Phosphorylation and subcellular distribution of calpastatin in human hematopoietic system cells.

Calpastatin is an endogenous inhibitor protein acting specifically on calpain (EC 3.4.22.17; Ca2(+)-dependent cysteine proteinase). The phosphorylation of calpastatin was investigated in human hematopoietic system cell lines. Microheterogeneity of calpastatin was observed, in which 118- and 116-kDa forms were named calpastatin a and b, respectively. The phosphorylation of both calpastatins was identified in all cell lines examined and occurred mainly at serine residues with trace amounts of phosphothreonine in vivo. The incubation of cells with 12-O-tetradecanoylphorbol-13-acetate increased the incorporation of 32P-orthophosphate into calpastatin a. Two-dimensional maps of 32P-labeled phosphopeptide from both calpastatins were identical except for additional minor spots for calpastatin a. [35S]methionine-labeled calpastatins a and b were localized mainly in the cytosol, and only 6% of cellular calpastatins were detected in the membrane fraction. By contrast, more than 30% of the 32P-labeled calpastatins a and b were distributed in the membrane fraction. Thus, the phosphorylation of calpastatin may be involved in regulating the calpain-calpastatin protein kinase system by its subcellular distribution.

and calpastatin are known to be distributed widely in various tissues and cells of different animal species, though unevenly in abundance (7)(8)(9)(10)(11). It has been suggested that calpain plays important roles in various cellular functions by limited proteolysis coupled with transient Ca2+ mobilization (1)(2)(3).
The biochemical and catalytic properties of calpains I and I1 and calpastatin have been extensively studied in many laboratories including ours (1)(2)(3)(4)(5)(6)12), and the nucleotide sequences of cDNAs coding for calpains I and I1 and calpastatin in several animal species have recently been determined (13)(14)(15)(16)(17)(18). Despite the rapid progress in molecular research on calpains and calpastatin, their physiological roles and regulatory mechanisms in cells remain obscure.
As activation or inactivation of endogeneous substrates by * This work was supported in part by grants from Uehara Memorial Foundation and the Yamada Science Foundation and by Grants-in-Aid for Scientific and Cancer Research from the Ministry of Education, Science and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. !I To whom correspondence should be addressed.
calpain is an irreversible reaction caused by cleavage of the peptide bond, the activity of calpain is thought to be regulated very strictly. First, calpain activities are regulated by the intracellular Ca2+ concentrations and the quantities of calpains I and I1 and calpastatin (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11). Second, qualitative alterations such as limited autoproteolysis of calpain are also known to regulate its activity (19,20). Furthermore, it is possible that specific activities, affinities for the interaction, and intracellular localization of calpains I and I1 and calpastatin are dependent on posttranslational protein modification. In order to explore the physiological roles and regulatory mechanisms of the calpain-calpastatin system, it is important to understand their state of modification in vivo. Among various kinds of protein modifications, protein phosphorylation is an important reaction to the control of a number of diverse biological processes such as signal transduction, gene expression, and cell differentiation (21). Therefore, we have focused on protein phosphorylation of calpains I and I1 and calpastatin.
Our previous studies have shown that both calpains I and I1 are not phosphorylated in vivo and that they do not possess kinase activity (22). The present study was undertaken to examine in uiuo phosphorylation of calpastatin and to explore possible regulatory mechanisms for the calpain-calpastatin system. For this purpose, we isolated 32P-metabolically labeled calpastatins from human hematopoietic system cell lines by immunoprecipitation and performed various biochemical analyses.
Cellulose-coated thin layer glass plates were purchased from Merck and Enlightning from Du Pont-New England Nuclear. Other reagent grade chemicals were obtained from Wako Pure Chemicals or Nakarai Chemicals. neous leukemia (27). All cell lines were routinely maintained in RPMI 1640 medium supplemented with 10% fetal calf serum (heat-inactivated at 56 "C for 30 min) at 37 "C in humidified air with 5% CO,. Labeling of cells with ~-[:'"S]methionine or [:"P]orthophosphate was conducted by incubating cells (1 X loG cells/ml) in methionine-free or phosphate-free medium supplemented with 10% fetal calf serum dialyzed against Tris-buffered saline (20 mM Tris-HCI, 0.15 M NaCI, pH '7.5) at 3'7 "C for 8 or 2 h, respectively, with various stimulations. The concentration of isotope added to the medium was 7.4 MBq/ml for ~-[""S]methionine or 18.5 MRq/ml for [:"P]orthophosphate. At the end of the incubation period, cells were washed with phosphatebuffered saline (10 mM phosphate, 0.14 M NaCI, pH 7.5).
Antibody-Antiserum against purified porcine heart calpastatin (107 kDa) was raised in rabbit. An immunoglobulin G (IgG) fraction of antiserum was applied onto an affinity column of porcine heart calpastatin as previously reported (10). The bound IgG was eluted and then dialyzed against phosphate-buffered saline.
Phosphoamino Acid Analysis-Phosphoamino acid analysis was performed essentially as described by Hunter and Sefton (29) with a slight modification. "'P-Labeled calpastatins were eluted from acrylamide gel sections by TPCK-trypsin digestion (50 pg/ml) in 50 mM NH,HCO:,. Recovered phosphopeptides were hydrolyzed in 6 N HCI a t 105 "C for 2 h. Resulting phosphoamino acid samples were mixed with standards (P-serine, P-threonine, and P-tyrosine; 5 pg each) and applied onto cellulose-coated thin layer glass plates (20 X 20 cm). High voltage electrophoresis was performed in pH 3.5 buffer (pyridinemetic acid:water, 1:10:189 (v/v)) a t 4 "C. '"P-Labeled phosphoamino acids were identified by autoradiography, and phosphoamino acid standards were identified by spraying with ninhydrin.
Two-dimensional Peptide Mapping-Two-dimensional mapping of phosphopeptides from ."P-labeled calpastatins was performed as described by Akiyama et al. (30). Bands of phosphorylated calpastatins were excised from gels and digested in 1 ml of 50 mM NH,HCOs with 50 pg of TPCK-treated trypsin at 37 "C for 18 h. After lyophilization, tryptic phosphopeptides were applied to cellulose plates (20 X 20 cm) for electrophoresis in 1% NH,HCOa, pH 8.9, a t 400 V for 2 h, followed by chromatography in the second dimension for 4 h in butanol-1:pyridine:acetic acid:wat.er (15:103:12, v/v). ?'P-Labeled peptides were identified by autoradiography. Subcelhlar Fractionation-Subcellular fractionation was performed according to the method of Courtneidge et a/. (31) with a slight modification. Labeled cells (4 X 10") and carrier cells (8 X 10') were mixed and disrupted by a tightly fitting Dounce homogenizer in homogenizing buffer (20 mM HEPES-NaOH containing 5 mM KC1 and 1 mM MgCI,, pH '7.1) until nuclei became microscopically free. Nuclei were pelleted by centrifugation a t 1,000 X g for 5 min and washed once with homogenizing buffer. Combined supernatants were ultracentrifuged a t 100,000 X g for 30 min. The supernatant and the pelleted debris were termed cytoplasm fraction and membrane fraction, respectively. Nuclei and membrane fractions were resuspended in homogenizing buffer. A portion of each fraction was removed for evaluation by subcellular fractionation. Each fraction was solubilized by the addition of 2 X radioimmune precipitation buffer and subjected to immunoprecipitation. All experimental operations were performed a t 4 " c or on ice. Evaluation of subcellular fractions was performed by lactate dehydrogenase assay (32), 5'-nucleotidase assay ( 3 3 ) , and DNA measurement (34).

