- Email: [email protected]
Purification and properties of protease F, a bacterial enzyme with chymotrypsin and elastase specificities
ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 225, No. 2, September, pp. 451-45’7, 1983
Purification and Properties with Chymotrypsin YEHUDITH
of Protease F, a Bacterial Enzyme and Elastase Specificities’
BIRK,**’ SHULAMITH
KHALEF,*
AND
MICHAEL
D. JIBSONY
‘Department of Agricultural Biochemistry, The Hebrew University of Jerusalem, Rehovot 76100, Ismel, and ~Hm-mone Research Lakratorg, University of California, San Francisco, Cal@brnia 9&g? Received March
17, 1983
It has been previously demonstrated that commercial bacterial fibrinolysin (EC 3.4.21.7) selectively cleaves the bond between Met-53 and Ala-54 in ovine prolactin (199 amino acids). A one-step purification procedure on DEAE-cellulose for Protease F, which is the active component of bacterial fibrinolysin, and properties of the purified enzyme are reported. The enzyme is homogeneous as judged by acrylamide gel electrophoresis. Its molecular weight, calculated from gel filtration experiments on Sephadex G-100, is around 13,800. Amino acid analyses do not reveal the presence of any halfcystines. The presence of one tryptophan residue per enzyme molecule was resolved from the fluorescence spectrum. Amino terminal analysis showed that leucine was at the amino terminal position. Protease F hydrolyzes casein and synthetic specific substrates for chymotrypsin and elastase esterases but not for trypsin esterases. It is fully inhibited by phenylmethylsulfonyl fluoride, by chicken ovoinhibitor, and by Chymotrypsin Inhibitor I from potatoes but not by the trypsin-chymotrypsin inhibitors from soybeans and chick peas or by tosyl-L-phenylalanine chloromethyl ketone. The enzyme is stable at room temperature and in the cold, it is not affected by dialysis or by freezing and thawing, but it is inactivated during freeze-drying. The circular dichroism spectra of Protease F indicate an -20% a-helix content of the enzyme with a considerable similarity to those of subtilisin, elastase, and /3-trypsin. The relatively low molecular weight of Protease F, the absence of intrachain disulfide bridges, and the fact that it is inhibited by several, but not all, chymotrypsin inhibitors suggest that it may differ phylogenetically from the known serine proteases.
Fibrinolysin (EC 3.4.21.7), a commercial bacterial protease, increased the activity of human growth hormone when reacted with the latter (1). A similar enzyme preparation was used for limited proteolysis of ovine prolactin (199 amino acids), resulting in cleavage of a single peptide bond (Met 53-Ala 54) and yielding two separable fragments, l-53 and 54-199, which could be recombined by a noncovalent interaction
(2). The intriguing property of this bacterial protease to hydrolyze selectively (under defined experimental conditions) a single peptide bond in the prolactin molecule, and the lack of a pronounced denaturation effect of the enzyme on the conformation of the prolactin (2,3), triggered the present study. In this article we describe the purification and properties of an active protease from bacterial fibrinolysin, henceforth designated Protease F. The specificity of Protease F toward different substrates and its susceptibility to synthetic and naturally occurring inhibitors of proteolysis are also reported.
1 Dedicated to Dr. Choh Hao Li on the occasion of his 70th birthday. * To whom correspondence should be addressed. 451
0003-9861/83 $3.00 Copyright All rights
0 1983 by Academic Press, Inc. of reproduction in any form reserved.
452
BIRK,
KHALEF,
EXPERIMENTAL Fibrinolysin (bacterial, e&erase free) was a product of Calbiochem (Cat. No. 34163). Bovine trypsin and a-chymotrypsin were purchased from Worthington Biochemical Corporation. The Bowman-Birk trypsin and chymotrypsin inhibitor from soybeans and inhibitor CI from chickpeas were prepared as described previously (4, 5). Chicken ovoinhibitor was isolated from ovomucoid by salt fractionation (6, 7). AcTyrOEt,S BzArgGEt, BzArgNan, PMSF, TosPheCK, and Sephadex G-200 were purchased from Sigma Chemical Company. Dansyl chloride and acrylamide were from Fluka AG. AcTyrNan, Coomassie brilliant blue (R-250), and Tris were from Serva. Casein (vitamin free) was from the Nutritional Biochemical Corporation. Naphthalene Black 12B (amid0 black) was from the British Drug Houses. DEAE-cellulose (DE32) was a product of Whatman. Chymotrypsin Inhibitor I from potatoes (8) was obtained from Dr. C. A. Ryan (Washington State University). Proteolytic activity was determined by Kunitz’s casein digestion method (9, 10). Trypsin and plasminlike e&erase activities were checked on the substrates TosArgOMe, BzArgOEt, and BzArgNan at 0.01 M substrate concentration in 0.046 M Tris/HCl buffer in 0.115 M CaClz at pH 8.1, 25”C, in a Gilford 2400 spectrophotometer (11). In a similar manner, chymotrypsin-like esterase activity was determined on AcTyrOEt (5), and elastase-like activity on AcAl@an (12). The capacity of any of the above-listed inhibitors to inhibit proteolysis or chymotrypsin and elastase e&erase was determined by preincubation of the enzyme with the inhibitor for 5 min. The reaction was started by addition of the appropriate substrate (AcTyrOEt for chymotrypsin, AcAla&Ian for elastase, and casein for general proteolysis) and the residual enzymatic activity was measured as described earlier (13). Protein content was estimated by the absorbance at 280 nm. A protein solution, giving an absorbance of 1.00 through a l-cm-path-length cell was defined as 1 absorbance unit (AU)/ml. Carbohydrate content was checked by the orcinol method (14). Amino acid analyses were performed on 1-mg samples hydrolyzed in 1 ml of 5.7 N HCl in ‘ua.cuo at 110°C for 22 h. For tryptophan determination, 24-h hydrolysis was performed in the presence of 4% thioglycolic acid (15). Norleucine was added as standard. The
3 Abbreviations used: AcTyrOEt, N-acetyl-L-tyrosine ethyl ester; TosArgGMe, tosyl-L-arginine methyl ester; BzArgOEt, N-benzoyl+tyrosine ethyl ester; BzArgNan, benzoyl-L-arginine pnitroanilide; PMSF, phenylmethylsulfonyl fluoride; TosPheCK, tosyl-Lphenylalanine chloromethyl ketone; AcTyrNan, Nacetyl-L-tyrosine pnitroanilide.
