A new thiol proteinase from rat liver.

An enzyme recently purified from rat liver (Gohda, E., and Pitot, H. C. (1980) J Biol. Chem., in press), which catalyzes the conversion of multiple forms of tyrosine aminotransferase, has been further characterized. The purified enzyme, termed a convertase, hydrolyzed oxidized ribonuclease A and acid-denatured hemoglobin in addition to Form I of tyrosine aminotransferase and azocasein. The enzyme exhibited little activity toward albumin, native ribonuclease A, cytochrome C, and oxidized insulin B chain. During the early stages of digestion of oxidized ribonuclease A with the purified convertase, two major cleavage products were detected. Oxidized ribonuclease A inhibited in a dose-dependent manner the conversion of Form I of tyrosine aminotransferase to Forms II and 111. Hydrolysis by the convertase of several synthetic substrates for exopeptidases was not detectable. During the purification of the convertase from rat liver, the activities of cathepsins B,, H, and L were eliminated from the convertase fractions, and the final preparation of the convertase showed no detectable activity of these cathepsins. Cathepsin D activity, measured as acid hemoglobin-hydrolyzing activity, was still detectable in the purified convertase fraction, but its yield and purification were much less than that of the convertase (about 2% of the convertase). The purified convertase was activated by sulfhydryl compounds and EDTA, and inhibited by sulfhydryl-reactive reagents. The convertase was also sensitive to leupeptin, antipain, and N-a-p-tosyl-Mysine chloromethyl ketone, but insensitive to phenylmethylsulfonyl fluoride and pepstatin. These characteristics of the purified convertase are consistent with the interpretation that the enzyme is a new thiol endopeptidase in rat liver lysosomes. We propose the name, cathepsin T, for this new proteinase.

catalyzes the conversion of multiple forms of tyrosine aminotransferase, has been further characterized. The purified enzyme, termed a convertase, hydrolyzed oxidized ribonuclease A and acid-denatured hemoglobin in addition to Form I of tyrosine aminotransferase and azocasein. The enzyme exhibited little activity toward albumin, native ribonuclease A, cytochrome C, and oxidized insulin B chain. During the early stages of digestion of oxidized ribonuclease A with the purified convertase, two major cleavage products were detected. Oxidized ribonuclease A inhibited in a dose-dependent manner the conversion of Form I of tyrosine aminotransferase to Forms II and 111. Hydrolysis by the convertase of several synthetic substrates for exopeptidases was not detectable. During the purification of the convertase from rat liver, the activities of cathepsins B,, H, and L were eliminated from the convertase fractions, and the final preparation of the convertase showed no detectable activity of these cathepsins. Cathepsin D activity, measured as acid hemoglobin-hydrolyzing activity, was still detectable in the purified convertase fraction, but its yield and purification were much less than that of the convertase (about 2% of the convertase). The purified convertase was activated by sulfhydryl compounds and EDTA, and inhibited by sulfhydryl-reactive reagents. The convertase was also sensitive to leupeptin, antipain, and N-a-p-tosyl-Mysine chloromethyl ketone, but insensitive to phenylmethylsulfonyl fluoride and pepstatin. These characteristics of the purified convertase are consistent with the interpretation that the enzyme is a new thiol endopeptidase in rat liver lysosomes. We propose the name, cathepsin T, for this new proteinase.
+ To whom all correspondence should be addressed.
In a previous report (13), we have purified the converting factor (termed "convertase") to homogeneity from rat liver. The purified convertase was a monomeric protein with a molecular weight of 33,500 to 35,000 and had a neutral optimum pH. The purified convertase catalyzed the conversion of Form I of tyrosine aminotransferase to Form 11, and subsequently to Form 111 in vitro with concomitant production of 48,000-Mr subunits from 52,000-Mr subunits of the aminotransferase. Those values are almost the same as the molecular weights of subunits of purified tyrosine aminotransferase forms reported by Hargrove et al. (7). The purified convertase also showed a potent azocaseinolytic activity. These results indicate that the convertase is a proteinase. In this report, we demonstrate that the convertase is a new thiol endopeptidase in rat liver lysosomes.
