- Email: [email protected]
Clarification of substrate specificity of papain by crystal analyses of complexes with covalent-type inhibitors
ELSEVIER
Biochimicaet BiophysicaActa 1208 (1994) 268-276
et Biophysica
........
/~
ta
Clarification of substrate specificity of papain by crystal analyses of complexes with covalent-type inhibitors Keita Matsumoto a, Mitsuo Murata a, Shigeyuki Sumiya a, Kunihiro Kitamura a,* Toshimasa Ishida b a Research Center, Taisho Pharmaceutical Co. Ltd., 1-403 Yoshino-cho, Ohmiya, Saitama 330, Japan b Osaka University of Pharmaceutical Sciences, 2-10-65 Kawai, Matsubara, Osaka 580, Japan
Received 19 April 1994
Abstract
In order to investigate the stereo specificity of papain S~ subsites (n = 1-4) at the atomic level, two kinds of covalent-type inhibitors were designed based on the previous results on papain-E-64 and papain-E-64-c interactions, and their complex crystals with papain were analyzed by X-ray diffraction. The results show that the hydrophobic regions consisting of Val-133, Val-157 and Asp-158 and of Tyr-61, Gly-66 and Tyr-67 residues interact with the hydrophobic P2 and P3 side chains of inhibitors, thus indicating the function of the former and latter binding pockets as S2 and S 3 subsites, respectively. Furthermore, the X-ray analysis suggests that the papain has no definite Sn subsite of n >_4, and the S3-P3 hydrophobic interaction is significantly affected by the Pn side chain (n > 4) of both the substrate and the inhibitor. Keywords: Papain;Agmatine; Irreversibleproteinaseinhibitor; Complex;X-ray analysis; Substrate specificity
1. Introduction
Cysteine proteinases having reactive Cys residue at the active center are abundant in living cells and play important roles in intracellular proteolysis. Since the imbalance of their enzymatic activities causes serious diseases such as muscular dystrophy [1], osteoporosis [2] and tumor invasion [3], the development of low-molecular inhibitors that can moderately control the activity has been desired as useful therapeutic drugs. E-64 (1 in Fig. 1), N-[N-(L-3-trar, s-carboxyoxirane-2carbonyl)-L-leucyl]agmatine, which was isolated as a natural product from Aspergillus japonicus [4], is a potent and irreversible (covalent-type) inhibitor of many cysteine proteinases, and its inhibitory mechanism at the atomic level has been investigated based on the spectroscopic [5], molecular dynamics [6] and X-ray structural [7,8] analyses
* Correspondingauthor. Fax: + 81 48 6527254. 0167-4838/94/$07.00 © 1994 ElsevierScienceB.V. All rights reserved SSD! 0167-4838(94)00098-2
of the proteinase-inhibitor complexes. Recently, we reported two different binding modes of E-64-c (2), a potent inhibitor developed from E-64 [9], at the active site of papain by X-ray crystal analyses [10,11]; their binding modes are stereoscopically shown in Fig. 2. Judging from this binding diversity of inhibitor and the relatively large binding pocket of enzyme, it was proposed [11] that, contrary to general concept for the binding pocket of proteinase [12], the substrate-binding site of papain should not be simply divided into each S~ subsite (n = 2 and 3), but rather considered as a large hydrophobic pocket. This hypothesis would be important for the development of potent inhibitors against papain and other proteinases. In order to confirm this hypothesis and further make clear the substrate specificity of papain Sn subsite of n > 4, this paper deals with the X-ray crystal analyses of papain complexed with 3 and 4. Compound 3 was designed to make clear which of L-Phe side chain or benzylamide moieties is most profitable for the binding with S 2 or S 3 site, and compound 4 would answer the question whether or not the S n subsites of n > 4 play important roles in the binding with substrate or inhibitor.
269
K. Matsumoto et al. / Biochimica et Biophysica Acta 1208 (1994) 268-276
H3C-cH-CH3 I
0 CH2 NH2 .~-- N- C~.H-C--N- CH,-CH2- CH2-CH2-NH-- C/+
HOOC~'~"/Hx/ H #-I & H 0 1
NH2
HaC..cH-CH3 I
O CH2 ,CH3 II Y ~C-- N-C.H- C-- N- CH2- CH,- CH • f H A II H " "~CH:~'~"--7~-~ H 0 "~ HOOC" \ / "H 2 0 25
2411~26 523~212"/
O CH2 13 14 II s - - e 10 11 1 ~ C--N--U.H- C--N- CH2--~ /)1$ H ~,t II H \ \ //
H I= ~ . _ _ ~ 4 18
01
3
OH
0 CH2 0 II -= II .C--N- C.H-C--N-CH-C-NH-CH= HO0 0
,4
/
\
I CH2 I CH2 I
NH I
HNOC'~NH
I
O~N÷.o.
© CHa
C~12
Cys-25 S - C H = - C - ~ N H - c - C H - N - - C - - O - - C H = ~ - ~ 5 Fig. 1. Chemical structures of E-64 (1), E-64-c (2) and their analogues 3 and 4, together with the atomic numbering of 3 used in this work. The chemical structure of ZPACK (5), an inhibitor analogous to substrate, is also given.
270
K. Matsumoto et al. / Biochiraica et Biophysica Acta 1208 (1994) 268-276 Table 1
2. M a t e r i a l s a n d m e t h o d s
NMR chemical shifts vs. DMSO-d6 (ppm)
2.1. Syntheses o f 3 and 4 According to a general approach developed for the synthesis of 2 [13], compounds 3 and 4 were synthesized by condensing ethyl p-nitrophenyl L-trans-epoxysuccinate with the H - P h e - B A ( b e n z y l a m i d e ) and H - T y r - ( N C nitro)Arg-BA, and hydrolyzing, respectively. H - P h e - B A and H - T y r - ( N C n i t r o ) A r g - B A were prepared from respective L-components by using a conventional method of peptide chemistry. The purities ( > 95%) were checked by HPLC. 1H-NMR spectra were determined on a Varian VXR-200 instrument using tetramethylsilane as internal standard. Chemical shifts are given in ppm, and J values are given in Hz. Mass spectra were obtained on a J E O L JMS-SX102. The physical properties of these compounds are summarized in Table 1.