AND
JIBSON
analyses were performed on a LKB 3201 amino acid analyzer by a single-column method. The number of tryptophan residues was also resolved from the fluorescence spectrum measured with a Hitachi PerkinElmer Model MPF4A fluorescence spectrophotometer. The excitation wavelength was 295 nm with a lo-nm bandpass, and the emission spectrum was scanned from 260 to 460 nm. The sample was dissolved in 0.046 M Tris/HCl buffer in 0.115 M CaClz, pH 8.1. Amino terminal residues were determined according to Woods and Wang (16). Electrophoresis on cellulose acetate membranes was performed in a Beckman Microzone Electrophoresis System Model R-100 in 0.008 M collidine acetate buffer, pH 7.0 (the molarity is that of the cationic component). Samples (0.25-l ~1) of -4% protein solution were applied to the membrane and the electrophoresis was performed for 20 min at 400 V. Staining was carried out in 0.2% amido black 12 in MeOH:HAc:HzO (9:2:9). In the destaining solution the amido black was omitted. Electrophoresis in polyacrylamide gels was performed on samples of the enzyme (lo-200 ng) according to Reisfeld et al. (17) in 0.35 M @-alanineacetate buffer, pH 4.5, in 7.5% polyacrylamide gels. The gels were stained with Coomassie brilliant blue (R-250). For determination of enzymatic activity in the gels, frozen, unstained gels were cut into l-mm slices using a gel slicer. Each slice was transferred into a tube containing 1 ml of 0.1 M Tris/HCl buffer, pH 8, with 0.025 M CaClz. The substrate, 50 ~1 of 0.01 M AcTyrNan in dimethylformamide, was then added. After 24 h at 37°C the reaction was terminated by 0.5 ml 30% acetic acid and the extent of hydrolysis could be determined at 410 nm. The molecular weight of Protease F was determined by the method of Laurent and Killander (18). The sample (-2 mg) was run on a Sephadex G-100 column (1.4 X 57 cm) upward flow, 7.2 ml/h, with transmittance monitored continuously at 280 nm. The column was previously calibrated with bovine serum albumin (crystalline, Miles), myoglobin (crystalline, MilesServac), and ovine prolactin (Dr. C. H. Li). The extinction coefficient of Protease F was determined by comparison of the uv absorption spectrum and quantitative amino acid analysis of the sample. In this case a Beckman DK-2A scanning spectrophotometer was used to determine the spectrum between 360 and 255 nm, with light scattering corrected by the method of Beavan and Holiday (19). The sample was then quantitatively hydrolyzed for 26 h by constant-boiling HCl and analyzed as described. The sample concentration was determined from the amino acid analysis, and comparison with sample absorption prior to hydrolysis allowed determination of absorptivity and extinction coefficient. Circular dichroism spectra were taken with a Cary Model 60 spectrophotometer, equipped with a Model
PURIFICATION
AND
PROPERTIES
6002 circular dichroism attachment. Spectra were taken with O.l- or l.O-cm cells. Sample concentrations were between 0.6 and 1.0 mg/ml. In no case was the dynode voltage allowed to exceed 600 V. Each sample spectrum was scanned at least three times, and each baseline at least twice. Data are expressed in terms of mean residue ellipticity, 0~s~. A mean residue weight average of 98.6 was used. Content of a-helix was deterjmined according to the method of Bewley et al. (20).
RESULTS
Pur$catim
AND
DISCUSSION
of Protease F
OF PROTEASE
453
F
Commercial bacterial fibrinolysin (100 mg) was suspended in 10 ml of 0.046 M Tris/HCl buffer in 0.115 M CaClz, pH 8.1, and stirred with a magnetic stirrer for 5 min. The suspension was centrifuged for 5 min in a clinical laboratory centrifuge. The supernatant was applied to a DEAEcellulose column (1 X 50 cm) equilibrated with the above buffer. Elution was performed with the same buffer. Fractions of 1.6 ml were collected at a constant flow rate of 16 ml/h. The active fraction was eluted between 30 and 42 ml elution volume and pooled.
The commercial bacterial protease fibrinolysin served as starting material. Activity assays on casein and on specific synthetic substrates indicated the presence of protease and of chymotrypsin and elastase esterase activities, but no trypsin esterase was detected. Preliminary separations of the enzyme activities by ion-exchange and exclusion chromatography showed that the chymotrypsin esterase is not separable from the elastase esterase, suggesting that the two activities abide in the same enzyme molecule. Of all the columns used, DEAE-cellulose was most satisfactory with respect to purification and recovery of enzymatic activity. A typical one-step purification procedure is summarized in Table I. The purification of the enzyme was followed by the spectrophotometric assay on the esteric substrate AcTyrOEt for chymotrypsin. The extent of hydrolysis was estimated from the change in absorbance at 237 nm and expressed in arbitrary units. In view of the stability of the enzymatic activity at room temperature, the purification was carried out at room temperature as follows.
