Three forms of thiol proteinase inhibitor from rat liver formed depending on the oxidation-reduction state of a sulfhydryl group.

Three forms of a thiol proteinase inhibitor were isolated from rat liver cytosol. The monomeric inhibitor (pI 5.2) (TPI-1) formed a complex with cathepsin H even in the absence of reducing agents. The inhibitor with pI 5.0 (TPI-2) was inactive in the absence of reducing agents but was converted to an active inhibitor on addition of reducing agents such as dithiothreitol, GSH, cysteine, or 2-mercaptoethanol. The dimeric inhibitor (TPI-D) with an intermolecular disulfide bridge was also inactive and was converted to the active monomeric inhibitor on addition of dithiothreitol. TPI-2 is most likely a mixed disulfide with glutathione. One (Cys-3) of two cysteine residues exposed on the surface of the molecule of TPI-2 is involved in the formation of a mixed disulfide, and the other cysteine residue (Cys-64) is buried in the molecule. The activity of rat liver thiol proteinase inhibitor may possibly be regulated by formation of a protein mixed disulfide or by reduction of the mixed disulfide.

polyacrylamide gel electrophoresis. In the present study, we examined the relation of these three forms to the oxidationreduction state of their sulfhydryl groups and to the inhibitor activity. Results showed that the monomeric inhibitor with a PI of 5.0 (TPI-Zl) is a mixed disulfide, possibly with glutathione. TPI-2 and the dimeric inhibitor with an intermolecular disulfide bridge were inactive and did not form a complex with cathepsin H in the absence of thiol compounds. The third form, TPI-1 was active even in the absence of thiol compounds.
This paper reports on the characteristics of the three forms of the inhibitor and discusses the possibility that changes in the oxidation-reduction state regulate activity of thiol proteinase inhibitor in intact cells.
Enzyme and Enzyme Assays-Cathepsin H was prepared from rat liver by a modification of the method of Kirschke et al. (12). Papain (Type 111) was obtained from Sigma. Cathepsin H and papain were assayed with benzoyl-arginine-2-naphthylamide as substrate by the method of Barrett (13). The reaction mixture in 1.0 ml, containing 5 pmol of substrate, added as a solution in 50 pl of dimethyl sulfoxide, 100 pmol of potassium phosphate buffer, pH 6.0,2 pmol of EDTA, 4 pmol of cysteine, and an appropriate amount of enzyme, was incubated at 37 "C for 10 min.

Three F o r m of Thiol Proteinase Inhibitor from Rat Liver
thiocyanate-inactivated column was used in previous studies to prevent proteolysis of the inhibitor during its elution (9). For complete prevention of proteolysis, we used a column of carboxymethylated papain-Sepharose 4B in the present study.
The purification steps included affinity chromatography on carboxymethylated papain, Sephadex G-75 chromatography, and DEAE-cellulose chromatography. The supernatant at 105,000 X g of a liver extract was directly applied to the affinity column. At the G-75 step, two peaks of inhibitor activity were obtained (Fig. 1). The major fraction of lower molecular weight was dialyzed overnight a t 4 "C against 20 mM Tris-HCI, pH 7.5, and applied to a DEAE-cellulose (DE-52) column (1 X 18 cm) equilibrated with 20 mM Tris-HCI, pH 7.5. The column was washed and then the inhibitor was eluted with 60 ml of a linear gradient of 0 to 0.1 M NaCl in 20 mM Tris-HCI, pH 7.5. Fractions with inhibitory activity against papain were pooled, concentrated, and stored a t -20 "C.
