown).
To ensure that the determined effects are related to the cation and not to the chloride anion, the influence of the following K ϩ salts was tested: KNO 3 , K 2 SO 4 , and KBr. These salts increased the level of phosphorylated OmpR to the same extent as KCl (data not shown).
The Effect of Various Solutes on the Autokinase Activity of EnvZ in RSO-MV-Another in vitro test system based on EnvZ in RSO-MV was applied. This system has the following advan-tages: (i) the orientation of EnvZ is that of whole cells, (ii) RSO-MV still behave like osmometers, and (iii) there are two compartments in which the buffer composition can be altered. Recently, this system was applied successfully for the KdpD/ KdpE system (37). To determine EnvZ autokinase activity, the membrane vesicles have to be made accessible for ATP, which was done by permeabilization with Mg 2ϩ . A comprehensive study that describes ATP accessibility was described earlier (37). In the first experiment the influence of NaCl and KCl on the autokinase activity of EnvZ was tested under isoosmolar conditions. In the presence of KCl at concentrations of 300 mM and higher, a significant increase of the autokinase activity of EnvZ was observed (Fig. 6). In contrast, no stimulatory effect was observed by increasing the concentration of NaCl. In the next experiments the buffer concentration inside of the vesicles was held constant, and the osmolality was raised outside. To obtain reasonable amounts of phosphorylated EnvZ and to mimic more physiological conditions, the buffer inside of the vesicles contained 300 mM KCl. For each test, RSO-MV were loaded with radiolabeled ATP, collected by centrifugation and resuspended in buffer of increasing osmolality. The influence of the ionic solutes NaCl and KCl as well as the nonionic solutes sucrose, glucose, and sorbitol was tested. However, none of these compounds influenced EnvZ autokinase activity significantly (data not shown). Thus, a rise of the osmolality outside of the vesicles did not affect EnvZ autokinase activity. DISCUSSION Here we described a procedure for the purification and reconstitution of full-length EnvZ. All known activities were detectable for the purified and reconstituted EnvZ under standard phosphorylation conditions, although rates were different compared with a truncated EnvZ (EnvZc). In case of full-length EnvZ, transfer of the phosphoryl group to OmpR was slow and not completed within 5 min. Earlier, similar results were obtained with EnvZ enriched in membrane vesicles (15). For EnvZc a very rapid transfer of the phosphoryl group was shown (9). EnvZ-mediated OmpRϳP dephosphorylation was also slower compared with the results described for EnvZc (9). These examples illustrate that interactions of the domains of the sensor kinase probably fine tune the enzymatic activities. This is in accord with the findings of Inouye and co-workers in regard to the importance of an interaction between domains A and B for the phosphatase activity (43).
The availability of purified components allowed the reconstruction of the whole signal transduction cascade in vitro. As shown before (9), the amount of phosphorylated OmpR was drastically increased in the presence of DNA fragments comprising the OmpR-binding site. Our studies have shown that the addition of increasing concentrations of KCl raised tremendously the initial rate of OmpR phosphorylation as well as the steady state accumulation of OmpRϳP. An activation of the signal transduction cascade, as detected by the increased accumulation of OmpRϳP, was also achieved in the presence of other monovalent cations such as Rb ϩ , Na ϩ , or NH 4 ϩ , although to a lower extent compared with KCl. Because increasing concentrations of a Tris/MES buffer did not stimulate the signal transduction cascade, it is unlikely that changes of the ionic strength influence EnvZ activity. The data also imply that the stimulation is not due to an increase of the osmolality, because sucrose or trehalose did not show any effect. Osmotic upshifted cells are able to take up substantial amounts of proline or glycine betaine, but none of these compounds exhibited a stimulatory effect on EnvZ autokinase activity. Recently, it was shown that chloride ions are important for the osmoregulation of the halophilic bacterium Halobacillus halophilus (44). To ascertain that the determined effects are due to the cation and not to the anion, we tested the influence of different salts. We found that the presence of KNO 3 , K 2 SO 4 , or KBr had the same stimulatory effects as KCl. Thus, K ϩ and related monovalent ions specifically enhance the production of OmpRϳP. Earlier, Mizuno and co-workers (25) characterized EnvZ in membrane vesicles, and they found also an increased accumulation of phosphorylated OmpR depending on an increase of K ϩ , Na ϩ , or Li ϩ . They attributed this effect to a decrease of the EnvZ phosphatase activity; however, they did not measure phosphatase activity directly. Our studies of the single enzymatic activities of EnvZ revealed that K ϩ and other monovalent cations specifically stimulate the autokinase activity, whereas the rates of phosphotransfer and phosphatase activities of EnvZ are unchanged. Therefore, we conclude that accumulation of OmpRϳP in the presence of K ϩ is due to an increased autokinase activity of EnvZ.