Enzyme Stability
The active eluate from the DEAE-cellulose column was stable for several days when kept at room temperature or in the cold. Dialysis against distilled water and repeated freezing and thawing did not affect enzymatic activity. Since the enzyme lost most of its activity during freeze-drying, a series of attempts was made to concentrate the activity of the pure enzyme preparation. The following means were found to be most satisfactory. The lo-ml portion of the active enzyme eluted from the column was transferred into a dialysis bag which was then rolled in dry Sephadex G-200 for concentration to a volume of -0.5 ml. The bag with the concentrated enzyme was placed in a lo-ml cylinder filled with a 50% solution of glycerol in water. After 24 h of equilibration the enzyme solution (amounting to -0.3 ml) was removed from the dialysis tubing and kept either at room temperature or in the cold. No activity was lost during this concentration process. The activity of the pure enzyme in glycerol
TABLE
I
PURIFICATION OF PROTEASE F
Step Crude extract Active effluent from DEAE-cellulose column
Total protein W-J)
Total activity (arbitrary units)
90
1500
0.7
1242
Specific activity (activity units/AU) 16.7 1800
Purification (fold) 108
Yield (o/o) 100 83
454
BIRK,
KHALEF,
could be fully retained for at least 1 month when kept at room temperature or in the cold (at 0 or -20°C). No sugar could be detected in the purified enzyme preparation. SpecQkity and Kinetic Properties The purified enzyme preparation from the DEAE-cellulose column was active on the substrates casein, AcTyrOEt, and AcAla3Nan. No activity was exhibited on TosArgOMe, BzArgOEt, or BzArgNan. The K, value for AcTyrOEt was 5.7 X 10m3M and for AcAlasNan was 6.8 X 10e3M. Reaction of ovine prolactin with the purified enzyme resulted in the same limited proteolysis pattern as achieved with the starting material-the commercial bacterial fibrinolysin (2). The pH maximum for the hydrolysis of AcTyrOEt was 8.8 and for the hydrolysis of AcAla3Nan-a plateau in the range of pH 7.6-9.4. Eflect of Inhibitors Inhibition of the proteolytic activity of Protease F as well as of the chymotrypsin and elastase esterase activities was achieved with PMSF, chicken ovoinhibitor, and Chymotrypsin Inhibitor I from potatoes. No inhibition was shown by either TosPheCK, even after 24 h incubation with the enzyme, or by the trypsin and chymotrypsin inhibitors from soybeans and chickpeas, at an enzyme to inhibitor ratio of 1:20 (mol/mol). These findings show that Protease F is a serine enzyme, but it differs, most likely, in the structure of its chymotrypsin binding site from the known chymotrypsins since it is inhibited by some but not all of the chymotrypsin inhibitors. The formation of a complex between Protease F and chicken ovoinhibitor is demonstrated by electrophoresis at pH 7.0 on cellulose acetate strips (Fig. 1). Criteria of Purity When a lOO-pg sample of the purified and concentrated enzyme preparation was submitted to polyacrylamide gel electrophoresis only a single band could be vi-
AND
JIBSON
;:: FIG. 1. Cellulose acetate electrophoresis at pH 7.0 of(E) Protease F isolated from commercial bacterial fibrinolysin; (I) chicken ovoinhibitor; and, (E + I) Protease F + chicken ovoinhibitor.
sualized after staining with Coomassie blue. Assay of gel segments, which had been incubated with 50 ~1 of 0.01 M AcTyrNan at 3’7°C for 24 h, indicated that the esterase activity coincided with the protein band. Molecular Wtight of Protease F The molecular weight was determined to be 13,800 from chromatography on Sephadex G-100 as described under Experimental, corresponding to a Stokes radius of 17.1. Amino Acid Composition Amino acid analysis was performed after acid hydrolysis as described under Experimental. The results are summarized in Table II. Of special interest is the absence of half-cystines. The complete lack of intrachain disulfide bridges undoubtedly plays an important role in the conformational behavior of the enzyme molecule. It is also self-evident that no free SH groups are involved in enzyme activity. Since estimation of tryptophan content from the amino acid analysis was not conclusive, the content of tryptophan in Protease F was resolved from the fluorescence spectrum. The latter showed an emission peak at 342 nm when excited at 295 nm. At pH 8.1 this spectrum may be attributed almost entirely to tryptophan residues. The number of tryptophan residues may be surmised by estimation of the absorptivity according to the method of Wetlaufer (21), based on the number of chromophoric side chains in the sequence. This method does not take into account contributions of con-
PURIFICATION TABLE
AND
PROPERTIES
II
AMINO ACID COMPOSITION OF PROTEASE F
Amino acid
Residues/molecule” (nearest integer)
Lysine Histidine Arginine Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine Half-cystine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Tryptophanb
7 3 3 15 8 13 11 7 16 13 none 11 1 6 8 5 3 1
Total residues
131
“Calculated on the basis of 11 residues of valine per molecule. b Resolved from the fluorescence spectrum.
formation to the absorptivity, and therefore consistently underestimates the absorptivity of rigid proteins (21). Calculation of absorptivity by this method, based on 5 tyrosines and 1 tryptophan, gives a value of 0.89 (0.79 for 4 tyrosines and 0.98 for 6 tyrosines), while use of 2 tryptophans raises the estimate to 1.29 (1.19 for 4 tyrosines and 1.39 for 6 tyrosines). The estimate for 2 tryptophans is higher than the empirical value, suggesting the presence of a single tryptophan in protease F. Amino Terminal Analysis Visualization of the dansylated residues of Protease F showed a major spot corresponding to leucine and traces of valine.
OF PROTEASE
F
455
are based on a molecular weight of 13,800, and will reflect any error in the molecular weight, as well as an uncertainty of approximately 5% arising from errors in the quantitative amino acid analysis. Circular L?ichr&sm Spectra The circular dichroism spectra of Protease F in glycerol buffer and PMSFtreated protease F in the near-uv region, are shown in Fig. 2A. Both spectra show a single major band with a negative maximum at 276-277 nm. PMSF-treated protease F also shows apparent fine structure between 255 and 285 nm. The far-uv circular dichroism spectra of Protease F in the two conditions are shown in Fig. 2B. Both spectra show a negative maximum at 218-220 nm. Neither spectrum was recorded below 210 nm because of an unfavorable dynode voltage. The near-uv circular dichroism spectra of both preparations is dominated by tyrosine, with fine structure apparent in the PMSF-treated sample between 270 and 285 nm. The additional fine structure between 255 and 270 nm may be assigned to weak phenylalanine transitions or may be due to PMSF alone. The 220-nm maximum of Protease F may be due to small amounts of helical or P-sheet structure. The maximum a-helix content of the enzyme is 20% as determined by the 220-nm band intensity. Thus, aperiodic structure dominates the Protease F backbone. These circular dichroism spectra are similar to those reported for subtilisin (22) and elastase (23), each of which is characterized by a single negative band around 275 nm and predominantly aperiodic secondary structure. In addition, the near-uv circular dichroism spectrum of fi-trypsin shows the same basic features as these spectra (24,25). These parallels suggest the possibility of considerable structural homology among these molecules. CONCLUSION
Extinction
Cweent
The extinction coefficient of Protease F was found to be 14,100, corresponding to an absorptivity (A:::) of 1.02. These values
Protease F is the active component of bacterial fibrinolysin with chymotrypsin and elastase esterase specificities. It amounts to about 1% of the commercial
456
BIRK,
FIBRINOLYSIN - - - FIBRINOLVSIN I -5 200 210
KHALEF,
AND
JIBSON
IN BUFFER + GLYCEROL + PMSF I I 220 230 WAVELENGTH (nm)
-I I 240
I
FIG. 2. Circular dichroism spectra of Protease F and PMSF-treated Protease F (A) in the nearuv region (side-chain absorption) and, (B) in the far-uv region (amide bond absorption).
