Phosphorylation of nodulin 26 on serine 262 affects its voltage-sensitive channel activity in planar lipid bilayers.

Nodulin 26 is an symbiosome membrane protein of soybean nodules that shows ion channel activity in planar lipid bilayers. Serine 262 of nodulin 26 is phosphorylated by calmodulin-like domain protein kinase. To study the effects of phosphorylation, nodulin 26 with Ser, Ala, or Asp at position 262 were expressed in Escherichia coli. The expressed protein possessed a histidine-rich leader sequence for purification by Nichelate fast protein liquid chromatography. Upon reconstitution into planar lipid bilayers, the recombinant proteins showed a large single channel conductance (3.1 nanosiemens (nS) in cis/trans and 1.6 nS in cis/trans) and weak anion selectivity, similar to native soybean nodulin 26. Nodulin 26 with Ser- or Ala-262 occupied the maximal open conductance state greater than 97% of the time (3.1 nS in cis/trans) regardless of applied voltage. However, nodulin 26 with Asp-262 showed increased gating and preferential occupancy of lower subconductance states (1.8 and 0.6 nS in cis/trans) at high applied voltages (e.g. 70 mV). In situ phosphorylation of Ser-262 of nodulin 26 by calmodulin-like domain protein kinase also resulted in increased voltage-dependent gating and preferential occupancy of lower subconductance states. These results suggest that phosphorylation of serine 262 of nodulin 26 modulates channel activity by conferring voltage sensitivity.

The establishment of symbioses between legumes and rhizobia bacteria represents a specialized developmental pathway that leads to the formation of a root nodule on the plant host. The bacteria infect this structure and become enclosed in intracellular organelles known as symbiosomes (1). The symbiosome membrane encloses the bacterium and controls the exchange of metabolites and nutrients between the host and the bacterial symbiont (2). During nodule formation, nodule-specific genes are induced that encode proteins that aid in the establishment and maintenance of the symbiosis. Among these is nodulin 26, which is a major integral symbiosome membrane protein of soybean nodules (3,4).
Nodulin 26 is a member of the MIP 1 channel protein family, but its role in symbiosome membranes remains unknown. Recently, the in vitro activity of purified soybean nodulin 26 was studied by reconstitution into planar lipid bilayers for single channel conductance measurements (5). Nodulin 26 formed channels with a large single channel conductance and weak anion selectivity (5). Furthermore, nodulin 26 channels showed sensitivity to high applied voltages, including more active gating and the tendency to occupy discreet lower subconductance states (5).
Previous work also showed that nodulin 26 is phosphorylated by a calcium-dependent protein kinase on the symbiosome membrane (4). This kinase has characteristics of the calmodulin-like domain protein kinase family (4). Members of this family possess a protein kinase catalytic domain fused to a calmodulin-like regulatory domain with four EF-hand calciumbinding sites (6,7). Based on protein sequence analysis, in vivo and in vitro phosphorylation of nodulin 26 occurs at only one residue, serine 262 within the hydrophilic, cytoplasmic COOHterminal domain (4,8).
The finding that nodulin 26 is phosphorylated by a symbiosome membrane CDPK suggests that calcium signaling may be involved in its regulation. A correlation between nodulin 26 phosphorylation and changes in metabolite transport have been observed with isolated symbiosomes (9), but a role for nodulin 26 and phosphorylation in symbiosome membrane transport is still not defined. To study the effect of phosphorylation on nodulin 26, we have investigated the channel activities of wild-type recombinant nodulin 26 before and after in situ phosphorylation by CDPK, as well as the activities of nodulin 26 mutant proteins with substitutions at position 262 that imitate the unphosphorylated or phosphorylated states.

