Role of tyrosine phosphorylation in potassium channel activation. Functional association with prolactin receptor and JAK2 tyrosine kinase.

Chinese hamster ovary (CHO) cells, stably transfected with the long form of the prolactin (PRL) receptor (PRL-R) cDNA, were used for PRL-R signal transduction studies. Patch-clamp technique in whole cell and cell-free configurations were employed. Exposure of transfected CHO cells to 5 nM PRL led to the increase of Ca2+- and voltage-dependent K+ channel (KCa) activity. The effect was direct as it was observed also in excised patch experiments. A series of tyrosine kinase inhibitors was studied to investigate the possible involvement of protein tyrosine kinases in KCa functioning and its stimulation by PRL. Genistein, lavendustin A, and herbimycin A decreased in a concentration and time-dependent manner the amplitude of the KCa current in whole cell and the open probability of KCa channels in cell-free experiments. The subsequent application of PRL was ineffective. The protein tyrosine phosphatase inhibitor orthovanadate (1 mM) stimulated KCa channel activity in excised patches, indicating that channels can be modulated in opposite directions by protein tyrosine kinase and protein tyrosine phosphatase. Moreover, in whole cell experiments as well as in excised patch recordings, anti-JAK2 tyrosine kinase antibody decreased the KCa conductance and the open probability of the KCa channels. Subsequent application of PRL was no longer able to stimulate KCa conductance. Immunoblotting studies using the same anti-JAK2 antibody, revealed the constitutive association of JAK2 kinase with PRL-R. Preincubation of anti-JAK2 antibody with the JAK2 Immunizing Peptide abolished the effects observed using anti-JAK2 antibody alone in both electrophysiological and immunoblotting studies. We conclude from these findings that these KCa channels are regulated through tyrosine phosphorylation/dephosphorylation; JAK2 tyrosine kinase, constitutively associated with PRL-R, is implicated in PRL stimulation of KCa channels.

Prolactin (PRL) 1 is a multifunctional pituitary hormone in-volved in the control of a wide variety of physiological processes in vertebrates, including lactation, reproduction, immune responses, and osmoregulation, as well as cell proliferation (1)(2)(3). The PRL receptor (PRL-R) belongs to the cytokine-growth factor receptor superfamily that includes receptors for growth hormone, erythropoietin, numerous hematopoietic interleukins (IL)-2, IL-3, IL-4, IL-5, IL-6,IL-7, IL-9, granulocyte colonystimulating factor, granulomacrophage colony-stimulating factor, and ciliary neurotrophic factor (4,5). This family of receptors possesses common structural motifs, both external (two disulfide loops and the WSXWS homology box) and internal (proline-rich homology box 1). Recent studies have been marked by considerable progress in understanding the mechanisms of intracellular signaling for the different members of this family, particularly for PRL-R. Most of the data were obtained in the PRL-dependent rat T lymphoma cell line Nb2. It has been shown by several groups that, following binding of PRL to the PRL-R in these cells, dimerization of the receptor occurs (6,7) prior to phosphorylation of an associated tyrosine kinase (JAK2). This represents the first event in the process of PRL-R signal transduction (8,9). Other studies demonstrated that PRL stimulation of Nb2 cells induced a concentration-and time-dependent activation of another protein tyrosine kinase, p59 fyn , from the Src protein tyrosine kinase family (10). On the other hand, more and more studies demonstrate an important role of ion channels in receptor signal transduction (11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22). Several different mechanisms have been proposed for channel involvement in signal transduction: direct agonist effect on the channel (15,16), second messenger participation in channel modulation (17)(18)(19), and regulation by kinases and phosphatases through channel phosphorylation/dephosphorylation processes (11, 20 -22). Moreover channel regulation by phosphorylation has been shown to play a key role in physiological processes such as proliferation and transformation (23)(24)(25). Nothing, however, is known about the putative role of channel phosphorylation in cytokine-growth factor receptor superfamily signal transduction and, in particular, in PRL-R signal transduction.
