Distinct Cytoplasmic Regions of the Prolactin Receptor Are Required for Prolactin-induced Calcium Entry*

Two cytoplasmic regions of the prolactin (PRL) receptor are well documented for their participation in PRL signal transduction, the membrane proximal box 1 and the COOH-terminal region. In order to study the role of these regions in PRL-induced Ca2+ increase, we use Chinese hamster ovary (CHO) cells stably transfected with mutated PRL receptor cDNA. These cells express the long form of PRL receptor deleted from box 1 (CHO Δ1 cells) or the 141 amino acids of the COOH-terminal region (CHO H3 cells). The patch-clamp technique in “whole-cell” configuration and microfluorimetric techniques were used singly or in combination. Data obtained for these cells were compared with those we have recently published using CHO cells expressing the wild-type long form of the PRL receptor (CHO TSE32). In contrast to CHO TSE32 cells, exposure of CHO Δ1 or H3 cells to PRL (0.05–50 nm) did not modify [Ca2+] i . We have previously shown that the PRL-induced calcium influx via voltage-insensitive, Ca2+ channels was due to the activation of tyrosine kinase-dependent K+ channels that hyperpolarize the CHO TSE32 cell membrane (hyperpolarization-driven Ca2+influx). Therefore, two events are involved in PRL-induced Ca2+ changes (i) JAK2-activation of K+ channels and (ii) intracellular messenger-opening of Ca2+ channels. In CHO Δ1 cells, PRL (0.05–50 nm) neither hyperpolarized the membrane potential nor stimulated the JAK2-dependent K+ current, confirming the pivotal role played by box 1/JAK2 in the PRL-induced activation of K+ channels. However, when these cells were voltage-clamped below the resting membrane potential, application of 5 nm PRL resulted in an increase in Ca2+ influx. Therefore, box 1/JAK2 was not involved in the opening of these Ca2+ channels. In CHO H3 cells, 5 nm PRL activated the K+ current and hyperpolarized the membrane potential without any effect on [Ca2+] i . Moreover, PRL was also ineffective on CHO H3 cells voltage-clamped below the resting membrane potential. Therefore, the COOH-terminal region is involved in the production of the intracellular messenger that opens voltage-independent Ca2+ channels. We conclude from these findings that box 1 and COOH-terminal regions are both needed for PRL-induced Ca2+ changes.

The peptide hormone, prolactin (PRL), 1 exerts pleiotropic biological effects in a wide variety of cells and tissues via membrane receptors (1). PRL receptors (PRL-R) belong genetically to the larger family of receptors known as the hematopoietic-cytokine superfamily (2). These receptors present common structural features in the extracellular domain, in particular two cysteine pairs and ws/ws box. In the cytoplasmic domain, a proline-rich motif (box 1) located in the membrane proximal region is common to several receptors in this superfamily (3). The intracellular domain of receptors in this family lacks intrinsic kinase activity and varies in both length and sequence. The short form of PRL-R (57 or 30 -50 amino acids) was at first only found in rat and mouse cells. However, recent reports presented some evidence for the expression of short and long forms of PRL-R in other species (4). While the physiological functions of the short form remain controversial, the long form has been found to be capable of activating transcription of genes involved in cell differentiation (5).
Recent studies have been marked by considerable progress in understanding the intracellular signaling mechanisms for the different members of this receptor superfamily. These receptors associate with and activate several cytoplasmic tyrosine kinases in the Janus tyrosine kinase family. It has been shown for PRL-R that box 1 was required for JAK2 association with this receptor and phosphorylation (6) but that it was not sufficient for signal transduction (7). Deletion of the carboxylterminal 141 amino acids results in partial loss of transcriptional signaling activity (8). In the same way, a recent report indicated that the carboxyl-terminal portion of the growth hormone receptor is also required to activate transcription (9). Proteins, such as signal transducers and activators of transcription (STAT), are assumed to associate, through their SH2 domains, with COOH-terminal phosphotyrosines of activated receptors, where they become phosphorylated and activated (10).
To study early events in PRL signal transduction, we have developed a CHO cell line (CHO TSE32) stably transfected with the cDNA of the long form of rabbit (rb) mammary PRL-R (11). These CHO-transfected cells respond to PRL by stimulating the co-transfected milk protein gene promoter (12), proving that such cells are fully capable of transmitting the PRL signal and that PRL-R is functional. Using this cell line, we demonstrated that exposure of cells to physiological concentrations of PRL (5 nM) resulted in an increase in the cytosolic free calcium concentration ([Ca 2ϩ ] i ) (13) by stimulating both Ca 2ϩ entry and mobilization from intracellular Ca 2ϩ stores. Electrophysiologi-cal techniques were used to improve characterization of the early effects of PRL on membrane ion conductances. We have recently shown that PRL-induced Ca 2ϩ entry was due to JAK2dependent stimulation of calcium and voltage-activated potassium channels (14,15). The resulting hyperpolarization stimulates Ca 2ϩ entry through voltage-insensitive Ca 2ϩ channels (14), which require PRL-induced production of a cytosolic messenger to be opened.