I n Vivo Phosphorylation of Calpastatin in Human Hematopoietic System
Cells-To test for immunoprecipitation of human calpastatin using anti-porcine heart calpastatin IgG, human T-cell leukemia virus type I-infected human T-cell line HUT-102, which was found to produce a large amount of calpastatin (lo), was metabolically labeled with :"S-methionine and subjected to immunoprecipitation. As shown in Fig.   U, the antibody against porcine heart calpastatin was able to immunoprecipitate human authentic calpastatin and nascent or degraded calpastatin. In addition, microheterogeneity of authentic calpastatin was identified, in which the 118-and 116-kDa forms were named calpastatin a and b, respectively.
When HUT-102 cells were metabolically labeled with ['"PI orthophosphate for 2 h, radioactive calpastatins a and b were immunoprecipitated (Fig. lB, lane 2). The radioactivity of calpastatin b was 4-fold higher than that of calpastatin a. However, when the cells were labeled with ["P]orthophosphate and pulsed with 100 ng/ml TPA for 15 min, the incorporation of :$'Pi into calpastatin a was preferentially augmented to the same level as calpastatin b (Fig. 1B, lane 4 ) . As the radioactivity associated with calpastatins a and b was sensitive for the alkaline phosphatase treatment (data not shown), the radioactivity was due to protein phosphorylation and not ADP-ribosylation. The relative amount of the radioactivity incorporated into calpastatins a and b was measured by densitometric scanning (Fig. IC) and is summarized in Table I.   3 and 4 ) or without (lunes 1 and 2) 100 ng/ml TI'A for 15 min. Cell lysates were subjected to immunoprecipitation with anticalpastatin IgG (lanes 2 and 4 ) or preimmune IgG (lunes 1 and 3).   The data were obtained from the experiment in Fig. 1. by densitometric scannine. The values are mean relative intensities. the presence or absence of TPA. The cell lysates were subjected to immunoprecipitation with anti-calpastatin IgG (Fig.  2). Although the apparent incorporations of "Pi into calpastatins showed a marked variation depending upon the cell lines, both calpastatins a and b in all cell lines tested were phosphorylated in vivo. The phosphorylated calpastatins a' and b' in K-562 cells were approximately 10 kDa smaller than the calpastatins a and b in the other cell lines as described in our previous report (10). Since human T-cell leukemia virus type I-infected T-cell lineage comprised a high content of calpastatin (lo), the high degrees of radioactivities associated with calpastatins in HUT-102 and MT-2 cells corresponded well with their cellular content of calpastatin. Furthermore, TPA treatment caused a rapid and preferential increase in phosphorylation of calpastatin a or a' in all cells examined.

