Involvement of a calpain-like protease in the processing of the murine interleukin 1 alpha precursor.

Interleukin (IL) 1 alpha is synthesized as a 33-kDa precursor that is enzymatically cleaved to the 15-17-kDa forms that are found in the culture supernatants of activated macrophages. We have explored the possibility that calcium might enhance IL-1 processing and secretion via the stimulation of a calcium-dependent protease. We have found that lysates prepared from human peripheral blood monocytes, the human histiocytic lymphoma cell line U937, and the murine macrophage cell line P388D1 contain a calcium-dependent IL-1 alpha processing activity that cleaves the IL-1 alpha precursor to its mature form. Although NIH 3T3 mouse fibroblast cell lysates also contain IL-1 processing activity, lysates from the murine thymoma EL-4, the human epidermoid cell line HEp-2, and the human foreskin fibroblast line FS-4 lack this activity. IL-1 processing activity is inhibited by leupeptin and exhibits a molecular mass of 80-110 kDa. The processing activity is also inhibited by a monoclonal antibody directed against calpain type I. These results indicate that the processing of the IL-1 alpha precursor is mediated, at least in part, by a member of the calpain family of proteases. Mixing experiments revealed that lysates from EL-4 or HEp-2 cells contain an inhibitor(s) of the calpain-like protease in macrophage extracts. It is, therefore, likely that many non-macrophage cell types are unable to process the IL-1 alpha precursor because the calpain present in these cells is only weakly active due to the presence of a specific inhibitor(s) such as calpastatin.


Involvement of a Calpain-like Protease in the Processing of the Murine Interleukin la Precursor*
(Received for publication, December 4, 1990) Lucy M. Carruth, Stephen DemczukS, and Steven B. Mizels Interleukin (IL) l a is synthesized as a 33-kDa precursor that is enzymatically cleaved to the 15-17-kDa forms that are found in the culture supernatants of activated macrophages. We have explored the possibility that calcium might enhance IL-1 processing and secretion via the stimulation of a calcium-dependent protease. We have found that lysates prepared from human peripheral blood monocytes, the human histiocytic lymphoma cell line U937, and the murine macrophage cell line P388D1 contain a calcium-dependent IL-la processing activity that cleaves the IL-la precursor to its mature form. Although NIH 3T3 mouse fibroblast cell lysates also contain IL-1 processing activity, lysates from the murine thymoma EL-4, the human epidermoid cell line HEp-2, and the human foreskin fibroblast line FS-4 lack this activity. IL-1 processing activity is inhibited by leupeptin and exhibits a molecular mass of 80-1 10 kDa. The processing activity is also inhibited by a monoclonal antibody directed against calpain type I. These results indicate that the processing of the IL-la precursor is mediated, a t least in part, by a member of the calpain family of proteases. Mixing experiments revealed that lysates from EL-4 or HEp-2 cells contain an inhibitor(s) of the calpain-like protease in macrophage extracts. It is, therefore, likely that many non-macrophage cell types are unable to process the IL-la precursor because the calpain present in these cells is only weakly active due to the presence of a specific inhibitor(s) such as calpastatin.
Interleukin (IL)' 1 is a peptide hormone produced by activated macrophages which stimulates a broad spectrum of immune and inflammatory responses including antigen and mitogen-dependent lymphocyte activation (1,2), fever (3), and the synthesis of several acute-phase response proteins (4). Cloning studies revealed the existence of two forms of human and murine IL-1, termed IL-la and IL-1P, that possess * This work was supported by National Institutes of Health Grant AI25836. 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. similar molecular masses but differ in isoelectric point and amino acid sequence (5)(6)(7)(8).