In this article, we report on the structure-function relationships of PRL-R: using microfluorimetry and electrophysiology on CHO cells expressing mutated PRL-R, we confirm that box 1 is required for activation of the Ca 2ϩ -and voltage-dependent Kϩ conductance and demonstrate, for the first time, that the COOH-terminal region of PRL-R is necessary for Ca 2ϩ entry.
oPRL Binding on Cell Membrane and Western Blot Analysis-"Whole cell" binding and Scatchard analysis were performed essentially as described previously (8). Purification of the rbPRL-R complexes, immunoprecipitation, and Western blot Analysis were previously described (8).
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 whole cell recordings. 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 resistance were compensated and 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 (Ϫ3dB) and digitized at 10 KHz for storage and analysis.
Spectrofluorimetric Assay of Cytosolic Calcium-These experiments were performed using the fluorescent probe indo-1, as already described (13). The cells were incubated with 5 M indopentaacetoxymethyl ester (indo-1/AM) and 0.02% Pluronic F127 (Molecular Probes, Eugene, OR) in Hank's solution for 30 min at 37 Ϯ 1°C, then washed and maintained at room temperature in the same saline solution before the fluorescence measurements. This procedure resulted in an intracellular indo-1 concentration of between 20 and 60 M, estimated from comparison with the loading via a patch pipette.
For single cell measurements the dual emission microspectrofluorimeter was constructed from a Nikon Diaphot inverted microscope (Nikon France, Paris, France) fitted with epifluorescence (ϫ 100 oil immersion fluorescence objective; numerical aperture, 1.3). For excitation of indo-1, a collimated light beam from a 100-watt mercury arc lamp (Nikon) was filtered at 355 nm and reflected from a dichroic mirror (380 nm). The emitted fluorescence signal was passed through a pinhole diaphragm slightly larger than the selected cell and directed onto another dichroic mirror (455 nm). Transmitted light was filtered at 480 nm, reflected light was filtered at 405 nm, and the intensities were recorded by separate photometers (P1, Nikon). Single photon currents were converted to voltage signals, divided on-line by a monolithic laser-trimmed two-quadrant divider (AD535, Analog Devices, Norwood, MA). Under these experimental conditions, the r ϭ F405/F480 ratio was recorded on-line as a voltage signal and was expressed as [Ca 2ϩ ] i using the formula derived by Grinkiewicz et al. (17). Ca 2ϩ calibrations were obtained under simultaneous whole cell clamp and microspectrofluorimetric measurements. The patch pipettes were filled with internal solution containing 10 mM EGTA (solution A), 10 mM CaCl 2 (solution B), or 9.2 mM EGTA and 5.4 mM CaCl 2 (solution C). Solutions A and B were used to estimate minimum and maximum values, R min and R max , respectively. Solution C was used to evaluate the product of the apparent dissociation constant (K d ) and the ratio of fluorescence of free indo-1 divided by the fluorescence of Ca 2ϩ -bound indo-1 with 355 nm excitation and 480 nm emission (␤). The latter solution had a free Ca 2ϩ of 300 nM, calculated using the stability constants and computer program of Fabiato and Fabiato (18). R min , R max , and K d x␤ averaged 0.039 Ϯ 0.01 (n ϭ 15), 0.65 Ϯ 0.08 (n ϭ 17), and 581 Ϯ 22 nM (n ϭ 15), respectively.
In other experiments [Ca 2ϩ ] i was measured on cell populations using a Hitachi F2000 spectrofluorometer. The glass coverslide carrying the cells was positioned on a plastic holder in a quartz cuvette. The indo-1 fluorescence response to the intracellular calcium concentration was calibrated from the ratio of 405/480 nm fluorescence values after subtraction of the background fluorescence of the cells at 405 and 480 nm as described by Grinkiewicz et al. (17). The dissociation constant for the indo-1⅐Ca 2ϩ complex was taken as 405 nM. The R max and R min values were calculated from measurements using 25 M digitonine and 5 mM EGTA.
Simultaneous Electrophysiological and Microfluorimetric Recordings-These experiments were performed using the fluorescent Ca 2ϩ probe indo-1. The cells were loaded with indo-1 pentasodium salt (30 M) via the patch pipette. Indo-1 salt diffused gradually through the patch pipette to the recorded cell. [Ca 2ϩ ] i was determined as described above.
