The role of the calpain-calpastatin system in thyrotropin-releasing hormone-induced selective down-regulation of a protein kinase C isozyme, nPKC epsilon, in rat pituitary GH4C1 cells.

We have examined the mechanism for the selective down-regulation of protein kinase C epsilon (nPKC epsilon) in rat pituitary GH4C1 cells responding to thyrotropin-releasing hormone (TRH) stimulation. Among various low molecular weight protease inhibitors examined, only a cysteine protease inhibitor (calpain inhibitor I, N-acetyl-Leu-Leu-norleucinal) blocked the down-regulation of nPKC epsilon. Furthermore, the introduction of a synthetic calpastatin peptide, an exclusively specific inhibitor of calpain, into the cells also reduced the down-regulation, suggesting the involvement of calpain among all the intracellular cysteine proteases in this process. In accordance, we observed TRH-induced translocation of m-calpain from the cytosol to the membrane and the concomitant up-regulation of calpastatin isoforms; presumably, the former represents activation of the protease initiating the kinase degradation, while the latter constitutes a negative feedback system protecting the cells from activated calpain. These results suggest that in GH4C1 cells, TRH mobilizes both protease (m-calpain) and inhibitor (calpastatin) as a strictly regulating system for the nPKC epsilon pathway mediating TRH signals.

Protein kinase C (PKC) 1 isozymes play pivotal roles as major serine/threonine kinases in signal transduction cascades involved in agonist-induced responses of various cells (1). The isozymes can be categorized into three groups: conventional PKC (cPKC), novel PKC (nPKC), and atypical PKC (aPKC), based on their structural and enzymatic properties (2). Although their distinct tissue distribution has suggested different functional roles for each isozyme (2,3), their isozymespecific functions in physiological terms have not yet been fully resolved (4). We have focused our attention on the cellular function of PKC isozymes in the responses elicited by thyrotropin-releasing hormone (TRH), a hypothalamic hormone, in rat pituitary GH 4 C 1 cells and demonstrated that nPKC⑀ provides a major rate-limiting step in the secretion of prolactin (PRL) in response to agonist stimulation (5)(6)(7)(8).
GH 4 C 1 cells possess at least six PKC isozymes, i.e. ␣ and ␤II cPKC, ␦, ⑀, and, nPKC, and aPKC (5,8). The suppression of their activities by PKC inhibitors results in a reduction in PRL secretion evoked by TRH (6), indicating a mediator role in stimulus-response coupling. In accordance, TRH induces the translocation of all isozymes except from the cytosol to membranes (8). However, only the ⑀ isozyme undergoes subsequent down-regulation (5,7,8), indicating that this isozyme is the most extensively involved. Consistently, the introduction of an expression vector carrying nPKC⑀ cDNA into the cells induces an increase in PRL secretion, whereas increasing the amount of other isozymes, ␣, ␤II, and ␦, has no effect (8). These data suggest that nPKC⑀ among all the PKC isozymes present in GH 4 C 1 cells plays a central role in the secretory process and that down-regulation following translocation is a hallmark for the thorough activation of the enzyme.
Intracellular down-regulation of PKC following activation in general seems to involve facilitated proteolysis rather than transcriptional suppression (9). The protease(s) responsible for this process, however, have not yet been identified, although previous studies on phorbol ester-evoked cPKC down-regulation implies the involvement of calpain (10,11), serine proteases (12,13), and/or possibly other proteases (14). Consequently, the physiological significance of this down-regulation remains uncertain. Possibly, the proteolysis of PKC may produce a signal affecting the cellular situation. For instance, an active kinase fragment may be produced from PKC by limited proteolysis (15) as an intermediate product in down-regulation, depending on the protease involved. This kinase that acts independently of cofactors, including phospholipids, would phosphorylate substrates that may not be accessible to intact PKC. Alternatively, the loss of a specific PKC activity may lead to an alteration in cellular functions. Due to our limited knowledge about the mechanism of down-regulation, these possibilities remain obscure.
