- Email: [email protected]
Similarity and dissimilarity in the stereogeometry of the active sites of thrombin, trypsin, plasmin and glandular kallikrein
THROMBOSIS RESEARCH 45; 451-462, 1987 0049-3848/87 $3.00 t .OO Printed in the USA. Copyright (c) 1987 Pergamon Journals Ltd. Al? rights reserved.
SIMILARITY OF THROMBIN,
Akiko
AND
DISSIMILARITY IN THE STEREOGEOMETRY OF THE ACTIVE SITES TRYPSIN, PLASMIN AND GLANDULAR KALLIKREIN
HIJIKATA-OKUNOMIYA*, Shosuke OKAMOTO**, Ryoji KIKUMOTO***, Yoshikuni TAMAO***, Kazuo OHKUBO***, Tohru TEZUKA***, Shinji TONOMLJRA*** and Osamu MATSUMOTO****
*School of Allied Medical Sciences, Kobe University, Kobe, Kobe University School of Medicine, **Department of Physiology, Chemical Industries Ltd., Center, Mitsubishi Kobe, ***Research and ****Faculty of Pharmaceutical Sciences, Yokohama, Kyoto University, Kyoto, Japan (Received 7.7.1986; Accepted in revised form 18.11.1986 by Editor J. Seghatchian)
ABSTRACT The relationship between chemical modifications of arginine derivatives and inhibitory activity to trypsin, plasmin and glandular kallikrein was investigated comparing with that of thrombin and concluded as follows: 1) The hydrophobic binding pocket, which has been reported previously to be stereogeometrically very similar in trypsin and thrombin, corresponded to the length of ethylpiperidine. 2) Concerning the site (termed the P site) next to the hydrophobic binding pocket, there were large differences in stereogeometry between trypsin and thrombin; the binding site of trypsin extended further to allow propyl and phenyl group attached to piperidine, while that of thrombin would be much narrower and unable 3) The P sites of plasmin and glandular to allow them. kallikrein resembled that of trypsin in being able to allow phenyl group. To substantialize the hydrophobic binding pocket and the P site, a (2R,4R)-MQPA-trypsin complex model was generated using the results of X-ray crystallography of (2R,CR)-MQPA and BPTI-trypsin complex by calculation to minimize van der Waals contacts, and it was of great use for understanding the geometry of the active sites of trypsin, thrombin, plasmin and glandular kallikrein.
Key words:
Thrombin, trypsin, plasmin, kallikrein, active site, structure-activity relationship. 451
452
ACTIVE SITES OF SERINE PROTEASES
INTRODUCTION A series of extremely potent and highly selective thrombininhibitors have been reported previously (l-3). As shown by one of the representatives such as (2R,4R)-MQPA, they are all Larginine derivatives characterized by the following tri-pod structure: 1) specific group (charged guanidinium) of Arg, 2) aromatic group covering e-NH2 group (N-terminal side), 3) hydrophobic group covering or-COOH group (C-terminal side). The tripod structure has the advantage that the structural change in each pod is available for gaining informations about the corresponding portion of the enzyme independently. Indeed, the experimental studies of thrombin inhibition by the arginine derivatives have partly revealed the stereogeometry of the inhibitor binding portion of thrombin (3). If the inhibitor binding portion is specified in the structure of the protein molecule, the structure-activity relationship would provide more detailed information about the structural features of the portion, which can help to elucidate the differences in the substrate specificity of serine proteases. Recently, study of the computer-generated ligand-enzyme complex models has been reported and it is useful for the viewing of the features of the ligand binding portion when the crystal structure of an appropriate protein is found as a prototype (4). BPTI-trypsin complex, whose three dimensional structure has been determined crystallographically (5-7), seemed to be a good prototype in computer-generating a complex model of (ZR,4R)-MQPAtrypsin, since (2R,4R)-MQPA exhibited competitive and fairly strong inhibition of trypsin (3). In this report, experimental results for the inhibition of trypsin by arginine derivatives are described and compared with those of thrombin (1,8,9). Some results for plasmin and glandular kallikrein are also described. The experiments clearly indicated similarities and dissimilarities between thrombin and the other three enzymes in their active site geometry corresponding to the C-terminal side of arginine derivatives. Furthermore, a model for the (PR,4R)-MQPA-trypsin complex was generated by docking (2R,4R)-MQPA and trypsin using the coordinates of simple crystals of (2R,4R)-MQPA (10) and BPTI-trypsin complex (11). Although the method involved only very simple and theoretical calculations, it provided one of the most reasonable models for correlating the various experimental results for the structureactivity relationship with the structural features of the enzymes.
Abbreviations: MQPA, 4-methyl-l-[N2-[(3-methyl-1,2,3,4tetrahydro-8-quinol~nyl)sulfonyl]-L-arginyl]-2-piperidinecarboxylit acid; QPA, 1-[N -[(3-methyl-1,2,3,4-tetrahydro-8-quinolinyl) sulfgnyl]-L-arginyll-2-piperidinecarboxylic acid; MQP, 4-methyll-[N -[(3-methyl-1,2,3,4-tet~ahydro-8-quinolinyl)sulfonyl]-Larginyl piperidine; QP, 1-[N -[(3-methyl-1,2,3,4-tetrahydro-8quinolinyl)sulfonyl]-L-arginyllpiperidine; BPTI, bovine pancreatic trypsin inhibitor.
Vol. 45, No. 5
ACTIVE SITES OF SERINE PROTEASES
MATERIALS
AND
453
METHODS
Materials Trypsin: bovine pancreas, 2 x crystallized ethanol precipitate, Sigma Chemical Co., St. Louis. Thrombin: bovine, purified from topical thrombin (Mochida Pharm. Co. Ltd., Tokyo) according to Plasmin: human, AB Lundblad (12) with sulfopropyl-Sephadex. Glandular kallikrein: porcine pancreas, 1330 KABI, Stockholm. Na-benzoyl-DL-arginine KU/mg, supplied by Bayer AG, Wuppertal. H-D-Phep-nitroanilideoHC1 (BANA): BDH Chemicals Ltd., Poole. Pip-Arg-pNA, H-D-Val-Leu-Lys-pNA and H-D-Pro-Phe-Arg-pNA: KABI Diagnostica, Stockholm. Tos-Gly-Pro-Arg-pNA: Pentapharm AG, Basel. Inhibitors of arginine derivatives were prepared as described previously (3,8,9). Other chemicals were of reagent grade. Enzyme
assay
Ki values were calculated from the hydrolysis of synthetic peptide substrates measured in the following reaction system: HD-Phe-Pip-Arg-pNA and Tris-imidazole buffer (pH 8.1) for trypsin, Tos-Gly-Pro-Arg-pNA and Tris-imidazole buffer (pH 8.1) containing 1% polyethyleneglycol #6000 for thrombin, H-D-Val-Leu-Lys-pNA and 0.05 M Tris-HCl buffer (pH 7.4) for plasmin and D-Pro-Phe-Arg-pNA and Tris-imidazole buffer (pH 8.1) for glandular kallikrein. After a reaction time of 1.5 min at 37"C, the hydrolysis (final volume of 1.25 ml) was stopped by adding 0.15 ml of acetic acid and the absorbance at 405 nm was measured. The rate of3hydrolysis of BANA by trypsin was measured at 410 nm using 10 M BANA in 0.1 M Tris-HCl buffer (pH 8.0) at 37°C. (2R,4R)-MQPA-trypsin
model
Atomic coordinates for (2R,4R)-MQPA determined by X-ray crystallography were kindly provided by Dr. T. Matsuzaki, Mitsubishi Chemical Ind., Ltd., Yokohama, and atomic coordinates for BPTI-trypsin complex (1PTC) determined by Huber et al. were supplied by the Institute for Protein Research, Osaka University, Osaka. Hydrogens were added to the carbon, oxygen and nitrogen atoms of the (PR,4R)-MQPA structure and the protein structure by using known average bond angles and bond lengths. The atoms of (2R,4R)-MQPA were numbered as follows:
‘24=;12=‘25 /
N19 \/a c18 u
/ ‘23
c21 \c2~clyf15
c13
O33
O34
454
Numbers
ACTIVE SITES OF SERINE PROTEASES
of amino
acid
residues
of enzymes
refer
Vol. 45, No. 5
to chymotrypsin.