Separation. of Multiple Forms of Tyrosine Aminotransferase by Hydroxylapatite Chromatography-Hydroxylapatite chromatography of multiple forms of tyrosine aminotransferase by use of a stepwise elution procedure was performed as described previously (13) a t 4"C, except that 0.5 M potassium phosphate buffer, p H 6.9, was used instead of 0.32 M buffer for the elution of Form 111 of the enzyme.
Purification and Assay of Tyrosine Ammotransferase-Purified and partially purified Form I of tyrosine aminotransferase was prepared from rat liver as described previously (13). The specific activities of the preparations were 729 and 8.82 units/mg of protein, respectively. Tyrosine aminotransferase activity was assayed by the modification of Diamondstone's method (14) described by Iwasaki and Pitot (3). One unit, of activity is defined as that amount of enzyme

2567
A New Thiol Proteinase that forms 1 pmol ofp-hvdroxyphenylpyruvate/min at 37°C. A value of 19,900 M I cm I was used for the molar extinction coefficient ofphydroxyphenylpyruvate (14).
Purification a n d Assay of the Conuertase-The convertase was purified from rat liver as described previously (13). Where the convertase free of sulfhydryl-reducing compounds was required, the purified convertase was dialyzed overnight against 50 mM potassium phosphate buffer, pH 6.5. containing 12.55; glycerol and 0.1 M KC1 (Buffer A). The convertase activity was determined by measuring the formation of Form 111 from Form I of tyrosine aminotransferase according to the method described (13). For the assay of the convertase activity during purification, partially purified Form I of tyrosine aminotransferase (1.27 mg. 11.2 units) was used. One unit of the convertase activity is defined as that amount of endyme forming 1 unit of Form 111 from partially purified Form I of tyrosine aminotransferase/min at 0°C. The activity of the purified convertase was also assayed by use of purified Form I of tyrosine aminotransferase (3.75 pg. 2.73 units). In both systems, no conversion of Form I of tyrosine aminotransferase to other forms was observed during incubation in the absence of a convertase preparation. All assays were controlled by heating the reaction mixture for 10 min a t 60°C immediately after adding a convertase preparation and albumin solution.
Digestion of Proteins a n d Peptides by the Purified Conuertase-Bovine serum albumin and acid-denatured hemoglobin (15) were incubated at 1% (w/v) concentration with the purified convertase in 50 mM Mes buffer (pH 6.5 for bovine serum albumin and pH 5.8 for acid-denatured hemoglobin) containing 2 mM dithiothreitol for 10-30 min a t 37OC. After termination of the reaction with trichloroacetic acid a t a final concentration of 5% and filtration of the mixture, the clear filtrate was used for the determination of released peptides by the method of Lowry et al. (16) with slight modifications (17). Other enzymes and peptides (ribonuclease A. oxidized ribonuclease A, cytochrome c, and oxidized insulin B chain) were incubated at 0.04% (w/v) concentration with the purified convertase in 50 mM Mops buffer, pH 7.0, containing 2 mM dithiothreitol for 30-60 min a t 0°C. After incubation, aliquots were heated in 20 mM sodium phosphate buffer, pH 7.0, containing 0.2% (w/v) SDS and 0.05% (v/v) 8-mercaptoethanol for 10 min at 65°C. MDPF conjugation of peptides and SIX-polyacrylamide gel electrophoresis of fluorescent-labeled peptides were performed according to the method described by Chen-Kiang et al. (18). The migration of MDPF-conjugated peptide bands was visualized with a UV lamp and photographed. Horse heart cytochrome c ( M , = 12,300). bovine pancreatic trypsin inhibitor (M, = 6,160). and bacitracin ( M , = 1,400) were used as molecular weight markers.