93-
Effects of Drugs on Phosphorylation of Calpastatin-To examine which kinase is related to the phosphorylation of calpastatin, we studied the effects of various drugs on its phosphorylation. Phorbol ester, TPA, and calcium ionophore AL':lIRi are known to activate cellular protein kinase C. Forskolin and db-CAMP are activators of adenylate cyclase and protein kinase A, respectively. HUT-102 cells were pulsed with 100 ng/ml TPA, 10 pM A231Ri, 10 p M forskolin, or 0.5 mM db-CAMP for 15 min at the end of "P-labeling period, and "'P-labeled calpastatins were immunoprecipitated. As shown in Fig. 3, there were no effects on the phosphorylation of calpastatin b by any treatment. By contrast, the incorporation of ,c2Pi into calpastatin a was preferentially and rapidly augmented by TPA treatment (Fig. 3, lane 2) but not by any other drug-treatments (lanes 3-5). The effects of drugs on phosphorylation of calpastatins a and b were measured by densitometric scanning and are summarized in Table 11. The TPA-dependent increase of phosphorylation in calpastatin a was approximately 4-fold higher than the control. Phosphoamino Acid Analysis of Calpastatin-It is known

Effects of drugs on phosphorylation of calpains a and b
The data were obtained from the experiment in Fig. 3. by densitometric scanning. The values are mean relative intensities.  that calpastatin consists of a leader domain and four repetitive functional domains (16)(17)(18). One functional domain contains three consensus sequences. Among them, the central consensus sequence, GKREVTIPPKYR, is essential for inhibition of calpain activity (36)(37)(38). As this sequence is phosphorylatable at the threonine and/or tyrosine residues, the phosphorylation may directly regulate its inhibitory activity. "'P-Labeled calpastatins from drug-treated HUT-102 cells (shown in Fig. 3) were eluted from the gel and hydrolyzed in 6 N HCl a t 105 "C for 2 h. Phosphoamino acids were separated by electrophoresis and autoradiographed. As shown in Fig 4, in all drug treatments both calpastatins a and b were phosphorylated mainly at the serine residue with trace amounts of phosphothreonine in vivo. In calpastatins from other cells, the major and minor phosphoamino acids were also phosphoserine and phosphothreonine, respectively (data not shown). Phosphotyrosine was not detected in any cells examined.