A (2R,4R)-MQPA-trypsin complex model was generated using Xray coordinates for (2R,4R)-MQPA crystal (10) and BPTI-trypsin complex (ll), based on two assumptions that arginine side chain of (2R,4R)-MQPA was to enter the specificity pocket in trypsin, since (2R,4R)-MQPA inhibited trypsin competitively as reported previously (3), and that N9 was to be placed at a position very near to Lys-151 N of BPTI, since, from the results in Table 1, it seemed reasonable to assume that N9 occupied a position very near to Lys-151 N in BPTI-trypsin complex. Applying these assumptions, (2R,4R)-MQPA was docked into trypsin as follows: C -C was superimposed on Lys-151 CA-CB of BPTI and N9 was brougftt $0 Lys-151 N of BPTI as near as possible. As the arginine side chain of (2R,4R)-MQPA was largely outside of the specificity pocket and the pipecolic acid portion came into collision with rotation of (PR,4R)-MQPA was trypsin in this form, intramolecular made to search for a suitable complex model with minimum van der Trypsin molecule was assumed to be rigid and Waals contacts. Ser-195 of trypsin was assumed to be allowed to contact with Clo, The following distances were used as the 011 of (2R,4R)-MQPA. minimum van der Waals radii and the sums of the minimum van der Waals radii: C, 1.35 A; 0, 1.30 A; S, 1.60 A; N, 1.30 b; C-H, 2.20 A; O-H, 2.20 b; and N-H, 2.20 A.
RESULTS I) Structure-activity
relationship
of arginine
derivatives
As shown in Table 1, methylation of Ng caused a marked This could be decrease in the inhibitory activity to trypsin. satisfactorily understood if it was assumed that, in MQPA-trypsin complex, Ng occupied a position corresponding to Lys-151 N in BPTI-trypsin complex where Lys-151 N was reported to interact Methylation of N9 also with Ser-214 0 with hydrogen bonding (7). caused a similar decrease in the inhibitory activity to thrombin, indicating that the portion of thrombin interacting with Ng was constructed similarly as in trypsin. Table 2 shows the structural modification of the methylpipecolic acid portion of (2R,4R)-MQPA and Ki values for trypsin. Comparison of the Ki values between (2R,4R)-MQPA and (2R)-QPA and between MQP and QP indicated that the 4-methyl group of (2R,4R)MQPA contributed to the interaction with the hydrophobic binding Furthermore, the finding that there was no pocket of trypsin. significant difference in Ki values between (2R,4R)-MQPA, (2R,4R)MQPA-Me and MQP and between (2R)-QPA and QP, indicated that the (2R)-COOH of (2R,4R)-MQPA neither contributed to nor interfered These observations were very with the interaction with trypsin. similar to the results for thrombin (3). Table 3 gives data for the relationship between chemical modifications of the C-terminal of arginine derivatives and inhiAll these compounds inhibited trypsin combition for trypsin.
ACTIVE SITES OF SERINE PROTEASES
Vol. 45, No. 5
455
TABLE 1 N-methylation and Inhibition for Trypsin and Thrombin N\
C-NH-(CH2)3-CH-CO-N
I
NH;
CH3
3
N-R
COOH
I
0CH3 I
R
-H -CH3
I
Ki for trypsln (M)
Ki for thrombin (H)
4.2 x lO-6
2 x 1o-7
>2.Q x 1o-4
6 x 1O-5
TABLE 2 MQPA-derivatives and Inhibition for Trypsin
I (2R,4R)-MQPA
!
Ki
(Ml
8.8 x 1o-6
(2R,4R)-MQPA-Me
1.5 x 1o-5
MQP
-5 1.1 x 10
(2R)-QPA
4.2 x lO-5
QP
5.3 x 1o-5
I5O values of compounds lwith H-D-Phe-Pip-Arg-pNA. 10 for thrombin were reported previously as 3.3, 1, 0.9, 1.5, It could be (8). 0.67, 1, 0.3, 0.1, 1 and 100 uM respectively values for thrombin would be proportional to Ki if the mode of inhibition is of the series of these inhibitors were reported to inhibit thrombin competitively with Thus, the ratio of the l/Ki or synthetic substrates (1,2,3). of each compound to that of compound 2 is given in parenl/I the2g.s as the relative inhibitory activity for easy comparison of the effects of the chemical modification on the inhibitory actiThe relative inhibitory activity between thrombin and trypsin. vities of compounds 1~8 for trypsin and thrombin were almost the petitively
456
ACTIVE SITES OF SERINE PROTEASES
TABLE Dansylarginine NH2\ NH/
a) b) c) d)
Derivatives
Vol. 45, No. 5
3 and
Inhibition
of Trypsin
C-NH-(CH2)3-CH-C0-~2
I
measured with H-D-Phe-Pip-Arg-pNA as substrate. relative inhibitory activity (R.I.A.). measured with 1 mM BANA as substrate. calculated from the values reported previously
(8).