MATERIALS AND METHODS
Molecular Cloning Techniques-A full-length nodulin 26 cDNA was obtained and cloned into M13mp19 as described previously (10), and site-directed mutagenesis was done by using a Bio-Rad mutagenesis kit. Mutagenesis primers were 5Ј-AAGAGTGCTGCTTTCCTCAA-3Ј for the Ser-262 3 Ala substitution, and 5Ј-AAGAGTGCTGATTTCCT-CAA-3Ј for the Ser-262 3 Asp substitution. Mutants were confirmed by dideoxynucleotide chain termination DNA sequencing with a Sequenase kit (U. S. Biochemical Corp. Expression and Purification of Recombinant Nodulin 26 -E. coli clones were cultured in 40 ml of LB medium, 50 g/ml carbenicillin, and 34 g/ml chloramphenicol at 37°C with shaking until the A 600 reached 0.6. The culture was placed at 4°C overnight. The cells were collected by centrifugation and were resuspended in 40 ml of fresh LB media. The cells were inoculated into 1 liter of LB medium, 50 g/ml carbenicillin, and 34 g/ml chloramphenicol and were grown at 37°C with shaking until the A 600 reached 0.6. Isopropyl-1-thio-␤-D-galactopyranoside (1 mM) was added, and the culture was incubated at 37°C with shaking for 2.5-3.0 h. The cells were harvested by centrifugation at 5000 ϫ g for 5 min at 4°C and were stored at Ϫ80°C.
The cell pellet was thawed at 22-25°C in 70 ml of extraction buffer (20 mM Tris-HCl, pH 7.9, 0.5 M NaCl, 1 M pepstatin A, 1 mM phenylmethylsulfonyl fluoride, 1 M leupeptin). MgCl 2 (10 mM) and DNase I (20 g/ml) were added, and the suspension was incubated at room temperature until the viscosity was reduced. The suspension was centrifuged at 100,000 ϫ g for 1 h at 4°C, the pellet was resuspended in 25 ml of 20 mM Tris-HCl, pH 7.9, 1 M KI, 1 M pepstatin A, 1 mM phenylmethylsulfonyl fluoride, 1 M leupeptin and was incubated for 30 min at 37°C. The sample was centrifuged at 100,000 ϫ g for 1 h at 4°C, and the pellet was washed with 25 ml of extraction buffer. Nodulin 26 was solubilized by resuspending the pellet in 10 ml of 20 mM Tris-HCl, pH 7.9, 0.5 M NaCl, 1 M pepstatin A, 1 mM phenylmethylsulfonyl fluoride, 1 M leupeptin, 2% (w/v) OG, and incubating for 20 h at 4°C with shaking. The mixture was centrifuged at 100,000 ϫ g for 1 h at 4°C, and the supernatant fraction was applied to a Ni 2ϩ -iminodiacetic acid-Superose column (1.3 cm ϫ 1.9 cm) attached to a Pharmacia FPLC system. The column was equilibrated with 20 mM Tris-HCl, pH 7.9, 0.5 M NaCl, 1% (w/v) OG (column buffer). The column was washed with column buffer until the A 280 reached base line, and nodulin 26 was eluted with a linear gradient of 0 -300 mM imidazole (⌬12 mM/ml) in column buffer. One-ml fractions were collected and screened for nodulin 26 by SDS-polyacrylamide gel electrophoresis and Western blot analysis (4,10). Fractions containing purified recombinant nodulin 26 were combined and stored at Ϫ80°C.
Nodulin 26 Reconstitution and Electrophysiological Methods-Nodulin 26 proteoliposomes were prepared by reconstitution into phosphatidylcholine bilayers by a dialysis method (5). Thirty-five g of purified protein was used for each preparation with a mole ratio of nodulin 26 to phosphatidylcholine of 1:9,000. This low nodulin 26 to lipid ratio helps insure the likelihood of the incorporation of a single channel into the bilayer (5). After dialysis, proteoliposomes were sedimented at 100,000 ϫ g for 1 h at 4°C, were resuspended in 0.75 ml of 0.2 M KCl, 20 mM MOPS-KOH, pH 7.4, and were stored on ice. The channel activity of these preparations were stable for up to 10 days.