2.2. Inhibitory activities o f 3 and 4 for papain Inhibitory activities (ICs0, concentration of 50% inhibition) of 3 and 4 against papain in vitro were determined by the methods of Barrett et al. [14]. Inhibitors were neutralized with sodium bicarbonate to dissolve and were diluted with distilled water containing dimethylsulfoxide before the use. The enzymatic activity of papain after the incubation for 5 min with inhibitor was measured. Furthermore, the second-order rate constants k2(M - i s - 1)for inactivation of papain were determined as a function of incubation time [14,15]. Inhibitors were dissolved in 0.1 ml of dimethylsulfoxide, and diluted with water to make a 1.0 ml stock solution. These were further diluted with 0.1% Brij35. The reaction mixture containing 10 nM papain, 50 n M - 1 0 0 nM inhibitors, 50 m M phosphate buffer (pH 6.8), 1 m M EDTA, 8 m M dithiothreitol, and 0.1% Brij-35 was incubated at 40°C. Aliquots were removed at intervals of 30 s for determination of the residual activity using 2 0 / z M
/'~G]rtZ9 159
Part 1: the physical properties and inhibitory activities of 3 2.85 (1H, dd, J = 13.6, 9.4 Hz) 3.06 (1H, dd, J = 13.6, 5.4 l-Iz) 3.29 (1n, d, J = 1.8 Hz) 3.59 (1n, d, J = 1.8 Hz) 4.28 (2H, d, J = 5.9 Hz) 4.60 (1H, dt, J = 8.5, 4.6 I-Iz) 7.00-7.46 (10H, m) 8.62 (1H, t, J = 5.9 Hz) 13.04-13.98(1H, br) Molecular formula High-resolution mass spectral data FAB (re~e) ICs0 (nM)
C 20H 20N205 369.1463 (M ÷ + 1) 27 0.66.105
k 2 (M- is- 1)
Part 2: the physical properties and inhibitory activities of 4 1.30-1.85 (4H, m) 2.67 (1H, dd, J = 10.0, 14.0 Hz) 2.93 (1H, dd, J = 4.1, 14.0 I-Iz) 3.05-3.25 (2H, m) 3.23 (1H, d, J = 1.8 Hz) 3.53 (1H, d, J = 1.8 Hz) 4.15-4.43 (3H, m) 4.46-4.62 (1H, m) 6.64 (2H, d, J = 8.5 Hz) 7.02 (2H, d, J = 8.5 I-lz) 7.18-7.39 (5H, m) 7.80-8.20 (2H, br) 8.31 (1H, d, J = 6.7 Hz) 8.37 (1H, t, J = 6.2 Hz) 8.44 (1H, d, J = 8.5 l-Iz) 8.35-8.75 (1H, br) 9.00-9.60 (1H, br) Molecular formula High-resolution mass spectral data FAB ( m / e ) IC50 (nM) k2 (M- is- 1)
C26H 31N7O9 586.2260 (M ÷ + 1) 15 1.23.105
/~Glnl9 ~/~a25
159
"" ~sp158 "bVa1157
Fig. 2. Schematic view of two different binding modes (thick and thin lines) of E-64-c (2) at the papain binding pocket. Possible hydrogen bonds and electronic short contacts are represented by dotted lines
.....
L
272
K. Matsumoto et aL/ Biochimica et Biophysica Acta 1208 (1994) 268-276 Table 2 Summary of crystal data and stereochemically restrained least-squares refinements of 3 and 4
Z-phe-Arg-NMee as a substrate. These results are also given in Table 1.
2.3. Preparation and crystallization of papain-3 and papain-4 complexes
Crystal data Space group Cell constant a, A b, .~ c, .~ Resolution, .~ Total reflections independent reflections Rmerge,%
The purification and preparation procedures of the papain-inhibitor complexes were the same as those described previously [10]. The complex crystals, the complete enzymarie inhibition of which was confirmed by assaying caseinolyrie activity [16], were grown within few weeks by the vapor diffusion of 1.5% ( w / v ) complex solution, which was dissolved in 76 m M NaCI buffer containing 64% methanol/ethanol (2:1, v / v ) and was equilibrated against 0.1 M aminoethanol-HCl buffer containing 64% methanol/ethanol (2:1, v / v ) solution at room temperature (20°C).
3
4
P212121
P212121
42.85(2) 96.19(5) 50.22(4) 1.9 25005 14734 7.7
42.93(4) 96.02(2) 50.15(3) 1.9 33190 15001 4.2
8-1.9 3 12664 9951 1990 35.49 0.028 0.160
8-1.9 3 13696 10051 2010 28.94 0.010 0.128
Least-squares refinement at final state Resolution,/~ Fo/o-(Fo) No. of reflections No. of parameters No. of atoms Average of (Fo - Fc), e / A 2 RMS shifts R
2.4. Data collection X-ray data collections of both complex crystals were carried out with the same manner. A single crystal with dimensions of ca. 0.5 × 0.2 × 0.2 mm 3 (3) or 0.6 × 0.2 ×
~Olnl9
~ln19
*~ e/as
,~
S~r2(5 ~a1157
cms
$er205 Vail57 (a)
°•Glnl9
°•Glnl9
°'~a~I~*
BlelS$
o.,,.,, re25
Bls159
•
155 Set205¥allS? (b) Fig. 4. Stereo views of the binding modes of 3 (a) and 4 (b) at the papain active site. Possible hydrogen bonds or electrostatic short contacts are shown by dotted lines. Respective amino acids composing the papain active site are also labelled.
IC Matsumoto et aL / Biochimica et Biophysica Acta 1208 (1994) 268-276
0.2 mm 3 (4), which was sealed in a glass capillary with a small amount of mother liquor, was mounted on a MacScience DIP100 .area detector. Data collection to 1.9 ,~ resolution was performed with Cu Ka radiation (45 kV, 250 mA). Crystal data are summarized in Table 2.
273
ular replacement method. The atomic positions of 3 and 4 were traced on the difference Fourier maps. The stereochemically restrained least squares method formulated by the PROLSQ program [17] was used for the structure refinement. The manual revision for the mispositioned residues and for the wrong conformations were done by means of the FRODO program [18]. Solvent molecules were gradually included in the model structure, as the structural refinement was in progress, where the electron densities greater than 0.40 e / ~ 3 on the map were selected as solvent molecules. The refinement at the final stage is also summarized in Table 2.
2.5. Structure solution and refinement
Since both of complex crystals were isomorphous to the l~apain-E-64-oc complex crystal [10] (a = 42.9 A, b = 96.1 A, c = 50.1 A, P212121) , they were solved by the molec-
3
(a)
Tyr67
(b) )
)
(
(
(c) Fig. 5. Stereo drawings of edge-to-face interactions of Phe (a) and benzylamide (b) aromatic rings of 3 and Tyr (c) aromatic ring of 4 (c)with the hydrophobic residues of papain S2 and S3 subsites.