portion as piperidine _p same, and changes in the C-terminal caused a gradual increase in methylpiperidine + ethylpiperidine On the other the relative inhibitory activity for both enzymes. inhibitory activities of compounds 9 and 10 hand, the relative different from each other. for these two enzymes were apparently
457
ACTIVE SITES OF SERINE PROTEASES
Vol. 45, No. 5
the relative inhibitory activity was In the case of thrombin, decreased from 10 to 1 with the structural change from ethylpiperidine (compound 8) to propylpiperidine (compound 9), and finally to 0.01 with phenylpiperidine (compound 10). In the case modifications caused less change in of trypsin, these structural the relative inhibitory activity, and there was, if any, a small effect in the opposite direction such as to increase the relative inhibitory activity. Effect of introduction of 4-phenyl group into piperidine was further studied as shown in Table 4. The mode of inhibition of these inhibitors for thrombin, trypsin, plasmin and glandular kallikrein was competitive with the synthetic peptide substrates used. 'Introduction of 4-phenyl group into piperidine in naphthalenesulfonylarginine compounds led to the same effect on the inhibitory activity for thrombin and trypsin as in dansyl arginine compounds shown in Table 3. In addition, introduction of 4-phenyl group into piperidine caused a similar effect on the inhibitory activity for plasmin and glandular kallikrein as for trypsin. These results provided valuable information about the active site geometry of these enzymes corresponding to the C-terminal portion of arginine compounds. First, the hydrophobic binding pocket, which was reported previously (3) and further confirmed in this report to be of very similar construction in trypsin and thrombin, was about 7.6 A long from the amide N corresponding to ethylpiperidine. Second, trypsin, plasmin and glandular kallikrein could allow a longer hydrophobic C-chain or phenyl group attached to piperidine but thrombin could not, indicating that
TABLE Phenyl
Introduction
into
the
4
Piperidine
Ring
and
Enzyme
Inhibition
NH,, ,C-NH-(CH2)3-YH-CO-R2 NH2
NH
I R1
Ki
R1 -So2 8 0 0
-
CM)
I
R2
thrombin
trvDsin __
plas ;min
kallikrein
3 N(;E3
,
-N%
0
2 x lo-5
1.2 x 1o-5
1 x
1o-3
3 x 10
-5
ACTIVE SITES OF SERINE
458
the portion next to the hydrophobic geometrically different in thrombin and it was termed the P site. II)
(2R,4R)-MQPA-trypsin
complex
Vol. 45, No. 5
binding pocket was stereofrom in the other enzymes,
model
The arginine side chain was placed into the specificity crystal pocket as follows. N5, C6, N7 and N8 of the (2R,4R)-MQPA were first oriented by superimposing C4-N5 of the (PR,4R)-MQPA Then, C -N crystal on CD-GE of Lys-151 of BPTI-trypsin complex. rotated to search for the position 4 a 2 and N5-C6 were subsequently which the atomic pair distances between N7 and Asp-189 ODl, N8 nearest to the and Asp-189 ODl, and N8 and Asp-189 0D2 approached Reavalues measured in Arg-151-BPTI-trypsinogen complex (13). and N were found at which the distances sonable positions for N to those in Argfrom Asp-189 ODl and ODZ were a8lmost identical 151-BPTI-trypsinogen complex. The (2R,4R)-4-methylpipecolic acid portion was rotated freely around Cl-Cl0 to search for a position with minimum van der and a very narrow range (9") was Waals contacts with trypsin, Next, the found to be allowed without van der Waals contacts. rotation angle around Clo-Na6 was checked, and it was found This (2R,4R)-4around C10-N26. possible to rotate up to 14 van der methylpipecolic acid position did not give intramolecular side chain (C -N ) conWaals contacts with Ng and the arginine ' *$ portion aci structed above. Thus, the (2R,4R)-4-methylpipecolic was fixed to the slight cavity constructed from Phe-41, Cys-42, of the His-57, Cys-58 and Ser-195, due to the extreme narrowness range of rotation angle. -0 ) was exThe tetrahydroquinolinesulfonyl portion (S amined for a possible position by rotating in t #I2 e or2a er of Cl-Ng, where it was in minimum contacts intraand S12-C N9-S12 acid molecularly with 'she argi nine side chain and methylpipecolic portion constructed above as well as intermolecularly with Three forms were found for the tetrahydroquinolinetrypsin.
Proposed
model I,
FIG. 1 for the (PR,4R)-MQPA-trypsin -, enzyme. inhibitor;
complex
459
ACTIVE SITES OF SERINE PROTEASES
Vol. 45, No. 5
(2R,4R) -MQPA-trypsin v
FIG. 2 complex model with , inhibitor; -,
4-phenyl enzyme.
group
in.trodu ted
sulfonyl portion without intra-molecular contacts. One was in contact with His-57, one was in contact with His-57, Trp-215 and Leu-99, and the other was in contact with Ser-214, Trp-215, Gly216 and Leu-99. Since His-57 was the constituent of the slight cavity to which the methylpipecolic acid portion came into close proximity, the last form which did not require any shift of His57 was taken as the most approvable model. The extent of contact was the largest at the atomic pairs of C and Trp-215 CB even where the distance was much longer than $8 e covalent bond length, and it would allow there to be no fatal collision considering the possibility that trypsin in the complex with (2R,4R)-MQPA might undergo a slight rearrangement from that in the complex with BPTI. Thus, the proposed model for the (PR,4R)-MQPA-trypsin complex was illustrated in Fig.1 which was constructed with a reasonably fitted arginine side chain, almost definitely fixed methylpipecolic acid portion and tetrahydroquinolinesulfonyl portion with minimum van der Waals contacts. In this model, 4methylpiperidine showed a good fit with the slight cavity constructed from Phe-41, Cys-42, His-57, Cys-58 and Ser-195, indicating that this site was the hydrophobic binding pocket. Introduction of 4-phenyl group into the piperidine ring in the (2R, 4R)-MQPA-trypsin model was shown in Fig.2. The 4-phenyl group extended to the trypsin surface constructed from Phe-41, His-57, Cys-58 and Tyr-59, indicating that this portion was the P site.
DISCUSSION The model of the (2R,4R)-MQPA-trypsin complex was generated by a simple, theoretical method in which van der Waals contacts were considered and the structure of trypsin was assumed to be rigid and identical to those in the complex with BPTI. Although
460
ACTIVE SITES OF SERINE PROTEASES
proteins tend to exhibit conformational motility in solution and trypsin complexed with (2R,4R)-MQPA might exhibit some rearrangement from BPTI-complexed trypsin, the method provided one of the most practically reasonable models which enabled us to visualize the binding site of trypsin for (2R,4R)-MQPA, to interpret the experimental data for the structure-activity relationship satisfactorily, and to substantialize the hydrophobic binding pocket and the P site. The amino acid sequences surrounding the His-57 loop and Ser-195 loop have been reported to be highly homologous in trypsin and thrombin (14). The present results strongly indicated that the cavities constructed from Phe(Leu)-41, Cys-42, His-57, Cys-58 and Ser-195 in these enzymes were very similarly arranged stereogeometrically. In the model, (2R)-COOH faced the solvent side, opposite to the hydrophobic binding pocket. The experimental finding that (PR)-COOH neither contributed to nor interfered with the interaction with the enzymes, was thus reasonably explained from the model. The structure-activity relationships as shown in Tables 3 and 4 suggested that the stereogeometry of the P site of thrombin might be much narrower than those of the other three serine Comparison of the amino acid sequences surrounding proteases. the P sites of these proteases had revealed that, only in thrombin, were there large insertions of 9 amino acids and carbohydthat the P site of rates next to 63 (14), raising the possibility thrombin had become much narrower than those of the other serine proteases owing to these bulky insertions. The fact that the arginine side chain of (2R,4R)-MQPA could be reasonably fitted in the specificity pocket of trypsin when Cl was superimposed on CA of Lys-151 of BPTI, implied that the side chain of Arg (P,) could enter into the specificity pocket of trypsin setting its CA at the same position as the CA of Lys-151 (P > of BPTI. In other words, it may safely be inferred that ei.ther Arg or Lys in P entered into the specificity pocket of trypsin setting each CA at almost the same position as the CA of Lys-151 of BPTI. portion was in In the model, the tetrahydroquinolinesulfonyl contact with Leu-99, Trp-215 and Gly-216 of trypsin and these Consequently, the were largely conserved in thrombin (14,15). tetrahydroquinolinesulfonyl portion of (2R,4R)-MQPA would fit well and interact with the binding portion of thrombin presumably constructed mainly from the amino acids involving Trp-215, resulting in increase in fluorescence intensity if it was substituted with dansyl group (16).