The signal transduction mechanism for the full-length PRL receptor has been studied using a CHO line stably transfected with the cDNA of the long form of rabbit mammary PRL-R (26). These CHO-transfected cells responded to PRL by stimulating the cotransfected milk protein gene promoter (27), proving that such cells are fully capable of transmitting the PRL signal and that PRL-R is functional. In a series of studies using patch clamp and microfluorimetric techniques, we analyzed the first steps of the PRL-R signal transduction pathway: an increase in intracellular Ca 2ϩ (28,29) and direct stimulation of calciumand voltage-activated potassium channels (K Ca ) by PRL (29). These observations suggested the existence of a regulatory complex involving a protein kinase tightly associated with K Ca channels and PRL-R. Furthermore, by immunoblotting studies we presented evidence for the tyrosine phosphorylation of this type of PRL-R, the association of JAK2 tyrosine kinase with the receptor, as well as changes in tyrosine phosphorylation of a number of cytoplasmic proteins (30).
In this article, we report on the very first steps in PRL-R signal transduction at the plasma membrane level: we demonstrate an endogenous large conductance K Ca channel as the primary ionic event triggered by PRL. These K Ca channels are constitutively regulated through tyrosine phosphorylation/dephosphorylation. We also show that JAK2 tyrosine kinase, associated with PRL-R, is implicated in the stimulation of K Ca channels by PRL.

EXPERIMENTAL PROCEDURES
Cell Cultures-We used CHO cells transfected with PRL-R-cDNA (CHO E3) as described previously (31). Different subclones were challenged for PRL binding, and one of them (E32), exhibiting the highest binding capacity (12% specific binding versus 4% for E3) was used in these experiments. The PRL receptor in E32 clone has a K a ϭ 10.8 ϫ 10 Ϫ9 M Ϫ1 which is higher than that of the parental E3 clone (32), but the same number of sites (about 9000). The cells were grown in Ham's F-12 medium (Seromed, Strasbourg, France) containing 10% (v/v) fetal calf serum (Life Technologies, Inc.). Medium was changed every 2-3 days. Cells were maintained at 37°C in a humified atmosphere gassed with 95% air, 5% CO 2 . In order to avoid occupancy of PRL receptors by lactogenic factors contained in the serum of the culture medium, 6 -24 h before the experiments cells were transferred into a serum-free medium (32). This medium was derived from the GC3 medium described by Gasser et al. (33) and is a 1:1 mixture of Dulbecco's modified Eagle's medium and Ham's F-12 (Seromed) supplemented with nonessential amino acids (Life Technologies, Inc.), insulin (Sigma; 80 milliunits/ml), glutamine (Sigma, 2.5 nM), and transferrin (Life Technologies, Inc., 10 mg/ml).
Electrophysiological Recordings-The cultures were viewed under phase contrast with a "Leitz-Diavert" (Leitz, Germany) inverted microscope. Electrodes were positioned with "Leitz" (Germany) micromanipulators. Grounding was through a silver chloride-coated silver wire inserted into an agar bridge (4% agar in electrode solution). An Axopatch-1D amplifier (Axon Instruments, Inc., Foster City, Ca) was used for tight seal, whole cell, and cell-free voltage clamping. Stimulus control and data acquisition and processing were carried out with a PC computer AT-80386 (Tandon, Moorpark, CA), fitted with a Labmaster TL-1 interface, using Pclamp 5.5.1 software (Axon Instruments, Inc., interface and software). Electrode offset was balanced before forming a "giga-seal." Leakage and capacitive current subtraction protocols were composed of four or five hyperpolarizing pulses, one-fourth or one-fifth pulse, respectively, and were applied from a holding potential before test pulses eliciting active responses. During data analysis, leak data were subtracted from the raw data. Series resistances were compensated and were calculated before and after compensation. Recordings where series resistance resulted in a 5 mV or greater error in voltage commands were discarded. Currents were low pass-filtered at 2 KHz with an eight-pole Bessel filter (-3dB) and digitized at 10 KHz for storage and analysis.
Data and Statistical Analysis-Peak currents in whole cell recordings were measured using the automatic peak detection function in the Clampan section of the Pclamp software. Late currents measured isochronally were taken before the end of the pulse to avoid capacitative transients spread out by digital filtering.
Single channel data analysis was performed after elimination of capacity transients and leak current by subtraction of record averages without channel activity from each current record. The openings and closings of the channel were detected using a criterion of a 50% excur-  3) for the pipette. Free Ca 2ϩ concentrations in the range of 10 nM to 1 M for the solutions, applied from the inner side of membrane, were buffered with 1.1 mM EGTA and were calculated using the method previously described by Abercrombie et al. (34). Calcium concentrations greater than 1 M were achieved by adding the desired amount of CaCl 2 . In all studies of K ϩ channel activity 0.1 mM Mg-ATP was added to the internal solution. To allow local drug application to the investigated cell an additional "pouring" pipette with a tip opening of 10 -30 m was used. This pipette was filled with the same extracellular saline as was in the bath and the drug under investigation added to it in appropriate concentrations. The pipette was brought close to the investigated cell at a distance of 50 -100 m. All experiments were performed at room temperature (20 -22°C).
