- Email: [email protected]
Refined 1.8Å resolution crystal structure of the porcine ϵ-trypsin
BB
Biochi~ic~a
et Biophysica A~ta
ELSEVIER
Biochimica et Biophysica Acta 1209 (1994) 77-82
o
Refined 1.8 A resolution crystal structure of the porcine e-trypsin
~r
Qichen Huang *, Zhiping Wang, Yili Li 1, Shenping Liu, Youqi Tang Department of Chemistry, Peking University, Beijing 100871, China Received 4 November 1993; revised 31 May 1994
Abstract
Porcine e-trypsin is a three-chain inactivated trypsin from the limited autolysis of porcine /3-trypsin. It is cleaved at positions Lys6°-Ser 61 and Lys145-Ser 146. The crystal structure has been determined by using the molecular replacement method, and refined at 1.8 A resolution. The R-value of final model is 0.184. Comparison with the electron density map of porcine fl-trypsin (PTRY) in complex (BBIT), and with that of native bovine/3-trypsin (HTNA), revealed no obvious changes except at the autolysis positions, and no changes at the active center were observed. The autolysis at positions Lys6°-Ser 61 and LyslaS-Ser 146 does not affect the conformation of His-57 in the active center and therefore cannot explain for a loss in porcine e-trypsin activity. Keywords: e-Trypsin; Autolysis; Crystal structure; Electron density map; Activity; (Porcine)
1. Introduction
The crystal structures of some porcine and bovine serine proteinase, such as porcine kallikrein [2], porcine elastase [3], porcine /3-trypsin [4], and bovine /3-trypsin [6], have been determined and refined at high resolution (1.8 •~). Porcine trypsine as a serine proteinase, its active center is a charge transfer catalytic triad which is composed of His-57, Ser-195 and Asp-102, and it plays an important role in the catalytic mechanism of serine proteinases [5]. Ser-195 is the active center and the imido of His-57 increase the activity of Ser-195, but there are still some arguments about whether there is a hydrogen bond between His-57 and hydroxyl of Asp-102 [6,7]. The crystal structures of proteinase and their complexes with inhibitors have been studied in detail [8]. Maroux et al. [9] had showed indirectly that at moderate conditions, the proteinase probably lost activity if the peptide bond Lys 6°-
Abbreviations: EPT, porcine 8-trypsin; PTRY, porcine B-trypsin; MCTI-A, trypsin inhibitor A from bitter gourd seeds; BBIT, complex formed between PTRY and MCTI-A; HTNA, native bovine fl-trypsin; r.m.s., root-mean-square; B, isotropic temperature factor. *This work was supported by National Science Foundation of China for 'Important Chemical Problems in Life Processes, III-3-2'. * Corresponding author. Fax: +86 1 2501725. 1 Present address: Shanghai Institute of Organic Chemistry, Academia Sinica, Shanghai 200032, China. 0167-4838/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSD1 0 1 6 7 - 4 8 3 8 ( 9 4 ) 0 0 1 2 5 - Z
Ser 6t was broken, the main reason being that the Lys-60 and Ser-61 residues may have been too close to His-57, which is one of the three residues at the active center. In addition, the changes of the bond length between Lys 6°Ser 61 might affect the conformation of the active center. However, Maroux et al. did not obtain the autolysate of trypsin and could not give the crystal structure of this kind of trypsin, so there were no detailed descriptions of the effect of the autolysis at peptide bond Lys6°-Ser 61 on the active center geometry of trypsin. In this paper, we report the crystal structure of porcine e-trypsin (EPT), an uninhibited, inactivated triple-chain trypsin, which arises from porcine fl-trypsin by a limited autolysis method [1,22]. As we know, most crystal structures of trypsin are inhibited enzymes, because during the course of crystallization, uninhibited enzymes will undergo autolysis and then cause the crystallization to fail. It is interesting to compare activating and inactivating native structures. Our final result shows reliable accuracies. From the final 2 F o - F c or F o - F¢ Fourier maps, we found that the EPT molecule was autolyzed at positions Lys6°-Ser61 and t y s 1 4 5 - S e r 146. Comparison was made between our refined structure and the highly resolution refined structure of PTRY (porcine /3-trypsin in BBIT [4]) or H T N A (uninhibited native bovine fl-trypsin [6]), illustrate the affection of autolysis at Lys6°-Ser 61 on the conformation of the active center and its proteinase activity. The final refined coordinates and structure factors have been deposited with
78
Q. Huang et al. / Biochimica et Biophysica Acta 1209 (1994) 77-82
Table 1 Reflection data Set
Resolution (,~)
Equipment
Reflections measured
Reflections observed
Observed/ possible (%)
Rmerg a (%)
1 2 Total
2.6 1.8 1.8
Siemens X-200B Weissenberg b camera with IP
38046 31303
4885 11023 11739
81.8 64.2 68.3
6.1 6.5 9.2
a Rmerg=~lii_ < l > l / ~ i r b BL6A2 synchrotron radiation beamline (0.9 .~, X-ray) in Photon Factory, KEK, Tsukuba, Japan.