Single channel analyses of recombinant nodulin 26 in planar lipid bilayers were performed as described in Ref. 5 with the following modifications. Muller-Rudin planar lipid bilayers were formed with synthetic phosphatidylethanolamine, phosphatidylserine, and phosphatidylcholine (5:3:2 (w/w/w); Avanti Polar Lipids). The standard recording solution for incorporation was cis 0. For the study of voltage-current relationships of recombinant nodulin 26 channels, a ramp protocol was used in which the voltage was increased from Ϫ90 to ϩ90 mV at 0.09 mV/ms (CLAMPEX program, PCLAMP software, Axon Instruments). For studies of voltage sensitivity, a standard pulsing protocol was used for recordings of channel currents at 30-and 70-mV voltage potentials. One-s pulses to the test voltage were done alternated with 24-ms rests at a holding potential of 0 mV. Analog data were filtered at 2 kHz through an eight-pole Bessel filter (Frequency Devices, model 902). Potential incorporation from either the cis and trans compartments and uncertainty regarding the orientation of the reconstituted nodulin 26 protein in liposomes precluded the determination of channel sidedness. However, the channel shows a linear voltage and current relationship (see Fig. 3) with no apparent rectification. Thus, for simplification and clarity, currents are expressed as absolute deflections from zero. The FETCHAN and PSTAT programs (PCLAMP software, Axon Instruments) were used to analyze the channel currents recorded by the pulsing protocols. The TRANSIT program (11) was used to determine the conductance amplitude histograms and to weight the events by the duration of the occupancy of the conductance states.
For studies of the effects of phosphorylation on nodulin 26 channel activity, the trans chamber was perfused with 0.2 M KCl, 20 mM MOPS-KOH, pH 7.4, and bovine serum albumin was added to a concentration of 3 mg/ml. The channel activity under these recording conditions was recorded at 30 and 70 mV using the standard pulsing protocol. MgCl 2 and ATP were added to both chambers to final concentrations of 1 mM and 100 M, respectively, and CDPK (KJM23-6H2, a kind gift from Dr. Jeffrey Harper, Scripps Research Institute, Ref. 12) was added to both cis and trans chambers to a final concentration of 4 -5 g of CDPK/ml recording solution. All reagents were added to both chambers because of the uncertainty of the sidedness of the reconstituted channel. Chamber solutions were stirred for 1 min and were incubated for 35-70 min before recording the current at 30 and 70 mV by using the standard pulsing protocol.
For dephosphorylation experiments, alkaline phosphatase from bovine intestinal mucosa (Sigma, lot no. 103H72502) was added after CDPK phosphorylation. Additions were symmetrical (40 units cis and 80 units trans). Chamber solutions were stirred for 1 min and were incubated for an additional 25 min prior to channel recording.

Expression and Purification of Recombinant Nodulin 26 -
For expression in E. coli, nodulin 26 cDNA was cloned into the pRSET A vector as a His-tag fusion under the control of the T7 promoter. Based on Western blot analysis, nodulin 26 was successfully expressed in E. coli containing these vectors and is associated with the membrane fraction. The His-tag at the amino terminus of nodulin 26 allowed it to be separated from other E. coli membrane proteins by Ni 2ϩ -chelate chromatography. Initial attempts to isolate recombinant nodulin 26 by stepwise Ni 2ϩ -chelate chromatography were not successful because of the contamination by other metal binding proteins in E. coli membranes (data not shown). However, FPLC Ni 2ϩchelate chromatography with the application of a linear gradient of imidazole resulted in excellent separation of recombinant nodulin 26 from other E. coli membrane proteins (Fig. 1). The final product was pure based on SDS-polyacrylamide gel electrophoresis (Fig. 1A), and its identity as nodulin 26 was confirmed by Western blot analysis (Fig. 1B).