Assay of Cathepsins-Cathepsin D activity was measured a t pH 3.5 and 37°C with bovine hemoglobin as substrate by the method of Barrett (17). One unit of enzyme activity is defined as that amount of enzyme releasing 1 pg of tyrosine/min. Cathepsin L activity was assayed with yeast glucose-6-phosphate dehydrogenase as substrate following the procedure of Towatari et al. (24). The reaction mixture contained 0.1 M potassium phosphate buffer, pH 6.0.5 mM 8-mercaptoethanol, 2.75 units of yeast glucose-&phosphate dehydrogenase (5.4 pg), and a proteinase preparation in a final volume of 200 pI. One unit of enzyme activity is defined as that amount of proteinase inactivating 50% of the substrate enzyme in 90 min a t 37°C. BANA-hydrolyzing activity was measured according to the method of Barrett (19) at 37OC. One unit of enzyme activity is defined as that quantity releasing 1 nmol of 8-naphthylamine/min.
Other Assays-Protein was measured by the method of Lowry et al. (16) with bovine serum albumin as the standard. For protein concentrations lower than 200 pg/ml, the modified Lowry protein assay was utilized (25). Glucose-&phosphate dehydrogenase activity was assayed by the method of Kuby and Noltmann (26). One unit of enzyme activity is defined as that quantity which catalyzes the reduction of 1 pmol of NAl>I'/min at pH 8.0 a t 30°C.

RESULTS
Proteolysis of Oxidized Ribonuclease A by the Conoertase-In the previous report (13), we demonstrated that the convertase purified from rat liver showed a potent azocaseinolytic activity that was about 6 times higher than that of cathepsin L, a thiol endopeptidase in rat liver lysosomes (24,27). This finding prompted us to investigate the activity of the convertase on other substrates of proteinases. Of the protein and peptide substrates tested, oxidized ribonuclease A was readily cleaved by the convertase even at O'C, and aciddenatured hemoglobin was also hydrolyzed. However, digestion of native bovine serum albumin was not detectable. The enzyme had little or no activity toward native ribonuclease A, cytochrome c, and oxidized insulin B chain at OOC. Fig. 1 demonstrates the electrophoretogram of oxidized ribonuclease A digested by the convertase at pH 7.0 and OOC. Two major cleaved products, observed during early (530 min) stages of incubation, had molecular weights of approximately 6,000 and 2,500. At later times (eg. 60 min) some other bands were also detected. This result suggests that cleavage of oxidized ribonuclease A by the convertase was by means of endopeptidase activity rather than exopeptidase activity.
As shown in Fig. 2, conversion of Form I of tyrosine aminotransferase to Forms I1 and 111 was inhibited in a dosedependent manner by the presence of oxidized ribonuclease A, but not by native ribonuclease A in the assay mixture. Equimolar concentrations of oxidized ribonuclease A to the subunits of Form I of tyrosine aminotransferase reduced this conversion to 40% of control. Concomitant studies of the same sample on SDS-polyacrylamide gel electrophoresis showed that the production of 48,000-Mr subunits of tyrosine aminotransferase from 52,000-Mr subunits was also inhibited by oxidized ribonuclease A in accordance with the inhibition in the conversion of the aminotransferase forms measured by Activity ofthe Convertase on Synthetic Substrates ofSmall Molecular Weight-The ability of the convertase to hydrolyze small molecular weight peptides and amides was examined by use of the following synthetic substrates: glycyl-Lphenylalanine-P-naphthylamide, L-seryl-L-tyrosine-P-naphthylamide, L-leucine-P-naphthylamide, L-arginine-/3-naphthylamide, L-leucinamide, hippuryl-L-arginine, hippuryl-Lphenylalanine, and N-benzoyl-t-tyrosine ethyl ester. No detectable cleavage of any of these compounds by the purified convertase was observed under the optimal conditions for the conversion of multiple forms of tyrosine aminotransferase (13 and see below).
Activity of Cathepsins B,, 0, H cathepsins BI, D, H, and L, have been found in rat liver lysosomes. In order to clarify whether the convertase was identical with one of these proteinases, activities of these cathepsins were followed during purification of the convertase from rat liver and compared with that of the convertase.