the Protein Data Bank, Brookhaven National Laboratory (reference: EPT).
2. Materials and methods Porcine 8-trypsin was prepared according to the reference [1] and [22]. Crystals were obtained using the 'hanging drop' vapour diffusion technique. EPT was dissolved in 0.01 M HAc-NaAc buffer (pH 4.3). The total protein concentration was 20 mg/ml. The precipitant and reservoir solution were 0.2 M sodium phosphate buffer containing 2 M ammonia sulfate (pH 6.0-7.4). The precipitant and protein solution were mixed into the hanging drop in ratios between 1:1 and 1:2 (v/v). After 3 days at 24°C, the precipitant were found in the drops. Prismatic crystals were obtained after about one month. The maximum size of the crystals.is 1.2 × 0.3 × 0.2 mm, and diffracts X-ray to almost 1.6 A. The cell constants are a = 76.9 A, b = 53.4 A, c = 46.6, a = f l = y = 9 0 ° . The space group is P212121, Z = 4 , V = 1.92.105 ,~3, Vm is 1.96 .~3/Da. The crystal contains one molecule per asymmetric unit. X-ray intensity data were collected with a Siemens X-200B area detector to 2.6 .~ and high-resolution intensitoy data were collected with a Weissenberg camera to 1.8 A at a synchrotron radiation beamline. The diffraction data were processed and merged by means of the PROTEIN program package by Steigemann [10] (see Table 1). The coordinates (PDB entry 1MCT) of refined 1.6 ~, resolution crystal structure of PTRY in its complex (BBIT) formed with MCIT-A [4] were used as the search model for molecular replacement calculations. By using the A L M N / C C P 4 program, a finer search for Euler angles space led to the Euler angles [11]: o
o
O1 = 95-0°, ~)2 = 52.5°, 03 = - 147-5°" The translation function was calculated using TFSG E N / C C P 4 program with a correctly oriented model molecule. The highest peak is about 4.40-times as high as the second peak and is proved to be the correct one by following refinement: o
o
o
8 X = 18.30A, B Y = 7.24A, 8Z = 18.64A The initial structural model was obtained with the rotation and translational operation according to the model
parameters for model molecule. Using a SGI 7 0 / G T graphic system and T O M / F R O D O software [12,13], the model was built and improved on the basis of 2 F o - F c and F o - F c Fourier maps. The models of two autolysis sites were based on the following procedure: (i) calculate Fc excluding the atoms of six residues around every terminals, (ii) calculate 2 F o - F~, F o - F c maps using above F~, (iii) fit a o n e - residue model for every terminal to these maps, (iv) calculate F¢ including the atoms of these new residues, and (v) repeat this procedure till whole model has been built. The model was refined with the energy restraint crystallographic refinement procedures in program X-PLOR [14,15]. After the crystallographic R-value had dropped to 0.333, group and restraint B-values were refined on 1.8 ,~ resolution data. Solvent molecules were inserted at stereochemically reasonable positions where the difference density exceeds 3. Finally the individual atomic B-values were also refined. The final R-value for 11691 reflections with 7.0-1.8 .~ resolution is 0.184. The final refinement characteristics is given in Table 2.