The K ϩ -specific effect on EnvZ autokinase activity was confirmed when we used another in vitro test system, which is based on RSO-MV. When the K ϩ concentration in the lumen was increased, autokinase activity of EnvZ was stimulated. This effect was not observed in the presence of NaCl or when outside of the vesicles the concentration of KCl, NaCl, sucrose, glucose, or sorbitol was increased. Thus, domains of EnvZ exposed to the luminal side, which are the cytoplasmic domains, are sensitive toward K ϩ .
It is known that expression of both ompF and ompC requires phosphorylated OmpR; preferential expression of one of these genes depends on the amount of phosphorylated OmpR (4). Here we demonstrated that the amount of OmpRϳP is sub-stantially increased by stimulating the autokinase activity of EnvZ with increasing concentrations of K ϩ ions. Uptake of K ϩ appears to be the earliest response of E. coli after an osmotic upshift (45). Therefore, it is suggested that the increased K ϩ concentration in osmotic stressed cells raises the autokinase activity of EnvZ, thereby increasing the amount of OmpRϳP and leading to ompC expression. Earlier, Epstein and co-workers (46) had shown the interdependence of K ϩ and glutamate accumulation in cells exposed to an osmotic upshift. In accord with a stimulation of EnvZ autokinase activity by K ϩ is the finding that ompC is not induced when the osmotic upshift is done in the presence of betaine, a condition that reduces the rate and the extent of K ϩ accumulation (24). Regulation of one of the enzymatic activities of EnvZ is in agreement with the numerous EnvZ derivatives with altered enzymatic properties that lead to altered ompF/ompC expression patterns. Furthermore, the sensitivity of the cytoplasmic domains toward K ϩ fits well with EnvZ mutants with altered sensing properties, which have single amino acid replacements within the linker or the X domain, which are both cytoplasmic (14,15). Despite these agreements, the correlation between the accumulation of K ϩ and the stimulation of EnvZ autokinase in vivo still has to be shown. In addition, in vitro studies indicated that maximal amounts of phosphorylated OmpR were already detected at physiological K ϩ concentrations. Therefore, an increase of the intracellular K ϩ concentration seems to be one primary stimulus perceived by EnvZ but not the only one. Integration host factor and DNA bending have already been shown to be involved in the transcriptional regulation of ompF (47)(48)(49).
The obtained results are clearly distinct from the activation mechanism of the osmosensor and transporter ProP of E. coli. Transport activity of purified ProP in proteoliposomes (50) or RSO-MV (42) was significantly increased upon an osmotic upshift imposed by NaCl or sucrose. In addition, the results obtained for EnvZ are different to KdpD, a sensor kinase, which responds to K ϩ limitation or an osmotic upshift imposed by salts. In the RSO-MV test system, KdpD autokinase activity was activated by an increase of the NaCl concentration and inhibited by K ϩ ions in the lumen of the vesicles (37). When the salt concentration outside of the vesicles was raised, a small but significant stimulation of autokinase activity of KdpD was observed, whereas no effects were seen for EnvZ under these conditions.
In summary, EnvZ catalyzes several reactions: its autophosphorylation, the transfer of the phosphoryl group to OmpR, and the dephosphorylation of OmpRϳP. Our results reveal that the autokinase activity of EnvZ is stimulated by monovalent cations, specifically K ϩ ions. The logical consequence of these results is the search for EnvZ mutants that lost the sensitivity toward K ϩ .