Reconstitution of Recombinant Nodulin 26 into Planar Lipid Bilayers and Measurement of Ion
Channel Activity-Purified recombinant nodulin 26 was reconstituted into proteoliposomes, and its single channel properties were investigated in planar lipid bilayers as described for soybean nodulin 26 (5). Addition of proteoliposomes to the bilayer chamber resulted in the appearance of ion channel activity, which is inhibited by the addition of the anti-nodulin 26 IgG but not preimmune IgG (Fig. 2). This observation suggests that the interaction of the antibody with nodulin 26 protein blocks the channel, and thus channel activity is due to the insertion of nodulin 26 into the bilayer.
A current voltage relationship for recombinant S262 nodulin 26 is shown in Fig. 3. The current voltage relationship is linear with conductance values of 3.1 nS (plot 1) or 1.6 nS (plot 2) calculated depending upon the ionic strength of the recording solutions (Fig. 3). Under asymmetric recording conditions (cis 0. Representative channel records and conductance amplitude histograms are shown in Fig. 4. The data shown are from a typical, representative channel incorporation, but several channel incorporations (17 separate S262 channels, eight S262A channels, and nine S262D channels) have been analyzed and show similar single-channel conductances and voltage-dependent behavior. Similar to native soybean nodulin 26, all recombinant nodulin 26 channel proteins show a maximal single channel conductance of 3.1 nS under standard recording conditions (Fig. 4) and at low applied voltages (e.g. 30 mV) showed a principal single channel conductance of 3.1 nS with only infrequent occupancy of lower subconductance levels (Fig.  4A). We showed previously that native soybean nodulin 26 from symbiosome membranes showed increased channel gating and a tendency to preferentially occupy lower conductance substates at high applied voltages (e.g. 70 mV, Ref. 5). In contrast, at 70-mV potentials, the recombinant S262 and S262A nodulin 26 channels still remained completely open with a principal single channel conductance of 3.1 nS and only infrequent transitions to lower conductance states (Fig. 4B). Based on the amplitude histogram (Fig. 4), these channels exist in the 3.1-nS state greater than 97% of the time. However, the recombinant S262D nodulin 26 channel shows more frequent gating to lower subconductance states at 70 mV (Fig. 4B). Based on the amplitude histogram, the S262D nodulin 26 channel shows three major conductance states at 70 mV: 3.1, 1.8, and 0.6 nS (Fig.   4B), with a preference for the lower substates. The percent occupancy times for the 3.1-, 1.8-, and 0.6-nS states were 13.1, 35.1, and 51.8%, respectively. The data suggest that a negative charge at residue 262 confers voltage-sensitive behavior on the channel and that the phosphorylation of Ser-262 of nodulin 26 by CDPK may regulate voltage-sensitive channel activity. This was tested by direct phosphorylation of S262 nodulin 26 by

CDPK.
Phosphorylation of Nodulin 26 in Situ in Planar Lipid Bilayers by CDPK-The effects of phosphorylation of the recombinant nodulin 26 channel were investigated using recombi-nant KJM23-6H2 CDPK (12). KJM23-6H2 is derived from the expression of an Arabidopsis cDNA clone in E. coli and has a substitution of six amino acids in its autoinhibitor site resulting in a highly active, constitutive enzyme activity (12). Because the KJM23-6H2 CDPK is easily prepared at high concentrations in a constitutively active form, we selected this enzyme for phosphorylation studies. Purified KJM23-6H2 CDPK readily phosphorylates CK-15 (a synthetic peptide substrate containing the CDPK recognition sequence and the unique phosphorylation site, Ser-262 of nodulin 26 (8)) showing hyperbolic kinetics, and an apparent K m of 142 M (Fig. 5), and thus has kinetic properties similar to the soybean nodule CDPK activity (4).
