A Pre-aspartate-specific Protease from Human Leukocytes That Cleaves Pro-interleukin- 18"

Interleukin-18 is a 17.4-kilodalton hormone derived from a 33-kilodalton inactive precursor produced by monocytes. We used the precursor as a substrate to detect proteolytic activities in peripheral blood mononuclear cell-conditioned medium that might be involved in interleukin-1/3 processing. We found that the conditioned medium, a biologically active frag- ment from the precursor that runs slightly higher than the mature hormone in sodium dodecyl sulfate-poly- acrylamide gel electrophoresis. The activity behaved as a single protein in ion exchange chro- matography. It was completely inhibited by metal ion chelators and not by inhibitors of serine, cysteine, or aspartate proteases, and it was dependent on both cal-cium (or magnesium) and zinc. The enzyme was not inhibited by three substrate-based metalloprotease inhibitors, phosphoramidon, benzyloxycarbonyl-Gly- Leu-NHa, and N-(2-carboxy-3-phenylpropionyl)-Leu. NHz-terminal sequence analysis that cleavage of the precursor occurred between a histidine and an aspartate residue, and digestion of synthetic peptides indicated that the protease is specific for pre-aspartate cleavages. a Microfuge, and the supernate was applied to a Vydac CIS HPLC column. Peptides were eluted with a gradient of acetonitrile, hydrolyzed, and analyzed for composition with an LKB a-amino acid analyzer.

terminus, another cleaves 3 amino acid residues upstream, and the third appears to cleave 13 residues upstream. HOWever, inhibitors of these activities failed to inhibit IL-18 production by PBMC cultures.
In the present work we therefore investigated human PBMC cultures themselves as a source of I1-1p processing activity. We concentrated on the conditioned medium from such cultures because the processing apparently does not occur intracellularly (8,9). We found a proteolytic activity that cleaves pro-IL-lp at the amino side of the aspartate residue preceding the mature NHz-terminal alanine. In this paper we describe the initial characterization of this activity.

EXPERIMENTAL PROCEDURES
Cell Culture and Preparation of Conditioned Medium-Human blood was obtained, and serum-free PBMC cultures were set up as described previously (IO), except that the medium used was RPMI 1640 buffered with 10 mM Hepes (pH 7.4). The cells were stimulated with 1% phytohemagglutinin (Gibco or Sigma) for 18 h. The conditioned medium was harvested as described previously (10) and concentrated 10-20-fold on an Amicon DC-10 L concentrator. The concentrated material was diluted 10-fold with 10 mM Tris-HC1 (pH 8.1) to reduce the molarity and then re-concentrated 10-fold. The final concentrate was then centrifuged at 4 "C for 30 min at 730 X g, and the supernate was passed through a 0.45-pm filter. For the experiment shown in Fig. 1, the material was then concentrated 20fold further with an Amicon YM-10 membrane.
Protease Assay-Five microliters of pro-IL-10 (10-50 pg/ml, prepared as in Ref. 3) were incubated with 10 pl of the material being tested for protease activity. The incubation was carried out at 37 "C for 1 h and was terminated by addition of 15 pl of 2 X SDS sample buffer (11) followed by boiling for 5 min. SDS-polyacrylamide gel electrophoresis (11) was carried out with 14% polyacrylamide gels, and Western blots were carried out as described previously utilizing the IL-10 COOH-terminal-specific monoclonal antibody 16F5 (3).
Mature IL-10 was prepared as described previously (lo), and approximately 20 ng was used for Western blots.
Chromatography-All chromatographic procedures were carried out at 4 "C. The ion exchange resins were pretreated with 0.1% Triton X-100 and 10% fetal calf serum to reduce nonspecific adsorption of protein. Fractions were assayed for protease activity as described above, for protein concentration with the Bio-Rad protein assay and ovalbumin as standard, and for salt concentration with a conductivity meter. PBMC conditioned medium was prepared as described above and applied at 50 ml/h to a 10 X 5-cm column of DEAE-Sephacel (Pharmacia LKB Biotechnology Inc.) which had been equilibrated in 10 mM Tris-HC1 buffer (pH 8.1). Material that flowed through the DEAE column was loaded at 30 ml/h onto a 30 X 1.6-cm column of blue agarose (Bethesda Research Laboratories) which had been equilibrated in 10 mM Tris-HC1 buffer (pH 8.1). The column was washed with 3 column volumes of the equilibration buffer, and material was then eluted with an increasing linear gradient of NaCl ranging from 0 to 800 mM in 10 mM Tris-HC1 buffer (pH 8.1) (3 column volumes). Thirty 5-ml fractions were collected.
Znhibitor Studies-Five microliters of a solution of the inhibitor, of the inhibitor plus divalent cation, or of an appropriate control solution (i.e. buffer or solvent without inhibitor) was added to 10 pl of protease, and the mixture was incubated for 10 min at 37 "C. Five microliters of pro-IL-10 was then added and the incubation was continued as described above. All inhibitors were obtained from Sigma except Z-Gly-Leu-NHn, which was obtained from Vega. The protease used in these experiments had been purified through the DEAE and blue agarose steps described above. It was then diluted 5-fold in 10 mM Tris-HCI (pH 8.1) to reduce the salt concentration, and the diluted material was concentrated 20-fold in an Amicon Centricon 10 concentrator.
NHz-terminal Sequencing-One milliliter of the material purified sequentially through DEAE and blue agarose (see above) was incu-