preparation. Since the proteolytic and esterolytic activities of Protease F were inseparable by ion-exchange and exclusion chromatography and by cellulose acetate and polyacrylamide gel electrophoresis, we assume that Protease F comprises one enzyme only. This assumption is supported by the fact that PMSF, chicken ovoinhibitor, and Chymotrypsin Inhibitor I from potatoes inhibit all the known enzymatic activities of Protease F. However, the possibility that Protease F contains two active sites, one for chymotrypsin and the other for elastase esterase activity, should not be ruled out. It has been shown that chicken ovoinhibitor can simultaneously inhibit trypsin, chymotrypsin, and elastase (7).
The similarity between Protease F and mammalian chymotrypsin and elastase, exemplified by specificity in the hydrolysis of specific synthetic substrates, is questioned with respect to the composition and structure of the active site. The fact that the double-headed trypsin and chymotrypsin inhibitors from soybeans and chickpeas, as well as TosPheCK, do not inhibit Protease F strongly suggests that the nature of the binding site of Protease F and/or the environment in its vicinity differ from those of the known serine proteases. Employment of active-site titrants, such as suitable chloromethyl ketones, should allow the mapping of the binding site and thus help in elucidation of this problem.
PURIFICATION
AND
PROPERTIES
So far, we have no information on the degree of homology of Protease F with the related proteolytic enzymes from phylogenetically diverse species (26). However, the relatively low molecular weight of Protease F (13,800) and the lack of intrachain disulfide bridges do not necessarily imply a high degree of similarity between this enzyme and the known serine protease. Determination of the primary structure of this single-chain protease should provide the missing information. Finally, it should be noted that Protease F originates from a commercial bacterial preparation. It might have undergone conformational and other unknown changes during the process of preparation, The data reported herewith were consistent for at least five different batches of bacterial fibrinolysin. Nevertheless, Protease F has an interesting potential for the use in selective, controlled, limited proteolysis as demonstrated by its action on ovine prolactin (2). REFERENCES 1. LEWIS, U. J., PENCE, S. J., SINGH, R. N. P., AND VANDERLAAN, W. P. (1975) B&hem. Biophys. Res. Cornmun 67, 617-624. 2. BIRK, Y., AND LI, C. H. (1978) Proc Nat. Acad Sci USA 75,2155-2159. 3. JIBSON, M. D., Lr, C. H., AND GLAZER, C. B. (1981) Proc. Natl. Acad Sci USA 78, 2830-2832. 4. BIRK, Y., AND GERTLER, A. (1968) in Biochemical Preparations (Lands, W. E. M., ed.), Vol. 12, pp. 25-29, Wiley, New York. 5. SMIRNOFF, P., KHALEF, S., BIRK, Y., AND APPLEBAUM, S. W. (1976) Biochem. J. 157, 745-751. 6. TOMIMATSU, Y., CLARY, J. J., AND BARTULOVICH, J. J. (1966) Arch Biochem. Biophys. 115, 536544.
OF PROTEASE
F
457
7. VERED, M., GERTLER, A., AND BURSTEIN, Y. (1981) Int. J. Pept. Prot. Res. 18, 169-179. 8. RYAN, C. A., AND KASSELL, B. (1970) in Methods in Enzymology (Perlmann, G. E., and Lorand, L., eds.), Vol. 19, pp. 883-889, Academic Press, New York. 9. KUNITZ, M. (1947) J. Gen Physid 30,291-310. 10. LASKOWSKI, M. (1955) in Methods in Enzymology (Colowick, S. P., and Kaplan, N. O., eds.), Vol. 2, pp. 8-54, Academic Press, New York. 11. HUMMEL, B. C. W. (1959) Caned J. B&hem. Physiol! 37, 1393-1399. 12. COHEN, T., GERTLER, A., ANLI BIRK, Y. (1981) Camp B&hem. Physid B 69, 639-653. 13. BIRK, Y. (1976) in Methods in Enzymology (Lorand, L., ed.), Vol. 45 pp. 695-739, Academic Press, New York. 14. FRANCOIS, C., MARSHALL, R. O., AND NEUBERGER, A. (1962) Biochem J. 83.335-341. 15. MATSUBARA, H., AND SASAKI, R. M. (1969) Biochem Biophys Res. Commun 35, 175-181. 16. WOODS, W. R., AND WANG, K. T. (1967) Biochim. Biophys. Acta 133, 369-370. 17. REISFELD, R. A., LEWIS, U. J., AND WILLIAMS, D. E. (1962) Nature (London) 195,281-283. 18. LAURENT, T. C., AND KILLANDER, J. (1964) J. Chromatog. 14.317-330. 19. BEAVAN, G. H., AND HOLIDAY, E. R. (1952) Advan Prot. Cfiem 7,319-386. 20. BEWLEY, T. A., BROVETTO-CRUZ, J., AND LI, C. H. (1969) Biochemistry 8,4701-4708. 21. WETLAUFER, D. B. (1962) Advan. Prot. Chem. 17, 304-390. 22. MYERS, B., II, AND GLAZER, A. N. (1971) J. Biol. Chem. 246,412-419. 23. GORBUNOFF, M. J., AND TIMASHEFF, S. N. (1972) Arch. Biochem. Biophys. 152,413-422. 24. VILLANUEVA, G. B., AND HERSKOVITS, T. T. (1971) Biochemistry 10,4589-4594. 25. BODE, W., AND HUBER, R. (1976) FEBS I&t. 68, 231-236. 26. DE-HAEN, C., NEURATH, H., AND TELLER, D. C. (1975) J. Mol. Biol 92,225-259.