In single cell experiments, a "pouring" pipette with a tip opening of 10 -20 m was used for local drug application to the investigated cell. This pipette was filled with the same extracellular saline as that used in the bath and the drug under investigation was added to it in appropriate concentrations. The pipette was brought close to the investigated cell at a distance of 50 -100 m. In cell population experiments the drugs and reagents were added directly to the cuvette under continuous stirring. All experiments were performed at room temperature (20 -22°C).
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 were measured isochronally before the end of the pulse. Results are expressed as mean Ϯ S.D. where appropriate. Each experiment was repeated several times. Student's t test was used for statistical comparison among means and differences with p Ͻ 0.05 were considered significant.

Characterization of rbPRL-R Deletion Mutants Expressed in CHO Stable
Transfectants-To analyze the functional domains of the rbPRL-R involved in prolactin-induced calcium entry, we designated two mutated forms of the rbPRL-R (Fig. 1A). One form is a carboxyl-terminal deleted mutant lacking the last 141 residues (H3/T451). The other form is an internal deletion mutant lacking the proline-rich homology region called box 1 (⌬1/⌬245-267). The corresponding cDNA were inserted in pECE expression vector (19). CHO cells were stably transfected with wild-type rbPRL-R (WT, CHO TSE32) and mutated forms, H3 and ⌬1. Binding studies (Fig. 1B) indicate that mutant receptor forms (H3 and ⌬1) have higher binding capacities for oPRL than the wild-type receptor. Scatchard analysis indicate that mutant receptor forms have similar binding affinity to wild-type receptor but are more expressed at the cell surface. The size of the different receptor forms and their expression were then determined in CHO stable transfectants. Solubilized proteins were immunoprecipitated by anti-receptor polyclonal antibody developed against recombinant rbPRL-R extracellular domain expressed in Escherichia coli, ␣102, and the blot was revealed with a second polyclonal antibody (␣46). As shown in Fig. 1C, the different CHO stable clones express receptor proteins of expected molecular sizes for the WT (ϳ100 kDa), ⌬1 (ϳ95 kDa), and H3 (ϳ80 kDa). We confirmed that more receptor proteins are expressed for the two receptor mutants than for WT receptor form.
Effects of PRL on [Ca 2ϩ ] i in CHO TSE32, CHO ⌬1, and CHO H3 Cells-We have recently used "single cell" microfluorimetry to show that physiological concentrations of PRL stimulate Ca 2ϩ entry and/or induce a mobilization of calcium ions stored in intracellular compartments in CHO TSE32 cells (13). In this paper, we confirm this result on cell populations. Fig. 2A shows the PRL-induced Ca 2ϩ increase in CHO TSE32. Continuous perfusion of 5 nM PRL resulted in an increase in [Ca 2ϩ ] i (amplitude: 263 Ϯ 14 nM, n ϭ 36). Under this experimental procedure, Ca 2ϩ returned very slowly to basal value. When PRL was removed from the cuvette, Ca 2ϩ returned more rapidly to basal value (about 200 s).
By contrast, application of the same concentration of PRL (5 nM) did not affect [Ca 2ϩ ] i in CHO ⌬1 (Fig. 2B) or in CHO H3 cells (Fig. 2C). On the contrary, PRL slightly decreased the [Ca 2ϩ ] i in both cell lines (CHO ⌬1: Ϫ19 Ϯ 5 nM, n ϭ 34; CHO H3: Ϫ22 Ϯ 6 nM, n ϭ 32). This effect persisted as long as PRL was present in the bath. On termination of the PRL application, [Ca 2ϩ ] i slowly returned to baseline levels (within 200 s).
As we have observed several types of response to PRL (see Ref. 13 7). Therefore, although a systematic dose-response study was not carried out, PRL did not increase [Ca 2ϩ ] i in CHO ⌬1 and CHO H3, whatever the concentration used.
Effects of PRL on Voltage-activated, Ca 2ϩ -dependent K ϩ Con- ductance of CHO ⌬1 and H3 Cells-To investigate the physiological action of PRL on K ϩ conductance, CHO cells were voltage-clamped at Ϫ40 mV, close to the mean resting membrane potential. Contamination of K ϩ current recordings with Na ϩ was avoided by the use of TTX (2 M) containing external solution.
Application of PRL (5 nM) to CHO ⌬1 cells had no effect on the amplitude or kinetics of the K ϩ current (I-V relationships and time course) in the majority of recorded cells (9 of 15 cells).