In the present study, we aimed to elucidate the mechanism of selective nPKC⑀ down-regulation in GH 4 C 1 cells by identifying the responsible protease as an initial step in addressing these questions in this particular system.

EXPERIMENTAL PROCEDURES
Materials-Leupeptin and casein were purchased from Microbial Research Institute and Merck, respectively. TRH was from Sigma. E-64d was generously provided by Dr. K. Hanada (Taisho Pharmaceuticals). Other reagents including protease inhibitors were purchased from WAKO Pure Chemicals, Nacalai Tesque, or Sigma. Antipeptidic antibodies to the carboxyl-terminal sequence of nPKC⑀ (8), to the amino-terminal sequence (22-mer) of -calpain (16), and to the aminoterminal sequence (18-mer) of m-calpain (17) were previously described.
An anti-calpastatin antibody was raised against a synthetic 16-mer peptide, CTIELDLISWLCFSVL, conjugated to keyhole limpet hemocyanin (Calibiochem) as described previously (16). This sequence corresponds to the carboxyl-terminal portion of rat calpastatin (18). A 27-mer synthetic calpastatin peptide (CS peptide, acetyl-DPMSSTYIEEL-GKREVTIPPKYRELLA-NH 2 ) corresponding to the minimum inhibitory segment in domain I of human calpastatin (19,20) and a control scramble peptide (acetyl-APRLEIVPTMYIYKLSPTGSEKLEDER-NH 2 ) were produced using an ACT396 peptide synthesizer (Advanced Chemtech) as described previously (16). A rhodamine-labeled CS peptide was produced as follows. An unacetylated CS peptide (25 mol) on a resin prior to deprotection was incubated with 100 mol of rhodamine isothiocyanate (Sigma) in 2 ml of dimethylformamide for 2 h at room temperature. The product was subjected to deprotection and purification as other peptides after extensive washing with the solvent.
Cell Culture-Rat pituitary GH 4 C 1 cells were maintained in Ham's F-10 medium (Life Technologies, Inc.) supplemented with 15% horse serum and 2.5% fetal bovine serum under a humidified atmosphere of 5% CO 2 at 37°C as described (21). Approximately 3.0 ϫ 10 6 cells were plated in 10-cm dishes and cultured for 5-7 days prior to use in each experiment.
Analysis of nPKC⑀ in GH 4 C 1 Cells-The cells were preincubated in F-12 medium (Life Technologies, Inc.) containing 1 mg/ml bovine serum albumin for 1 h prior to inhibitor treatment to minimize the influence of serum. The cells were then incubated in media containing protease inhibitors, synthetic peptides, or vehicle (0.1% dimethyl sulfoxide) for 1 h and stimulated with 200 nM TRH for 6 h unless stated otherwise in the figure legends. To examine down-regulation, cells were washed with ice-cold Mg 2ϩ -, Ca 2ϩ -free phosphate-buffered saline, collected in SDSpolyacrylamide gel electrophoresis solubilization buffer, and subjected to Western blot analysis as described previously (16). A Fab fragment of donkey anti-rabbit IgG antibody conjugated to peroxidase (Amersham) was used as a secondary antibody. The blots were visualized using an ECL kit (Amersham). To examine the localization of nPKC⑀, cells were collected in homogenization buffer (20 mM Tris-HCl, pH 7.5, 250 mM sucrose, 2.5 mM EGTA, 2.5 mM MgCl 2 , 230 M leupeptin, 2 mM phenylmethanesulfonyl fluoride, 50 mM 2-mercaptoethanol) and homogenized by sonication (Branson Sonifier 185, power 10, 15 s ϫ 4 times) at 4°C. The homogenates were ultracentrifuged at 350,000 ϫ g for 15 min. The supernatants and pellets were collected as cytosol and membrane fractions, respectively, and subjected to Western blot analysis as above.