Purification and Properties with Chymotrypsin YEHUDITH
of Protease F, a Bacterial Enzyme and Elastase Specificities’
BIRK,**’ SHULAMITH
KHALEF,*
AND
MICHAEL
D. JIBSONY
‘Department of Agricultural Biochemistry, The Hebrew University of Jerusalem, Rehovot 76100, Ismel, and ~Hm-mone Research Lakratorg, University of California, San Francisco, Cal@brnia 9&g? Received March
17, 1983
It has been previously demonstrated that commercial bacterial fibrinolysin (EC 3.4.21.7) selectively cleaves the bond between Met-53 and Ala-54 in ovine prolactin (199 amino acids). A one-step purification procedure on DEAE-cellulose for Protease F, which is the active component of bacterial fibrinolysin, and properties of the purified enzyme are reported. The enzyme is homogeneous as judged by acrylamide gel electrophoresis. Its molecular weight, calculated from gel filtration experiments on Sephadex G-100, is around 13,800. Amino acid analyses do not reveal the presence of any halfcystines. The presence of one tryptophan residue per enzyme molecule was resolved from the fluorescence spectrum. Amino terminal analysis showed that leucine was at the amino terminal position. Protease F hydrolyzes casein and synthetic specific substrates for chymotrypsin and elastase esterases but not for trypsin esterases. It is fully inhibited by phenylmethylsulfonyl fluoride, by chicken ovoinhibitor, and by Chymotrypsin Inhibitor I from potatoes but not by the trypsin-chymotrypsin inhibitors from soybeans and chick peas or by tosyl-L-phenylalanine chloromethyl ketone. The enzyme is stable at room temperature and in the cold, it is not affected by dialysis or by freezing and thawing, but it is inactivated during freeze-drying. The circular dichroism spectra of Protease F indicate an -20% a-helix content of the enzyme with a considerable similarity to those of subtilisin, elastase, and /3-trypsin. The relatively low molecular weight of Protease F, the absence of intrachain disulfide bridges, and the fact that it is inhibited by several, but not all, chymotrypsin inhibitors suggest that it may differ phylogenetically from the known serine proteases.
Fibrinolysin (EC 3.4.21.7), a commercial bacterial protease, increased the activity of human growth hormone when reacted with the latter (1). A similar enzyme preparation was used for limited proteolysis of ovine prolactin (199 amino acids), resulting in cleavage of a single peptide bond (Met 53-Ala 54) and yielding two separable fragments, l-53 and 54-199, which could be recombined by a noncovalent interaction
(2). The intriguing property of this bacterial protease to hydrolyze selectively (under defined experimental conditions) a single peptide bond in the prolactin molecule, and the lack of a pronounced denaturation effect of the enzyme on the conformation of the prolactin (2,3), triggered the present study. In this article we describe the purification and properties of an active protease from bacterial fibrinolysin, henceforth designated Protease F. The specificity of Protease F toward different substrates and its susceptibility to synthetic and naturally occurring inhibitors of proteolysis are also reported.
1 Dedicated to Dr. Choh Hao Li on the occasion of his 70th birthday. * To whom correspondence should be addressed. 451
0003-9861/83 $3.00 Copyright All rights
0 1983 by Academic Press, Inc. of reproduction in any form reserved.
452
BIRK,
KHALEF,
EXPERIMENTAL Fibrinolysin (bacterial, e&erase free) was a product of Calbiochem (Cat. No. 34163). Bovine trypsin and a-chymotrypsin were purchased from Worthington Biochemical Corporation. The Bowman-Birk trypsin and chymotrypsin inhibitor from soybeans and inhibitor CI from chickpeas were prepared as described previously (4, 5). Chicken ovoinhibitor was isolated from ovomucoid by salt fractionation (6, 7). AcTyrOEt,S BzArgGEt, BzArgNan, PMSF, TosPheCK, and Sephadex G-200 were purchased from Sigma Chemical Company. Dansyl chloride and acrylamide were from Fluka AG. AcTyrNan, Coomassie brilliant blue (R-250), and Tris were from Serva. Casein (vitamin free) was from the Nutritional Biochemical Corporation. Naphthalene Black 12B (amid0 black) was from the British Drug Houses. DEAE-cellulose (DE32) was a product of Whatman. Chymotrypsin Inhibitor I from potatoes (8) was obtained from Dr. C. A. Ryan (Washington State University). Proteolytic activity was determined by Kunitz’s casein digestion method (9, 10). Trypsin and plasminlike e&erase activities were checked on the substrates TosArgOMe, BzArgOEt, and BzArgNan at 0.01 M substrate concentration in 0.046 M Tris/HCl buffer in 0.115 M CaClz at pH 8.1, 25”C, in a Gilford 2400 spectrophotometer (11). In a similar manner, chymotrypsin-like esterase activity was determined on AcTyrOEt (5), and elastase-like activity on AcAl@an (12). The capacity of any of the above-listed inhibitors to inhibit proteolysis or chymotrypsin and elastase e&erase was determined by preincubation of the enzyme with the inhibitor for 5 min. The reaction was started by addition of the appropriate substrate (AcTyrOEt for chymotrypsin, AcAla&Ian for elastase, and casein for general proteolysis) and the residual enzymatic activity was measured as described earlier (13). Protein content was estimated by the absorbance at 280 nm. A protein solution, giving an absorbance of 1.00 through a l-cm-path-length cell was defined as 1 absorbance unit (AU)/ml. Carbohydrate content was checked by the orcinol method (14). Amino acid analyses were performed on 1-mg samples hydrolyzed in 1 ml of 5.7 N HCl in ‘ua.cuo at 110°C for 22 h. For tryptophan determination, 24-h hydrolysis was performed in the presence of 4% thioglycolic acid (15). Norleucine was added as standard. The
3 Abbreviations used: AcTyrOEt, N-acetyl-L-tyrosine ethyl ester; TosArgGMe, tosyl-L-arginine methyl ester; BzArgOEt, N-benzoyl+tyrosine ethyl ester; BzArgNan, benzoyl-L-arginine pnitroanilide; PMSF, phenylmethylsulfonyl fluoride; TosPheCK, tosyl-Lphenylalanine chloromethyl ketone; AcTyrNan, Nacetyl-L-tyrosine pnitroanilide.