274
K Matsumotoet aL/ Biochimicaet BiophysicaActa 1208(1994)268-276
3. Results and discussion
3.1. Inhibitory activities of compounds 3 and 4
The isopropyl and isobutyl groups of E-64-c were replaced with benzene tings in 3 to further confirm the effect of the hydrophobic interactions with papain S n (n -- 2 and 3) subsites. Nearly the same k 2 order of 3 as that of E-64-c (k 2 = 2.05. 105M-is - 1, iC50 = 13 nM) indicates that (a) the function of E-64-c P2 and P3 aliphatic side chains is the same as that of benzene rings of 3, and (b) the S 2 and S 3 subsites of papain, which provide the binding pockets for the P2 and P3 side chains, respectively, have the same hydrophobic binding abilities (or capacities), although the bulkiness of P2 side chain has been proposed to be dependent on the activity [19]. On the other hand, it appears important upon considering the characteristic of papain subsites that 4 exhibits nearly the same k 2 value as that of E-64-c. This may reflect that (i) papain has a binding pocket which accommodates the P4 side chain of inhibitor or (ii) the substratespecificity (or inhibitory activity) of papain is dominated by only the S 1 ~ S 3 subsites, and the Sn subsite of n > 4 is not directly related with the activity. 3.2. Binding mode of compound 3 to papain
No notable difference was observed concerning the overall structure of papain, when the present complex was compared with the previous ones [10,11]. The electrondensity map of 3 is shown in Fig. 3(a) and also its binding mode at papain active site is shown in Fig. 4(a). Significant interactions of 3 with papain are summarized in Table 3. Similar to E-64 [7] and E-64-c [10], a covalent bond (C-S = 1.80 ]k) was formed between the epoxy C2 atom of' 2[ and the Cys-25 S~'H of papain, thus forming the R-confignrational C2 adduct of 3 with a free OH group at C3 atom. The carboxyl group in 3 formed three hydrogen
Table 3
bonds with the catalytic His-159 N ~1, Gin-19 N ~2 and Cys-25 N atoms, and the 05 and N6 atoms of 3 were participated in an hydrogen bond with Gly-66 N atom and an electrostatic short contact with Asp-158 O atom, respectively. Since similar atomic interactions have also been observed in the papain-E-64 and papain-E-64-c complexes, they could be said as an essential situation in the binding of E-64 analogue to papain. Judging from the X-ray crystal analyses of two polymorphic papain-E-64-c complexes [10,11], it has been proposed that the papain could not simply be divided into each of S 2 and S a subsites, because they can form a relatively large hydrophobic pocket which allows the interconversion of P2 and P3 residues of E-64-c (Fig. 2). Present X-ray crystal analysis of papain-3 complex shows, however, that the benzene rings of Phe (P2) and terminal benzylamide (P3) of 3 were located at the respective hydrophobic pockets consisting of Val-133, Va1-157 and Asp-158 residues and of Tyr-61, Gly-66 and Tyr-67 residues, corresponding to S 2 and S 3 subsites, respectively. It appears interesting to note that the S2-P 2 and S3-P 3 interaction modes are very similar to those of ZPACK (a covalent-type inhibitor analogous to substrate) in the papain complex [20], irrespective of the reverse peptide bonding between both compounds (Fig. 1) and of thus the different hydrogen bondings in the papain binding pocket. This indicates that in addition to significant interactions listed in Table 3, the hydrophobic interactions in which the S n of n = 2 and 3 subsites participate are also essential for the binding with substrate or inhibitor. The two benzene rings of 3 were both held through the edge-to-face interactions, an well-recognized mode in aromatic-aromatic interaction [21]. As is shown in Fig. 5(a), Val-133 isobutyl side chain is located on the Phe benzene ring with a distance of 4.41 .~ (center-to-center) and an angle of 56.9 °. Similar edge-to-face interaction has also been observed in the Phe residue of ZPACK [20]. On the other hand, the benzene ring of benzylamide group in 3 is held by edge-to-face interaction with two neighboring Tyr-61 and Tyr-67 as is shown in Fig. 5(b); the center-tocenter distance = 5.11 .A for Tyr-61 and 5.90 ~, for Tyr-67 and dihedral angle = 81.9 ° for Tyr-61 and 89.8 ° for Tyr-67. In the case of ZPACK, the terminal benzene ring is held by the face-to-face interaction with Tyr-61.
Sjm~iflcant atomicdistances(A) of 3 and 4 with papain activesite
Typeof interaction
Atom Residue of inlfibitor of papain
3
4
Hydrogenbond 05 O19 020 020 032
N(GIy-66) 2.8 9 Nal(His-159) 2.8 4 N(Cys-25) 3.4 7 Nt2(Gin-19) 2.6 3 O~'H(Ser-205) -
2.6 9 2.9 1
N6
O(Asp-158) 3.6 9 3.6 2
C2
$~Cys-~)
Electrmtaticshortcontact Covalentbond 1.80 1.83
3.3. Binding mode of 4 to papain
Compound 4 was designed to obtain the information concerning the S4-P4 binding mode. The Tyr residue was used as P2 residue, because it was suggested from the crystal analysis of 3 that the Ser-205 side chain is located at the position near to Val-133 and could form an hydrogen bond with the polar atom of P2 side chain, which would contribute greatly to the fixation of P2 residue at papain S e subsite. Contrary to our expectation, the crystal structure of the complex did not show the entire structure
K. Matsumoto et al. / Biochimica et Biophysica Acta 1208 (1994) 268-276