Drugs
Phosphopeptide Mapping of Calpastatin-To examine the number of phosphorylation sites and the differences between calpastatin a and b, we analyzed phosphopeptide maps of both calpastatins. '"P-Labeled calpastatins were digested with TPCK-treated trypsin, and two-dimensional phosphopeptide maps were prepared by electrophoresis and chromatography (Fig. 5). In both phosphopeptide maps, a few major phosphopeptides with minor spots were observed. In the molecules, their main phosphorylation sites were limited to a few regions. Moreover, both phosphopeptide maps were identical except for five additional minor spots for calpastatin a. These data suggest that the mode of phosphorylation in calpastatins and b was almost identical. Subcellular Distribution of Calpastatin-We examined the subcellular distribution of phosphorylated calpastatin. Labeled HUT-102 cells were fractionated into membrane, cytoplasm, and nuclei according to the method of Courtneidge et al. (31) with a slight modification. Labeled calpastatins in each subcellular fraction were immunoprecipitated and subjected to SDS-PAGE followed by fluorography or autoradiography. The relative levels of labeled calpastatin in each subcellular fraction were measured by densitometric scanning. Evaluation of subcellular fractions is summarized in Table  111. Assays for 5'-nucleotidase (33), which provides a biochemical marker for the membrane, indicated that 9.6 and 11.2% of the total cellular activity were in the cytoplasmic and nuclear fractions, respectively. Lactate dehydrogenase (321, which is a marker for cytoplasmic contamination of the membrane and nuclear fractions, revealed that less than 7% of the total cellular activity was in the membrane and nuclear fractions. Moreover, the nuclei remained intact, as most of the DNA is retained in the nuclear fraction. The subcellular distribution of labeled calpastatin is shown in Fig. 6. Assuming that distribution of [:''SS]methioninelabeled calpastatin reflected that of the total calpastatin, the majority of calpastatins a and b was located in the cytosol. By contrast, more than 30% of the total '"P-labeled calpastatin a and b were located in the membrane. There was no difference in the behaviors of calpastatins a and b as to subcellular location (data not shown). These data suggest the possibilities that (i) the membrane-associated calpastatins were preferentially phosphorylated and that (ii) the phosphorylated calpastatins tended to associate with the membrane.

DISCUSSION
We have explored the possibility that phosphorylation of the calpain-calpastatin system could regulate the system according to the conversion of their activities and affinities for interactions. Our previous studies have shown that both calpains I and I1 were not phosphorylated in vivo and did not possess kinase activity themselves (22). The data described in this paper clarify that calpastatin is phosphorylated in vivo. The in vivo phosphorylation of calpastatins a and b was observed in all human hematopoietic system cell lines tested, HUT-102, MT-2, CCRF-CEM, Jurkat, and K-562. As TPAdependent augmentation of phosphorylation was observed, protein kinase C may be involved in a portion of calpastatin a phosphorylation. The major phosphoamino acids of phosphorylated calpastatin were phosphoserine with a trace amount of phosphothreonine. Tyrosine kinase was not involved as phosphotyrosine was not detected. This also negated the possibility of in vivo phosphorylation at the tyrosine and threonine residues in the central consensus sequence which is essential for proteinase inhibitor activity. However, there are another two consensus sequences (DDALDKLSDSLG and IDALSGDL) and several phosphorylatable serine residues besides the central consensus sequence (16)(17)(18). Phosphorylation may occur in these regions. As a few major '"Plabeled peptides with several minor spots were observed on tryptic peptide maps, the main phosphorylation sites were concentrated to a few regions of calpastatin molecules.
Although Mellgren and Carr (39) observed no significant changes in calpastatin activity by in vitro phosphorylation using protein kinase A, the phosphorylation of calpastatin by other kinase(s) might affect its inhibitory activity by alteration of affinity for protein interaction, intracellular localization, and/or protein conformation. We examined the subcellular distribution of labeled calpastatin. beled calpastatin was mainly localized in the cytosol, and only 6% of total calpastatin molecules was detected in the membrane fraction. This distribution of calpastatin corresponded well with our previous data (40). By contrast, more than 30% of the 32P-labeled calpastatin was observed in the membrane fraction. There are two possible explanations for these findings: (i) the membrane-associated calpastatins were preferentially phosphorylated, or (ii) the phosphorylated calpastatins tended to associate with the membrane. Those membrane-bound phosphorylated calpastatin molecules may associate with other membrane proteins, including calpains and their substrate proteins, whose affinity, stability, and intracellular localization may be altered.
Among endogeneous substrates of calpain (41), the limited proteolysis of protein kinase C (38, 42-46), protein kinase A (47), phosphorylase b kinase (48), and epidermal growth factor receptor (49) by calpain have been demonstrated in uitro and in uiuo. These protein kinases are thought to play a central role in the induction of cellular responses and signal transduction. In this paper, it is demonstrated that calpastatin is one of the endogeneous substrates of protein kinase(s) in uiuo. It is likely that there is an interaction among calpain, calpastatin, and protein kinase in cells. Although calpastatin can inhibit calpain activity, a portion of calpastatin is also phosphorylated by protein kinase, which is one of the endogeneous substrates of calpain. Thus, these three Ca2+-regulated enzymes and inhibitor interact and regulate each other.