Reduction and Alkylation of Thiol Proteinase Inhibitor-The inhibitor was treated with 10 mM dithiothreitol in 0.1 M Tris-HC1 buffer, pH 8.0, for 20 min a t room temperature in the absence and presence of 6 M guanidine hydrochloride and then with 18 mM iodoacetate or iodoacetamide for 20 min. The resulting solution was dialyzed for 24 h against five changes of 20 mM potassium phosphate buffer, pH 7.0, containing 1 mM EDTA, a t 4 "C. Incorporation of carboxymethyl groups was monitored by measuring S-carboxymethylcysteine after hydrolysis with 6 M HCI in uacuo. Amino acid analyses were performed on a Hitachi model 835-30 amino acid analyzer.
Titration of SH Groups of Thiol Proteinase Inhibitor with DTNB-The SH groups of thiol proteinase inhibitor were titrated by the method of Ellman (14). The reaction mixture contained 0.05 M potassium phosphate buffer, pH 8.0, 0.1 mM DTNB and thiol proteinase inhibitor in a final volume of 1.0 ml.
Polyacrylamide Gel Electrophoresis-Analytical native polyacrylamide gel electrophoresis was carried out at pH 8.0 as described by Williams and Reisfeld (15). SDS-polyacrylamide gel electrophoresis was performed by the method of Laemmli in 15% gel (16). Gels were stained with 0.2% Coomassie Brilliant Blue R-250 in methanol/acetic acid/water (25:7:68, v/v/v). Incorporation of r3H]GSSG into Thiol Proteinase Inhibitor-TPI-2 was reduced with 1.0 mM dithiothreitol and incubated with [glycine-2-3H]GSSG in the presence of 20 mM unlabeled GSSG. The incubation mixture was then applied to a column of Sephadex G-25. The radioactive peak, coinciding with the inhibitory activity, was concentrated and subjected to polyacrylamide gel electrophoresis without SDS. After electrophoresis the gel was cut into 2-mm sections, which were incubated with 20 mM Tris-HCI buffer, pH 7.5. The extracts were then dissolved in ACS-I1 (Amersham) and radioactivity was measured in a scintillation counter.
Isolation of Glutathione Bound to Thiol Proteinase Inhibitor-2-TPI-2 was reduced with 2 mM dithiothreitol. After the pH of the solution was adjusted to 2 with HCl, it was applied to a column of Sephadex G-25 (1.0 X 50 cm) equilibrated with 20 mM HCI. S H groups in each fraction of eluate were detected with DTNB. The positive fractions, which were devoid of inhibitory activity, were collected. After the solution was lyophilized, the sample was hydrolyzed a t 110 "C for 24 h in 6 N HCI in u a c o . The HCI was removed under reduced pressure, and the residue was dissolved in 0.5 ml of 20 mM HCl. Amino acid analyses were performed on a Hitachi Model 835-30 amino acid analyzer.

Isolation of Three Forms of Thiol Proteinase Inhibitor from
Rat Liuer-The thiol proteinase inhibitor was purified from rat liver on a carboxymethylated papain affinity column. of thiol proteinase inhibitor. Inhibitors were purified from the cytosol fraction of rat liver (see "Experimental Procedures") and then applied to a Sephadex G-75 column (2.4 X 96 cm) equilibrated with 20 mM potassium phosphate buffer, pH 7.0, containing 0.15 M NaCI. The flow rate was 20 ml/h and fractions of 2.0 ml were collected. Aliquots (10 pl) were assayed for inhibitory activity on papain. Fractions 90-110 (peak a) and fractions 120-140 (peak b) were pooled and concentrated by ultrafiltration on YM 5. Inset shows the results on polyacrylamide gel electrophoresis of T P I from peak a and peak b. Polyacrylamide gel electrophoresis was carried out in 7.0% acrylamide gels at pH 8.0. Both samples contained 10 pg of inhibitor. B, DEAE-cellulose column chromatography of peak b. Peak b was dialyzed against 20 mM Tris-HCl buffer, p H 7.5, and applied to a DEAE-cellulose column (1.0 X 18 cm) equilibrated with 20 mM Tris-HCI buffer, pH 7.5. All the inhibitory activity was retained on the ion exchanger. The column was washed with the same buffer until the eluate showed no absorbance a t 230 nm. The absorbed inhibitors were then eluted with a linear gradient of 60 ml of 0-0.1 M NaCl in 20 mM Tris-HCI buffer, pH 7.5, at a flow rate of 20 ml/h and fractions of 1.0 ml were collected. lA shows a typical elution pattern on Sephadex G-75. Two peaks of inhibitory activity (peak a and peak b) were obtained in positions corresponding to M , = 24,000 and 12,000, respectively. On native polyacrylamide gel electrophoresis, peak a gave one protein