of 4, as is obvious from the electron-density map in Fig. 3(b). Since the present X-ray accuracy (R value = 0.128) by the structural refinement using 13 696 independent reflections is in an usual range, the unobserved moiety could be interpreted as being in the disordered state. The binding mode of the clearly identified moiety of 4 is shown in Fig. 4(b). Similar to 3, the 0 5 and N6 of 4 were participated in a hydrogen bond with Gly-66 N and an electrostatic short contact with Asp-158 O atoms, respectively (Table 3). The Tyr residue was located at the S 2 subsite, in which the edge-to-face interaction was formed between the phenol ring and Val-133 isopropyl group with a distance of 4.18 A (center-to-center) and an angle of 61.6 ° (Fig. 5(c)). The phenol O was further hydrogen-bonded to the Ser-205 OVH, as was expected. Since X-ray analysis of papain-E-64 complex [7] showed the binding of 4-guanidinobutane moiety of E-64 (Fig. 1) to S 3 hydrophobic pocket, the pocket size of S 3 subsite could be large enough to accommodate either of 4-(N ~nitro)Arg or benzylamide residue of 4. Thus, the disappearance of this moiety would be interpreted as that the definite S 4 pocket is lacking in papain and this destabilizes the P3-$3 hydrophobic interaction, significantly, thus leading to high flexibility of this moiety. On the other hand, the carboxyl group of 4 was not observed in the electron-density map (Fig. 3(b)). As judged from (i) the thermal factor of the C2 atom, which is covalently bound to the Cys-25 S:' atom ( C - S = 1.80 ,~), is usual (B = 12.5 ,~2) and (ii) in the case of 3, the carboxyl group was stably fixed by three hydrogen bonds with neighboring His-159 N ~1, Gln-19 N "2 and Cys-25 N atoms, it would be unreasonable to consider that not observing this group results from its high flexibility. As an alternative reason, it may be presumed that a decarboxylation is taken place during the covalent bond formation of 4 with papain Cys-25 residue. We have no satisfactory explanation for it at present, and one should await a further detailed analysis. 3.4. B i o c h e m i c a l implication
Early studies on the substrate-specificity of papain suggested that the specificity is primarily governed by the S2-P 2 hydrophobic interaction [22]. The present results, together with previous X-ray analyses of papain-E64 [7] and papain-E-64-c [10,11] complex crystals, gave this insight a structural background in such a meaning that the S 2 subsite is a relatively large hydrophobic pocket consisting of Va1-133, Va1-157 and Asp-158, and the edge-to-face hydrophobic interaction is important for the fixation of P2 aromatic ring of substrate or inhibitor. Also it was made clear that the substitution of Tyr for Phe residue at P2 position is further stably locked at the S 2 subsite by an additional hydrogen-bond formation of Tyr O • • • Ser-205 OVH. On the other hand, S 3 subsite consisting of Tyr-61,
275
Gly-66 and Tyr-67 residues forms a hydrophobic binding pocket with nearly the same size as S 2 subsite. The binding with subsite P3 residue is carded out with the face-to-face or edge-to-face hydrophobic interaction. This S3-P 3 interaction could be as strong as the S2-P 2 interaction, when the inhibitor has no residue longer than PaHowever, the former interaction is largely affected by the structure of P~ in the case of n >_ 4. X-ray analysis of papain-4 complex suggests that papain has no definite Sn binding pocket of n >_ 4, and the P4 residue of 4 protrudes from the papain binding pocket and interfaces with solvent molecules without any specific interaction. Inhibitors 3 and 4 exhibited nearly the same bioactivities as E-64-c. Therefore, it could be concluded that the specificity and activity of substrate/inhibitor are mainly determined by the S~-P n interaction pattern of n _< 3.
Acknowledgments
We thank Mr. K. Isogai of our research center for measurements of the second-order rate constants of the used inhibitors to papain.
References
[1] Katunuma, N. and Kominami, E. (1987) Rev. Physiol. Biochem. Pharmacol. 108, 1-20. [2] Delaisse, J.M., Eeckhout, Y. and Vaes, G. (1984) Biochem. Biophys. Res. Commun. 125, 441-447. [3] Denhardt, D., Greenberg, A.H., Egan, S.E., Hamilton, R.T. and Wright, J.A. (1987) Oncogene2, 55-59. [4] Hanada, K., Tamai, M., Yamagishi, S., Ohmura, S., Sawada, J. and Tanaka, I. (1978) Agric. Biol. Chem. 42, 523-528. [5] Yabe, Y., Guillaume, D. and Rich, D.H. (1988) J. Am. Chem. Soc. 110, 4043-4044. [6] Yamamoto, D., Ohishi, H., Ishida, T., Inoue, M., Sumiya, S. and Kitamura, K. (1990) Chem. Pharm. Bull. 38, 2339-2343. [7] Varughese,K.I., Ahmed, F.R., Carey,P.R., Hasnain, S., Huber, C.P. and Storer, A.C. (1989) Biochemistry28, 1330-1332. [8] Vamghese, K.I., Su, Y., Cromwell, D., Hasnain, S. and Xuong, N. (1992) Biochemistry31, 5172-5176. [9] Miyahara, T., Shimojo, S., Toyohara, K., Imai, T., Miyajima, M., Honda, H., Kamegai, M., Ohzeki, M. and Komatsu, J. (1985) Rinshoyakuri 16, 537-546. [10] Yamamoto, D., Matsumoto, IC, Ohishi, H., Ishida, T., Inoue, M., Kitamura, K. and Mizuno, H. (1991) J. Biol. Chem. 266, 1477114777. [11] Kim, M.-J., Yamamoto,D., Matsumoto, K., Inoue, M., Ishida, T., Mizuno, H., Sumiya, S. and Kitamura, K. (1992) Biochem. J. 287, 797- 803. [12] Schechter, I. and Berger, A. (1967) Biochem. Biophys. Res. Commun. 27, 157-162. [13] Tamai, M., Yokoo, C., Murata, M., Oguma, K., Sota, K., Sato, E. and Kanaoka, Y. (1987) Chem. Pharm. Bull. 35, 1098-1104. [14] Barrett, A.J., Kembhavi,A.A., Brown, M.A., Kirschke, H., Knight, C.G., Tamai, M. and Hanada, IC (1982) Biochem. J. 201, 189-198. [15] Kit.z,R., Wilson, I.B. (1962) J. Biol. Chem. 237, 3245-3249. [16] Amon, R. (1970) Methods Enzymol. 19, 226-244. [17] Hendrickson,W.A. and Konnert,J.H. (1980) in BiomolecularStruc-
276
K. Matsumoto et al. / Biochimica et Biophysica Acta 1208 (1994) 268-276
ture, Function, Conformation and Evolution (Srinivasan, R., ed.), Vol. 1, pp. 43-57, Pergamon, Oxford. [18] Jones, T.A. (1978) J. Appi. Crystallogr. 11, 268-272. [19] Yamamoto, D., Matsumoto, K., Ohishi, H., Ishida, T., Inoue, M., Kitamura, K. and Hanada, K. (1990) FEBS Lett. 263, 134-136.
[20] Drenth, J., Kalk, K.H. And Swen, H.M. (t976) Biochemistry 15, 3731-3738. [21] Burley, S.K. and Petsko, G.A. (1985) Science 229, 23-28. [22] Berger, A. and Schechter, I. (1970) Phil. Trans. R. Soc. London, Ser. B257, 249-264.