Previous studies have shown that both forms of human and murine IL-1 are synthesized as 31,000-33,000-Da precursor proteins that are enzymatically cleaved to the lower molecular mass forms found in the culture supernatant of activated macrophages (7,9). Sequence analysis of the IL-la and IL-l@ precursors revealed that they do not contain a conventional N-terminal hydrophobic signal sequence (7,9). Furthermore, the presence of an internal signal sequence appears unlikely since the murine IL-la precursor (pIL-la) synthesized in an in vitro reticulocyte lysate system in the presence of dog pancreas microsomes does not translocate into the lumen of the microsomes (9). In addition, immunocytochemical studies on stimulated human monocytes demonstrated that newly synthesized pIL-lP is not localized in the Golgi apparatus but rather in the cytoplasm of these cells (10,ll). Taken together, these results are consistent with the notion that IL-la and IL-lP might be processed and secreted from cells by a novel pathway that bypasses the Golgi apparatus. In a recent study we found that the processing and secretion of human and murine macrophage-derived IL-la and IL-16 are dramatically enhanced by the addition of the calcium ionophores, ionomycin and A23187 (12). This finding raised the possibility that one or more calcium-dependent proteases might be involved in the secretory pathway(s) for IL-1. In the current study we have extended our observations on the effects of calcium on IL-la processing and secretion and have characterized the involvement of a calpain-like protease in the maturation of murine pIL-la.

MATERIALS AND METHODS
Cells-The murine macrophage cell line P388D1, the murine thymoma cell line EL-4, and the human histiocytic lymphoma cell line U937 were grown in suspension culture at 37 "C in RPMI 1640 (GIBCO) supplemented with 10% fetal bovine serum (KC Biologicals) and 50 +g/ml gentamicin sulfate (U. S. Biochemical Corp). The human foreskin fibroblast cell line FS-4, NIH 3T3 murine fibroblasts, and the human epidermoid carcinoma line HEp-2 were grown as monolayers in RPMI 1640 supplemented with 10% fetal bovine serum and gentamicin sulfate. Cells were routinely tested for mycoplasma contamination and were found to be negative.
Preparation of Radiolabeled Murine pIL-la-The plasmid pI270mIL-1 containing a 1.1-kilobase cDNA for murine pIL-la was kindly provided by Dr. Peter Lomedico (Hoffmann-LaRoche, Nutley, NJ). The 1.1-kilobase pIL-la cDNA was excised and inserted into the HincII site of pGEM-3z (Promega) and amplified in Escherichia coli using standard methods (13). Purified plasmid was linearized with Hind111 or EcoRI and then transcribed using SP6 or T7 polymerase in an in vitro transcription system (Promega). The mRNAs were processed according to the manufacturer's instructions. In vitro translations were performed using micrococcal nuclease-treated rabbit reticulocyte extracts (Promega Biotec) supplemented with in vitro synthesized mRNA and 25 +Ci/ml ["S]methionine.
Preparation of Cell-free Extracts-Cell-free extracts were prepared by hypotonic lysis of cells. Adherent cells were washed twice with phosphate-buffered saline (10 mM phosphate, pH 7.4, 150 mM NaCI) and once with lysis buffer (20 mM HEPES, pH 7.4, 10 mM KC1, and 1.5 mM MgC1,) prior to scraping with a rubber policeman. Adherent and nonadherent cells were harvested by centrifugation at 500 X g and then swelled in lysis buffer for 20 min on ice. Lysis of cell suspensions was accomplished by 35 strokes in a Dounce homogenizer. Lysates were then spun in a Beckman table-top ultracentrifuge at 100,000 X g in a TLA-45 rotor for 15 min. The supernatants (termed lysates) were frozen a t -70 "C until used in cleavage assays. I n Vitro Cleauage Assay for pIL-la-pIL-la cleavage assays were performed by mixing 1 pl of a 1:4 dilution of rabbit reticulocyte extract containing [:'"S]methionine-labeled pIL-la with 50 or 100 pl of cell lysate in the presence or absence of 1 mM CaCL Reactions were incubated 37 'C for 1 h. In some experiments, cell lysates from different cell types were mixed with macrophage-derived lysate to determine if the absence of processing activity in some cell types was due to the presence of an inhibitor(s) of processing activity. Specifically, U937 lysate (20 pl) was incubated with varying amounts of HEp-2 or EL-4 cell lysates (0-80 pl), and the cleavage of pIL-la was assessed. The protein concentration of the mixtures was held constant by the addition of bovine serum albumin. As a control in these experiments we evaluated the effect of the nonprocessing HEp-2 lysate (see "Results") on trypsin-catalyzed pIL-la cleavage. pIL-ln was added to each mixture and incubated as described above.