3. Results and discussion The overall structure features of EPT are very similar to those of PTRY [4], and of HTNA [6] (PDB entry 1TLD). Comparison of these structures was done using the computer program QUANTA. If these molecules are expressed
Table 2 Final model parameters of the EPT structure Number of active protein atoms Number of solvent atoms Number of metal ions Standard deviation (r.m.s.) from target values Bond length (,~) Bond angle (°) Torsion angle (o) Improper (,~) Resolution range used for refinement (/~) No. of unique reflections used for refinement R-value Estimated mean coordinate error from R-value according to Luzzati [18] (A) a Without omitting reflection data.
1527 72 1 0.010 1.26 27.52 2.08 7.0-1.8 11691 a 0.184 0.25
79
Q. Huang et al. / Biochimica et Biophysica Acta 1209 (1994) 77-82
in equivalent coordinate systems, their main-chain fold shows a high level of agreement, and the r.m.s, deviation between EPT and PTRY is 0.40 ,~ for all main-chain atoms, theolargest deviation which occurs in position Asn25 is 2.7 A except the autolysis positions. Those between EPT and HTNA are 0.47 A, Gly-78, 3.9 A, respectively. Fig. l(a,b) shows the comparisons of Ca backbone atoms of the EPT and PTRY, and EPT and HTNA. There are many solvent molecules in protein crystals, and they have much affect on the function of biological molecules. In an EPT crystal, solvent molecules occupy 39% of the entire crystal volume. Compared with HTNA [6], a low packing density orthogonal bovine fl-trypsin crystal (including 57% solvent molecules), and high packing density orthogonal bovine trypsin crystal [16] (including 44% solvent molecules), the EPT molecule is packed very tightly in the unit cell. This confirms the low Vrn value of the EPT crystal (only 1.96 ,~3/Da). The quantity of solvent molecules involved is almost the same as trigonal bovine trypsin crystal [17]. In our final model, 72 reliable water molecules and one Ca ion were found, with 18 water molecules common to both structures (deviations
less than 1.0 ,~). Every EPT molecule in the unit cell has contacts with 12 neighbour molecules. The number of independent contacts sites is 6. Fig. 2(a,b) shows the autolysis sites of EFT at Lys 6°Ser 61 and Lys145-Ser146. The density maps showed Lys-60 and Lys-145 were not in contact with Ser-61 and Ser-146, respectively. Although at the autolysis site Lys145-Ser146, the electron density map is noisy and broken, considering that this is a disordered region of trypsin, it is reasonable. In EPT and in HTNA the peptide from Ser-146 to Ser-150 has highest B-factors, (see Fig. 3 and ref. [6], Fig. 1), and in PTRY has same high B-factors value (see ref. [4], Fig. 4). Fig. 3 shows the B-values distribution of every aminoacid residue of EPT. The average B-values of EPT are little higher than that of PTRY or HTNA. This illustrate that the EPT has more disordered areas than those of PTRY or HTNA. It is related to the autolysis of peptide bond of Lys145-Ser146 and Lys6°-Ser 61. Peptide msp74-Asn 79 and L e u l l 4 - A r g 117 show high B values, these are also same as PTRY and HTNA. In our final refinement results, His-57, Ser-195 and Asp-102 have smaller temperature factors, so the accuracy of these residues can be reliably
(b) 99
!
99
Fig. 1. Main-chain atoms (Ca, C, N, O) overlay of (a) EPT (thick line) on PTRY in BBIT (thin line), and (b) EPT (thick line) on HTNA (thin line).
Q. Huang et al. / Biochimica et Biophysica Acta 1209 (1994) 77-82
80
accepted. Furthermore, the analysis about the conformation of the active center, especially His-57, is accurate. In 1969, Maroux and Desnulle found the specific peptide bond Arg117-Va1118 can break in bovine trypsin at moderate conditions, and this break did not decrease the activity of bovine trypsin. They also indirectly illustrated that some commercial crystallized bovine trypsins still had activity when peptide bonds tys145-tys 146 and Arg llTVal ~8 were both broken, but once the peptide bond between L y s 6 ° - S e r 61 breaks, the trypsin might lose activity. This is possibly because the peptide bond between Lys 6°-
(a)
is in such close a vicinity to His-57 that it may affect catalysis through the conformational changes. However, they did not obtain trypsins cleaved at positions Arg117-Va1118, t y s 1 4 5 - S e r 146 and/or Lys6°-Ser 61. In 1968, Schroeder and Shaw [19] used ion chromatography to separate commercial bovine trypsin, and got single-chain /3-trypsin and double-chain a-trypsin. They found the peptide bond Lys145-Ser 146 of a-trypsin had broken, and the activity did not have clear changes compared with /3-trypsin. In 1969, Smith and Shaw [20] made a-trypsin autolyze at pH 7.8 in the presence of CaC12, and the S e r 612
J
©
©
(b) f
TI-
Fig. 2. The 2 F o - F c maps based on the final model of two autolysis sites of EFT. (a) Site Lys~-Ser 61, (b) Site Lys145-Ser 146.