ACKNOWLEDGEMENTS We wish to express our deep thanks to Dr. T. Matsuzaki of us with the Mitsubishi Chemical Ind. Co. Ltd., for providing coordinates of the (2R,4R)-MQPA crystal prior to publication and
Vol. 45, No. 5
ACTIVE SITES OF SERINE PROTEASES
461
Dr. M. to Dr. N. Yasuoka of Himeji Institute of Technology, Matsushima of Osaka Medical College, Dr. H. Mori of Kobe Univerand a number of friends in sity, Dr. T. Taga of Kyoto University We are also our laboratories for their invaluable suggestions. the manuscript. indebted to Mrs. H. Nishi for preparing
REFERENCES 1.
OKAMOTO, S., HIJIKATA, A., KINJO, K., KIKUMOTO, R., OHKUBO, A novel series of synthetic TONOMURA, S. and TAMAO, Y. K thiombin-inhibitors having extremely potent and highly Kobe J. Med. Sci., 21, 43-51, 1975. selective action.
2.
HIJIKATA, A., OKAMOTO, S., MORI, E., KINJO, K., KIKUMOTO, Kinetic studies R TONOMURA, S., TAMAO, Y. and HARA, H. th;? selectivity of synthetic thrombin-inhibitor using Thrombos. Haemostas. synthetic peptide substrates. (Stuttgart), 2, 1039-1045, 1979.
on
3.
KIKUMOTO, R., TAMAO, Y., TEZUKA, T., TONOMURA, S., HARA, H., A. and OKAMOTO, S. Selsctive NINOMIYA, K., HIJIKATA, inhibition of thrombin by (2R,4R)-4-methyl-l-[N -[(3-methyl1,2,3,4-tetrahydro-8-quinolinyl)sulfonyl]-L-arginyl)]-2Biochemistry, G, 85-90, 1984. piperidinecarboxylic acid.
4.
FELDMANN, R.J., BING, D.H., FURIL, B.C. and FURIL, B. Interactive computer surface graphics approach to study of the active site of bovine trypsin. Biochemistry, 75, 54095412, 1978.
5.
RUEHLMANN, A., KUKLA, D., SCHWAGER, P., BARTELS, K. and Structure of the complex formed by bovine HUBER, R. trypsin and bovine pancreatic trypsin inhibitor: Crystal structure determination and stereochemistry of the contact J. Mol. Biol., 77, 417-436, 1973. region.
6.
HUBER, R., KUKLA, D., BODE, W., SCHWAGER, P., BARTELS, K., DEISENHOFER, J. and STEIGEMANN, W. Structure of the complex formed by bovine trypsin and bovine pancreatic trypsin inhibitor. II. Crystallographic refinement at 1.9 angstroms J. Mol. Biol., 89, 73-101, 1974. resolution.
7.
HUBER, R., KUKLA, D., STEIGEMANN, W., DEISENHOFER, J. and JONES, A. Structure of the complex formed by bovine trypsin and bovine pancreatic trypsin inhibitor: Refinement of the crystal structure analysis. in: Proteinase Inhibitors.
A short description of the possession of the hydrophobic binding pocket in common in trypsin, thrombin, acetylated thromto Cys-Cys-His in the bin, plasmin and FXa and its correspondence histidine loop, was presented at the 59th Annual Meeting of the Physiological Society of Japan, 1982 (17).
462
ACTIVE SITES OF SERENE PROTEASES
Frite, (Eds.)
H., Tschesche, H., Greene, New York: Springer Verlag,
Vol. 45, No. 5
L.J. and Truscheit, 1974, pp.497-512.
E.
8.
KIKUMOTO, R., TAMAO, Y., OHKUBO, K., TEZUKA, T., TONOMURA, OKAMOTO, S., FUNAHARA, Y. and HIJIKATA, A. S Thrombin of Ne-substituted Linhibitors. 2. Amide derivatives J. Med. Chem., 3, 830-836, 1980. arginine.
9.
KIKUMOTO, R., TAMAO, Y., OHKUBO, K., TEZUKA, T., TONOMURA, Thrombin inhibitors. S OKAMOTO, S. and HIJIKATA, A. 3:'Carboxyl-containing amide derivatives of No-substituted J. Med. Chem., 23, 1293-1299, 1980. L-arginine.
10.
MATSUZAKI,
11.
Obtained from the Protein Data Research, Osaka Univ., Osaka.
Bank,
12.
LUNDBLAD, R.L. A rapid method bovine thrombin and inhibition phenylmethylsulfonyl fuloride. 1971.
for the purification of of the purified enzyme with Biochemistry, 2, 2501-2506,
13.
BODE, W., WALTER, J., HUBER, R., WENZEL, H.R. and TSCHESCHE, H. The refined 2.2-A (0.22-nm) X-ray crystal structure of the ternary complex formedlky bovine valine-valine and the Arg trypsinogen, analogue of bovine pancreatic trypsin inhibitor. Eur. J. Biochem., 144, 185190, 1984.
14.
R.J. and MANN, K.G. ELION, J., DOWNING, M.R., BUTKOWSKI, Structure of human thrombin: Comparison with other serine proteases. in: Chemistry and Biology of Thrombin. Lundblad, B.L., Fenton II, J.W. and Mann, K.G. (Eds.) Ann Arbor Science, 1977, pp.97-111. Michigan:
15.
MAGNUSSON, S., PETERSON, T.E., SOTTRUP-JENSEN, L. and Complete primary structure of prothrombin: CLAEYS, H. structure and reactivity of ten carboxylated isolation, glutamic acid residues and regulation of prothrombin Proteases and biological activation by thrombin. in: Reich, E., Rifkin, D.B. and Shaw, E. (Eds.) Cold control. Spring Harbor Laboratory, 1975, pp. 123-149.
16.
NESHEIM, M.E., PRENDERGAST, F.G. and MANN, K.G. Interactions of a fluorescent active-site-directed inhibitor Dansylarginine N-(3-ethyl-1,5-pentanediyl) of thrombin: Biochemistry, 18, 996-1003, 1979. amide.
17.