Membrane Stripping and Rehybridization-In order to rehybridize the membranes with other antibodies, they were stripped 30 mn at 60°C in 62.5 mM Tris (pH 6.7), 2% SDS, 100 mM ␤-mercaptoethanol. After extensive washes, the membranes were then processed as described before.
Chemicals-PRL (o-PRL-19) and anti-rat PRL antibody were kindly provided by the NIDDK (National Hormone and Pituitary Program, University of Maryland School of Medicine, Baltimore, MD). Mg-ATP, TTX, genistein, genistin, and orthovanadate were from Sigma. Charybdotoxin (CTX) was obtained from Latoxan (Rosans, France). Lavendustin A and herbimycin A were from Life Technologies, Inc. Antibody anti-JAK2 and JAK2 immunizing peptide were purchased from Upstate Biotechnology, Inc. Normal rabbit serum was from Sera Lab (London, United Kingdom).

RESULTS
As was demonstrated by our previous studies (35), a 210picosiemens K ϩ conductance, dependent on voltage and intracellular Ca 2ϩ , was revealed by patch-clamp experiments in CHO cells. To check whether the K Ca channels could be phosphorylated, we studied the effect of ATP, which serves as a substrate for protein kinase, on the activity of the K Ca channels in CHO cells. Experiments using Mg-ATP (10 -100 M), applied to the cytoplasmic side of the membrane, showed an increase in the open probability of the channel, proving that a protein kinase is tightly associated with it. Fig. 1 shows the very low K Ca channel open probability values in the absence of ATP in the internal solution and the much higher values in the presence of ATP. The open probability values in the absence of ATP were so low that it was impossible to carry out the statistical analysis required to establish the effect of PRL on K ϩ channel activity under these conditions. In whole cell experiments without ATP in the patch pipette, no effect of PRL on K ϩ total current was observed. ATP hydrolysis was required for channel modulation, because the nonhydrolyzable ATP analog, AMP-PNP, was ineffective (not shown). As, on the one hand, tyrosine kinase was found to be a primary target of PRL in PRL-R signal transduction (8,9) and, on the other hand, the K ϩ channels stimulated by PRL were found to be associated with protein kinase, we assumed the existence of a PRL-R-K Ca channeltyrosine kinase regulatory complex. To investigate whether the K ϩ channels stimulated by PRL may be modulated by an endogenous tyrosine kinase, we examined the effects of different tyrosine kinase inhibitors on K Ca channel activity, using both whole cell and cell-free modes of the patch-clamp technique. In these studies we used two types of experiments differing in the duration of drug application Short Application of the Drug for Periods Varying from 15 s to Several Minutes-Application was performed from an additional pipette directly on the cell membrane (in whole cell patch-clamp configuration) or pieces excised from the membrane, containing one or more ion channels (excised-patch configuration).
Bath application (3 min) of the protein tyrosine kinase inhibitor genistein (36, 37) caused a progressive reduction in the K ϩ current ( Fig. 2A). Further application of PRL was ineffective. Genistein concentrations lower than 50 M were ineffective under these experimental conditions (n ϭ 6). After genistein was washed out, K Ca currents gradually recovered, indicating that the depression was reversible. Genistin, an analog of genistein that lacks protein tyrosine kinase inhibitory activity (36), had no effect on K Ca current (n ϭ 4; Fig. 9A). We also tested two structurally distinct protein tyrosine kinase inhibitors: herbimycin A (9, 38) and lavendustin A (36,39). Herbimycin A (1.5 M) and lavendustin A (10 M) depressed K Ca currents to 57 Ϯ 8% (n ϭ 8) and to 54 Ϯ 5% (n ϭ 6) of control, respectively. Fig. 2B shows an example of K Ca current inhibition induced by 3-min application of 10 M lavendustin A and current-voltage relationships for this effect, where current amplitudes were plotted at different test potentials. K Ca currents were evoked by 160-ms test pulses from a holding potential of Ϫ40 to ϩ10 mV.