Biochimicaet BiophysicaActa 1208 (1994) 268-276
et Biophysica
........
/~
ta
Clarification of substrate specificity of papain by crystal analyses of complexes with covalent-type inhibitors Keita Matsumoto a, Mitsuo Murata a, Shigeyuki Sumiya a, Kunihiro Kitamura a,* Toshimasa Ishida b a Research Center, Taisho Pharmaceutical Co. Ltd., 1-403 Yoshino-cho, Ohmiya, Saitama 330, Japan b Osaka University of Pharmaceutical Sciences, 2-10-65 Kawai, Matsubara, Osaka 580, Japan
Received 19 April 1994
Abstract
In order to investigate the stereo specificity of papain S~ subsites (n = 1-4) at the atomic level, two kinds of covalent-type inhibitors were designed based on the previous results on papain-E-64 and papain-E-64-c interactions, and their complex crystals with papain were analyzed by X-ray diffraction. The results show that the hydrophobic regions consisting of Val-133, Val-157 and Asp-158 and of Tyr-61, Gly-66 and Tyr-67 residues interact with the hydrophobic P2 and P3 side chains of inhibitors, thus indicating the function of the former and latter binding pockets as S2 and S 3 subsites, respectively. Furthermore, the X-ray analysis suggests that the papain has no definite Sn subsite of n >_4, and the S3-P3 hydrophobic interaction is significantly affected by the Pn side chain (n > 4) of both the substrate and the inhibitor. Keywords: Papain;Agmatine; Irreversibleproteinaseinhibitor; Complex;X-ray analysis; Substrate specificity
1. Introduction
Cysteine proteinases having reactive Cys residue at the active center are abundant in living cells and play important roles in intracellular proteolysis. Since the imbalance of their enzymatic activities causes serious diseases such as muscular dystrophy [1], osteoporosis [2] and tumor invasion [3], the development of low-molecular inhibitors that can moderately control the activity has been desired as useful therapeutic drugs. E-64 (1 in Fig. 1), N-[N-(L-3-trar, s-carboxyoxirane-2carbonyl)-L-leucyl]agmatine, which was isolated as a natural product from Aspergillus japonicus [4], is a potent and irreversible (covalent-type) inhibitor of many cysteine proteinases, and its inhibitory mechanism at the atomic level has been investigated based on the spectroscopic [5], molecular dynamics [6] and X-ray structural [7,8] analyses
* Correspondingauthor. Fax: + 81 48 6527254. 0167-4838/94/$07.00 © 1994 ElsevierScienceB.V. All rights reserved SSD! 0167-4838(94)00098-2
of the proteinase-inhibitor complexes. Recently, we reported two different binding modes of E-64-c (2), a potent inhibitor developed from E-64 [9], at the active site of papain by X-ray crystal analyses [10,11]; their binding modes are stereoscopically shown in Fig. 2. Judging from this binding diversity of inhibitor and the relatively large binding pocket of enzyme, it was proposed [11] that, contrary to general concept for the binding pocket of proteinase [12], the substrate-binding site of papain should not be simply divided into each S~ subsite (n = 2 and 3), but rather considered as a large hydrophobic pocket. This hypothesis would be important for the development of potent inhibitors against papain and other proteinases. In order to confirm this hypothesis and further make clear the substrate specificity of papain Sn subsite of n > 4, this paper deals with the X-ray crystal analyses of papain complexed with 3 and 4. Compound 3 was designed to make clear which of L-Phe side chain or benzylamide moieties is most profitable for the binding with S 2 or S 3 site, and compound 4 would answer the question whether or not the S n subsites of n > 4 play important roles in the binding with substrate or inhibitor.
269
K. Matsumoto et al. / Biochimica et Biophysica Acta 1208 (1994) 268-276
H3C-cH-CH3 I
0 CH2 NH2 .~-- N- C~.H-C--N- CH,-CH2- CH2-CH2-NH-- C/+
HOOC~'~"/Hx/ H #-I & H 0 1
NH2
HaC..cH-CH3 I
O CH2 ,CH3 II Y ~C-- N-C.H- C-- N- CH2- CH,- CH • f H A II H " "~CH:~'~"--7~-~ H 0 "~ HOOC" \ / "H 2 0 25
2411~26 523~212"/
O CH2 13 14 II s - - e 10 11 1 ~ C--N--U.H- C--N- CH2--~ /)1$ H ~,t II H \ \ //
H I= ~ . _ _ ~ 4 18
01
3
OH
0 CH2 0 II -= II .C--N- C.H-C--N-CH-C-NH-CH= HO0 0
,4
/
\
I CH2 I CH2 I
NH I
HNOC'~NH
I
O~N÷.o.
© CHa
C~12
Cys-25 S - C H = - C - ~ N H - c - C H - N - - C - - O - - C H = ~ - ~ 5 Fig. 1. Chemical structures of E-64 (1), E-64-c (2) and their analogues 3 and 4, together with the atomic numbering of 3 used in this work. The chemical structure of ZPACK (5), an inhibitor analogous to substrate, is also given.
270
K. Matsumoto et al. / Biochiraica et Biophysica Acta 1208 (1994) 268-276 Table 1
2. M a t e r i a l s a n d m e t h o d s
NMR chemical shifts vs. DMSO-d6 (ppm)
2.1. Syntheses o f 3 and 4 According to a general approach developed for the synthesis of 2 [13], compounds 3 and 4 were synthesized by condensing ethyl p-nitrophenyl L-trans-epoxysuccinate with the H - P h e - B A ( b e n z y l a m i d e ) and H - T y r - ( N C nitro)Arg-BA, and hydrolyzing, respectively. H - P h e - B A and H - T y r - ( N C n i t r o ) A r g - B A were prepared from respective L-components by using a conventional method of peptide chemistry. The purities ( > 95%) were checked by HPLC. 1H-NMR spectra were determined on a Varian VXR-200 instrument using tetramethylsilane as internal standard. Chemical shifts are given in ppm, and J values are given in Hz. Mass spectra were obtained on a J E O L JMS-SX102. The physical properties of these compounds are summarized in Table 1.