The effects of phosphorylation of S262 nodulin 26 were studied by in situ phosphorylation with CDPK after incorporation into planar lipid bilayers (Fig. 6). S262A nodulin 26, which possesses an Ala-262, was used as a negative control. Experiments were performed in symmetric 0.2 M KCl. Under these conditions, both channels show a maximum single channel conductance of 1.6 nS (Figs. 3 and 6). Addition of MgCl 2 and ATP did not affect channel properties (data not shown). However, subsequent addition of CDPK resulted in changes in the gating behavior of S262 nodulin 26 at 70 mV (Fig. 6B). At 70 mV, CDPK-treated S262 nodulin 26 showed several conductance substates including 1.6 nS (28.3%), 1.0 nS (24.4%), and 0.6 nS (42.6%), as well as a completely closed state (4.7%) (Fig. 7A). Conductance of S262 nodulin 26 at low voltage potentials (e.g. 30 mV) was not significantly affected by CDPK treatment. Furthermore, CDPK appears to mediate this effect on nodulin 26 by phosphorylation of Ser-262. This is supported by the control experiments that show that S262A nodulin 26 only occupies the fully open 1.6 nS conductance state even after prolonged treatment with CDPK (Fig. 7B).
If phosphorylation is responsible for the change in the voltage sensitivity of nodulin 26, then the effect should be reversed by dephosphorylation of Ser-262. In previous work (9) it was shown that nodulin 26 can be dephosphorylated in vitro by alkaline phosphatase. Alkaline phosphatase treatment of phosphorylated S262 nodulin 26 results in the restoration of voltage-insensitive behavior (Fig. 8). Furthermore, this appears to be the result of removal of phosphate from S262 nodulin 26 since phosphatase treatment of S262D nodulin 26 (Asp-262) has no effect on its voltage sensitivity (data not shown). These data show that phosphorylation at Ser-262 of nodulin 26 by CDPK modulates its channel activity by affecting its voltage sensitivity. DISCUSSION We have purified recombinant nodulin 26 derivatives expressed in E. coli by Ni 2ϩ -chelate chromatography, and have shown that they form channels in planar lipid bilayers with large single channel conductances and weak anion selectivity, similar to nodulin 26 from soybean symbiosome membranes (5). However, nodulin 26 proteins with serine or alanine at residue 262 showed no voltage sensitivity, whereas nodulin 26 with aspartate 262 showed voltage-sensitive behavior that included more active gating and a tendency to preferentially occupy lower subconductance states. Nodulin 26 with serine 262 was converted to a similar voltage-sensitive state by CDPK phosphorylation and this effect was reversed by dephosphorylation with alkaline phosphatase. Overall, the data suggest that the presence of a negatively charged residue at position 262 confers voltage-sensitive behavior and that the phosphorylation of Ser-262 of nodulin 26 by CDPK modulates nodulin 26 channel activity.
The data show that the recombinant His-tag nodulin 26 derivatives have the same maximal single channel conductance and ion selectivity values as soybean nodulin 26. These results imply that recombinant nodulin 26 expressed in E. coli is structurally and functionally homologous to the native nodulin 26 molecule and that the presence of the His-tag sequence does not affect its conductance properties. Furthermore, the His-tag allows the purification of nodulin 26 in one step by FPLC on nickel chelate resins by using gradient elution conditions. The use of this system should allow the production of other sitedirected mutations to further probe the nodulin 26 structure and function. Another advantage of expression in E. coli, which lacks CDPK, is the generation of nodulin 26 that is not phosphorylated on Ser-262. This is an important consideration for planar lipid bilayer studies since nodulin 26 purified from soybean probably exists as a mixture of phosphorylated and unphosphorylated forms, and it is unclear whether singlechannel data represent the insertion of an unphosphorylated or phosphorylated nodulin 26 molecule. From the present study, it can be concluded that voltage-sensitive gating is observed only upon phosphorylation of Ser-262. Interestingly, all soybean nodulin 26 channels examined previously showed voltage-sensitive behavior similar to S262D and phosphorylated S262 recombinant nodulin 26 (5). Thus, a major population of nodulin 26 isolated from symbiosome membranes appears to be phosphorylated before or during purification. This is supported by the observation that alkaline phosphatase treatment of soybean nodulin 26 results in a channel that is less sensitive to  voltage (data not shown).