The effect of calcium-activated neutral protease (CANP) on pIL-ICY processing was assessed by incubating pIL-la in lysis buffer containing 1 unit/ml CANP in the presence or absence of 1 mM CaC12. In some experiments, a protease inhibitor (specific for CANP) was present a t a concentration of 5 units/ml. The calpain inhibitor trans-epoxysuccinyl-~-leucylamido-(4-guanidino)butane (E-64) was used a t a concentration of 50 pglml. A murine ascites containing a monoclonal antibody against human type I calpain (the gift of Dr. J. Eke, Queen's University) was used at a dilution of 1500. This antibody was capable of blocking 1 unit/ml CANP activity. In related experiments using trypsin the antibody was ineffective in blocking cleavage of pIL-la. After incubation, all samples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as described previously (9). For quantitative analysis (percent conversion), autoradiographs were scanned using an LKB Brommo Ultroscan XL laser densitometer (LKB).
Gel Filtration Chromatography of the pIL-la Processing Activity in Lysate of U937 Cells-U937 cell lysate (300 pl, 1 X 10" cells/ml of lysate) was loaded onto a 20-ml Sephadex G-150 column equilibrated in lysis buffer and chromatographed a t a flow rate of 4 ml/h. Fractions (0.25 ml) were collected and individually assayed for cleavage activity by incubation with pIL-la as described above.
Western Blot Analysis of Cell Lysates-Lysate and column fractions were assayed for the presence of calpain using Western blots. Briefly, 10 p1 of cell lysate or column fractions was subjected to SDS-PAGE on 7.5% gels, and the proteins were blotted to nitrocellulose membranes (Schleicher & Schuell) using a Mini Trans-blot Electrophoretic Transfer Cell (Bio-Rad) (100 V, 15 min). The membranes were washed with Tris-buffered saline (TBS) (10 mM Tris-HCI, pH 8.0, 150 mM NaCI) containing 5% nonfat dry milk to block nonspecific binding and then washed with TBS prior to incubation with a mouse anti-human monoclonal against human type I calpain. Following overnight incubation a t 4 "C, the membranes were washed with TBS containing 0.05% Tween-20 and incubated for 2 h with a 1:lOOO dilution of a rabbit anti-mouse IgG alkaline phosphatase-conjugated antibody. The membranes were washed and incubated with nitro blue tetrazolium and 5-bromo-4-chloro 3-indoyl phosphate (Promega) in alkaline phosphatase buffer (0.1 M Tris, pH 9.5, 0.1 M NaC1, 5 mM MgCl,). The reaction was stopped by the addition of TE (10 mM Tris-HCI, pH 8.0, and 1 mM EDTA).
Site-directed Mutagenesis of pIL-la-The pIL-la cDNA was excised from pI270mIL-1 with BamHI and then ligated into the HincII site of the &galactosidase gene in the M13 cloning vector M13mp19. Point mutations were made a t desired sites using M13 site-directed mutagenesis as described by Kunkel et al. (14). Oligonucleotides (18mers) encoding specific amino acid substitutions were synthesized on an Applied Biosystems DNA according to established protocols. Mutations were confirmed by dideoxy sequencing (15) using the Sequenase Version 2.0 kit (U. S. Biochemical Corp.).