On the other hand, PRL (5 nM) considerably increased the amplitude of the steady-state K ϩ current in 7 of 12 CHO H3 cells. In these cells, the mean amplitude of the K ϩ current (for a step from Ϫ40 to ϩ60 mV) was not significantly modified during hormone application (t 0 ), whereas it was markedly enhanced from 28 Ϯ 4.6 pA under control conditions to 115 Ϯ 23 pA 200 s after the end of the hormone application (t 0 ϩ 200) (Fig. 4C). This PRL-induced increase was very slow (time to peak 169 Ϯ 56 s) with an incomplete return to basal level after 8 -12 min (Fig. 4D). These results for CHO H3 cells are similar to those previously reported for CHO TSE32 cells (14). In both cases the outward current was characterized by application of calcium-dependent K ϩ channel inhibitors (50 nM charybdotoxin or 50 nM iberiotoxin) at the end of the recording (Fig. 4, B and  D).
Effects of PRL on Membrane Potential of CHO ⌬1 and H3 Cells-Local application of PRL (5 nM) to 6 of 10 CHO ⌬1 cells current clamped at resting potential caused a prolonged depolarization (15-25 mV) associated with a 150% increase in input resistance, as can be seen in Fig. 5A, a. These responses to PRL were reversible within 3-5 min after the peptide application was terminated. A similar response was obtained when we applied 50 nM CTX to CHO cells, instead of PRL (data not shown) or during PRL-induced depolarization (Fig. 5A, b).
When the same concentration of PRL was applied to CHO H3 cells, 5 out of 9 cells that hyperpolarized (10 to 20 mV) in response to the peptide remained so for a prolonged period during which the membrane conductance was markedly increased (Fig. 5B, a). These effects took considerably longer to reverse in these cells. Application of CTX (50 nM) blocked PRL-induced membrane hyperpolarization (Fig. 5B, b).
Simultaneous Monitoring of [Ca 2ϩ ] i , Membrane Potential, and Current in CHO ⌬1 and H3 Cells Stimulated with 5 nM PRL-As demonstrated by our previous studies (13,14) PRL stimulates Ca 2ϩ entry through voltage-insensitive nonspecific channels by hyperpolarizing the membrane potential of CHO cells expressing the long form of PRL receptor. To further investigate the action mechanism of PRL in CHO cells expressing the mutated receptors, the relation between intracellular calcium concentration changes and calcium influx via voltageindependent Ca 2ϩ channels was systematically investigated, using the patch-clamp technique combined with dual emission [Ca 2ϩ ] i recordings obtained with indo-1. These experiments were conducted under voltage-clamp conditions, including inhibitors of both Na ϩ (tetrodotoxin, extracellularly applied) and K ϩ channels (NMG gluconate in the recording pipette) to isolate Ca 2ϩ currents for monitoring. Two approaches were used (i) PRL was applied to the recorded cell after the patch of membrane was disrupted and control recordings were made (Figs. 6A and 7A) and (ii) PRL was applied a few minutes before the establishment of whole cell recordings (Figs. 6B and 7B), therefore without control recordings on the same cell. These experiments made it possible to vary the holding potential in one cell and follow the subsequent changes in [Ca 2ϩ ] i . In unstimulated CHO ⌬1 cells, maintenance of the holding potential of Ϫ40 mV resulted in a stable current base line and equally a stable [Ca 2ϩ ] i level of about 150 nM (range: 120 -180 nM) (Fig. 6A). Upon shifting the holding potential to Ϫ60 or Ϫ80 mV, no current or [Ca 2ϩ ] i changes were triggered in the absence of PRL (data not shown). On the other hand, application of 5 nM PRL to 7 out of 12 CHO ⌬1 cells voltage-clamped at Ϫ40 mV induced a long-lasting inward current of modest amplitude (Ͻ10 pA, Fig. 6A, middle trace). This current displayed slow activation and very slow, if any, inactivation. It was associated with a slight increase in the conductance amplitude (about 25%) and a distinguishable increase in background noise in the current trace. Concomitantly to the onset of the inward current, a slowly developing [Ca 2ϩ ] i rise became detectable in the indo-1 recording (Fig. 6A, lower trace). Within 100 -120 s after hormone application, a new steady state [Ca 2ϩ ] i of 210 nM (range, 170 -245 nM) was reached (plateau). Successive 20-mV step decreases in holding voltage below Ϫ40 mV resulted in gradual stepwise increases in [Ca 2ϩ ] i (Fig. 6A, lower trace). At In PRL-pretreated CHO ⌬1 cells, exploration of membrane potential from 0 to Ϫ80 mV stimulated Ca 2ϩ influx in 8 out of 8 cells tested. Hyperpolarization to Ϫ40, Ϫ60, and then Ϫ80 mV led to subsequent Ca 2ϩ rises, increasing with the amplitude of the hyperpolarizing potential, reaching a maximal value (683 Ϯ 89 nM) at Ϫ80 mV (Fig. 6B). Depolarization to 0 mV again decreased [Ca 2ϩ ] i to basal level. The percentage of responding cells and the response amplitudes were lower in the first experimental approach. This discrepancy may be due to a longer delay between the rupture of the patch membrane and the examination of the PRL effects in this case. Moreover, in the second experimental approach, when the exploration of the membrane potential was performed more than 20 min after establishing the whole cell recording, the hyperpolarization-driven calcium entry was drastically reduced or absent. Taken together, these data confirm the involvement of a cytosolic, diffusible second messenger in the PRL-induced Ca 2ϩ entry (14), and show that this messenger can be produced in response to PRL in CHO ⌬1 cells.