Nodulin 26 is a member of a structurally homologous family of membrane channel proteins (16). In addition to nodulin 26, some other family members are phosphorylated by various protein kinases (17)(18)(19)(20)(21). Of particular interest is the similarity between the lens MIP and nodulin 26 with respect to the functional effects of phosphorylation. Similar to nodulin 26, MIP forms channels in planar lipid bilayers with a large unitary conductance and similar ion selectivity (22). Both proteins have an unique phosphorylation site (Ser-262 for nodulin 26 and Ser-243 for MIP) at homologous positions within their COOH-terminal domains (8,20). However, whereas nodulin 26 is phosphorylated by CDPK, MIP is phosphorylated at Ser-243 by the cAMP-dependent protein kinase (20). Similar to our findings with nodulin 26, unphosphorylated MIP forms a voltage-insensitive channel, and phosphorylation with cAMPdependent protein kinase results in voltage-sensitive gating behavior and partial channel closure (23). This suggests that phosphorylation within the COOH-terminal region of these proteins results in a similar change in their structure and function as manifested by their channel behavior in planar lipid bilayers.
The mechanism through which phosphorylation affects these proteins is not yet clear. Phosphorylation is a common mechanism for controlling ion channel activities, including through the regulation of channel gating (24). In the case of nodulin 26 and MIP, one possibility is that phosphorylation of the COOHterminal domain results in the interaction of this part of the protein with the channel pore, resulting in a change in the kinetic properties of the channel. Evidence is accumulating that cytosolic "gating domains" at the termini of channel proteins interact with the channel pore, resulting in a change in gating kinetics (34). For example, site-directed mutagenesis studies show that removal of the cytosolic amino-terminal domain of the Shaker K ϩ channel results in an open channel that cannot be inactivated (25). Inactivation can be restored by the addition of a synthetic peptide corresponding to the missing residues (25). Similar cytosolic gate domains have been demonstrated in other channels (for review, see Ref. 26), and this may be a common mechanism for controlling channel activity. Evidence that this may be the case in MIP comes from the observation that proteolytic removal of the COOH-terminal region results in a channel with the same single channel conductance that is no longer voltage-sensitive (23). However, the interaction of the phosphorylated COOH-terminal region with MIP and nodulin 26 is likely to be complex since phosphorylation results in the appearance of at least two to three well defined subconductance states and a closed state, rather than just simple channel closure.
Nodulin 26 is an in vivo target of calcium-dependent phosphorylation by a calmodulin-like domain protein kinase on the symbiosome membrane of soybean nodules (4,8). In light of previous results and the present findings, it is attractive to propose a role for calcium-dependent phosphorylation in the regulation of nodulin 26 channel activity in response to membrane potentials. This is further supported by the finding of an electrogenic H ϩ -pumping ATPase on the symbiosome membrane, which is capable of producing large transmembrane potentials (27) that could affect the activity of phosphorylated nodulin 26. However, several potential factors will need to be taken into consideration before assessing the role of nodulin 26 and phosphorylation in symbiosome membrane function. First, the single-channel conductance of nodulin 26 in planar lipid membranes is very large and complete closure is infrequent, even with the phosphorylated form, a condition that may not be likely in vivo (2). However, other endogenous symbiosome membrane lipids or proteins that are absent from the reconstituted planar lipid bilayer system also may contribute to the modulation of nodulin 26 activity along with symbiosome membrane potentials. For example, it has been found that certain membrane lipids (e.g. cholesterol) can attenuate the conductance levels of MIP channels (28). Furthermore, many members of the MIP family are reported to form water channels (29) or channels for uncharged solutes such as glycerol (30). Although it has been reported that MIP is not a water channel (31), other recent evidence suggests that MIP can form a low activity water channel upon heterologous expression in Xenopus (32,35). Regardless of these considerations, the planar lipid bilayer experiments have revealed that a fundamental change in the structure and function of nodulin 26 and MIP occurs upon phosphorylation. Further work, possibly in situ with symbiosome membranes, may provide further insight into the biological role of nodulin 26 phosphorylation.  Fig. 7; C, the same channel from panel B after the symmetrical addition of bovine intestinal alkaline phosphatase (50 units/ml).