Existence of the hydrophobic HIJIKATA, A. and OKAMOTO, S. pocket in common in the active sites of some serine proteases. J. Physiol. Sot. Japan, 44, 314, 1982.
T.
Private
communication,
1982. Institute
for Protein
SIMILARITY OF THROMBIN,
Akiko
AND
DISSIMILARITY IN THE STEREOGEOMETRY OF THE ACTIVE SITES TRYPSIN, PLASMIN AND GLANDULAR KALLIKREIN
HIJIKATA-OKUNOMIYA*, Shosuke OKAMOTO**, Ryoji KIKUMOTO***, Yoshikuni TAMAO***, Kazuo OHKUBO***, Tohru TEZUKA***, Shinji TONOMLJRA*** and Osamu MATSUMOTO****
*School of Allied Medical Sciences, Kobe University, Kobe, Kobe University School of Medicine, **Department of Physiology, Chemical Industries Ltd., Center, Mitsubishi Kobe, ***Research and ****Faculty of Pharmaceutical Sciences, Yokohama, Kyoto University, Kyoto, Japan (Received 7.7.1986; Accepted in revised form 18.11.1986 by Editor J. Seghatchian)
ABSTRACT The relationship between chemical modifications of arginine derivatives and inhibitory activity to trypsin, plasmin and glandular kallikrein was investigated comparing with that of thrombin and concluded as follows: 1) The hydrophobic binding pocket, which has been reported previously to be stereogeometrically very similar in trypsin and thrombin, corresponded to the length of ethylpiperidine. 2) Concerning the site (termed the P site) next to the hydrophobic binding pocket, there were large differences in stereogeometry between trypsin and thrombin; the binding site of trypsin extended further to allow propyl and phenyl group attached to piperidine, while that of thrombin would be much narrower and unable 3) The P sites of plasmin and glandular to allow them. kallikrein resembled that of trypsin in being able to allow phenyl group. To substantialize the hydrophobic binding pocket and the P site, a (2R,4R)-MQPA-trypsin complex model was generated using the results of X-ray crystallography of (2R,CR)-MQPA and BPTI-trypsin complex by calculation to minimize van der Waals contacts, and it was of great use for understanding the geometry of the active sites of trypsin, thrombin, plasmin and glandular kallikrein.
Key words:
Thrombin, trypsin, plasmin, kallikrein, active site, structure-activity relationship. 451
452
ACTIVE SITES OF SERINE PROTEASES
INTRODUCTION A series of extremely potent and highly selective thrombininhibitors have been reported previously (l-3). As shown by one of the representatives such as (2R,4R)-MQPA, they are all Larginine derivatives characterized by the following tri-pod structure: 1) specific group (charged guanidinium) of Arg, 2) aromatic group covering e-NH2 group (N-terminal side), 3) hydrophobic group covering or-COOH group (C-terminal side). The tripod structure has the advantage that the structural change in each pod is available for gaining informations about the corresponding portion of the enzyme independently. Indeed, the experimental studies of thrombin inhibition by the arginine derivatives have partly revealed the stereogeometry of the inhibitor binding portion of thrombin (3). If the inhibitor binding portion is specified in the structure of the protein molecule, the structure-activity relationship would provide more detailed information about the structural features of the portion, which can help to elucidate the differences in the substrate specificity of serine proteases. Recently, study of the computer-generated ligand-enzyme complex models has been reported and it is useful for the viewing of the features of the ligand binding portion when the crystal structure of an appropriate protein is found as a prototype (4). BPTI-trypsin complex, whose three dimensional structure has been determined crystallographically (5-7), seemed to be a good prototype in computer-generating a complex model of (ZR,4R)-MQPAtrypsin, since (2R,4R)-MQPA exhibited competitive and fairly strong inhibition of trypsin (3). In this report, experimental results for the inhibition of trypsin by arginine derivatives are described and compared with those of thrombin (1,8,9). Some results for plasmin and glandular kallikrein are also described. The experiments clearly indicated similarities and dissimilarities between thrombin and the other three enzymes in their active site geometry corresponding to the C-terminal side of arginine derivatives. Furthermore, a model for the (PR,4R)-MQPA-trypsin complex was generated by docking (2R,4R)-MQPA and trypsin using the coordinates of simple crystals of (2R,4R)-MQPA (10) and BPTI-trypsin complex (11). Although the method involved only very simple and theoretical calculations, it provided one of the most reasonable models for correlating the various experimental results for the structureactivity relationship with the structural features of the enzymes.
Abbreviations: MQPA, 4-methyl-l-[N2-[(3-methyl-1,2,3,4tetrahydro-8-quinol~nyl)sulfonyl]-L-arginyl]-2-piperidinecarboxylit acid; QPA, 1-[N -[(3-methyl-1,2,3,4-tetrahydro-8-quinolinyl) sulfgnyl]-L-arginyll-2-piperidinecarboxylic acid; MQP, 4-methyll-[N -[(3-methyl-1,2,3,4-tet~ahydro-8-quinolinyl)sulfonyl]-Larginyl piperidine; QP, 1-[N -[(3-methyl-1,2,3,4-tetrahydro-8quinolinyl)sulfonyl]-L-arginyllpiperidine; BPTI, bovine pancreatic trypsin inhibitor.
Vol. 45, No. 5
ACTIVE SITES OF SERINE PROTEASES
MATERIALS
AND
453
METHODS
Materials Trypsin: bovine pancreas, 2 x crystallized ethanol precipitate, Sigma Chemical Co., St. Louis. Thrombin: bovine, purified from topical thrombin (Mochida Pharm. Co. Ltd., Tokyo) according to Plasmin: human, AB Lundblad (12) with sulfopropyl-Sephadex. Glandular kallikrein: porcine pancreas, 1330 KABI, Stockholm. Na-benzoyl-DL-arginine KU/mg, supplied by Bayer AG, Wuppertal. H-D-Phep-nitroanilideoHC1 (BANA): BDH Chemicals Ltd., Poole. Pip-Arg-pNA, H-D-Val-Leu-Lys-pNA and H-D-Pro-Phe-Arg-pNA: KABI Diagnostica, Stockholm. Tos-Gly-Pro-Arg-pNA: Pentapharm AG, Basel. Inhibitors of arginine derivatives were prepared as described previously (3,8,9). Other chemicals were of reagent grade. Enzyme
assay
Ki values were calculated from the hydrolysis of synthetic peptide substrates measured in the following reaction system: HD-Phe-Pip-Arg-pNA and Tris-imidazole buffer (pH 8.1) for trypsin, Tos-Gly-Pro-Arg-pNA and Tris-imidazole buffer (pH 8.1) containing 1% polyethyleneglycol #6000 for thrombin, H-D-Val-Leu-Lys-pNA and 0.05 M Tris-HCl buffer (pH 7.4) for plasmin and D-Pro-Phe-Arg-pNA and Tris-imidazole buffer (pH 8.1) for glandular kallikrein. After a reaction time of 1.5 min at 37"C, the hydrolysis (final volume of 1.25 ml) was stopped by adding 0.15 ml of acetic acid and the absorbance at 405 nm was measured. The rate of3hydrolysis of BANA by trypsin was measured at 410 nm using 10 M BANA in 0.1 M Tris-HCl buffer (pH 8.0) at 37°C. (2R,4R)-MQPA-trypsin
model
Atomic coordinates for (2R,4R)-MQPA determined by X-ray crystallography were kindly provided by Dr. T. Matsuzaki, Mitsubishi Chemical Ind., Ltd., Yokohama, and atomic coordinates for BPTI-trypsin complex (1PTC) determined by Huber et al. were supplied by the Institute for Protein Research, Osaka University, Osaka. Hydrogens were added to the carbon, oxygen and nitrogen atoms of the (PR,4R)-MQPA structure and the protein structure by using known average bond angles and bond lengths. The atoms of (2R,4R)-MQPA were numbered as follows:
‘24=;12=‘25 /
N19 \/a c18 u
/ ‘23
c21 \c2~clyf15
c13
O33
O34
454
Numbers
ACTIVE SITES OF SERINE PROTEASES
of amino
acid
residues
of enzymes
refer
Vol. 45, No. 5
to chymotrypsin.