A similar enhancement of steady state [Ca 2ϩ ] i levels upon PRL application was observed under conditions where neither Na ϩ nor K ϩ were blocked (in 4 out of 9 cells tested, data not shown). Under these conditions, obviously no inward current was detectable. There was, however, a small outward current , two calcium-dependent K ϩ channel inhibitors. Symbols represent K ϩ current amplitudes measured by 100 ms pulses at the end of a step depolarization to ϩ20 mV from Vh of Ϫ40 mV. due, at least in part, to the activation of a Ca 2ϩ -dependent K ϩ current as shown in CHO TSE32 cells (14) in parallel with the calcium increase.
In PRL-pretreated CHO H3 cells, continuous variation of holding potential from 0 to Ϫ80 mV had very little effect on [Ca 2ϩ ] i at the beginning of the recording (Fig. 7B), comparable to the recordings in unstimulated CHO cells. Higher (50 nM) or lower (0.5 and 0.05 nM) concentrations of PRL had no greater effect on the [Ca 2ϩ ] i of these cells (data not shown). From these data, it appears that, in CHO H3 cells, PRL did not produced the intracellular messenger which opened the voltage-insensitive Ca 2ϩ channels.
Effects of PRL on Voltage-dependent Inward Currents in CHO ⌬1 and H3 Cells-We have previously shown that CHO cells possess two voltage-dependent inward currents: (i) a TTXsensitive Na ϩ channel and (ii) an L-type Ca 2ϩ channel (20). PRL (5 nM) had no effect on the Na ϩ channel but slightly inhibited the voltage-dependent Ca 2ϩ current in CHO TSE32 cells (14). When the outward current was completely blocked by replacing intracellular K ϩ with N-methyl-D-glucamine, and the L-type Ca 2ϩ current by adding nifedipine (0.5 M), PRL (0.5 to 50 nM) caused no change in the Na ϩ current in CHO ⌬1 cells or in CHO H3 cells (data not shown), whatever the amplitude of step depolarizations (current-voltage relationships) or the delay between hormone application and current eliciting (time course).
In order to record voltage-activated Ca 2ϩ current, nifedipine was replaced by TTX (2 M). Repetitive (every 20 s) voltage steps (time course) from Ϫ40 to ϩ20 mV elicited Ca 2ϩ currents whose amplitudes were not affected by PRL (5 nM) in the majority of CHO ⌬1 (7 of 12 cells) and CHO H3 (9 of 15 cells) cells (Fig. 8B). In the responding CHO ⌬1 (Fig. 8A) and CHO H3 cells, PRL induced a slight increase in Ca 2ϩ current amplitude (CHO ⌬1, control: Ϫ28.8 Ϯ 4.5 pA, PRL, Ϫ36 Ϯ 5.7 pA, n ϭ 5; CHO H3, control: Ϫ30.8 Ϯ 5.3 pA, PRL, Ϫ37.2 Ϯ 4.5 pA). These effects of PRL were observed in both cell types at all potentials where the current was activated (data not shown). DISCUSSION In our previous research we established that, in CHO cells, transfected with cDNA of the long form of rbPRL-R (CHO TSE32), the primary ionic events in PRL-R signal transduction were the activation of Ca 2ϩ -dependent K ϩ channels and Ca 2ϩ entry (13,14). More recently, our data have suggested that at least one of the kinases involved in channel stimulation by PRL ] i was associated with activation of an inward current and an increase in membrane conductance. Following hormone application, membrane potential was stepped down from Ϫ40, Ϫ60, and Ϫ80 mV, until [Ca 2ϩ ] i reached a plateau. Finally, membrane potential was depolarized to 0 mV. B, cells were pretreated with 5 nM PRL before the establishment of whole cell recording and variations in the holding potential were performed shortly after this in order to avoid or to minimize the effects of prolonged dialysis of whole cell recording. Under these conditions, hyperpolarization-driven calcium entry was greater than in A, suggesting the involvement of an intracellular second messenger in this mechanism. m.p., membrane potential; Im, whole cell membrane current. Traces are representative of 20 experiments. may be JAK2 tyrosine kinase (15). The mutagenesis of cloned receptors is a powerful tool that can help elucidate the mechanism whereby the ligand binding signal is tranduced and interpreted by a responding cell. So, in these experiments, we used CHO cells expressing wild type and two deletion mutants of the rbPRL-R, to elucidate the structure-function relationships of PRL receptors. Stably transfected CHO cell lines were characterized by PRL binding analysis and the receptor size was confirmed by Western blot analysis.