Analysis of Calpain and Calpastatin-Calpain and calpastatin activities were assayed using 14 C-succinyl-casein ([ 14 C]casein) as follows. Cells (5-6 ϫ 10 7 cells) collected in homogenization buffer were sonicated and ultracentrifuged as described above. The supernatant was applied to a Mono-Q column (volume, 1 ml) connected to a Pharmacia fast protein liquid chromatography system. The adsorbed proteins were eluted with a linear gradient of 0 -0.5 M NaCl in 5 mM Tris-HCl, pH 7.5, 5 mM EDTA, and 5 mM 2-mercaptoethanol. For calpain assay, aliquots (50 l) of each fraction were incubated with 65 g of [ 14 C]casein (about 28,000 dpm) for 20 min at 30°C in 50 l of reaction buffer (50 mM Tris-HCl, pH 7.5, 5 mM CaCl 2 , 25 mM 2-mercaptoethanol), and the trichloroacetic acid-soluble radioactivity was measured in a scintillation counter as described previously (22). The calpain activity was calibrated against known amounts of purified calpain and expressed in units as previously defined (23). For the assay of calpastatin, aliquots (25 l) were heated at 100°C for 3 min and, after cooling, incubated for 40 min with 65 g of [ 14 C]casein and purified rabbit m-calpain (0.15 units) in 25 l of reaction buffer. The ability to inhibit the exogenously added calpain activity was defined as calpastatin activity.
To examine the localization of calpain, the cells were stimulated with TRH for various periods, harvested in Ca 2ϩ /EGTA buffer (pCa of 7, 20 mM Pipes/KOH, pH 7.2, 115 mM NaCl, 2 mM MgCl 2 , 1 mM CaCl 2 , 2.5 mM EGTA, 230 M leupeptin, 2 mM phenylmethanesulfonyl fluoride, 50 mM 2-mercaptoethanol), and fractionated as above. Cytosol and membrane fractions were subjected to Western blot analysis using anti-calpain isozyme antibodies. The localization of calpastatin was examined in a similar manner except that "homogenization buffer" was employed instead of Ca 2ϩ /EGTA buffer.

Effect of Various Protease Inhibitors on the TRH-induced
Down-regulation of nPKC⑀-Upon stimulation of GH 4 C 1 cells with TRH, the amount of nPKC⑀ decreased indicating downregulation (Fig. 1, lanes 1 and 2), as previously reported (5,7). To identify the class of protease(s) involved in this catabolic process, we examined the effects of various protease inhibitors.
Cells were preincubated with the protease inhibitors and subsequently stimulated with TRH (lanes 3-10); only ALLNal, a cell-permeable cysteine protease inhibitor (24), blocked the down-regulation of nPKC⑀ (lane 9). Phosphoramidon, pepstatin, N␣-tosyl-L-lysyl chloromethyl ketone, N-tosyl-L-phenylalanyl chloromethyl ketone, leupeptin, and E-64d had no effect. O-Phenanthroline seemed to be toxic to cells at the concentration used, causing cell damage and resulting in the apparent promotion of down-regulation. These results indicate that a cysteine protease, not a metallo, serine, or aspartic protease, is involved in the down-regulation of nPKC⑀. Lysosomal cysteine proteases do not seem to be involved since leupeptin had no effect (25). We do not know why E-64d, another cell-permeable cysteine protease inhibitor, did not inhibit the down-regulation. We presume that it may have been quickly degraded extra-or intracellularly and failed to reach sufficient concentrations to be inhibitory in the cytoplasm. This is because E-64d needs to enter cells in the ester form and is then hydrolyzed by an intracellular esterase to yield E-64 (26).
The effect of ALLNal was dose dependent (Fig. 2A). The concentration of inhibitor giving 50% inhibition was about 5 M, consistent with previous studies (27,28). The possibility that ALLNal influences the translocation of nPKC⑀ from the cytosol to membranes that occurs prior to down-regulation was excluded because the inhibitor had no effect on the amount of nPKC⑀ in the membrane fraction of TRH-treated cells (Fig. 2B).