A (2R,4R)-MQPA-trypsin complex model was generated using Xray coordinates for (2R,4R)-MQPA crystal (10) and BPTI-trypsin complex (ll), based on two assumptions that arginine side chain of (2R,4R)-MQPA was to enter the specificity pocket in trypsin, since (2R,4R)-MQPA inhibited trypsin competitively as reported previously (3), and that N9 was to be placed at a position very near to Lys-151 N of BPTI, since, from the results in Table 1, it seemed reasonable to assume that N9 occupied a position very near to Lys-151 N in BPTI-trypsin complex. Applying these assumptions, (2R,4R)-MQPA was docked into trypsin as follows: C -C was superimposed on Lys-151 CA-CB of BPTI and N9 was brougftt $0 Lys-151 N of BPTI as near as possible. As the arginine side chain of (2R,4R)-MQPA was largely outside of the specificity pocket and the pipecolic acid portion came into collision with rotation of (PR,4R)-MQPA was trypsin in this form, intramolecular made to search for a suitable complex model with minimum van der Trypsin molecule was assumed to be rigid and Waals contacts. Ser-195 of trypsin was assumed to be allowed to contact with Clo, The following distances were used as the 011 of (2R,4R)-MQPA. minimum van der Waals radii and the sums of the minimum van der Waals radii: C, 1.35 A; 0, 1.30 A; S, 1.60 A; N, 1.30 b; C-H, 2.20 A; O-H, 2.20 b; and N-H, 2.20 A.
RESULTS I) Structure-activity
relationship
of arginine
derivatives
As shown in Table 1, methylation of Ng caused a marked This could be decrease in the inhibitory activity to trypsin. satisfactorily understood if it was assumed that, in MQPA-trypsin complex, Ng occupied a position corresponding to Lys-151 N in BPTI-trypsin complex where Lys-151 N was reported to interact Methylation of N9 also with Ser-214 0 with hydrogen bonding (7). caused a similar decrease in the inhibitory activity to thrombin, indicating that the portion of thrombin interacting with Ng was constructed similarly as in trypsin. Table 2 shows the structural modification of the methylpipecolic acid portion of (2R,4R)-MQPA and Ki values for trypsin. Comparison of the Ki values between (2R,4R)-MQPA and (2R)-QPA and between MQP and QP indicated that the 4-methyl group of (2R,4R)MQPA contributed to the interaction with the hydrophobic binding Furthermore, the finding that there was no pocket of trypsin. significant difference in Ki values between (2R,4R)-MQPA, (2R,4R)MQPA-Me and MQP and between (2R)-QPA and QP, indicated that the (2R)-COOH of (2R,4R)-MQPA neither contributed to nor interfered These observations were very with the interaction with trypsin. similar to the results for thrombin (3). Table 3 gives data for the relationship between chemical modifications of the C-terminal of arginine derivatives and inhiAll these compounds inhibited trypsin combition for trypsin.
ACTIVE SITES OF SERINE PROTEASES
Vol. 45, No. 5
455
TABLE 1 N-methylation and Inhibition for Trypsin and Thrombin N\
C-NH-(CH2)3-CH-CO-N
I
NH;
CH3
3
N-R
COOH
I
0CH3 I
R
-H -CH3
I
Ki for trypsln (M)
Ki for thrombin (H)
4.2 x lO-6
2 x 1o-7
>2.Q x 1o-4
6 x 1O-5
TABLE 2 MQPA-derivatives and Inhibition for Trypsin
I (2R,4R)-MQPA
!
Ki
(Ml
8.8 x 1o-6
(2R,4R)-MQPA-Me
1.5 x 1o-5
MQP
-5 1.1 x 10
(2R)-QPA
4.2 x lO-5
QP
5.3 x 1o-5
I5O values of compounds lwith H-D-Phe-Pip-Arg-pNA. 10 for thrombin were reported previously as 3.3, 1, 0.9, 1.5, It could be (8). 0.67, 1, 0.3, 0.1, 1 and 100 uM respectively values for thrombin would be proportional to Ki if the mode of inhibition is of the series of these inhibitors were reported to inhibit thrombin competitively with Thus, the ratio of the l/Ki or synthetic substrates (1,2,3). of each compound to that of compound 2 is given in parenl/I the2g.s as the relative inhibitory activity for easy comparison of the effects of the chemical modification on the inhibitory actiThe relative inhibitory activity between thrombin and trypsin. vities of compounds 1~8 for trypsin and thrombin were almost the petitively
456
ACTIVE SITES OF SERINE PROTEASES
TABLE Dansylarginine NH2\ NH/
a) b) c) d)
Derivatives
Vol. 45, No. 5
3 and
Inhibition
of Trypsin
C-NH-(CH2)3-CH-C0-~2
I
measured with H-D-Phe-Pip-Arg-pNA as substrate. relative inhibitory activity (R.I.A.). measured with 1 mM BANA as substrate. calculated from the values reported previously
(8).