The major findings in this study are that (i) if the membrane proximal, proline-rich region called box 1 is deleted (CHO ⌬1), PRL stimulates neither K ϩ channels nor the subsequent Ca 2ϩ entry through voltage-independent, hyperpolarization-driven Ca 2ϩ channels and (ii) in the absence of the COOH-terminal region (CHO H3), PRL does not induce Ca 2ϩ influx though it activates K ϩ channels. We conclude from these findings that box 1 and the COOH-terminal region of rbPRL-R are both required for PRL-induced calcium entry.

Box 1 Region of PRL-R Is Necessary for K ϩ Channel Activation but Not Sufficient for Ca 2ϩ Entry Stimulation-Previous
works have shown that box 1 is essential for PRL signal transduction (19). For example, the box 1 region and, particularly, the last proline was shown to be critical for JAK2 phosphorylation and association with PRL-R (21). Treatment of CHO TSE32 cells with a series of tyrosine kinase inhibitors or intracellular dialysis of an anti-JAK2 tyrosine kinase antibody via a patch pipette, completely inhibits the basal and PRL stimulated activity of Ca 2ϩ -dependent K ϩ channels (15). From these findings it was concluded that JAK2 tyrosine kinase, constitutively associated with PRL-R, was implicated in basal activity and PRL stimulation of Ca 2ϩ -dependent K ϩ channels. However, it could not be excluded that these effects could be due to direct effects of JAK2 on K ϩ channels that prevented further activation by PRL. In unstimulated CHO cells expressing box 1-deleted PRL-R (⌬1), we show that the amplitude of the K ϩ current was not significantly modified as compared with CHO TSE32 or native CHO K1 cells. This observation suggests that all K ϩ channels are controlled by JAK2 tyrosine kinase, but only a small part of them are coupled to PRL-R. We have recently shown that PRL-induced K ϩ channel activation resulted in a transient, slow hyperpolarization in CHO TSE32 (14). Conversely, in CHO ⌬1 cells PRL decreased the same K ϩ current and depolarized the membrane potential. These effects may be due to the activation of a protein phosphatase that dephosphorylates K ϩ channels. This could explain the PRLinduced decrease in [Ca 2ϩ ] i we have observed in CHO ⌬1 cells. Protein phosphatases actively participate in transduction pathways (22,23). Furthermore, receptor or protein kinase downregulation and ion channel activity involve protein phosphatase activities (24,25). We have recently shown that, like PRL, orthovanadate, a tyrosine phosphatase inhibitor, was capable of increasing K ϩ channel activity in CHO TSE32 cells (15). From this study we have concluded that the functioning of PRL-stimulated K ϩ channels is modulated by protein tyrosine kinases and protein tyrosine phosphatases and thus regulated by constitutive tyrosine phosphorylation/dephosphorylation. What protein tyrosine phosphatase could be involved in this effect? A complex of PRL-R⅐JAK2⅐PTP1D appears to be necessary for initiating PRL-R signaling (26). PTP1D acts as a positive regulator of PRL-R-dependent induction of ␤-casein gene transcription. Their involvement is unlikely in our case since, in CHO ⌬1 cells, PRL inhibited K ϩ current, although JAK2 can neither associate with PRL-R nor form the complex in these cells. Other phosphatases appear to be implicated in cytokine receptor signaling (27). Their participation in the PRL-induced modulation of ion channels is under investigation in our laboratory.