AND
JIBSON
analyses were performed on a LKB 3201 amino acid analyzer by a single-column method. The number of tryptophan residues was also resolved from the fluorescence spectrum measured with a Hitachi PerkinElmer Model MPF4A fluorescence spectrophotometer. The excitation wavelength was 295 nm with a lo-nm bandpass, and the emission spectrum was scanned from 260 to 460 nm. The sample was dissolved in 0.046 M Tris/HCl buffer in 0.115 M CaClz, pH 8.1. Amino terminal residues were determined according to Woods and Wang (16). Electrophoresis on cellulose acetate membranes was performed in a Beckman Microzone Electrophoresis System Model R-100 in 0.008 M collidine acetate buffer, pH 7.0 (the molarity is that of the cationic component). Samples (0.25-l ~1) of -4% protein solution were applied to the membrane and the electrophoresis was performed for 20 min at 400 V. Staining was carried out in 0.2% amido black 12 in MeOH:HAc:HzO (9:2:9). In the destaining solution the amido black was omitted. Electrophoresis in polyacrylamide gels was performed on samples of the enzyme (lo-200 ng) according to Reisfeld et al. (17) in 0.35 M @-alanineacetate buffer, pH 4.5, in 7.5% polyacrylamide gels. The gels were stained with Coomassie brilliant blue (R-250). For determination of enzymatic activity in the gels, frozen, unstained gels were cut into l-mm slices using a gel slicer. Each slice was transferred into a tube containing 1 ml of 0.1 M Tris/HCl buffer, pH 8, with 0.025 M CaClz. The substrate, 50 ~1 of 0.01 M AcTyrNan in dimethylformamide, was then added. After 24 h at 37°C the reaction was terminated by 0.5 ml 30% acetic acid and the extent of hydrolysis could be determined at 410 nm. The molecular weight of Protease F was determined by the method of Laurent and Killander (18). The sample (-2 mg) was run on a Sephadex G-100 column (1.4 X 57 cm) upward flow, 7.2 ml/h, with transmittance monitored continuously at 280 nm. The column was previously calibrated with bovine serum albumin (crystalline, Miles), myoglobin (crystalline, MilesServac), and ovine prolactin (Dr. C. H. Li). The extinction coefficient of Protease F was determined by comparison of the uv absorption spectrum and quantitative amino acid analysis of the sample. In this case a Beckman DK-2A scanning spectrophotometer was used to determine the spectrum between 360 and 255 nm, with light scattering corrected by the method of Beavan and Holiday (19). The sample was then quantitatively hydrolyzed for 26 h by constant-boiling HCl and analyzed as described. The sample concentration was determined from the amino acid analysis, and comparison with sample absorption prior to hydrolysis allowed determination of absorptivity and extinction coefficient. Circular dichroism spectra were taken with a Cary Model 60 spectrophotometer, equipped with a Model
PURIFICATION
AND
PROPERTIES
6002 circular dichroism attachment. Spectra were taken with O.l- or l.O-cm cells. Sample concentrations were between 0.6 and 1.0 mg/ml. In no case was the dynode voltage allowed to exceed 600 V. Each sample spectrum was scanned at least three times, and each baseline at least twice. Data are expressed in terms of mean residue ellipticity, 0~s~. A mean residue weight average of 98.6 was used. Content of a-helix was deterjmined according to the method of Bewley et al. (20).
RESULTS
Pur$catim
AND
DISCUSSION
of Protease F
OF PROTEASE
453
F
Commercial bacterial fibrinolysin (100 mg) was suspended in 10 ml of 0.046 M Tris/HCl buffer in 0.115 M CaClz, pH 8.1, and stirred with a magnetic stirrer for 5 min. The suspension was centrifuged for 5 min in a clinical laboratory centrifuge. The supernatant was applied to a DEAEcellulose column (1 X 50 cm) equilibrated with the above buffer. Elution was performed with the same buffer. Fractions of 1.6 ml were collected at a constant flow rate of 16 ml/h. The active fraction was eluted between 30 and 42 ml elution volume and pooled.
The commercial bacterial protease fibrinolysin served as starting material. Activity assays on casein and on specific synthetic substrates indicated the presence of protease and of chymotrypsin and elastase esterase activities, but no trypsin esterase was detected. Preliminary separations of the enzyme activities by ion-exchange and exclusion chromatography showed that the chymotrypsin esterase is not separable from the elastase esterase, suggesting that the two activities abide in the same enzyme molecule. Of all the columns used, DEAE-cellulose was most satisfactory with respect to purification and recovery of enzymatic activity. A typical one-step purification procedure is summarized in Table I. The purification of the enzyme was followed by the spectrophotometric assay on the esteric substrate AcTyrOEt for chymotrypsin. The extent of hydrolysis was estimated from the change in absorbance at 237 nm and expressed in arbitrary units. In view of the stability of the enzymatic activity at room temperature, the purification was carried out at room temperature as follows.