portion as piperidine _p same, and changes in the C-terminal caused a gradual increase in methylpiperidine + ethylpiperidine On the other the relative inhibitory activity for both enzymes. inhibitory activities of compounds 9 and 10 hand, the relative different from each other. for these two enzymes were apparently
457
ACTIVE SITES OF SERINE PROTEASES
Vol. 45, No. 5
the relative inhibitory activity was In the case of thrombin, decreased from 10 to 1 with the structural change from ethylpiperidine (compound 8) to propylpiperidine (compound 9), and finally to 0.01 with phenylpiperidine (compound 10). In the case modifications caused less change in of trypsin, these structural the relative inhibitory activity, and there was, if any, a small effect in the opposite direction such as to increase the relative inhibitory activity. Effect of introduction of 4-phenyl group into piperidine was further studied as shown in Table 4. The mode of inhibition of these inhibitors for thrombin, trypsin, plasmin and glandular kallikrein was competitive with the synthetic peptide substrates used. 'Introduction of 4-phenyl group into piperidine in naphthalenesulfonylarginine compounds led to the same effect on the inhibitory activity for thrombin and trypsin as in dansyl arginine compounds shown in Table 3. In addition, introduction of 4-phenyl group into piperidine caused a similar effect on the inhibitory activity for plasmin and glandular kallikrein as for trypsin. These results provided valuable information about the active site geometry of these enzymes corresponding to the C-terminal portion of arginine compounds. First, the hydrophobic binding pocket, which was reported previously (3) and further confirmed in this report to be of very similar construction in trypsin and thrombin, was about 7.6 A long from the amide N corresponding to ethylpiperidine. Second, trypsin, plasmin and glandular kallikrein could allow a longer hydrophobic C-chain or phenyl group attached to piperidine but thrombin could not, indicating that
TABLE Phenyl
Introduction
into
the
4
Piperidine
Ring
and
Enzyme
Inhibition
NH,, ,C-NH-(CH2)3-YH-CO-R2 NH2
NH
I R1
Ki
R1 -So2 8 0 0
-
CM)
I
R2
thrombin
trvDsin __
plas ;min
kallikrein
3 N(;E3
,
-N%
0
2 x lo-5
1.2 x 1o-5
1 x
1o-3
3 x 10
-5
ACTIVE SITES OF SERINE
458
the portion next to the hydrophobic geometrically different in thrombin and it was termed the P site. II)
(2R,4R)-MQPA-trypsin
complex
Vol. 45, No. 5
binding pocket was stereofrom in the other enzymes,
model
The arginine side chain was placed into the specificity crystal pocket as follows. N5, C6, N7 and N8 of the (2R,4R)-MQPA were first oriented by superimposing C4-N5 of the (PR,4R)-MQPA Then, C -N crystal on CD-GE of Lys-151 of BPTI-trypsin complex. rotated to search for the position 4 a 2 and N5-C6 were subsequently which the atomic pair distances between N7 and Asp-189 ODl, N8 nearest to the and Asp-189 ODl, and N8 and Asp-189 0D2 approached Reavalues measured in Arg-151-BPTI-trypsinogen complex (13). and N were found at which the distances sonable positions for N to those in Argfrom Asp-189 ODl and ODZ were a8lmost identical 151-BPTI-trypsinogen complex. The (2R,4R)-4-methylpipecolic acid portion was rotated freely around Cl-Cl0 to search for a position with minimum van der and a very narrow range (9") was Waals contacts with trypsin, Next, the found to be allowed without van der Waals contacts. rotation angle around Clo-Na6 was checked, and it was found This (2R,4R)-4around C10-N26. possible to rotate up to 14 van der methylpipecolic acid position did not give intramolecular side chain (C -N ) conWaals contacts with Ng and the arginine ' *$ portion aci structed above. Thus, the (2R,4R)-4-methylpipecolic was fixed to the slight cavity constructed from Phe-41, Cys-42, of the His-57, Cys-58 and Ser-195, due to the extreme narrowness range of rotation angle. -0 ) was exThe tetrahydroquinolinesulfonyl portion (S amined for a possible position by rotating in t #I2 e or2a er of Cl-Ng, where it was in minimum contacts intraand S12-C N9-S12 acid molecularly with 'she argi nine side chain and methylpipecolic portion constructed above as well as intermolecularly with Three forms were found for the tetrahydroquinolinetrypsin.
Proposed
model I,
FIG. 1 for the (PR,4R)-MQPA-trypsin -, enzyme. inhibitor;
complex
459
ACTIVE SITES OF SERINE PROTEASES
Vol. 45, No. 5
(2R,4R) -MQPA-trypsin v
FIG. 2 complex model with , inhibitor; -,
4-phenyl enzyme.
group
in.trodu ted
sulfonyl portion without intra-molecular contacts. One was in contact with His-57, one was in contact with His-57, Trp-215 and Leu-99, and the other was in contact with Ser-214, Trp-215, Gly216 and Leu-99. Since His-57 was the constituent of the slight cavity to which the methylpipecolic acid portion came into close proximity, the last form which did not require any shift of His57 was taken as the most approvable model. The extent of contact was the largest at the atomic pairs of C and Trp-215 CB even where the distance was much longer than $8 e covalent bond length, and it would allow there to be no fatal collision considering the possibility that trypsin in the complex with (2R,4R)-MQPA might undergo a slight rearrangement from that in the complex with BPTI. Thus, the proposed model for the (PR,4R)-MQPA-trypsin complex was illustrated in Fig.1 which was constructed with a reasonably fitted arginine side chain, almost definitely fixed methylpipecolic acid portion and tetrahydroquinolinesulfonyl portion with minimum van der Waals contacts. In this model, 4methylpiperidine showed a good fit with the slight cavity constructed from Phe-41, Cys-42, His-57, Cys-58 and Ser-195, indicating that this site was the hydrophobic binding pocket. Introduction of 4-phenyl group into the piperidine ring in the (2R, 4R)-MQPA-trypsin model was shown in Fig.2. The 4-phenyl group extended to the trypsin surface constructed from Phe-41, His-57, Cys-58 and Tyr-59, indicating that this portion was the P site.
DISCUSSION The model of the (2R,4R)-MQPA-trypsin complex was generated by a simple, theoretical method in which van der Waals contacts were considered and the structure of trypsin was assumed to be rigid and identical to those in the complex with BPTI. Although
460
ACTIVE SITES OF SERINE PROTEASES
proteins tend to exhibit conformational motility in solution and trypsin complexed with (2R,4R)-MQPA might exhibit some rearrangement from BPTI-complexed trypsin, the method provided one of the most practically reasonable models which enabled us to visualize the binding site of trypsin for (2R,4R)-MQPA, to interpret the experimental data for the structure-activity relationship satisfactorily, and to substantialize the hydrophobic binding pocket and the P site. The amino acid sequences surrounding the His-57 loop and Ser-195 loop have been reported to be highly homologous in trypsin and thrombin (14). The present results strongly indicated that the cavities constructed from Phe(Leu)-41, Cys-42, His-57, Cys-58 and Ser-195 in these enzymes were very similarly arranged stereogeometrically. In the model, (2R)-COOH faced the solvent side, opposite to the hydrophobic binding pocket. The experimental finding that (PR)-COOH neither contributed to nor interfered with the interaction with the enzymes, was thus reasonably explained from the model. The structure-activity relationships as shown in Tables 3 and 4 suggested that the stereogeometry of the P site of thrombin might be much narrower than those of the other three serine Comparison of the amino acid sequences surrounding proteases. the P sites of these proteases had revealed that, only in thrombin, were there large insertions of 9 amino acids and carbohydthat the P site of rates next to 63 (14), raising the possibility thrombin had become much narrower than those of the other serine proteases owing to these bulky insertions. The fact that the arginine side chain of (2R,4R)-MQPA could be reasonably fitted in the specificity pocket of trypsin when Cl was superimposed on CA of Lys-151 of BPTI, implied that the side chain of Arg (P,) could enter into the specificity pocket of trypsin setting its CA at the same position as the CA of Lys-151 (P > of BPTI. In other words, it may safely be inferred that ei.ther Arg or Lys in P entered into the specificity pocket of trypsin setting each CA at almost the same position as the CA of Lys-151 of BPTI. portion was in In the model, the tetrahydroquinolinesulfonyl contact with Leu-99, Trp-215 and Gly-216 of trypsin and these Consequently, the were largely conserved in thrombin (14,15). tetrahydroquinolinesulfonyl portion of (2R,4R)-MQPA would fit well and interact with the binding portion of thrombin presumably constructed mainly from the amino acids involving Trp-215, resulting in increase in fluorescence intensity if it was substituted with dansyl group (16).