In CHO H3 cells (lacking the last 141 amino acids of the COOH-terminal region) PRL was always able to activate the K ϩ channels and hyperpolarize the membrane potential to the same extent as in CHO TS32 cells (Figs. 4, C and D, 5B). In addition, intracellular dialysis of these cells with anti-JAK2 tyrosine kinase antibody by the patch pipette decreased K ϩ conductance and prevented PRL-stimulation of the K ϩ channels (data not shown), as we have previously shown in CHO cells expressing the wild-type long form rbPRL-R (15). These results are not surprising, as we have previously shown (8) that PRL was still able to induce JAK2 tyrosine phosphorylation in CHO cells expressing the carboxyl-terminal truncated PRL-R (named T451 in this paper).
Data shown in this study further implicates the box 1-JAK2 couple in PRL-induced activation of K ϩ currents and demonstrate that the COOH-terminal region does not participate to the cascade of events leading to the activation of K ϩ channels.
At least in some cell types, proliferative signals seem only to require box 1 (6,28). A number of studies have shown that the mechanisms promoting proliferation and tumorization processes in various cell models (melanoma, breast cancer, prostate cancer, lymphocytes, neuroblastoma, brown fat cells, etc.) often involve the activation of voltage-and/or calcium-activated K ϩ channels (29 -33). Most of these studies were performed on T-lymphocytes. Indeed, the proliferation of these cells induced by various mitogenic stimuli, such as PRL, is modulated by potassium current blockers (34 -36). Because malignant (Nb2) lymphocytes proliferate independently of calcium influx and extracellular calcium concentrations have no influence on PRLinduced Nb2 proliferation (36 -38), Wang et al. (39) have proposed that potassium current per se rather than potassium current modulation of calcium influx mediates prolactin-induced proliferation of Nb2 cells (39). The predominant form of PRLR expressed in this cell line is a natural deletion mutant of the long form lacking a 198-amino acid segment of the cytoplasmic domain. However, this mutant receptor retains box 1 and box 2 and tyrosine phosphorylation participates in the anti-apoptotic effect of PRL in Nb2 cells (36). These studies are in agreement with the present work, showing that PRL-induced K ϩ channel activation requires box 1 but not the COOHterminal region. To our knowledge, no other studies have shown the effect of PRL on K ϩ conductance in another cell model.
To our knowledge, no previous studies have shown the participation of K ϩ channels in the signal transduction of the other members of the cytokine receptor superfamily. However, it has been reported that growth hormone (GH) can cause an increase in [Ca 2ϩ ] i , independently of JAK2 (40) or other tyrosine kinases (41). This Ca 2ϩ increase was attributed to a stimulation of voltage-dependent Ca 2ϩ channels (40). We have shown (14,42,43) that CHO cells possess L-type voltage-dependent Ca 2ϩ channels, but they can be only recorded with extracellular medium containing 60 mM Ca 2ϩ and, even under these particular conditions, the maximal amplitude of the voltage-dependent Ca 2ϩ current remains very low (10 -20 pA). According to our findings, it was not possible to detect any [Ca 2ϩ ] i increase in response to an electrical depolarization in CHO cells by combining electrophysiology and microspectrofluorimetry on the same cell. 2 Furthermore, more recently, we have shown that (i) GH inhibited L-type voltage-dependent Ca 2ϩ channels, and (ii) the GH-induced Ca 2ϩ increase was not affected by dihydropyridines in CHO cells expressing rabbit growth hormone receptor (43). Using these cells, we completely blocked Ca 2ϩ response to GH by using pretreatment with tyrosine kinase inhibitors. 3 Tyrosine kinase inhibitors blocked the erythropoietin-induced increase in [Ca 2ϩ ] i (44), suggesting a role for tyrosine phosphorylation but not for serine-threonine kinases in the erythropoietin modulation of intracellular calcium. The tyrosine kinase-ion channel coupling remains unclear in the cytokine receptor superfamily and requires further investigation.
COOH-terminal Region of the PRL-R but Not Box 1 Is Necessary for PRL Activation of the Voltage-independent Ca 2ϩ Channels-We have previously shown that PRL-induced Ca 2ϩ entry resulted from the activation of hyperpolarization-driven channels (14). Since the effect of PRL on these channels diminished during the electrophysiological recordings (patch-clamp configuration) because the saline patch pipette solution diluted the intracellular medium, we have postulated that one or more diffusible cytosolic second messengers and/or enzymatic activities were required. Earlier studies have shown the presence of a variety of second messenger-operated voltage-insensitive Ca 2ϩ conductances in many cell types (45,46). Inositol 1,3,4,5tetrakisphosphate (IP 4 ) is known to operate voltage-independent Ca 2ϩ conductance and to induce Ca 2ϩ entry (45). We have recently shown in CHOTSE32 cells that: (i) PRL induces rapid increases in two inositol phospholipids, phosphoinositol 4,5-P 2 and phosphoinositol 3,4,5-P 3 (47), (ii) PRL also increases the production of IP 4 without any significant increase in inositol 1,4,5-trisphosphate (48), (iii) IP 4 causes an increase in the open probability of a voltage-independent Ca 2ϩ channel (48), and (iv) PRL activates the same channel in "cell attached" configuration of the patch-clamp technique (48). PRL has no effect on voltage-independent Ca 2ϩ channels in cell-free configurations, confirming the involvement of a cytosolic messenger. Taken together, all these results demonstrate that PRL stimulates Ca 2ϩ entry through voltage-independent Ca 2ϩ channels by stimulating IP 4 production. The mechanism by which PRL stimulates IP 4 production remains unclear. To our knowledge, a metabolic pathway from phosphoinositol 3,4,5-P 3 to inositol 1,3,4,5-P 4 has not yet been described.