2.2. Inhibitory activities o f 3 and 4 for papain Inhibitory activities (ICs0, concentration of 50% inhibition) of 3 and 4 against papain in vitro were determined by the methods of Barrett et al. [14]. Inhibitors were neutralized with sodium bicarbonate to dissolve and were diluted with distilled water containing dimethylsulfoxide before the use. The enzymatic activity of papain after the incubation for 5 min with inhibitor was measured. Furthermore, the second-order rate constants k2(M - i s - 1)for inactivation of papain were determined as a function of incubation time [14,15]. Inhibitors were dissolved in 0.1 ml of dimethylsulfoxide, and diluted with water to make a 1.0 ml stock solution. These were further diluted with 0.1% Brij35. The reaction mixture containing 10 nM papain, 50 n M - 1 0 0 nM inhibitors, 50 m M phosphate buffer (pH 6.8), 1 m M EDTA, 8 m M dithiothreitol, and 0.1% Brij-35 was incubated at 40°C. Aliquots were removed at intervals of 30 s for determination of the residual activity using 2 0 / z M
/'~G]rtZ9 159
Part 1: the physical properties and inhibitory activities of 3 2.85 (1H, dd, J = 13.6, 9.4 Hz) 3.06 (1H, dd, J = 13.6, 5.4 l-Iz) 3.29 (1n, d, J = 1.8 Hz) 3.59 (1n, d, J = 1.8 Hz) 4.28 (2H, d, J = 5.9 Hz) 4.60 (1H, dt, J = 8.5, 4.6 I-Iz) 7.00-7.46 (10H, m) 8.62 (1H, t, J = 5.9 Hz) 13.04-13.98(1H, br) Molecular formula High-resolution mass spectral data FAB (re~e) ICs0 (nM)
C 20H 20N205 369.1463 (M ÷ + 1) 27 0.66.105
k 2 (M- is- 1)
Part 2: the physical properties and inhibitory activities of 4 1.30-1.85 (4H, m) 2.67 (1H, dd, J = 10.0, 14.0 Hz) 2.93 (1H, dd, J = 4.1, 14.0 I-Iz) 3.05-3.25 (2H, m) 3.23 (1H, d, J = 1.8 Hz) 3.53 (1H, d, J = 1.8 Hz) 4.15-4.43 (3H, m) 4.46-4.62 (1H, m) 6.64 (2H, d, J = 8.5 Hz) 7.02 (2H, d, J = 8.5 I-lz) 7.18-7.39 (5H, m) 7.80-8.20 (2H, br) 8.31 (1H, d, J = 6.7 Hz) 8.37 (1H, t, J = 6.2 Hz) 8.44 (1H, d, J = 8.5 l-Iz) 8.35-8.75 (1H, br) 9.00-9.60 (1H, br) Molecular formula High-resolution mass spectral data FAB ( m / e ) IC50 (nM) k2 (M- is- 1)
C26H 31N7O9 586.2260 (M ÷ + 1) 15 1.23.105
/~Glnl9 ~/~a25
159
"" ~sp158 "bVa1157
Fig. 2. Schematic view of two different binding modes (thick and thin lines) of E-64-c (2) at the papain binding pocket. Possible hydrogen bonds and electronic short contacts are represented by dotted lines
.....
L
272
K. Matsumoto et aL/ Biochimica et Biophysica Acta 1208 (1994) 268-276 Table 2 Summary of crystal data and stereochemically restrained least-squares refinements of 3 and 4
Z-phe-Arg-NMee as a substrate. These results are also given in Table 1.
2.3. Preparation and crystallization of papain-3 and papain-4 complexes
Crystal data Space group Cell constant a, A b, .~ c, .~ Resolution, .~ Total reflections independent reflections Rmerge,%
The purification and preparation procedures of the papain-inhibitor complexes were the same as those described previously [10]. The complex crystals, the complete enzymarie inhibition of which was confirmed by assaying caseinolyrie activity [16], were grown within few weeks by the vapor diffusion of 1.5% ( w / v ) complex solution, which was dissolved in 76 m M NaCI buffer containing 64% methanol/ethanol (2:1, v / v ) and was equilibrated against 0.1 M aminoethanol-HCl buffer containing 64% methanol/ethanol (2:1, v / v ) solution at room temperature (20°C).
3
4
P212121
P212121
42.85(2) 96.19(5) 50.22(4) 1.9 25005 14734 7.7
42.93(4) 96.02(2) 50.15(3) 1.9 33190 15001 4.2
8-1.9 3 12664 9951 1990 35.49 0.028 0.160
8-1.9 3 13696 10051 2010 28.94 0.010 0.128
Least-squares refinement at final state Resolution,/~ Fo/o-(Fo) No. of reflections No. of parameters No. of atoms Average of (Fo - Fc), e / A 2 RMS shifts R
2.4. Data collection X-ray data collections of both complex crystals were carried out with the same manner. A single crystal with dimensions of ca. 0.5 × 0.2 × 0.2 mm 3 (3) or 0.6 × 0.2 ×
~Olnl9
~ln19
*~ e/as
,~
S~r2(5 ~a1157
cms
$er205 Vail57 (a)
°•Glnl9
°•Glnl9
°'~a~I~*
BlelS$
o.,,.,, re25
Bls159
•
155 Set205¥allS? (b) Fig. 4. Stereo views of the binding modes of 3 (a) and 4 (b) at the papain active site. Possible hydrogen bonds or electrostatic short contacts are shown by dotted lines. Respective amino acids composing the papain active site are also labelled.
IC Matsumoto et aL / Biochimica et Biophysica Acta 1208 (1994) 268-276
0.2 mm 3 (4), which was sealed in a glass capillary with a small amount of mother liquor, was mounted on a MacScience DIP100 .area detector. Data collection to 1.9 ,~ resolution was performed with Cu Ka radiation (45 kV, 250 mA). Crystal data are summarized in Table 2.
273
ular replacement method. The atomic positions of 3 and 4 were traced on the difference Fourier maps. The stereochemically restrained least squares method formulated by the PROLSQ program [17] was used for the structure refinement. The manual revision for the mispositioned residues and for the wrong conformations were done by means of the FRODO program [18]. Solvent molecules were gradually included in the model structure, as the structural refinement was in progress, where the electron densities greater than 0.40 e / ~ 3 on the map were selected as solvent molecules. The refinement at the final stage is also summarized in Table 2.
2.5. Structure solution and refinement
Since both of complex crystals were isomorphous to the l~apain-E-64-oc complex crystal [10] (a = 42.9 A, b = 96.1 A, c = 50.1 A, P212121) , they were solved by the molec-
3
(a)
Tyr67
(b) )
)
(
(
(c) Fig. 5. Stereo drawings of edge-to-face interactions of Phe (a) and benzylamide (b) aromatic rings of 3 and Tyr (c) aromatic ring of 4 (c)with the hydrophobic residues of papain S2 and S3 subsites.