band but peak b gave two (Fig. lA, inset). The inhibitors of lower molecular weight in peak b) were further separated by DEAE-cellulose chromatography, as shown in Fig. 1B. We named these materials TPI-1 and TPI-2 in the order of elution from DEAE-cellulose. TPI-1 and TPI-2 each gave a single band in a position corresponding to M , = 12,000 on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Fig. 2, lanes 1 and 2). Mixtures of TPI-1 and TPI-2 also gave one protein band (Fig. 2, lane 3). Peak a obtained by Sephadex G-75 filtration (Fig. lA) gave a single band corresponding to M, = 12,000 on SDS-polyacrylamide gel electrophoresis in the presence of 2-mercaptoethanol ( Fig.  2, lanes 4 and 5 ) , but in the absence of 2-mercaptoethanol it gave two protein bands, corresponding to M , = 12,000 and 24,000, (Fig. 2, lane 6). Since thiol proteinase inhibitor contains two cysteine residues/mol ( l l ) , disulfide bond exchange may occur in the presence of SDS. Therefore, SDS-polyacrylamide gel electrophoresis was carried out in the presence of iodoacetate. Fig. 2, lane 7, shows that the main protein band

Effect of Thiol Compounds on Interconversions of Different Forms of Inhibitor and Formation of a Complex between the Inhibitor and Cathepsin
H-TPI-1 was incubated with or without 2 mM dithiothreitol at 25 "C for 10 min and then the mobility on nondenaturating polyacrylamide gel electrophoresis was examined. As shown in Fig. 3A, the mobility of TPI-1 was not affected (lane 2). The activity of TPI-1 was examined in the presence or absence of 2 mM dithiothreitol by testing its ability to form a complex with cathepsin H. Results mM dithiothreitol in 50 mM Tris-HC1, pH 7.0, were incubated a t 25 "C for 10 min and then applied to a Sephadex G-75 column (1 X 80 cm) equilibrated with 50 mM Tris-HCI, pH 7.0. One major and one minor protein peak were observed in positions corresponding to M, = 37,000 and 12,000, respectively (data not shown). Neither cathepsin H nor inhibitor activity was detected in the major peak, indicating that it was that of the complex, but inhibitor activity was detected in the minor peak. From its apparent molecular weight, the complex seems to be formed between cathepsin H and monomeric inhibitor, not the dimer inhibitor formed in the presence of dithiothreitol.
Cysteine Contents of the Three Forms of Inhibitor-To determine whether the sulfhydryl groups of the three forms of the inhibitor are involved in the interconversions and inhibitory activities of these forms, we measured the numbers of cysteine residues in the inhibitors. When TPI-1, TPI-2, and TPI-D were titrated with DTNB in the presence of 0.5% SDS, one S H group/mol of monomer was titrated in each form of the inhibitor. Next, the three forms of the inhibitor were reduced and carboxymethylated in the absence and presence of 6 M guanidine hydrochloride and their Cm-cysteine contents measured by amino acid analysis. As shown in Table I, there were no significant differences in the amino acid compositions of the three forms of the inhibitor. The Cm-cysteine contents of Cm-TPI-1 in the absence and presence of the denaturant were 0.2 and 1.1 residues/mol, respectively. But in Cm-TPI-2 and Cm-TPI-D about one Cmcysteine residue/mol of monomer was found in the absence of guanidine hydrochloride and about two residues/monomer inhibitor in the presence of the denaturant. These results suggest that one cysteinyl residue exposed on the surface of TPI-2 is carboxymethylated after reduction and that the other cysteinyl residue of TPI-2 is buried in the molecule because it was carboxymethylated only in the presence of guanidine hydrochloride. Results also indicated that one cysteine residue on the surface of the TPI-1 molecule was barely reactive even after treatment with dithiothreitol.  boxymethylated Inhibitors-On polyacrylamide gel electrophoresis without SDS, TPI-1, carboxymethylated in the presence of guanidine hydrochloride (denatured Cm-TPI-l), moved faster than untreated TPI-1 and in the position of the band of untreated TPI-2 (Fig. 4A, lane 2); the presence of dithiothreitol did not affect its mobility (Fig. 4A, lane 3). Denatured Cm-TPI-1 could form a complex with cathepsin H (Fig. 44, lane 5), like untreated TPI-1 (Fig. 4A, lane 4), although the former moved faster on the gel than the latter.