Radiosequencing of the Cleauage Product of plL-In-The ["'S] methionine-labeled IL-1 protein was subjected to SDS-PAGE and transferred electrophoretically to polyvinylidene difluoride membrane (Immobilon, Millipore) (16). Following transfer, bands were excised from the membrane using an autoradiograph of the membrane as a guide. Amino acid sequence analysis was performed using an Applied Biosystems model 475 protein sequenator. Amino acid derivatives released following each cycle of Edman degradation were analyzed for radioactivity to determine the cycle in which [""Slmethionine was released.

Calcium-dependent in Vitro
Processing of pIL-la-In a recent study we found that calcium ionophores such as ionomycin and A23187 dramatically enhanced the processing and secretion of newly synthesized IL-la from human and murine macrophages (12). This observation prompted us to suggest that a calcium-dependent protease might be involved in the processing and secretion of IL-la. To test this possibility formally we sought to establish an in uitro system for the processing of pIL-la to its mature form (15)(16)(17). Radiolabeled murine pIL-la was prepared by in uitro translation of pIL-la mRNA in the presence of ["'S]methionine and used a source of substrate. When radiolabeled pIL-la was added to lysates prepared from the murine macrophage cell line P388D1 and incubated in the absence of calcium, no detectable cleavage of the pIL-la was observed (Fig. 1, lune I).
However, if calcium was added to the cleavage reaction (lune Z), the pIL-la was completely converted to its mature form. Calcium-dependent processing of pIL-la was inhibited in the presence of 5 mM EGTA (lune 4 ) . Cleavage activity was also inhibited by the serine protease inhibitor, leupeptin (lune 5 ) , but not by aprotinin, phenylmethylsulfonyl fluoride, benzamidine, or a-macroglobulin (lunes [6][7][8][9]. Trypsin (lune 3 ) , in the absence of calcium, cleaved the pIL-la to a form that comigrated with the cleavage product produced by incubation of pIL-la with P388D1 cell lysate.
In addition to P388D1 cells, the human histiocytic lymphoma cell line U937, human peripheral blood monocytes, and NIH 3T3 mouse fibroblasts contained pIL-la processing activity (Fig. 2). However, FS-4 cells, a human foreskin fibroblast line that can synthesize but not secrete IL-1 (12, 17), did not contain detectable pIL-la processing activity (Fig. 2,  lune 5 ) . Lysates prepared from the murine thymoma cell line EL-4 and the human epidermoid carcinoma cell line HEp-2 also lacked pIL-la processing activity as measured in the presence or absence of calcium (Fig. 3). These results indicate that although the ability to process pIL-la is not macrophagespecific, it does not appear to be a universal property of all mammalian cells.   Characterization of the Protease(s) Responsible for the in Vitro Processing of pIL-la-Having established that the pILla processing activity is calcium dependent and leupeptin sensitive, we used Sephadex gel filtration chromatography to establish a molecular mass for the protease(s). A concentrated lysate prepared from U937 cells was chromatographed on a Sephadex G-150 column, and the fractions were assayed for cleavage activity against pIL-la. As shown in Fig. 4, the cleavage activity eluted with molecules with molecular masses in the range of 80-110 kDa.