Q. Huang et al. /Biochimica et Biophysica Acta 1209 (1994) 77-82 80.0
Table 3 Comparison of geometry around the trypsin active center PTRY
HTNA
2.75 3.42 2.77 2.93 2.80 2.93 2.91
2.68 3.44 2.72 2.91 3.10 2.94 2.79
EFT
Trypsinogen
H-Bond distances (,~)
~-~ 60.0
His 57 ND1-Asp 1°20D2 His 57 ND1-Asp 1°20D1 Asp 1°20D2-Ser 214 OG His 57 N-Asp 1°20D1 His 57 NE2-Ser 195 OG Asp 189 OD1-Gly TM N Ser 19° N-Asp i s 9 0 D 2
< < 40.0
>
81
2.70 3.49 2.81 2.76 3.10 2.86 2.90
2.56 3.24 2.71 2.77 2.60 8.26 3.77
H-Bond angle (°)
20.0
I
o 0
25
75
50
100
RESIDUE
125
150
175
200
225
NUMBER
Fig. 3. Average main-chain atoms B-factor distribution of EFT.
peptide chain broke at the position kys188-Asp 189, the autolysate was named @trypsin which is a three-chain trypsin. Gabel and Kasche [21] showed that the specific activity of @trypsin was clearly lower than that of /3trypsin. Guo et al. [22] had isolated three autolysates from porcine /3-trypsin: a double-chain &trypsin which was
Ser 195 CB-OG-His 57 NE2
96.9
87.8
83.1
95.0
of Leu-99 (~2)
15.2
10.8
2.2
4.5
cleaved at position A.rgl17-vall18, a three-chain "y-trypsin, which breaks at the positions Lys159-Ala160 and Arg 117Va1118, and a three-chain g-trypsin which was cleaved at positions Lys159-Ala160 and Lys145-Ser 146. These three autolysates still had trypsin activities (for BAEE, 15000 U / m g , 12000 U / m g , 12000 U / m g , respectively). But until now, EFT is the first trypsin autolysate which break at positions L y s 6 ° - S e r 61 and LyslgS-Ser 146, and was separated, and the first crystal structure of uninhibited native inactivating trypsin.
,8,
Fig. 4. Comparison of the active center of (a) EFT (thick line) and PTRY (thin line), (b) EFT (thick line) and ItTNA (thin line).
82
Q. Huang et al. /Biochimica et Biophysica Acta 1209 (1994) 77-82
Because the refined structure is accurate, the present results enable to describe the conformation of the active center in EPT. From the detailed comparison and analysis of EPT with highly refined PTRY and HTNA structures, the conformation parameters of the active center are found to be similar to those of PTRY and HTNA, and in particular the conformation of His-57 showed little changes in this respect. So we conclude that autolysis at the two sites does not change the conformation of the active center and so cannot be held responsible for EPT lose its activity. We have noticed that residues Leu-99 of EPT have lower average B-factors of main-chain atoms, which are just at the entrance of the active slit of the enzyme, and this also appears in trypsinogen [23] (PDB entry 1TGN), an inactivated zymogen. On the other hand, those in PTRY and in HTNA, inhibited and uninhibited active enzymes are not so low. Probably, some flexibility of residue Leu-99 is important for the substrate to enter the active slit of the enzyme (see Fig. 1 and Table 3).
Acknowledgements We thank Prof. N. Sakabe and Dr. T. Nakagawa at PF, KEK in Tsukube, Japan for their help in collecting our X-ray data.