These results indicate that carboxymethylation of the inhibitor with iodoacetic acid caused an increase in its negative charge.
Interconuersion of TPI-2 and TPZ-D-When TPI-D was incubated with 20 mM GSSG after reduction with 1 mM dithiothreitol, it migrated in the same position as TPI-2 on nondenaturing polyacrylamide gel electrophoresis. Conversely, extensive dialysis of TPI-2 after reduction with 2 mM dithiothreitol resulted in conversion of TPI-2 to TPI-D (not shown).
Relation of Electrophoretic Mobility of TPI-2 to Inhibitor Activities-TPI-2 was reacted with iodoacetamide after reduction with dithiothreitol in the absence of 6 M guanidine hydrochloride. Carboxamidomethylated TPI-2 migrated in the same position as TPI-1 (Fig. 5, lane 2). It was active and formed a complex with cathepsin H (Fig. 5, lane 4). These results indicate that the inhibitor activity is closely related to the charge of the cysteine residue on the surface of TPI-2. To confirm this conclusion, we reacted TPI-2 with 20 mM cystamine (Fig. 6, lane 2), 20 mM homocystine (Fig. 6, lane 3) and 20 mM oxidized glutathione (Fig. 6, lane 4) after reduction with 1 mM dithiothreitol and examined the change in electrophoretic mobility by polyacrylamide gel electrophoresis without SDS. The electrophoretic mobility of the inhibitor differed depending on the charge of the disulfides. Results indicated that the free sulfhydryl group on TPI-2 produced by reduction reacted with the various added reagents to form mixed disulfides. TPI-2, incubated with cystamine and homocystine, was active and formed a complex with cathepsin H, but like TPI-  2, the inhibitor incubated with oxidized glutathione was inactive in the absence of dithiothreitol (Fig. 6, hnes 5, 6, and  7).
The formation of a mixed disulfide between TPI-2 and glutathione was also examined by measuring incorporation of radioactivity into the protein after incubation with 3 H -o~idized glutathione. After 20-min incubation of reduced TPI-2 with GSSG containing [3H]GSSG, the incubation mixture was subjected to gel filtration on a Sephadex G-25 column (1 X 45 cm), and the radioactivity and inhibitory activity of fractions of the eluate were measured. The peak position of the inhibitory activity measured in the presence of 4 mM cysteine coincided with the first peak of radioactivity. The second peak of radioactivity was that of remaining GSSG, as demonstrated by assay of GSH after reduction. When the fractions containing the inhibitor were pooled, concentrated, and subjected to polyacrylamide gel electrophoresis under nondenaturating conditions, radioactivity was obtained in a protein band that migrated in the same position as TPI-2 on a parallel gel (Fig. 7A). When 2 mM dithiothreitol was included in the incubation mixture, the radioactivity was recovered not in the protein but in the position of the marker dye, corresponding to authentic CSH. On addition of dithio- Fractions of proteinbound radioactivity, which appeared in the first radioactive peak on the column, were pooled and concentrated. Panel A, 3H-labeled inhibitor (8 pg) was subjected to electrophoresis on nondenaturating 7% polyacrylamide gels. Then one gel was stained for protein, and the other gel was assayed for radioactivity and for inhibitory activity on papain. Panel B, 3H-labeled inhibitor (8 pg) was incubated with 2 mM dithiothreitol and then subjected to electrophoresis as for p a n e l A. threitol, the inhibitor was recovered in the same position as TPI-1 (Fig. 7 B ) . Thus, after reduction of the mixed disulfide of TPI-2, GSSG reacted with the free thiol group on TPI-2 to form a mixed disulfide and the inhibitory activity was lost again. Addition of reducing agents such as dithiothreitol cleaved the mixed disulfide and caused the change in electrophoretic mobility and recovery of inhibitor activity.