Preincubation of All Cells by Addition of the Drug to the Bath Solution for a Time Varying from 30 min to Several Hours-
The concentration dependence of current depression caused by three tyrosine kinase inhibitors is shown in Fig. 2C for cells preincubated with the drugs for 6 h. Under these conditions, even when cells were pretreated with low concentrations of inhibitors (n ϭ 5 for cells treated with 10 M of genistein, n ϭ 4 for 100 nM of herbimycin, and n ϭ 5 for 1 M of lavendustin), the application of 5 nM PRL on these cells was ineffective (data not shown).
Cell-free experiments demonstrated that the effects of PRL and protein tyrosine kinase inhibitors were not mediated by intracellular processes, as they could be also observed in detached patches. Fig. 3 shows K ϩ channel activity stimulation by PRL and its inhibition by genistein (n ϭ 7/9 patches). PRL (5 nM) caused an increase in the open probability of the channels (Fig. 3B), displaying the half-maximum increase in the open probability within 3.6 Ϯ 1.3 min. The open probability of the channel after the addition of PRL was not constant, but oscillated between lower and higher open probability values (Fig. 3B). Subsequent addition of 100 M genistein inhibited this K ϩ channel activity almost completely within 7 Ϯ 2 min (Fig. 3B). The amplitude histograms (Fig. 3C) for K ϩ channels in control (mean ϭ 8.48 Ϯ 0.12 pA) and in the presence of PRL (mean ϭ 8.34 Ϯ 0.18 pA) demonstrate that PRL does not activate additional conductances. Moreover, prolactin did not stimulate K ϩ channel activity in the presence of 30 nM CTX, a K Ca channel inhibitor in CHO cells (20), indicating that PRL activated the CTX-sensitive K Ca channels and no other type of outward channels (not shown). As in whole cell experiments, genistin was ineffective (Fig. 9B). Application of 1.5 M herbimycin A also decreased channel open probability without affecting single-channel conductance (n ϭ 4/5 patches, Fig. 4). The subsequent addition of PRL was ineffective.
As protein tyrosine kinase inhibitors in high concentrations are known to be able to inhibit not only protein tyrosine kinase but also protein kinase C and protein kinase A kinases in some cell types (36), we checked the putative involvement of protein kinase C and protein kinase A in the mechanisms studied. We tested both activators and inhibitors of protein kinase C and protein kinase A (10Ϫ8 M phorbol 12-myristate 13-acetate application, as protein kinase C activator; 10Ϫ6 M phorbol 12myristate 13-acetate 24-h incubation, as protein kinase C inhibitor; 250 M phloretin, as protein kinase C inhibitor; 2 M forskolin and 1 mM 8-bromo-cAMP, as protein kinase A activators) on K ϩ channel activity and on the stimulated effect of PRL. None of these drugs had any effect. A protein kinase C biochemical assay (40) was also carried out. 50 -100 M genistein had no effect in these studies.
The possibility that these channels may be regulated by protein tyrosine phosphatase as well as by protein tyrosine kinase was investigated by application of the protein tyrosine phosphatase inhibitor, sodium orthovanadate (41), to the cytoplasmic side of the cell-free patch. Orthovanadate (1 mM) increased the open probability of the channels in a time-dependent manner (Fig. 5). On average, activity increased by 198 Ϯ 27% within 5 min (n ϭ 6/9 patches). Fig. 5B shows the duration histograms for the channel prior to application of orthovanadate, then 2 and 8 min afterwards, respectively. In the presence of orthovanadate, channel openings were longer, as indicated by an increase in the relative number of events and the time constant to for the open state.
The preceding results strongly support the conclusion that the functioning of PRL-stimulated K Ca channels is modulated by protein tyrosine kinases and protein tyrosine phosphatases and thus regulated by constitutive tyrosine phosphorylation/ dephosphorylation. It has already been shown that tyrosinephosphorylated PRL-R is associated with JAK2 kinase (30). Additional immunoblotting experiments were carried out to find out if this association was constitutive. Solubilized proteins from E32 cells incubated with or without 500 ng/ml of prolactin were immunoprecipitated with anti-PRL-R antibody

FIG. 4. The effect of herbimycin A on K Ca channel activity.