Enzyme Stability
The active eluate from the DEAE-cellulose column was stable for several days when kept at room temperature or in the cold. Dialysis against distilled water and repeated freezing and thawing did not affect enzymatic activity. Since the enzyme lost most of its activity during freeze-drying, a series of attempts was made to concentrate the activity of the pure enzyme preparation. The following means were found to be most satisfactory. The lo-ml portion of the active enzyme eluted from the column was transferred into a dialysis bag which was then rolled in dry Sephadex G-200 for concentration to a volume of -0.5 ml. The bag with the concentrated enzyme was placed in a lo-ml cylinder filled with a 50% solution of glycerol in water. After 24 h of equilibration the enzyme solution (amounting to -0.3 ml) was removed from the dialysis tubing and kept either at room temperature or in the cold. No activity was lost during this concentration process. The activity of the pure enzyme in glycerol
TABLE
I
PURIFICATION OF PROTEASE F
Step Crude extract Active effluent from DEAE-cellulose column
Total protein W-J)
Total activity (arbitrary units)
90
1500
0.7
1242
Specific activity (activity units/AU) 16.7 1800
Purification (fold) 108
Yield (o/o) 100 83
454
BIRK,
KHALEF,
could be fully retained for at least 1 month when kept at room temperature or in the cold (at 0 or -20°C). No sugar could be detected in the purified enzyme preparation. SpecQkity and Kinetic Properties The purified enzyme preparation from the DEAE-cellulose column was active on the substrates casein, AcTyrOEt, and AcAla3Nan. No activity was exhibited on TosArgOMe, BzArgOEt, or BzArgNan. The K, value for AcTyrOEt was 5.7 X 10m3M and for AcAlasNan was 6.8 X 10e3M. Reaction of ovine prolactin with the purified enzyme resulted in the same limited proteolysis pattern as achieved with the starting material-the commercial bacterial fibrinolysin (2). The pH maximum for the hydrolysis of AcTyrOEt was 8.8 and for the hydrolysis of AcAla3Nan-a plateau in the range of pH 7.6-9.4. Eflect of Inhibitors Inhibition of the proteolytic activity of Protease F as well as of the chymotrypsin and elastase esterase activities was achieved with PMSF, chicken ovoinhibitor, and Chymotrypsin Inhibitor I from potatoes. No inhibition was shown by either TosPheCK, even after 24 h incubation with the enzyme, or by the trypsin and chymotrypsin inhibitors from soybeans and chickpeas, at an enzyme to inhibitor ratio of 1:20 (mol/mol). These findings show that Protease F is a serine enzyme, but it differs, most likely, in the structure of its chymotrypsin binding site from the known chymotrypsins since it is inhibited by some but not all of the chymotrypsin inhibitors. The formation of a complex between Protease F and chicken ovoinhibitor is demonstrated by electrophoresis at pH 7.0 on cellulose acetate strips (Fig. 1). Criteria of Purity When a lOO-pg sample of the purified and concentrated enzyme preparation was submitted to polyacrylamide gel electrophoresis only a single band could be vi-
AND
JIBSON
;:: FIG. 1. Cellulose acetate electrophoresis at pH 7.0 of(E) Protease F isolated from commercial bacterial fibrinolysin; (I) chicken ovoinhibitor; and, (E + I) Protease F + chicken ovoinhibitor.
sualized after staining with Coomassie blue. Assay of gel segments, which had been incubated with 50 ~1 of 0.01 M AcTyrNan at 3’7°C for 24 h, indicated that the esterase activity coincided with the protein band. Molecular Wtight of Protease F The molecular weight was determined to be 13,800 from chromatography on Sephadex G-100 as described under Experimental, corresponding to a Stokes radius of 17.1. Amino Acid Composition Amino acid analysis was performed after acid hydrolysis as described under Experimental. The results are summarized in Table II. Of special interest is the absence of half-cystines. The complete lack of intrachain disulfide bridges undoubtedly plays an important role in the conformational behavior of the enzyme molecule. It is also self-evident that no free SH groups are involved in enzyme activity. Since estimation of tryptophan content from the amino acid analysis was not conclusive, the content of tryptophan in Protease F was resolved from the fluorescence spectrum. The latter showed an emission peak at 342 nm when excited at 295 nm. At pH 8.1 this spectrum may be attributed almost entirely to tryptophan residues. The number of tryptophan residues may be surmised by estimation of the absorptivity according to the method of Wetlaufer (21), based on the number of chromophoric side chains in the sequence. This method does not take into account contributions of con-
PURIFICATION TABLE
AND
PROPERTIES
II
AMINO ACID COMPOSITION OF PROTEASE F
Amino acid
Residues/molecule” (nearest integer)
Lysine Histidine Arginine Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine Half-cystine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Tryptophanb
7 3 3 15 8 13 11 7 16 13 none 11 1 6 8 5 3 1
Total residues
131
“Calculated on the basis of 11 residues of valine per molecule. b Resolved from the fluorescence spectrum.
formation to the absorptivity, and therefore consistently underestimates the absorptivity of rigid proteins (21). Calculation of absorptivity by this method, based on 5 tyrosines and 1 tryptophan, gives a value of 0.89 (0.79 for 4 tyrosines and 0.98 for 6 tyrosines), while use of 2 tryptophans raises the estimate to 1.29 (1.19 for 4 tyrosines and 1.39 for 6 tyrosines). The estimate for 2 tryptophans is higher than the empirical value, suggesting the presence of a single tryptophan in protease F. Amino Terminal Analysis Visualization of the dansylated residues of Protease F showed a major spot corresponding to leucine and traces of valine.
OF PROTEASE
F
455
are based on a molecular weight of 13,800, and will reflect any error in the molecular weight, as well as an uncertainty of approximately 5% arising from errors in the quantitative amino acid analysis. Circular L?ichr&sm Spectra The circular dichroism spectra of Protease F in glycerol buffer and PMSFtreated protease F in the near-uv region, are shown in Fig. 2A. Both spectra show a single major band with a negative maximum at 276-277 nm. PMSF-treated protease F also shows apparent fine structure between 255 and 285 nm. The far-uv circular dichroism spectra of Protease F in the two conditions are shown in Fig. 2B. Both spectra show a negative maximum at 218-220 nm. Neither spectrum was recorded below 210 nm because of an unfavorable dynode voltage. The near-uv circular dichroism spectra of both preparations is dominated by tyrosine, with fine structure apparent in the PMSF-treated sample between 270 and 285 nm. The additional fine structure between 255 and 270 nm may be assigned to weak phenylalanine transitions or may be due to PMSF alone. The 220-nm maximum of Protease F may be due to small amounts of helical or P-sheet structure. The maximum a-helix content of the enzyme is 20% as determined by the 220-nm band intensity. Thus, aperiodic structure dominates the Protease F backbone. These circular dichroism spectra are similar to those reported for subtilisin (22) and elastase (23), each of which is characterized by a single negative band around 275 nm and predominantly aperiodic secondary structure. In addition, the near-uv circular dichroism spectrum of fi-trypsin shows the same basic features as these spectra (24,25). These parallels suggest the possibility of considerable structural homology among these molecules. CONCLUSION
Extinction
Cweent
The extinction coefficient of Protease F was found to be 14,100, corresponding to an absorptivity (A:::) of 1.02. These values
Protease F is the active component of bacterial fibrinolysin with chymotrypsin and elastase esterase specificities. It amounts to about 1% of the commercial
456
BIRK,
FIBRINOLYSIN - - - FIBRINOLVSIN I -5 200 210
KHALEF,
AND
JIBSON
IN BUFFER + GLYCEROL + PMSF I I 220 230 WAVELENGTH (nm)
-I I 240
I
FIG. 2. Circular dichroism spectra of Protease F and PMSF-treated Protease F (A) in the nearuv region (side-chain absorption) and, (B) in the far-uv region (amide bond absorption).