References [1] Li, Yili (1992) Ph.D. thesis, Peking University, Beijing, China. [2] Bode, W., Chen, Z., Bartels, K.S., Kutzbach, C., Schmidt-Kastner, G. and Huber, R. (1983) J. Mol. Biol. 164, 237-282.
[3] Mayer, E., Cole, G., Radhakrishnan, R. and Epp, O. (1988) Acta Crystallogr., B44, 26-38. [4] Huang, Q., Liu, S. and Tang, Y. (1993) J. Mol. Biol. 229, 10221036. [5] Blow, D.M. (1976) Acc. Chem. Res. 9, 145-152. [6] Bartunik, H.D., Snmmess, L.J. and Bartsh, H.H. (1989) J. Mol. Biol. 210, 813-828. [7] Warshel, A. and Russel, S. (1986) J. Am. Chem. Soc. 108, 65696579. [8] Chi, C.W., Tan, F.L., Chu, H.M. (1982) in Proteins in Biology and Medicine (Bradshaw, R.A., Hill, R.L., Tang, J., Liang, C., Tsao, T. and Tsou, C., eds.), pp. 341-362, Academic Press, New York. [9] Maroux, D. and Desnulle, P. (1969) Biochim. Biophys Acta 181, 59-72. [10] Steigmann, W. (1974) Ph.D. thesis, Technical Univ. Munich, Germany. [11] Crowther, R.A. (1972) in The Molecular Replacement Method (Rossmann, R.G., eds.), Int. Sci. Rev. Ser. No. 13, pp. 173-183, Gordon and Breach, New York. [12] Jones, T.A. (1978) J. Appl. Crystallogr. 11, 262-272. [13] Phigrath, J., Saper, M. and Quiocho, F.A. (1984) Methods and Application in Crystallographic Computing (Hall, S. and Ashida, T., eds.), pp. 403-407, Clarendon Press, London. [14] Brunger, A., Kuriyan, J. and Karplus, M. (1987) Science 235, 458-460. [15] Jack, A. and Levitt, M. (1978) Acta Crystallogr. A34, 931-935. [16] Walter, J., Steigemann, W. and Singh, T.P. (1982) Acta CrystaUogr. B38, 1462-1472. [17] Fehlhammer, H., Bode, W. and Huber, R. (1977) J. Mol. Biol. 111, 415-438. [18] Luzzati, V. (1952) Acta Crystallogr. 5, 802-810. [19] Schroeder, D.D. and Shaw, E. (1968) J. Mol. Chem. 243, 2943-2949. [20] Smith, R.L. and Shaw, E. (1969) J. Mol. Chem. 244, 4704-4707. [21] Gabel, D. and Kasche, V. (1973) Acta Chem. Scand. 27. 1971-1981. [22] Guo, H., Guan, Y. and Zhang, L. (1985) Chinese Biochem. J. 1, 53-59. [23] Kossiakoff, A.A., Chambers, J.L., Kay, L.M. and Stroud, R.M. (1977) Biochemistry 16, 654-660.