Representative records of single K ϩ currents in excised (outside-out) membrane patch (top). K ϩ channel activity in this patch was obtained with 0.5 M internal-free Ca 2ϩ at a membrane potential of ϩ20 mV. Records are shown for the control solution, 10 min after 1.5 M herbimycin application and 3 min after the subsequent addition of 5 nM PRL. The open probability of the K ϩ channel versus time is presented in the lower panel. 46 and analyzed for the presence of JAK2 in the complex. As shown in Fig. 6A, a protein of 130 kDa was detected in the blot hybridized with anti-JAK2, corresponding to JAK2 kinase. This protein was revealed in the presence or absence of stimulation by ovine PRL, showing that this protein is constitutively associated with the PRL receptor. A rehybridization of the same blot with S46 shows the presence of the same amount of receptor in each line. In Fig. 6B, we demonstrate the specificity of the recognition of JAK2 by the antibody. Cell extracts were immunoprecipitated with anti-JAK2 antibody or anti-PRL-R antibody and subjected to SDS-PAGE and Western blotting. Each blot was incubated either with anti-JAK2 antibody or with anti-JAK2 antibody preincubated with a peptide corresponding to the amino acid residues 758 -776 of murine JAK2 (JAK2 immunizing peptide). In JAK2 immunoprecipitates we observed that a protein of 130 kDa was specifically displaced by the presence of the peptide. Nonspecific bands were always present following incubation in the presence of the peptide. The same finding was observed in 46 immunoprecipitates. The 130-kDa band was not revealed in the presence of the peptide. Therefore, the 130-kDa protein revealed on the blot is the tyrosine kinase JAK2, and this kinase is constitutively associated with the prolactin receptor. To investigate whether K Ca channels are also constitutively associated with JAK2 kinase we studied the effect of the anti-JAK2 kinase antibody on the of the K Ca channel activity. Anti-JAK2 antibody (diluted 1:1000 from the indicated antibody) was introduced into the internal solution of the patch pipette. In whole cell experiments, the amplitude of the K Ca current gradually decreased following rupture of the seal. In control conditions of whole cell experiments, the K Ca current was stable under internal perfusion. No decrease in current (or "run-down") was observed during recordings lasting 20 min or more. Fig. 7 shows the time dependence of the decrease in K Ca current caused by anti-JAK2 antibody. Within approximately 10 min the current was almost completely inhibited (n ϭ 5). Subsequent application of PRL was ineffective (Fig. 7). Anti-JAK2 antibody applied to the cytoplasmic side of the membrane in inside-out studies decreased the open probability of the channel to 73 Ϯ 18% (n ϭ 4/4 patches) of control (Fig. 8). Subsequent application of PRL did not stimulate channel activity (Fig. 8). When anti-JAK2 antibody was preincubated with a peptide corresponding to the amino acid residues 758 -776 of murine JAK2 (JAK2 immunizing peptide) and this mixture applied to the cytoplasmic side of the membrane, the decrease in K ϩ current in the whole cell experiments (Fig. 7) and in the open probability of the channel in inside-out experiments (Fig. 8) were not observed. As the anti-JAK2 antibody was obtained from Upstate Biotechnology, Inc. in rabbit serum, we checked the effect of non-immune rabbit serum on K Ca conductance as a control. Rabbit serum at the same dilution (1:1000) was ineffective in both whole cell and inside-out experiments (not shown). An immune serum antibody (anti-rat PRL antibody) was used as an additional control. At the same dilution it was also ineffective. Fig. 9 presents a summary of the effects of all the drugs studied on normalized peak currents, obtained by whole cell experiments (Fig. 9A) and normalized P o , obtained by excised patch experiments (Fig. 9B). DISCUSSION We conclude from these findings that the activation of K Ca channels by PRL in CHO cells, transfected with cDNA of the long form of PRL-R, is the primary ionic event in PRL-R signal transduction. K Ca channels are constitutively regulated through tyrosine phosphorylation/dephosphorylation. Stimulation of the channels by PRL possibly occurs through phosphorylation of protein tyrosine residues of the channel or of one or more associated proteins. The observation that K Ca channel stimulation by PRL does not occur in the presence of anti-JAK2 kinase antibody suggests that at least one of the kinases involved in the channel stimulation by PRL may be JAK2 tyrosine kinase.