T o gain additional insights into the properties of the calcium-dependent protease(s), we sought to define the site(s) at which it cleaves the pIL-la. Site-directed mutagenesis was used to create mutations in the pIL-la protein at selected sites, and the mutant proteins were then tested for susceptibility to calcium-dependent cleavage. Previous studies on the murine IL-la purified from the culture supernatant of stimulated P388D1 cells demonstrated that the mature protein exhibits microheterogeneity (18). Six species of IL-la were detected which exhibited slight differences in size and isoelectric point. Sequence analysis of the major form of mature IL-ICY revealed that serine 115 of the precursor served as the Nterminal amino acid. Since residue 114 in the pIL-la is an arginine, we decided to generate a mutant pIL-la lacking an arginine at this position and test if it was susceptible to cleavage by the calcium-dependent protease(s) found in macrophage lysates. Site-directed mutagenesis was used to convert Arg"' to an isoleucine in the pIL-la cDNA. In a number of instances, endoproteases involved in hormone processing have been found to cleave a t di-, and occasionally tetrabasic residues, the cleavage occurring on the carboxyl side of these basic residues (19)(20)(21)(22). Since the pIL-la contains a tetrabasic region (Lys-Lys-Arg-Arg) at residues 85-89, the suggestion was made that this region might be a primary processing site (5). Thus, a second set of mutant pIL-la proteins was produced in which LysAG was converted to a threonine, and A r p was changed to an leucine. Finally, a "triple" mutant was created with all three substitutions described above (ThrRG, Leu%, and Ile"'). The mutant cDNAs were transcribed in vitro, and the resultant mutant pIL-la mRNAs were translated in a rabbit reticulocyte lysate in the presence of [:%] methionine. As shown in Fig. 5, all five mutant pIL-la proteins were cleaved to the same molecular mass as the wildtype pIL-la in the presence of P388D1 lysate. As expected, calcium was required for the cleavage of the mutant proteins; only background levels of cleavage occurred in the absence of calcium (data not shown). Thus, neither the tetrabasic region a t residues 85-89 nor Arg"" is required for processing of pILla in this in vitro system. Radiosequencing analysis of the mature form of the wild-type pIL-la generated by in uitro processing revealed that the amino-terminal sequence was Gln-Ser-Asp, thus establishing the preferred site of cleavage between Tyr-120 and Gln-121.
Inuoluement of a Calpain-like Protease inpIL-la Processing in Vitro-In view of the fact that members of the calpain family of proteases are calcium dependent, leupeptin sensitive, and possess a molecular mass in the range of 80-110 kDa (23), we examined the possibility that the pIL-la processing activity in macrophage lysates was a member of this family of proteases. The 80-110-kDa range of the pIL-la processing activity (Fig. 4) corresponds to the molecular mass of the large subunit (80 kDa) and heterodimeric (110 kDa) forms of calpain. As a first step, we used a monoclonal antibody against the 80-kDa subunit of human calpain type I in Western blots to determine if this enzyme was present in the Sephadex G-150 column fractions that contained pIL-la cleavage activity. We did indeed find that those fractions with cleavage activity were positive for calpain (data not shown). Furthermore, the fractions possessing the highest level of cleavage activity (Fig.  4) also contained the highest levels of calpain. It is important to note that in SDS gels, the molecular mass of the calpain in each column fraction was 80 kDa. This observation may provide an explanation for the relatively broad peak of pILla processing activity. During chromatography, a portion of the 80-kDa large subunit of calpain may have dissociated from the 30-kDa subunit, giving rise to the observed rather broad elution profile.
Additional evidence for the involvement of a calpain-like protease in the in uitro processing of pIL-la was obtained in two types of experiments in which the anti-calpain monoclonal antibody was used to inhibit calpain activity present in the lysates of U937 cells, and calpain itself was tested for pIL-la processing activity. As presented in Table I, the anticalpain monoclonal antibody significantly diminished the cleavage activity present in U937 lysates. Furthermore, a protease inhibitor (specific for CANP) also reduced the conversion of pIL-la to its mature form. E-64, another specific calpain inhibitor, also reduced the conversion of pIL-la by U937 lysate. When purified CANP was added to pIL-la in the absence of any cell lysate, a mature form corresponding in size to cell lysate-generated mature IL-la was obtained. The activity of CANP was blocked by the addition of the anticalpain monoclonal antibody (data not shown). Thus, on the basis of its requirement for calcium, leupeptin and E-64 sensitivity, and inhibition by a calpain inhibitor as well as an anti-calpain monoclonal antibody, it is quite likely that the pIL-la processing activity in macrophage lysates is mediated by a member of the calpain family of proteases. Radiosequence analysis of mature IL-1 generated by incubation with CANP revealed two possible amino-terminal species of the molecule at Gln-Ser-Asp and Ser-Ala-Pro. The first sequence corresponds to the amino terminus generated by the U937 lysate cleavage reaction. The second sequence corresponds to a species of mature IL-1 seen in the culture supernatant of activated macrophages (9,12,18). The results of this sequence analysis are consistent with a calpain-like enzyme being involved in the processing of IL-1.