Digestion of TPI with Cyanogen Bromide-The inhibitor consists of 98 amino acid residues and its two cysteine residues are residues 3 and 64 (11). It seemed interesting to determine which cysteine residue was involved in formation of the mixed disulfide. On cyanogen bromide digestion of the inhibitor, three major fragments are obtained (residues 11-98, residues 3-10, and an N-terminal fragment). The amino acid sequence of rat liver T P I (11) is: acetyl-MMCGAPSATMPATTETQ EIADKVKSQLEEKANQKFDVFKAISFRRQVVAGTNFFI KVDVGEEK C VHLRVFEPLPHENKPLTLSSYQTDKEK HDELTY F Thus, when a digest of the inhibitor with cyanogen bromide is applied to a Sephadex G-25 column, a large fragment containing cysteine 64 and a small fragment containing cysteine 3 should separate. TPI-2 was reduced and alkylated with ['4C]iodoacetic acid in the absence (Fig. 8A) and presence (Fig. 8B) of guanidine hydrochloride and treated with cyanogen bromide, and the digest was applied to a Sephadex G-25 column. Radioactivity was mainly recovered in the second peak of Cm-TPI-2 in the absence of guanidine hydrochloride, whereas it was recovered in both the first and second peaks of Cm-TPI-2 in the presence of guanidine hydrochloride. Thus the mixed disulfide is formed between cysteine 3 of TPI-2 and a thiol compound.
Isolation of Glutathione from TPI-2"To determine the identity of the thiol compound bound to TPI-2, TPI-2 was reduced with 2 mM dithiothreitol and the products were applied to a column of Sephadex G-25 equilibrated with 20 mM HCl. Inhibitory activity and SH groups in fractions of the eluate were measured. Fractions reacting with DTNB but containing no inhibitory activity were pooled. The solution was lyophilized, and the products were hydrolyzed at 110 "C for 24 h in 6 N HC1. Amino acid analysis of the hydrolysate showed that it contained 0.9, 1.1, 0.4, 0.2, and 0.2 residues/ mol of glutamic acid/glutamine, glycine, cystine, aspartic acid/asparagine, serine, respectively. Other amino acids detected were at less than 0.1 residue/mol. The value of cystine is not accurate, since cysteine and cystine were not determined after oxidation. The result suggests that the thiol compound bound to TPI-2 is glutathione.

DISCUSSION
Thiol proteinase inhibitor from rat liver consists of 98 amino acid residues and contains cysteine residues at positions 3 and 64 (11). Other thiol proteinase inhibitors purified from human leucocytes (18) and rat skin (19) contain no cysteine. We obtained three forms (TPI-1, TPI-2, and TPI-D) of the inhibitor during its purification from rat liver. This study showed that formation of these three forms was due to differences in the oxidation-reduction state of a specific cysteine residue in the inhibitor.