Inhibition of Down-regulation by Calpastatin Peptide-Besides the inhibitory effect of ALLNal, originally developed as a calpain inhibitor (24), on the down-regulation of nPKC⑀ as shown above, the facts that TRH causes an elevation in intracellular calcium concentration (29) and that nPKC⑀ is an in vitro substrate of calpain (30) imply the possible involvement of calpain in this process. However, ALLNal may have affected other proteases present in the cell because it is not strictly specific for calpain (24,31). We therefore employed a more specific method to inhibit intracellular calpain activity using a synthetic CS peptide corresponding to the inhibitory segment of calpastatin (Fig. 3). A scramble peptide with an identical amino acid composition and a random sequence was employed as a negative control.
We first examined the ability of the CS peptide to inhibit purified calpain activity and to permeate into cells. In test tubes, the CS peptide inhibited the proteolytic activity of purified calpain with high affinity; 50% inhibition was achieved at a peptide concentration of 20 nM (panel A). In contrast, approximately 10 3 -fold more scramble peptide was necessary to suppress calpain activity. The permeability of the synthetic peptide into cells was confirmed using a rhodamine-labeled peptide (panel B); fluorescence was observed only in cells incubated with the peptide. Although this observation does not quantitate the actual concentration of CS peptide inside GH 4 C 1 cells, it indicates that at least a fraction of the extracellularly administered CS peptide entered the cells and remained in cytoplasm.
Confirming the potent inhibitory effect of the CS peptide on calpain and its cell permeability, we examined its effect on intact cells (panel C). The CS peptide significantly reduced the down-regulation of nPKC⑀, whereas the control scramble peptide had no effect. Because no proteases other than calpain have been shown to be inhibited by calpastatin or calpastatin peptide to our knowledge (32), these data suggest that it is indeed calpain, among various intracellular cysteine proteases, that is involved in the proteolytic process.
Calpain and Calpastatin Activities in GH 4 C 1 Cells-Although the experiments using protease inhibitors indicated the involvement of calpain in the down-regulation of nPKC⑀ in GH 4 C 1 cells, previous studies have identified neither calpain nor calpastatin in either enzymatic or immunochemical terms in these cells. We therefore analyzed the calpain and calpastatin activities of GH 4 C 1 cells by anion exchange HPLC (Fig. 4) since cellular calpain activities are measurable only after sep- aration from calpastatin (33). The proteolytic activities eluted at the NaCl concentrations of 0.15 and 0.4 M were identified as -calpain and m-calpain, respectively (panel A), as indicated by Western blot analysis using human isozyme-specific antibodies (panel C). GH 4 C 1 cells seem to have comparableand mcalpain activities. These data, in turn, confirm that the antibodies employed are specific enough to distinguish between rat and m-calpain.
The activity to inhibit calpain (panel B) and the anti-calpastatin antibody immunoreactivity showing 100 -110-kDa bands on Western blotting (panel C) eluted identically, i.e. at 0.25 M NaCl, confirming the specificity of the antibody. Notably, the calpastatin activity present in the cells exceeded the sum of the calpain activities, suggesting that the protease activity is regulated rather stringently in these cells as previously reported in KB cells (22).
Mobilization of the Calpain-Calpastatin System Induced by TRH-To obtain clues as to which of the calpain isozymes is involved in the down-regulation of nPKC⑀, we examined the effect of TRH on the cellular localization ofand m-calpains. We employed a Ca 2ϩ /EGTA buffer (pCa of 7) in fractionating the cytoplasm and membranes because the interaction of calpain with cell membranes depends on the Ca 2ϩ concentration (34). As shown in Fig. 5, the amount of m-calpain in the membrane fraction increased within 1 h after stimulation, whereas the amount of -calpain remained essentially unchanged. This indicates that only m-calpain translocates from the cytosol to membranes in response to TRH treatment.
TRH also induced alterations in the calpastatin levels of GH 4 C 1 cells. The levels of both the cytoplasmic and membraneassociated calpastatin molecules increased gradually within 1 h (Fig. 6), and the increased levels were sustained for more than 12 h (data not shown). The amount of calpastatin increased by about 1.7-fold 1 h after stimulation as quantified by densitometry. In particular, the appearance of the high molecular form, indicated in Fig. 6 by the open triangle, was notable (see "Discussion").