274
K Matsumotoet aL/ Biochimicaet BiophysicaActa 1208(1994)268-276
3. Results and discussion
3.1. Inhibitory activities of compounds 3 and 4
The isopropyl and isobutyl groups of E-64-c were replaced with benzene tings in 3 to further confirm the effect of the hydrophobic interactions with papain S n (n -- 2 and 3) subsites. Nearly the same k 2 order of 3 as that of E-64-c (k 2 = 2.05. 105M-is - 1, iC50 = 13 nM) indicates that (a) the function of E-64-c P2 and P3 aliphatic side chains is the same as that of benzene rings of 3, and (b) the S 2 and S 3 subsites of papain, which provide the binding pockets for the P2 and P3 side chains, respectively, have the same hydrophobic binding abilities (or capacities), although the bulkiness of P2 side chain has been proposed to be dependent on the activity [19]. On the other hand, it appears important upon considering the characteristic of papain subsites that 4 exhibits nearly the same k 2 value as that of E-64-c. This may reflect that (i) papain has a binding pocket which accommodates the P4 side chain of inhibitor or (ii) the substratespecificity (or inhibitory activity) of papain is dominated by only the S 1 ~ S 3 subsites, and the Sn subsite of n > 4 is not directly related with the activity. 3.2. Binding mode of compound 3 to papain
No notable difference was observed concerning the overall structure of papain, when the present complex was compared with the previous ones [10,11]. The electrondensity map of 3 is shown in Fig. 3(a) and also its binding mode at papain active site is shown in Fig. 4(a). Significant interactions of 3 with papain are summarized in Table 3. Similar to E-64 [7] and E-64-c [10], a covalent bond (C-S = 1.80 ]k) was formed between the epoxy C2 atom of' 2[ and the Cys-25 S~'H of papain, thus forming the R-confignrational C2 adduct of 3 with a free OH group at C3 atom. The carboxyl group in 3 formed three hydrogen
Table 3
bonds with the catalytic His-159 N ~1, Gin-19 N ~2 and Cys-25 N atoms, and the 05 and N6 atoms of 3 were participated in an hydrogen bond with Gly-66 N atom and an electrostatic short contact with Asp-158 O atom, respectively. Since similar atomic interactions have also been observed in the papain-E-64 and papain-E-64-c complexes, they could be said as an essential situation in the binding of E-64 analogue to papain. Judging from the X-ray crystal analyses of two polymorphic papain-E-64-c complexes [10,11], it has been proposed that the papain could not simply be divided into each of S 2 and S a subsites, because they can form a relatively large hydrophobic pocket which allows the interconversion of P2 and P3 residues of E-64-c (Fig. 2). Present X-ray crystal analysis of papain-3 complex shows, however, that the benzene rings of Phe (P2) and terminal benzylamide (P3) of 3 were located at the respective hydrophobic pockets consisting of Val-133, Va1-157 and Asp-158 residues and of Tyr-61, Gly-66 and Tyr-67 residues, corresponding to S 2 and S 3 subsites, respectively. It appears interesting to note that the S2-P 2 and S3-P 3 interaction modes are very similar to those of ZPACK (a covalent-type inhibitor analogous to substrate) in the papain complex [20], irrespective of the reverse peptide bonding between both compounds (Fig. 1) and of thus the different hydrogen bondings in the papain binding pocket. This indicates that in addition to significant interactions listed in Table 3, the hydrophobic interactions in which the S n of n = 2 and 3 subsites participate are also essential for the binding with substrate or inhibitor. The two benzene rings of 3 were both held through the edge-to-face interactions, an well-recognized mode in aromatic-aromatic interaction [21]. As is shown in Fig. 5(a), Val-133 isobutyl side chain is located on the Phe benzene ring with a distance of 4.41 .~ (center-to-center) and an angle of 56.9 °. Similar edge-to-face interaction has also been observed in the Phe residue of ZPACK [20]. On the other hand, the benzene ring of benzylamide group in 3 is held by edge-to-face interaction with two neighboring Tyr-61 and Tyr-67 as is shown in Fig. 5(b); the center-tocenter distance = 5.11 .A for Tyr-61 and 5.90 ~, for Tyr-67 and dihedral angle = 81.9 ° for Tyr-61 and 89.8 ° for Tyr-67. In the case of ZPACK, the terminal benzene ring is held by the face-to-face interaction with Tyr-61.
Sjm~iflcant atomicdistances(A) of 3 and 4 with papain activesite
Typeof interaction
Atom Residue of inlfibitor of papain
3
4
Hydrogenbond 05 O19 020 020 032
N(GIy-66) 2.8 9 Nal(His-159) 2.8 4 N(Cys-25) 3.4 7 Nt2(Gin-19) 2.6 3 O~'H(Ser-205) -
2.6 9 2.9 1
N6
O(Asp-158) 3.6 9 3.6 2
C2
$~Cys-~)
Electrmtaticshortcontact Covalentbond 1.80 1.83
3.3. Binding mode of 4 to papain
Compound 4 was designed to obtain the information concerning the S4-P4 binding mode. The Tyr residue was used as P2 residue, because it was suggested from the crystal analysis of 3 that the Ser-205 side chain is located at the position near to Val-133 and could form an hydrogen bond with the polar atom of P2 side chain, which would contribute greatly to the fixation of P2 residue at papain S e subsite. Contrary to our expectation, the crystal structure of the complex did not show the entire structure
K. Matsumoto et al. / Biochimica et Biophysica Acta 1208 (1994) 268-276