From the following observations it appears that TPI-2 is most likely to be a mixed disulfide with glutathione. (a) Incubation of the inhibitor with a reducing agent caused a change in electrophoretic mobility (Fig. 3), which was reversed by reduction and carboxymethylation (Fig. 4). (b) When TPI-2 was reacted with low-molecular-weight disulfides after reduction, its mobility on polyacrylamide gel electrophoresis depended on the charge of the disulfide (Fig. 6), indicating disulfide exchange between the protein sulfhydryl group and the disulfide. Incubation with GSSG did not affect the mobility of TPI-2. (c) Reaction of TPI-2 with radioactive disulfide in the presence of a reducing agent led to incorporation of the isotope into the inhibitor, and this radioactivity in the protein was lost on subsequent treatment with reducing agent (Fig.  7). Measurement of carboxymethylcysteine in TPI-2 after alkylation in the absence and presence of guanidine hydrochloride showed that one cysteine residue is on the surface of the molecule and the other is buried ( Table I). The sulfhydryl group on the surface of the molecule can interact with disulfide and is related to the formation of a mixed disulfide. The fact that Cys-3 is at the surface was demonstrated by analyzing CNBr digests of '*C-labeled Cm-TPI-2 (Fig. 8). (d) When isolated TPI-2 was subjected to amino acid analysis without reduction, it contained 0.4, 0.9, and 0.7 residues of cystine, glycine, and glutamic acid, respectively, more than reduced and carboxymethylated TPI-2 (data not shown), and in fact, glutathione was isolated from TPI-2 after reduction.
Isolated TPI-2 cannot form a complex with cathepsin H in the absence of a thiol compound, and this seems to be related to the charge of the protein mixed disulfide. The inhibitor became inactive on disulfide exchange with a negatively charged disulfide (GSSG), but reaction with cystamine or homocystine restored its activity (Fig. 6). This conclusion is supported by the observation that carboxymethylated TPI-2 was inactive, but that the carboxamidomethylated inhibitor was active (Fig. 5). These results indicate that a free sulfhydryl group on Cys-3 is not essential for formation of a complex with cathepsin H, and that introduction of a negative charge at the cysteine residue interferes with complex formation.
TPI-D was also inactive and the reductive cleavage of its intermolecular disulfide bond restored activity (Fig. 3). TPI-D and TPI-2 could be interconverted in uitro, but conversion of TPI-D or TPI-2 to TPI-1 was unsuccessful. Cys-3, the surface sulfhydryl group in TPI-1, is not free, and not reduced or carboxymethylated (Fig. 3, Table I). It may have some special structure, the nature of which is unknown, which is resistant to reduction, or the amino-terminal sequence containing Cys-3 may not be present in TPI-1. The second possibility seems very unlikely because the content of methionine, located at residues 1, 2, and 10 (11), is the same in TPI-1, TPI-2, and TPI-D (Table I).
Both cysteine Cys-3 and Cys-64 of the thiol proteinase inhibitor are probably free immediately after synthesis. Three forms of the thiol proteinase inhibitor are found after papainaffinity chromatography. Removal of low-molecular-weight thiol compounds may be a major factor in formation of these three forms. The very low pH used in elution of the inhibitor from the column may also promote their formation. However, the fact that GSH is isolated from TPI-2 after reduction suggests the formation of TPI-2 in uiuo.
The activities of many enzymes (17, 20-25) have been suggested to be regulated by the formation of protein mixed disulfides, which has a dramatic effect on their catalytic activities. The present results indicate that the thiol proteinase inhibitor from rat liver is a typical example of this type of protein. Since it is found mainly in the soluble fraction (91, the cytosolic level of GSH and the GSH/GSSG ratio may influence the level of the active thiol proteinase inhibitor in the liver. In fact, mixed disulfides between protein and glutathione or cysteine are present in various mammalian tissues (26), and the amounts of these modified proteins may vary diurnally (27) or in response to the feeding cycle of the animal (28). However, the physiological significance of regulation of the inhibitory activity by formation of mixed disulfides and by formation of a dimer through an intermolecular disulfide bridge awaits further investigation.