5323
Leukocyte Pre-aspartate Protease bated with 0.5 ml of pro-IL-lp (50 pg/ml) at 37 "C for 2 h. The sample was then dialyzed three times against 1 liter of Hz0 at 4 "C. After dialysis, the material was concentrated to dryness in a Speed-Vac and then dissolved in SDS sample buffer. Following SDS-polyacrylamide gel electrophoresis, proteins were transferred to a polyvinylidene difluoride membrane and stained with Coomassie Blue (12). The product of interest was cut out and analyzed for sequence in an Applied Biosystems model 470A protein sequencer. Preparation of Peptides-Peptides were synthesized with an Applied Biosystems model 430A peptide synthesizer. The NH1 termini of the peptides were modified by the addition of either an acetyl group or 8-alanine, and the COOH termini were amidated. After cleavage from the resin by HF, the peptides were applied to a Vydac CIS HPLC column (25 X 1.0 cm) and eluted with a gradient of acetonitrile in 0.1% trifluoroacetic acid.
Digestion of Peptides and Analysis of Products-Generally, 0.5 ml of peptide solution (approximately 5 mg/ml in 10 mM Tris-HC1 (pH 8)) was mixed with 0.5 ml of protease solution, and the mixture was incubated at 37 "C. The protease used in these experiments was purified through the DEAE and blue agarose steps described above, and was pre-incubated 15 min with 1 mM PMSF and 1 pg/ml leupeptin to eliminate contaminating proteolytic activities that might have complicated the cleavage pattern. Fifty-microliter aliquots were removed from the incubation mixture periodically for analytical HPLC runs to monitor the progress of the reaction. When a substantial portion of the starting peptide had been cleaved (5-24 h into the incubation), the reaction was stopped by adding an equal volume of 10% trichloroacetic acid. Precipitated protein was removed by centrifuging in a Microfuge, and the supernate was applied to a Vydac CIS HPLC column. Peptides were eluted with a gradient of acetonitrile, hydrolyzed, and analyzed for composition with an LKB a-amino acid analyzer.