ACKNOWLEDGEMENTS We wish to express our deep thanks to Dr. T. Matsuzaki of us with the Mitsubishi Chemical Ind. Co. Ltd., for providing coordinates of the (2R,4R)-MQPA crystal prior to publication and
Vol. 45, No. 5
ACTIVE SITES OF SERINE PROTEASES
461
Dr. M. to Dr. N. Yasuoka of Himeji Institute of Technology, Matsushima of Osaka Medical College, Dr. H. Mori of Kobe Univerand a number of friends in sity, Dr. T. Taga of Kyoto University We are also our laboratories for their invaluable suggestions. the manuscript. indebted to Mrs. H. Nishi for preparing
REFERENCES 1.
OKAMOTO, S., HIJIKATA, A., KINJO, K., KIKUMOTO, R., OHKUBO, A novel series of synthetic TONOMURA, S. and TAMAO, Y. K thiombin-inhibitors having extremely potent and highly Kobe J. Med. Sci., 21, 43-51, 1975. selective action.
2.
HIJIKATA, A., OKAMOTO, S., MORI, E., KINJO, K., KIKUMOTO, Kinetic studies R TONOMURA, S., TAMAO, Y. and HARA, H. th;? selectivity of synthetic thrombin-inhibitor using Thrombos. Haemostas. synthetic peptide substrates. (Stuttgart), 2, 1039-1045, 1979.
on
3.
KIKUMOTO, R., TAMAO, Y., TEZUKA, T., TONOMURA, S., HARA, H., A. and OKAMOTO, S. Selsctive NINOMIYA, K., HIJIKATA, inhibition of thrombin by (2R,4R)-4-methyl-l-[N -[(3-methyl1,2,3,4-tetrahydro-8-quinolinyl)sulfonyl]-L-arginyl)]-2Biochemistry, G, 85-90, 1984. piperidinecarboxylic acid.
4.
FELDMANN, R.J., BING, D.H., FURIL, B.C. and FURIL, B. Interactive computer surface graphics approach to study of the active site of bovine trypsin. Biochemistry, 75, 54095412, 1978.
5.
RUEHLMANN, A., KUKLA, D., SCHWAGER, P., BARTELS, K. and Structure of the complex formed by bovine HUBER, R. trypsin and bovine pancreatic trypsin inhibitor: Crystal structure determination and stereochemistry of the contact J. Mol. Biol., 77, 417-436, 1973. region.
6.
HUBER, R., KUKLA, D., BODE, W., SCHWAGER, P., BARTELS, K., DEISENHOFER, J. and STEIGEMANN, W. Structure of the complex formed by bovine trypsin and bovine pancreatic trypsin inhibitor. II. Crystallographic refinement at 1.9 angstroms J. Mol. Biol., 89, 73-101, 1974. resolution.
7.
HUBER, R., KUKLA, D., STEIGEMANN, W., DEISENHOFER, J. and JONES, A. Structure of the complex formed by bovine trypsin and bovine pancreatic trypsin inhibitor: Refinement of the crystal structure analysis. in: Proteinase Inhibitors.
A short description of the possession of the hydrophobic binding pocket in common in trypsin, thrombin, acetylated thromto Cys-Cys-His in the bin, plasmin and FXa and its correspondence histidine loop, was presented at the 59th Annual Meeting of the Physiological Society of Japan, 1982 (17).
462
ACTIVE SITES OF SERENE PROTEASES
Frite, (Eds.)
H., Tschesche, H., Greene, New York: Springer Verlag,
Vol. 45, No. 5
L.J. and Truscheit, 1974, pp.497-512.
E.
8.
KIKUMOTO, R., TAMAO, Y., OHKUBO, K., TEZUKA, T., TONOMURA, OKAMOTO, S., FUNAHARA, Y. and HIJIKATA, A. S Thrombin of Ne-substituted Linhibitors. 2. Amide derivatives J. Med. Chem., 3, 830-836, 1980. arginine.
9.
KIKUMOTO, R., TAMAO, Y., OHKUBO, K., TEZUKA, T., TONOMURA, Thrombin inhibitors. S OKAMOTO, S. and HIJIKATA, A. 3:'Carboxyl-containing amide derivatives of No-substituted J. Med. Chem., 23, 1293-1299, 1980. L-arginine.
10.
MATSUZAKI,
11.
Obtained from the Protein Data Research, Osaka Univ., Osaka.
Bank,
12.
LUNDBLAD, R.L. A rapid method bovine thrombin and inhibition phenylmethylsulfonyl fuloride. 1971.
for the purification of of the purified enzyme with Biochemistry, 2, 2501-2506,
13.
BODE, W., WALTER, J., HUBER, R., WENZEL, H.R. and TSCHESCHE, H. The refined 2.2-A (0.22-nm) X-ray crystal structure of the ternary complex formedlky bovine valine-valine and the Arg trypsinogen, analogue of bovine pancreatic trypsin inhibitor. Eur. J. Biochem., 144, 185190, 1984.
14.
R.J. and MANN, K.G. ELION, J., DOWNING, M.R., BUTKOWSKI, Structure of human thrombin: Comparison with other serine proteases. in: Chemistry and Biology of Thrombin. Lundblad, B.L., Fenton II, J.W. and Mann, K.G. (Eds.) Ann Arbor Science, 1977, pp.97-111. Michigan:
15.
MAGNUSSON, S., PETERSON, T.E., SOTTRUP-JENSEN, L. and Complete primary structure of prothrombin: CLAEYS, H. structure and reactivity of ten carboxylated isolation, glutamic acid residues and regulation of prothrombin Proteases and biological activation by thrombin. in: Reich, E., Rifkin, D.B. and Shaw, E. (Eds.) Cold control. Spring Harbor Laboratory, 1975, pp. 123-149.
16.
NESHEIM, M.E., PRENDERGAST, F.G. and MANN, K.G. Interactions of a fluorescent active-site-directed inhibitor Dansylarginine N-(3-ethyl-1,5-pentanediyl) of thrombin: Biochemistry, 18, 996-1003, 1979. amide.
17.
Existence of the hydrophobic HIJIKATA, A. and OKAMOTO, S. pocket in common in the active sites of some serine proteases. J. Physiol. Sot. Japan, 44, 314, 1982.
T.
Private
communication,
1982. Institute
for Protein