Exploration of the membrane potential during simultaneous measurements of [Caϩ] i and ionic currents shows that, in PRL-R, the last 141 amino acids in the COOH-terminal regions necessary for the activation of the voltage-independent Ca 2ϩ channels, but box 1 is not (Figs. 6 and 7). The effects of PRL on the phosphoinositide metabolism and inositol phosphate production in CHO H3 cells are currently under investigation in our laboratory. It can be postulated that the three tyrosine residues in the 141 amino acids play a pivotal role. In order to confirm the role of these residues and specify which tyrosine is involved in channel activation, we will work on CHO cells expressing point mutated PRL-R. PRL-R intracytoplasmic domain can be used as a bait to clone potential PRL-R interacting proteins by the yeast two-hybrid interacting cloning strategy (23,49). This strategy may help us to specify how IP 4 is produced in response to PRL.
To our knowledge, none of the studies have shown a direct effect of PRL on Ca 2ϩ influx in cells expressing natural form and we have found only two studies of Nb2 cells, where the role of Ca 2ϩ in PRL-induced events was inferred from biochemical and pharmacological experiments (37,38). Furthermore, in our experiments, PRL failed to induce any [Ca 2ϩ ] i increase in Nb2 cells. 4 We have developed an in vitro functional test consisting of the co-transfection in CHO cells of PRL-R cDNAs with a chimeric gene encompassing a milk protein gene promoter (␤-lactoglobulin) fused to the coding sequence of the chloramphenicol acetyltransferase gene (5). Using this test, we have recently shown that PRL was still able to activate the transcription of the target gene in CHO cells expressing the rb-PRL-R, where the last 141 amino acids of the COOH-terminal region have been deleted (T451 mutant) (8). However, the T451 mutant was less active (2.0-fold induction) than the wild-type receptor (4.5-fold induction). Thus, the tyrosine phosphorylation of the carboxyl-terminal part of the rbPRL-R appears to be a positive modulator in signal transduction to milk protein genes. On the other hand, the T451 mutant is also about half as active in activating STAT5. In addition, recent findings indicate that serine/threonine phosphorylation of STAT proteins increases DNA binding (50,51). MAPK is a candidate for this action. As its activity is dependent on Ca 2ϩ influx (52), the lower activation of STAT5 and the consecutive lower transcriptional activity of the T451 mutant may be explained by the absence of an increase in [Ca 2ϩ ] i in response to PRL in CHO cells expressing this receptor.
The COOH-terminal region (amino acids 454 -506) of the growth hormone receptor is also required for Ca 2ϩ signaling (40). Calcium is an ubiquitous intracellular messenger and regulator of cellular activities, including proliferation, mitosis, muscle contraction, energy metabolism, secretion, etc. For example, increases in intracellular Ca 2ϩ has been associated with the induction of proliferation-associated immediate early genes, including c-fos and c-jun (53). Calcium influx through store-operated Ca 2ϩ channels also seems to play a role in the activation of MAPK pathways (52). Billestrup et al. (40) have reported that a GH-induced rise in [Ca 2ϩ ] i is required for GH-stimulated insulin gene transcription in RIN 5AH cells.
In conclusion, the data presented in this article confirm the involvement of box 1/JAK2 tyrosine kinase in PRL-induced K ϩ conductance activation. For the first time we show that the COOH-terminal part of the long form rbPRL-R is involved in the opening of the voltage-independent Ca 2ϩ channels expressed in CHO cells. Thus, both box 1 and the last 141 amino acids of the COOH-terminal region of the PRL-R are required to produce the PRL-induced rise in [Ca 2ϩ ] i . Mutated PRL receptors may help us to further investigate the early effects of PRL on ion conductance and [Ca 2ϩ ] i and to determine if these effects converge upon MAPK and STAT5 activation and milk gene transcription.