Cleavage of Pro-IL-lp by PBMC Conditioned Medium-
Conditioned medium from an 18-h phytohemagglutinin-stimulated PBMC culture was concentrated as described under "Experimental Procedures," and a portion of it was applied to a DEAE-Sephacel column. Both the crude material and the DEAE flow-through were incubated with pro-IL-lp, and the incubation mixtures were then analyzed by Western blot, utilizing an antibody specific for the COOH terminus of IL-18 (3). The crude material generated a number of products including one that migrated close to but distinctly above mature IL-lp, while the DEAE flow-through produced a different set of fragments including one that ran slightly above the mature protein (Fig. 1). The remainder of this communication focuses on the activity in the DEAE flow-through responsible for generating the product slightly larger than mature IL-lp (the "DEAE flow activity").
Chromatography of DEAE Flow Activity-To determine if the DEAE flow activity was due to a single protease or to a combination of enzymes, we subjected the flow-through material to a number of chromatographic procedures. Fig. 2 shows the results with a blue agarose column: all the activity was retained and was subsequently eluted with 0.2-0.3 M NaCl in a single symmetrical peak. Similar results were obtained with other chromatographic materials, including phenyl-Sepharose, Polybuffer Exchanger 118 (Pharmacia LKB Biotechnology Inc.), and Procion Red-agarose. Gel filtration with Sephadex G-75 indicated a molecular mass of between 20 and 30 kilodaltons. Inhibitor Sensitivity-Further characterization was carried out with a pool of the active fractions eluted from a blue agarose column. PMSF (1 mM), E-64 (0.1 mM), and pepstatin (5 pg/ml) had no effect on the activity, while 1 mM EDTA and 1 mM 1,lO-phenanthroline completely inhibited it (Fig  3). Full activity was restored to EDTA-treated material by the addition of 1.25 mM CaC12 (Fig. 4); MgC12 also fully restored activity at 1.25-2.5 mM (data not shown). CaC12 was inhibitory at concentrations above 1.25 mM. Neither CaC12 nor MgC12 restored activity to the phenanthroline-treated protease, but ZnCl2 did so, at 23 PM; higher concentrations were inhibitory (Fig. 4). Three commonly used substratebased inhibitors of metalloproteases, phosphoramidon, Z-Gly-Leu-NH2, and N-(2-carboxy-3-phenylpropionyl)-Leu, had little or no effect on the protease (Fig. 3).
To characterize further the specificity of the enzyme, two peptides were synthesized, composed of residues 109-121 and 103-113 of the precursor (with modified ends as described under "Experimental Procedures"). Each peptide was incubated with the partially purified protease, and the resulting fragments were separated by HPLC and analyzed for amino acid composition (Fig. 5). This analysis indicated that all the significant products resulted from cleavages at the amino side of the two aspartic acid residues in each peptide (Fig. 6, A  and B ) . DISCUSSION We have identified a previously undescribed proteolytic activity from human peripheral blood mononuclear cells that appears to cleave proteins and peptides prior to aspartate residues. Most if not all known mammalian metalloendopro- teases cleave before or after hydrophobic amino acids (13). The recovery of this activity in a single peak upon ion exchange chromatography suggests that it is due to a single protease. This conclusion is supported by the finding that metal chelators eliminate all of the activity, while inhibitors of non-metalloproteases have no effect. The failure of common substrate-based metalloprotease inhibitors to affect the activity further indicates a unique enzyme and is consistent with the unusual cleavage specificity indicated by the protein and peptide products of the protease. We detected this activity by its ability to cleave pro-IL-lP near the mature IL-la NH2 terminus, and it is interesting that this enzyme is much more active with respect to pro-IL-10 than other proteases we have tested. Based on a maximal estimate of how much of the protease is present, it converts the precursor to lower molecular weight products at one-tenth of the concentration required with the proteases used in our previous work (3). The apparent inhibition of the protease in crude PBMC conditioned medium is also consistent with a role in generating mature IL-10, because processing by mono-cytes appears to occur at the cell surface, where the protease could be shielded from inhibitors. (Since cleavage could occur during or immediately after secretion, an extracellular protease is as likely to be involved as a cell-associated one.) Apart from monocytes, keratinocytes also produce pro-IL-lP, but these cells are unable to convert it to the mature form (14). In this case, a circulating protease would have to be involved in activating the hormone. It should be noted, however, that the protease we have described generates a form of IL-lP one amino acid longer than the form purified from monocytic cultures. There is an aminopeptidase in human blood that removes NH2-terminal aspartate residues and thus could complete the processing (15). Another issue requiring further investigation is whether there are intermediate forms of IL-1p in vivo that correspond to the larger fragments of the precursor apparently produced by the enzyme. If the protease is not involved in generating mature IL-1P, it will be interesting to determine what function it does perform. The highly restricted cleavage specificity observed, with no cleavage even at glutamate or asparagine residues, suggests a role in processing rather than in general protein degradation. The development of a specific, nontoxic inhibitor will be required to pursue the function of this protease.