These results are shown in Table I. Since cathepsin H was reported by Kirschke et al. (28) to hydrolyze BANA rapidly, BANA was used as the substrate of both cathepsins BI and H. The final preparation of the convertase exhibited a 1490fold purification over the mitochondrial-lysosomal extract with a 20.5% yield in the convertase activity. This preparation of the convertase was homogeneous as judged by polyacrylamide gel electrophoresis in the presence or absence of SDS, as reported previously (13). Most of BANA hydrolase activity was separated from the convertase on CM-cellulose chromatography. On the last three steps, no activity of BANA hydrolase was detectable. Activity of cathepsin L was greatly reduced on hydroxylapatite chromatography and was not detectable after this step.
On the other hand, cathepsin D activity measured as acid hemoglobin-degrading activity was still detectable in the final fraction of DEAE-cellulose chromatography, although hydroxylapatite chromatography removed much of its activity. The yield and purification of cathepsin D were only 0.427% and 30.8-fold, respectively; these values were much lower than those of the convertase activity. Therefore, it is likely that the convertase exhibits hemoglobin-degrading activity at this pH. These results strongly suggest that the convertase is a proteinase different from cathepsins Bi, D, H, and L.
Influence of Group-specific Reagents and Proteinase Inhibitors on the Conuertase Activity-We reported previously (10) that the convertase activity in extracts of crude lysosomes from rat liver was markedly inhibited by sulfhydryl-reactive reagents, such as iodoacetate and p-chloromercuribenzoate. The effects of sulfhydryl compounds and EDTA were examined by incubating the convertase with these reagents for 15 min at 0°C prior to assay of the convertase. As shown in Table   11, the enzyme was activated by dithiothreitol at an optimal concentration of 2 mM. EDTA could activate the enzyme to TABLE I Activities of lysosomal endopeptidases duringpurification of the convertase from rat liver The convertase was purified from livers of 100 rats fasted overnight as described previously (13). The fractions from each step after extraction of mitochondrial-lysosomal fraction were used for the assay of activities of the convertase, BANA hydrolase, and cathepsins D and L.  The purified convertase dialyzed overnight against Buffer A was preincubated for 15 min at 0°C with the indicated concentration of reagents in Buffer A. After preincubation, aliquots were used for the assay of the convertase activity. The convertase activity was assayed in the presence of the indicated concentration of effectors at O0C with purified Form I of tyrosine aminotransferase used as substrate. The effects of sulfhydryl-reactive compounds and proteinase inhibitors on the convertase activity are summarized in Table 111. p-Chloromercuribenzoate, N-ethylmaleimide, and 4,4'-dipyridyl disulfide were powerful inhibitors of the enzyme. The purified convertase was also sensitive to iodoacetate and iodoacetamide. Of active site-specific reagents for serine proteinases, PMSF showed no effect on the enzyme activity after additional incubation with excess dithiothreitol, whereas TLCK inhibited the enzyme under the same conditions. The convertase was insensitive to pepstatin, primarily effective against carboxyl proteinases, and to soybean trypsin inhibitor. The purified convertase was inhibited by leupeptin and antipain, although relatively high concentrations of the inhibitors were required for inhibition. These results obtained with activators and inhibitors are compatible with the hypothesis that the convertase is a thiol proteinase.