DISCUSSION
In the present study, we have shown that calpain inhibitors with broad and narrow specificities, ALLNal and CS peptide, inhibit the TRH-induced down-regulation of nPKC⑀ in GH 4 C 1 cells and that TRH causes both the translocation of m-calpain and an increase in the calpastatin level. These observations indicate that the calpain-calpastatin system is a component in TRH-induced signal transduction and that calpain is the major protease involved in down-regulation. This is the first demonstration of calpain involvement in PKC down-regulation associated with physiological stimulus-response coupling in contrast to previous studies employing phorbol ester or calcium ionophore to induce down-regulation (10,11).
The present results, however, do not exclude the possible involvement of other intracellular proteases in down-regulation (12)(13)(14). Because the effect of the CS peptide was only partial (Fig. 3), a secondary protease other than calpain may participate in the proteolytic process; a candidate is proteasome, which was recently shown to be inhibited by ALLNal (31,35). The presence of a secondary protease is also implied by the fact that we do not observe an active nPKC⑀ fragment corresponding to the catalytic fragment ( Fig. 1) produced in vitro by calpain-catalyzed limited proteolysis (30). Apparently, degradation of the active fragment proceeds faster than its production, as shown previously in the case of cPKC␣ (36). Our observations also indicate that nPKC⑀, which is enzymatically independent of calcium (30), could be regulated by the calcium signaling cascade through the proteolytic action of calpain. Whether or not proteolysis of nPKC⑀ bears any specific physiological signals in these cells remains to be elucidated.
The data showing the translocation of m-calpain, but not -calpain, from the cytosol to membranes in response to TRH treatment (Fig. 5) and suggesting the involvement of m-calpain in nPKC⑀ down-regulation were rather unexpected because -calpain, which requires lower calcium concentrations for activation in vitro, has been considered as the more probable candidate for intracellular proteolytic phenomena (37). However, the present observation agrees with our previous reports, demonstrating that phorbol ester induces the synthesis and translocation of m-calpain in COS and K562 cells (38,39). Possibly, the mobilization of m-calpain is controlled by the PKC pathway. It is therefore likely that in some cases, m-calpain is more involved in cellular stimulus-response coupling than -calpain. There may exist an unknown cellular factor associated with membranes that contributes to the specific activation of m-calpain.
A possible reason for the selective down-regulation of nPKC⑀ among other PKC isozymes (see the Introduction) may lie in the difference in the time course of translocation to membrane. In contrast to cPKC␣ and ␤II, which are translocated transiently and are rapidly dissociated from the membrane within 1 min (data not shown), the majority of nPKC⑀ remains associated with the membrane in a more sustained manner favoring colocalization with m-calpain. Furthermore, this membrane association indicates that PKC exists under activating conditions, in which the kinase is more susceptible to proteolytic attack (15,40). Presumably, other PKC isozymes escape from calpain action due to their relatively short periods of association with membranes.
The up-regulation of calpastatin isoforms, particularly of the high molecular weight form in the membrane fraction, that is induced by TRH treatment (Fig. 6) seems to be part of the cellular negative feedback system counteracting activated calpain and protecting cellular components. Because the excessive and random degradation of proteins associated with membranes and the cytoskeleton would be highly toxic to cells and because the too rapid proteolysis of calpain substrates including nPKC⑀ might interfere with cellular functions, the strict regulation of calpain activity by calpastatin should be vitally important. The three isoforms with different relative molecular weights probably represent different phosphorylation levels (41) or differential alternative splicing (42); each isoform may possess distinct properties affecting cellular localization and interaction with calpain. Each might provide specific protection to a group of proteins that would otherwise be proteolyzed by calpain. Both qualitative and quantitative changes in calpastatin expression may play important roles, as shown for differentiation and secretion in other cells (43)(44)(45). The precise role of calpastatin in PRL secretion from GH 4 C 1 cells remains to be elucidated. In conclusion, we propose that physiologically stimulated nPKC⑀ is strictly regulated at the cell membrane by the calpain-calpastatin system.