Although the association of PRL-R with JAK2 (8,9,30), JAK1 (42), or Fyn (10) tyrosine kinases has already been clearly demonstrated, the cascade of ionic events induced by PRL and the nature of ion channels involved has not yet been studied. Our earlier studies have characterized a membrane hyperpolarization, caused by K Ca channel stimulation, and Ca 2ϩ influx among the first detectable responses to PRL-R activation (29). The underlying mechanisms are, however, not very clearly understood. In the present study, we applied the patch-clamp recording technique in order to unravel the mechanism of K Ca channel activation by PRL and to identify the nature of the associated protein kinase.
In our experiments ATP (10 -100 M) increased the open probability of the K Ca channel, therefore showing that a protein kinase is involved in the regulation of channel activity. Moreover PRL was unable to stimulate this activity when ATP was absent from the internal solution, demonstrating that kinase phosphorylation is needed for channel stimulation by PRL. Thus, we concluded that protein kinase is closely associated with K Ca channel and PRL-R in a regulatory complex.
It was recently demonstrated that in murine fibroblast cell lines, transfected with Ras or Raf plasmids, the K Ca , CTXsensitive channel is up-regulated by oncogenic p21 ras and that this regulation appears to be due to raf kinase-dependent induction of channel expression (43). The application of either epidermal growth factor or platelet-derived growth factor to nontransfected cells caused a time-dependent induction of K Ca channels, obviously, through activation of endogenous cellular p21 ras . Epidermal growth factor induction of the K Ca channel was blocked by the tyrosine kinase inhibitors lavendustin A (1 M) or genistein (50 M). However, application of genistein to cells transfected by oncogenic ras had no effect on K ϩ current density, indicating that genistein had no direct inhibitory effect on the K ϩ channel. These results suggest that ras regulates the K Ca channels through serine/threonine kinase and not through protein tyrosine kinase. Our experiments using tyrosine kinase inhibitors show a distinct mechanism of K Ca channel regulation in CHO cells transfected with cDNA of the long form of PRL-R: the direct regulation of K Ca channels by tyrosine kinase (as this activity was inhibited directly in whole cell and single channel experiments by three distinct tyrosine kinase inhibitors) and modulation of this activity by PRL (as PRL was no longer able to stimulate K Ca channel activity when the cells were treated with protein tyrosine kinase inhibitors). Experiments using protein kinase C and protein kinase A activators and inhibitors showed that these kinases are not involved in the channel regulation mechanisms.
For epidermal growth factor receptor, which has intrinsic tyrosine kinase (44), the first event is activation of voltageindependent Ca 2ϩ channels defined as direct receptor-operated channels (12). This in turn causes the activation of Ca 2ϩ -dependent K ϩ channels, sensitive to charybdotoxin (12,13), resulting in delayed membrane hyperpolarization and leading to the activation of a second class of hyperpolarization-sensitive Ca 2ϩ channels (14). We did not observe the Ca 2ϩ conductance activation prior to K Ca channel stimulation (29). Conversely, our results demonstrate that the first ionic event in PRL-R signal transduction is K Ca channel activation, since this activation is observed in excised patches. Based on the observed inhibitory effects of protein tyrosine kinase inhibitors on the  activity of PRL-stimulated K Ca channels, we propose that tyrosine kinase is involved in the positive regulation of these channels. The direct tyrosine phosphorylation of the delayed rectifier K ϩ channel has also been proposed for the m1 muscarinic acetylcholine receptor (11). This tyrosine kinase regulation is obviously an essential link in PRL signal transduction as it was recently found that the tyrosine kinase inhibitor herbimycin A was able to block a substantial portion of the prolactin signal to the milk protein gene promoter, ␤-lactoglobulin (30). In Nb2 cells it was shown that herbimycin A could also abolish the JAK2 kinase and receptor phosphorylation (9).