of 4, as is obvious from the electron-density map in Fig. 3(b). Since the present X-ray accuracy (R value = 0.128) by the structural refinement using 13 696 independent reflections is in an usual range, the unobserved moiety could be interpreted as being in the disordered state. The binding mode of the clearly identified moiety of 4 is shown in Fig. 4(b). Similar to 3, the 0 5 and N6 of 4 were participated in a hydrogen bond with Gly-66 N and an electrostatic short contact with Asp-158 O atoms, respectively (Table 3). The Tyr residue was located at the S 2 subsite, in which the edge-to-face interaction was formed between the phenol ring and Val-133 isopropyl group with a distance of 4.18 A (center-to-center) and an angle of 61.6 ° (Fig. 5(c)). The phenol O was further hydrogen-bonded to the Ser-205 OVH, as was expected. Since X-ray analysis of papain-E-64 complex [7] showed the binding of 4-guanidinobutane moiety of E-64 (Fig. 1) to S 3 hydrophobic pocket, the pocket size of S 3 subsite could be large enough to accommodate either of 4-(N ~nitro)Arg or benzylamide residue of 4. Thus, the disappearance of this moiety would be interpreted as that the definite S 4 pocket is lacking in papain and this destabilizes the P3-$3 hydrophobic interaction, significantly, thus leading to high flexibility of this moiety. On the other hand, the carboxyl group of 4 was not observed in the electron-density map (Fig. 3(b)). As judged from (i) the thermal factor of the C2 atom, which is covalently bound to the Cys-25 S:' atom ( C - S = 1.80 ,~), is usual (B = 12.5 ,~2) and (ii) in the case of 3, the carboxyl group was stably fixed by three hydrogen bonds with neighboring His-159 N ~1, Gln-19 N "2 and Cys-25 N atoms, it would be unreasonable to consider that not observing this group results from its high flexibility. As an alternative reason, it may be presumed that a decarboxylation is taken place during the covalent bond formation of 4 with papain Cys-25 residue. We have no satisfactory explanation for it at present, and one should await a further detailed analysis. 3.4. B i o c h e m i c a l implication
Early studies on the substrate-specificity of papain suggested that the specificity is primarily governed by the S2-P 2 hydrophobic interaction [22]. The present results, together with previous X-ray analyses of papain-E64 [7] and papain-E-64-c [10,11] complex crystals, gave this insight a structural background in such a meaning that the S 2 subsite is a relatively large hydrophobic pocket consisting of Va1-133, Va1-157 and Asp-158, and the edge-to-face hydrophobic interaction is important for the fixation of P2 aromatic ring of substrate or inhibitor. Also it was made clear that the substitution of Tyr for Phe residue at P2 position is further stably locked at the S 2 subsite by an additional hydrogen-bond formation of Tyr O • • • Ser-205 OVH. On the other hand, S 3 subsite consisting of Tyr-61,
275
Gly-66 and Tyr-67 residues forms a hydrophobic binding pocket with nearly the same size as S 2 subsite. The binding with subsite P3 residue is carded out with the face-to-face or edge-to-face hydrophobic interaction. This S3-P 3 interaction could be as strong as the S2-P 2 interaction, when the inhibitor has no residue longer than PaHowever, the former interaction is largely affected by the structure of P~ in the case of n >_ 4. X-ray analysis of papain-4 complex suggests that papain has no definite Sn binding pocket of n >_ 4, and the P4 residue of 4 protrudes from the papain binding pocket and interfaces with solvent molecules without any specific interaction. Inhibitors 3 and 4 exhibited nearly the same bioactivities as E-64-c. Therefore, it could be concluded that the specificity and activity of substrate/inhibitor are mainly determined by the S~-P n interaction pattern of n _< 3.
Acknowledgments
We thank Mr. K. Isogai of our research center for measurements of the second-order rate constants of the used inhibitors to papain.
References
[1] Katunuma, N. and Kominami, E. (1987) Rev. Physiol. Biochem. Pharmacol. 108, 1-20. [2] Delaisse, J.M., Eeckhout, Y. and Vaes, G. (1984) Biochem. Biophys. Res. Commun. 125, 441-447. [3] Denhardt, D., Greenberg, A.H., Egan, S.E., Hamilton, R.T. and Wright, J.A. (1987) Oncogene2, 55-59. [4] Hanada, K., Tamai, M., Yamagishi, S., Ohmura, S., Sawada, J. and Tanaka, I. (1978) Agric. Biol. Chem. 42, 523-528. [5] Yabe, Y., Guillaume, D. and Rich, D.H. (1988) J. Am. Chem. Soc. 110, 4043-4044. [6] Yamamoto, D., Ohishi, H., Ishida, T., Inoue, M., Sumiya, S. and Kitamura, K. (1990) Chem. Pharm. Bull. 38, 2339-2343. [7] Varughese,K.I., Ahmed, F.R., Carey,P.R., Hasnain, S., Huber, C.P. and Storer, A.C. (1989) Biochemistry28, 1330-1332. [8] Vamghese, K.I., Su, Y., Cromwell, D., Hasnain, S. and Xuong, N. (1992) Biochemistry31, 5172-5176. [9] Miyahara, T., Shimojo, S., Toyohara, K., Imai, T., Miyajima, M., Honda, H., Kamegai, M., Ohzeki, M. and Komatsu, J. (1985) Rinshoyakuri 16, 537-546. [10] Yamamoto, D., Matsumoto, IC, Ohishi, H., Ishida, T., Inoue, M., Kitamura, K. and Mizuno, H. (1991) J. Biol. Chem. 266, 1477114777. [11] Kim, M.-J., Yamamoto,D., Matsumoto, K., Inoue, M., Ishida, T., Mizuno, H., Sumiya, S. and Kitamura, K. (1992) Biochem. J. 287, 797- 803. [12] Schechter, I. and Berger, A. (1967) Biochem. Biophys. Res. Commun. 27, 157-162. [13] Tamai, M., Yokoo, C., Murata, M., Oguma, K., Sota, K., Sato, E. and Kanaoka, Y. (1987) Chem. Pharm. Bull. 35, 1098-1104. [14] Barrett, A.J., Kembhavi,A.A., Brown, M.A., Kirschke, H., Knight, C.G., Tamai, M. and Hanada, IC (1982) Biochem. J. 201, 189-198. [15] Kit.z,R., Wilson, I.B. (1962) J. Biol. Chem. 237, 3245-3249. [16] Amon, R. (1970) Methods Enzymol. 19, 226-244. [17] Hendrickson,W.A. and Konnert,J.H. (1980) in BiomolecularStruc-
276
K. Matsumoto et al. / Biochimica et Biophysica Acta 1208 (1994) 268-276
ture, Function, Conformation and Evolution (Srinivasan, R., ed.), Vol. 1, pp. 43-57, Pergamon, Oxford. [18] Jones, T.A. (1978) J. Appi. Crystallogr. 11, 268-272. [19] Yamamoto, D., Matsumoto, K., Ohishi, H., Ishida, T., Inoue, M., Kitamura, K. and Hanada, K. (1990) FEBS Lett. 263, 134-136.
[20] Drenth, J., Kalk, K.H. And Swen, H.M. (t976) Biochemistry 15, 3731-3738. [21] Burley, S.K. and Petsko, G.A. (1985) Science 229, 23-28. [22] Berger, A. and Schechter, I. (1970) Phil. Trans. R. Soc. London, Ser. B257, 249-264.