DISCUSSION
In a previous report (13) we demonstrated that the conversion of Form I of tyrosine aminotransferase to Forms I1 and I11 by the purified convertase was accompanied by the concomitant production of 48,000-Mr subunits from 52,000-Mr subunits of the enzyme. From these results and the observation that the purified convertase exhibited a potent azocaseinolytic activity, we concluded that the conversion of tyrosine aminotransferase forms was a proteolytic process (13). In this study, oxidized ribonuclease A was found to be readily cleaved by the purified convertase (Fig. 1). Furthermore, oxidized ribonuclease A strongly inhibited the conversion of the aminotransferase forms as well as the production of 48,000-M , subunits of the enzyme (Fig. 2 ) . These results further

TABLE I11
Effects of various inhibitors on the convertase activity In experiment A, the purified convertase dialyzed overnight against Buffer A was preincubated for 15 min at 0°C with the indicated concentration of inhibitors in Buffer A. After preincubation, aliquots were used for the assay of the convertase activity, which was performed at 0°C in the presence of the indicated concentration of effectors, but without dithiothreitol. In experiment B, the purified convertase was preincubated for 15 min at 0°C with the indicated concentration of inhibitors in Buffer A containing 1 0 p~ dithiothreitol. After further preincubation with 5 m M dithiothreitol for 15 min at 0"C, aliquots were used for the assay of the convertase activity. The convertase activity was assayed under usual conditions without further addition of inhibitors. Stock solutions of PMSF (20 mM), TLCK (IO mM), and pepstatin (6 mM) were prepared in methanol, and appropriate controls for the vehicles were included. In both experiments, purified Form I of tyrosine aminotransferase was used as substrate of the convertase. No changes of total activity of tyrosine aminotransferase were detected on the treatments described. Soybean trypsin inhibitor 100 pg/ml 124 support our previous conclusion that the conversion of the tyrosine aminotransferase forms is a proteolytic process. High activity of the convertase toward proteins such as azocasein and oxidized ribonuclease A suggests an endopeptidase nature of the enzyme. The following observations also demonstrate endopeptidase activity of the convertase. 1) During the early stages of the digestion of oxidized ribonuclease A by the convertase, only two major polypeptide bands were observed as cleavage products on gels of SDS-polyacrylamide electrophoresis (Fig. 1) 3) During the cleavage of purified serine dehydratase by the convertase a t O"C, two major polypeptides, whose molecular weights were approximately 24,000 and 10,000, were produced from 34,000-M,. subunits of the enzymes (29) with concomitant decrease of the enzyme activity. 2 4) No detectable hydrolysis by the purified convertase of synthetic substrates for exopeptidases tested was observed.
Several characteristics of the purified convertase are consistent with the interpretation that the enzyme is a thiol proteinase. Typical of other thiol proteinases (30), the convertase activity was stimulated by sulfhydryl-reducing compounds or EDTA ( Table 11). The enzyme was sensitive to E:. Gohda, M. Himeno, and H. C. Pitot, unpublished observations. sulfhydryl-reactive compounds such as p-chloromercuribenzoate, N-ethylmaleimide, etc. The convertase was also inactivated by TLCK, which inhibited a range of thiol proteinases irreversibly by alkylating their essential sulfhydryl groups (31,32), although this compound is usually regarded as an active site-specific reagent for trypsin. Aviram and Hershko (33) reported that the convertase activity in a particulate fraction from rat hepatoma HTC cells was inhibited by PMSF. In our experiment, however, PMSF showed no effect on the purified convertase activity after addition of excess dithiothreitol prior to assay of the convertase activity. PMSF has been reported to inhibit thiol proteinases, but this inhibition, unlike that of serine proteinases, was reversed by excess dithiothreitol (30).