In our study the effects of PRL and protein tyrosine kinase inhibitors were observed in excised-patch experiments, indicating that the effects are not controlled by cellular metabolism, but are direct and that protein tyrosine kinase remains closely associated with K Ca channel activity. In this context it was of interest to check the effect of anti-JAK2 antibody on K Ca channel activity. The effectiveness of using antibodies in patch-clamp experiments was previously shown by Schweizer et al. (45). When anti-JAK2 antibody was introduced into the patch pipette in whole cell experiments, the K Ca conductance was almost completely inhibited within an average of 15 min, and PRL was no longer able to stimulate K Ca conductance. Anti-JAK2 antibody also decreased the open probability of K Ca channels when it was applied to the cytoplasmic side of the membrane in the inside-out patch mode. In immunoblotting experiments JAK2 kinase was revealed in the presence or absence of stimulation by PRL, showing that it is constitutively associated with the PRL-R. On the other hand, electrophysiological studies using anti-JAK2 antibody showed that this kinase is also constitutively associated with the K Ca channel. The results with JAK2 immunizing peptide, demonstrating the suppression of the effects observed using anti-JAK2 antibody alone in both electrophysiological and immunoblotting studies, show that the effects of the anti-JAK2 antibody are specific. The ability of the constitutively active kinases to stimulate cellular responses has previously been shown for MAP and phosphatidylinositol 3-kinase (46). On the other hand, there is growing evidence for the existence of a protein tyrosine kinase activation mechanism that functions indirectly by second messengers. For example, in the brain, membrane depolarization, which causes an increase in intracellular Ca 2ϩ levels, increases protein tyrosine kinase (47) and protein mitogen-activated protein kinase (48) activity. Extracellular signals (in our case PRL) appear to stimulate the activity of protein tyrosine kinase, but it may also be regulated by other factors (e.g. membrane potential and intracellular Ca 2ϩ ). This constitutive activity of JAK2 kinase is the reason for K Ca channel inhibition by the anti-JAK2 antibody in the absence of PRL stimulation. Therefore, our results demonstrate the functional involvement of JAK2 kinase in constitutive K Ca channel activity and the stimulation of these channels by PRL. However, it was shown by immunoblotting experiments that genistein (100 -500 M) was not able to inhibit the JAK2 kinase phosphorylation induced by growth hormone but it blocked the tyrosyl phosphorylation of other proteins (intermediary tyrosine kinases) (49,50). This fact may be explained by the various examples of interactions at the level of the same protein; many phosphoproteins are phosphorylated at the same or at distinct residues by more than one protein kinase: in the case of tyrosine hydroxylase, the nicotinic acetylcholine receptor, synapsin 1, or Ca 2ϩ channel of the L type (47). This multisite protein phosphorylation appears to be the rule rather than the exception (47). Therefore, we cannot completely exclude the possible involvement of other tyrosine kinases in K Ca channel functions.
The fact that the channel was activated by the protein tyrosine phosphatase inhibitor orthovanadate in excised patches suggests that the channel can be modulated in opposite directions by protein tyrosine kinase and protein tyrosine phosphatase. This modulation may be due to phosphorylation/dephosphorylation of the channels themselves or of regulatory protein(s) associated with these channels and the PRL receptor. The protein tyrosine phosphatase endogenous to CHO cells was observed in most excised patches and thus may be a member of the transmembrane class of protein tyrosine phosphatase (51,52) or of the cytoplasmic protein tyrosine phosphatase which possess an SH2 domain or another means of association with membrane-bound proteins (53,54). These studies suggest that tyrosine phosphorylation/dephosphorylation systems can modify K Ca channels rapidly, producing flexible changes in PRL-R signaling. Similar types of complexes that contain regulatory protein kinases and phosphatases have previously been shown for a variety of different ion channels (20 -22, 55-57) and for K ϩ channels in particular (20,21,56,57). But the participation of such protein tyrosine kinase modulation of K ϩ channels in PRL signal transduction has never been shown. To our knowledge this study provides the first example of the regulation of K ϩ channels by PRL and protein tyrosine kinase in cell-free preparations. Our study is in accordance with a hypothesis previously proposed (57) about such submembrane complexing of ion channel proteins with modulatory enzymes like kinases and phosphatases as a common means by which cells achieve highly localized regulation of ion channel function by otherwise ubiquitous biochemical processes. Of particular interest is the finding that the modulation described here is due to a constitutive JAK2 tyrosine kinase activity. These results take on added significance because the PRL receptors were expressed in a heterologous system. However, remarkably, the endogenous JAK2 tyrosine kinase and endogenous K Ca channels in the host cell associate closely with heterologously expressed PRL-R and this complex retains its ability to be stimulated by PRL in a detached membrane patch.
As, on the one hand, PRL is known to stimulate the process of proliferation (58 -61) and, on the other hand, K Ca channels have also been shown to be involved in the control of cell proliferation (25,43,62), the mechanism of PRL stimulation of K Ca channels through tyrosine phosphorylation, presented in this work, could provide a clue to understanding the regulation of cell proliferation by PRL.