Although our findings are consistent with the notion that a calpain protease is responsible for the processing of pIL-la, we were confronted by the initially puzzling problem of why a cell such as FS-4 that possesses calpain (as determined by Western blot analysis; data not shown) is unable to process pIL-la in uitro. It has recently been shown that the calpain activity in a particular cell may be tightly regulated by the natural inhibitor of calpain, calpastatin (24, 25). Thus, it was possible that cell lines such as FS-4, EL-4, and HEp-2 might contain a higher level of a calpain inhibitor than P388D1, U937, human monocytes, or NIH 3T3 cells and thus would have insufficient levels of active calpain to process pIL-la in uitro. To examine this possibility, mixing experiments were performed to test the effect of HEp-2 and EL-4 cell lysates on the processing activity present in lysates from U937 cells. When increasing amounts of either Hep-2 or EL-4 lysate (both nonprocessors) were mixed with a constant amount of U937 lysate, a concentration-dependent inhibition of pIL-la cleavage was observed (Table 11). A dilution of the lysate with lysis buffer alone had no effect. Furthermore, the processing of pIL-la by CANP was also blocked, in a concentrationdependent manner, by HEp-2 lysates. HEp-2 lysate had no effect on trypsin-mediated cleavage of pIL-la (Table 11). Thus, it is quite likely that the ability to process pIL-la is dependent upon a favorable balance between the levels of calpain and one or more specific inhibitors of calpain activity.

DISCUSSION
The results from several recent studies indicate that IL-1 may not follow a conventional secretory pathway that involves cotranslational processing and transit through the Golgi apparatus (10)(11)(12)26). The mature forms of IL-1 (15)(16)(17) are not detectable in lysates from stimulated, IL-1-secreting macrophages, suggesting that processing is a coor postsecretory event. However, it is possible that IL-1 may be processed intracellularly and secreted in such a rapid and efficient manner that little, if any, mature IL-1 remains in the cell. Although most of the IL-1 secreted by macrophages is in the mature form, the presence of intact IL-1 precursor in the culture medium of activated macrophages indicates that processing may not be an obligatory event in the release of at least a portion of the newly synthesized IL-1 (12). The results of our earlier studies on the effect of calcium ionophores on IL-la secretion from murine and human macrophages established a role for calcium in the processing and secretion of this cytokine (12). In the current study, we have obtained evidence for the involvement of a member of the calcium-dependent calpain family of proteases in the processing of murine pIL-la by murine and human macrophages. The pIL-la processing activity was shown to be calcium dependent, E-64 and leupeptin sensitive, and inhibited by a monoclonal antibody directed against type I calpain. In addition, the U937 cell-derived cleavage activity eluted from a Sephadex G-150 column in a broad peak with molecules possessing a molecular mass of 80-110 kDa. This type of chromatographic behavior is consistent with the 110-kDa heterodimeric and 80-kDa catalytic subunit forms of calpain. The level of cleavage activity in each column fraction was proportional to the level of calpain (as detected in Western blots). Furthermore, radiosequencing data of both U937-and CANP-generated mature IL-1 revealed an identical aminoterminal species. Taken together, these observations indicate that calpain is responsible for the in uitro cleavage activity in the macrophage lysates and raises the possibility that in uiuo, calpain may play a major role in the processing of pIL-la in these cells.