The convertase was insensitive to pepstatin, a powerful inhibitor of all carboxyl proteinases. Leupeptin and antipain, which readily inhibit thiol proteinases (34), inactivated the convertase at relatively high concentration. Of thiol endopeptidases in rat liver lysosomes, cathepsins L and B1 were quite sensitive to leupeptin and antipain (24,27), but cathepsin H is relatively insensitive to leupeptin (28). Four lysosomal endopeptidases have been found in rat liver. Three of them (cathepsins B1, H, and L) are thiol endopeptidases and the fourth is a carboxyl endopeptidase, cathepsin D. We reported the molecular weight of the convertase to be 33,500 to 35,000 (13), which is greater than that of cathepsins B, ( M r = 22,500) (35), H ( M r = 28,000) (28), and L (Mr = 23,000-24,OOO or 22,000) (24,27). During purification of the convertase from a mitochondrial-lysosomal fraction of rat liver, the activity of cathepsins B1 and H, measured as BANAhydrolyzing activity, and of cathepsin L were completely removed from the convertase fraction after hydroxylapatite chromatography and Sephadex G-75 gel filtration respectively (Table I). Thus, the last DEAE-cellulose fraction demonstrated no detectable activity of these cathepsins. Cathepsin D activity, however, measured as acid hemoglobin-hydrolyzing activity, was still measurable on the final preparation of the homogeneous convertase. Since the recovery and purification of cathepsin D activity were much lower than those of the convertase activity, and the convertase was insensitive to pepstatin (Table 111) (an effective inhibitor of cathepsin D) and markedly sensitive to sulfhydryl-reactive reagents (Table   III), it is likely that the convertase itself possesses hemoglobinhydrolyzing activity at this pH rather than that the convertase is identical with cathepsin D. The convertase showed about the same molecular weight as BANA amidohydrolase, an enzyme distinct from cathepsin BI in rat liver, as reported by de Lumen and Tappel (36). BANA, however, was not hydrolyzed by the purified convertase, as mentioned above. Collagenolytic cathepsin, which degraded soluble and insoluble collagen at acid pH, and was separated from cathepsin B1 in bovine spleen by Etherington (37, 38), has recently been named cathepsin N (39). This enzyme was found to be a thiol proteinase having a molecular weight of 18,000 to 20,000 (39), which is smaller than that of the convertase from rat liver (13). His group has partially purified a similar thiol proteinase from human placenta (40) and has also named this enzyme cathepsin N (41), although the human enzyme showed some properties different from those of bovine spleen cathepsin N, such as molecular weight, isoelectric point, and specific activity (39,40). Cathepsin N partially purified from human placenta was reported to have a molecular weight of 34,600 (40), which is very close to that of the purified convertase of rat liver (13) when estimated on a Sephadex G-100 column. Our preliminary experiments have shown potent collagenolytic activity of the convertase a t acid pH and 28°C. However, the degradation of collagen is suggested to be a general property of thiol proteinases (42). Most protein substrates other than collagen remained apparently undegraded by cathepsin N (43), whereas the convertase was able to digest common proteins, such as azocasein, hemoglobin, and oxidized ribonuclease A in addition to cleaving Form I of tyrosine aminotransferase. Moreover, cathepsin N from human placenta was shown to be stable at pH 7.4 (43), but the convertase was extremely unstable at this pH, unless glycerol was present at a concentration greater than 10% (13). An additional difference between cathepsin N and the convertase is the fact that crude cathepsin N preparation could be dialyzed against 10 mM sodium phosphate buffer, pH 6.5, without appreciable loss of enzyme activity (40). Under those conditions the convertase in crude preparations from rat liver was unstable and only partially soluble or aggregated.:' Therefore, the convertase is likely to be a hitherto unreported endopeptidase in rat liver lysosomes.
Cathepsin L was reported to account for approximately 50% of the proteinase activity of lysosomal extract from rat liver (27) and to play an important role in intracellular protein breakdown (44). The convertase showed a specific activity for hydrolyzing azocasein about 6 times higher than cathepsin L (13). From yields of both the convertase activity (6.1%) and azocaseinolytic activity (1.4%) (13), the convertase was calculated to be responsible for about one-fourth of the total azocaseinolytic activity of rat liver homogenates. Thus, this new thiol proteinase might play an important role in lysosomal protein degradation.
One might raise the question why Kirschke et,aZ. (27,28) were not able to detect this proteinase in spite of using azocasein as substrate during purification of cathepsins BI, H, and L from rat liver lysosomes. In their experiments, the convertase, if extracted from lysosomes, would probably have been inactivated during chromatography on Sephadex G-75. Their conditions made use of 10 mM potassium phosphate buffer, pH 6.9, containing 150 mM KC1 as an elution buffer, in which the convertase was unstable as described previously (13). On the basis of the results described in this report, we propose the name of cathepsin T for this new thiol endopeptidase in rat liver lysosomes.