preparation. Since the proteolytic and esterolytic activities of Protease F were inseparable by ion-exchange and exclusion chromatography and by cellulose acetate and polyacrylamide gel electrophoresis, we assume that Protease F comprises one enzyme only. This assumption is supported by the fact that PMSF, chicken ovoinhibitor, and Chymotrypsin Inhibitor I from potatoes inhibit all the known enzymatic activities of Protease F. However, the possibility that Protease F contains two active sites, one for chymotrypsin and the other for elastase esterase activity, should not be ruled out. It has been shown that chicken ovoinhibitor can simultaneously inhibit trypsin, chymotrypsin, and elastase (7).
The similarity between Protease F and mammalian chymotrypsin and elastase, exemplified by specificity in the hydrolysis of specific synthetic substrates, is questioned with respect to the composition and structure of the active site. The fact that the double-headed trypsin and chymotrypsin inhibitors from soybeans and chickpeas, as well as TosPheCK, do not inhibit Protease F strongly suggests that the nature of the binding site of Protease F and/or the environment in its vicinity differ from those of the known serine proteases. Employment of active-site titrants, such as suitable chloromethyl ketones, should allow the mapping of the binding site and thus help in elucidation of this problem.
PURIFICATION
AND
PROPERTIES
So far, we have no information on the degree of homology of Protease F with the related proteolytic enzymes from phylogenetically diverse species (26). However, the relatively low molecular weight of Protease F (13,800) and the lack of intrachain disulfide bridges do not necessarily imply a high degree of similarity between this enzyme and the known serine protease. Determination of the primary structure of this single-chain protease should provide the missing information. Finally, it should be noted that Protease F originates from a commercial bacterial preparation. It might have undergone conformational and other unknown changes during the process of preparation, The data reported herewith were consistent for at least five different batches of bacterial fibrinolysin. Nevertheless, Protease F has an interesting potential for the use in selective, controlled, limited proteolysis as demonstrated by its action on ovine prolactin (2). REFERENCES 1. LEWIS, U. J., PENCE, S. J., SINGH, R. N. P., AND VANDERLAAN, W. P. (1975) B&hem. Biophys. Res. Cornmun 67, 617-624. 2. BIRK, Y., AND LI, C. H. (1978) Proc Nat. Acad Sci USA 75,2155-2159. 3. JIBSON, M. D., Lr, C. H., AND GLAZER, C. B. (1981) Proc. Natl. Acad Sci USA 78, 2830-2832. 4. BIRK, Y., AND GERTLER, A. (1968) in Biochemical Preparations (Lands, W. E. M., ed.), Vol. 12, pp. 25-29, Wiley, New York. 5. SMIRNOFF, P., KHALEF, S., BIRK, Y., AND APPLEBAUM, S. W. (1976) Biochem. J. 157, 745-751. 6. TOMIMATSU, Y., CLARY, J. J., AND BARTULOVICH, J. J. (1966) Arch Biochem. Biophys. 115, 536544.
OF PROTEASE
F
457
7. VERED, M., GERTLER, A., AND BURSTEIN, Y. (1981) Int. J. Pept. Prot. Res. 18, 169-179. 8. RYAN, C. A., AND KASSELL, B. (1970) in Methods in Enzymology (Perlmann, G. E., and Lorand, L., eds.), Vol. 19, pp. 883-889, Academic Press, New York. 9. KUNITZ, M. (1947) J. Gen Physid 30,291-310. 10. LASKOWSKI, M. (1955) in Methods in Enzymology (Colowick, S. P., and Kaplan, N. O., eds.), Vol. 2, pp. 8-54, Academic Press, New York. 11. HUMMEL, B. C. W. (1959) Caned J. B&hem. Physiol! 37, 1393-1399. 12. COHEN, T., GERTLER, A., ANLI BIRK, Y. (1981) Camp B&hem. Physid B 69, 639-653. 13. BIRK, Y. (1976) in Methods in Enzymology (Lorand, L., ed.), Vol. 45 pp. 695-739, Academic Press, New York. 14. FRANCOIS, C., MARSHALL, R. O., AND NEUBERGER, A. (1962) Biochem J. 83.335-341. 15. MATSUBARA, H., AND SASAKI, R. M. (1969) Biochem Biophys Res. Commun 35, 175-181. 16. WOODS, W. R., AND WANG, K. T. (1967) Biochim. Biophys. Acta 133, 369-370. 17. REISFELD, R. A., LEWIS, U. J., AND WILLIAMS, D. E. (1962) Nature (London) 195,281-283. 18. LAURENT, T. C., AND KILLANDER, J. (1964) J. Chromatog. 14.317-330. 19. BEAVAN, G. H., AND HOLIDAY, E. R. (1952) Advan Prot. Cfiem 7,319-386. 20. BEWLEY, T. A., BROVETTO-CRUZ, J., AND LI, C. H. (1969) Biochemistry 8,4701-4708. 21. WETLAUFER, D. B. (1962) Advan. Prot. Chem. 17, 304-390. 22. MYERS, B., II, AND GLAZER, A. N. (1971) J. Biol. Chem. 246,412-419. 23. GORBUNOFF, M. J., AND TIMASHEFF, S. N. (1972) Arch. Biochem. Biophys. 152,413-422. 24. VILLANUEVA, G. B., AND HERSKOVITS, T. T. (1971) Biochemistry 10,4589-4594. 25. BODE, W., AND HUBER, R. (1976) FEBS I&t. 68, 231-236. 26. DE-HAEN, C., NEURATH, H., AND TELLER, D. C. (1975) J. Mol. Biol 92,225-259.