We also found pIL-la processing activity in lysates of NIH 3T3 mouse fibroblasts, a cell line that produces and secretes IL-1 following stimulation with tumor necrosis factor a and LPS (27). In contrast, lysates prepared from EL-4, HEp-2, and FS-4 cells did not contain pIL-la processing activity, demonstrating that the ability to process pIL-la is not a property of all mammalian cell types. It is important to note that although FS-4 cells synthesize pIL-1, they are unable to process and secrete the protein, even in the presence of calcium ionophores (12). Thus, the absence of pIL-la processing activity in FS-4 lysates is consistent with the results of our earlier studies on IL-1 synthesis in this cell line. The results of mixing experiments support the view that although FS-4, EL-4, and HEp-2 cells contain calpain, the protease is prevented from processingpIL-la by the presence of a specific inhibitor, perhaps calpastatin (24, 25). The nonprocessing HEp-2 lysates inhibited the processing activity of purified CANP and U937 cell lysates, but not trypsin, thus demonstrating the calpain specificity of the inhibitory activity.
During the course of our studies, Kobayashi et al. (28) reported that the human IL-la precursor could be converted to its mature form by a protease in macrophage lysates or purified CANP. Based on the characteristics of the lysate activity, i.e. calcium dependence and sensitivity to leupeptin and E-64, Kobayashi and colleagues also attributed the cleavage activity in the macrophage lysates to a calpain-like protease. These investigators determined that a single cleavage between Phe"" and Leu11g results in the generation of the mature form of human IL-la. Although human and murine IL-la exhibit substantial amino acid sequence homology (67%; Ref. 29), murine pIL-la does not contain a Phe-Leu sequence in a region in which cleavage would generate a 17-kDa species. Our results indicate that in uitro, murine pIL-la is cleaved between T y P " a n d Gln'". Although the substrate specificity for the calpains is not well-defined, it has been reported that the P2 position of the subsite specificity outlined for calpain frequently contains hydrophobic or bulky residues. There is a proline (Pro"") in the P2 position of one of the cleavage sites of pIL-1 (between Argil4 and Ser1I5) generated by U937 lysate and CANP. Another substrate of calpain, acrystallin, contains a threonine in the P2 position (24). A threonine (Thr"') residue is present at the P2 position of the alternate site of CANP-generated mature pIL-1 (between TyrI2' and Gin'"). These sequences are consistent with the observed trend of a bulky residue present at the P2 position in calpain-mediated cleavage.
Although our results as well as those of Kobayashi et al. (28) demonstrate that a calpain-like protease may be involved in the processing of pIL-la, two important questions regarding IL-1 secretion remain to be answered. Where in the cell does calpain process IL-la? Is the cleavage reaction directly linked to the actual release of IL-la from the cell? With regard to the first question, there is evidence to suggest that calpain can translocate to the inner surface of the plasma membrane where it may make contact with substrates such as protein kinase C (30,31). Does IL-1 also translocate to the plasma membrane, perhaps in response to changes in the intracellular level of calcium?
Regarding the possible linkage between processing and secretion, the results of Krasney et al. (32) and Douvdenani et al. (27) indicate that processing is not, by itself, a sufficient stimulus for the secretion of IL-1. Although 3T3 cells possess pIL-la processing activity (Fig. 2), they are apparently still unable to secrete IL-1 (27). In addition, Krasney et al. found that the transfection of an expression plasmid containing a cDNA encoding the mature form of human IL-1 into a nonmacrophage cell type did not result in the secretion of mature IL-1 (32). The mature protein simply accumulated in the cytosol. Based on this finding as well as the accumulated evidence on IL-1 processing and secretion from other studies (26,29,(33)(34)(35), two fundamental mechanisms of secretion may be considered. First, secretion may be directly linked to processing in that the site of processing is physically associated with the secretory mechanism in such a way that only those IL-1 molecules that are processed at this site can utilize the secretory pathway. Thus, a mature form of IL-1 synthesized from information encoded in a truncated cDNA would not have access to the secretory pathway because it is unable to undergo processing. Alternatively, secretion is not directly linked to processing, but a carrier molecule(s) is required to transport the mature form of IL-1 out of cells. Many nonmacrophage cell types may lack this transport molecule(s) and thus are unable to release IL-1, even if the protein is already in a mature form. The results of future experiments should allow us to determine if either of these possibilities is operative in IL-1-secreting cells.