Biochi~ic~a
et Biophysica A~ta
ELSEVIER
Biochimica et Biophysica Acta 1209 (1994) 77-82
o
Refined 1.8 A resolution crystal structure of the porcine e-trypsin
~r
Qichen Huang *, Zhiping Wang, Yili Li 1, Shenping Liu, Youqi Tang Department of Chemistry, Peking University, Beijing 100871, China Received 4 November 1993; revised 31 May 1994
Abstract
Porcine e-trypsin is a three-chain inactivated trypsin from the limited autolysis of porcine /3-trypsin. It is cleaved at positions Lys6°-Ser 61 and Lys145-Ser 146. The crystal structure has been determined by using the molecular replacement method, and refined at 1.8 A resolution. The R-value of final model is 0.184. Comparison with the electron density map of porcine fl-trypsin (PTRY) in complex (BBIT), and with that of native bovine/3-trypsin (HTNA), revealed no obvious changes except at the autolysis positions, and no changes at the active center were observed. The autolysis at positions Lys6°-Ser 61 and LyslaS-Ser 146 does not affect the conformation of His-57 in the active center and therefore cannot explain for a loss in porcine e-trypsin activity. Keywords: e-Trypsin; Autolysis; Crystal structure; Electron density map; Activity; (Porcine)
1. Introduction
The crystal structures of some porcine and bovine serine proteinase, such as porcine kallikrein [2], porcine elastase [3], porcine /3-trypsin [4], and bovine /3-trypsin [6], have been determined and refined at high resolution (1.8 •~). Porcine trypsine as a serine proteinase, its active center is a charge transfer catalytic triad which is composed of His-57, Ser-195 and Asp-102, and it plays an important role in the catalytic mechanism of serine proteinases [5]. Ser-195 is the active center and the imido of His-57 increase the activity of Ser-195, but there are still some arguments about whether there is a hydrogen bond between His-57 and hydroxyl of Asp-102 [6,7]. The crystal structures of proteinase and their complexes with inhibitors have been studied in detail [8]. Maroux et al. [9] had showed indirectly that at moderate conditions, the proteinase probably lost activity if the peptide bond Lys 6°-
Abbreviations: EPT, porcine 8-trypsin; PTRY, porcine B-trypsin; MCTI-A, trypsin inhibitor A from bitter gourd seeds; BBIT, complex formed between PTRY and MCTI-A; HTNA, native bovine fl-trypsin; r.m.s., root-mean-square; B, isotropic temperature factor. *This work was supported by National Science Foundation of China for 'Important Chemical Problems in Life Processes, III-3-2'. * Corresponding author. Fax: +86 1 2501725. 1 Present address: Shanghai Institute of Organic Chemistry, Academia Sinica, Shanghai 200032, China. 0167-4838/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSD1 0 1 6 7 - 4 8 3 8 ( 9 4 ) 0 0 1 2 5 - Z
Ser 6t was broken, the main reason being that the Lys-60 and Ser-61 residues may have been too close to His-57, which is one of the three residues at the active center. In addition, the changes of the bond length between Lys 6°Ser 61 might affect the conformation of the active center. However, Maroux et al. did not obtain the autolysate of trypsin and could not give the crystal structure of this kind of trypsin, so there were no detailed descriptions of the effect of the autolysis at peptide bond Lys6°-Ser 61 on the active center geometry of trypsin. In this paper, we report the crystal structure of porcine e-trypsin (EPT), an uninhibited, inactivated triple-chain trypsin, which arises from porcine fl-trypsin by a limited autolysis method [1,22]. As we know, most crystal structures of trypsin are inhibited enzymes, because during the course of crystallization, uninhibited enzymes will undergo autolysis and then cause the crystallization to fail. It is interesting to compare activating and inactivating native structures. Our final result shows reliable accuracies. From the final 2 F o - F c or F o - F¢ Fourier maps, we found that the EPT molecule was autolyzed at positions Lys6°-Ser61 and t y s 1 4 5 - S e r 146. Comparison was made between our refined structure and the highly resolution refined structure of PTRY (porcine /3-trypsin in BBIT [4]) or H T N A (uninhibited native bovine fl-trypsin [6]), illustrate the affection of autolysis at Lys6°-Ser 61 on the conformation of the active center and its proteinase activity. The final refined coordinates and structure factors have been deposited with
78
Q. Huang et al. / Biochimica et Biophysica Acta 1209 (1994) 77-82
Table 1 Reflection data Set
Resolution (,~)
Equipment
Reflections measured
Reflections observed
Observed/ possible (%)
Rmerg a (%)
1 2 Total
2.6 1.8 1.8
Siemens X-200B Weissenberg b camera with IP
38046 31303
4885 11023 11739
81.8 64.2 68.3
6.1 6.5 9.2
a Rmerg=~lii_ < l > l / ~ i r b BL6A2 synchrotron radiation beamline (0.9 .~, X-ray) in Photon Factory, KEK, Tsukuba, Japan.
the Protein Data Bank, Brookhaven National Laboratory (reference: EPT).
2. Materials and methods Porcine 8-trypsin was prepared according to the reference [1] and [22]. Crystals were obtained using the 'hanging drop' vapour diffusion technique. EPT was dissolved in 0.01 M HAc-NaAc buffer (pH 4.3). The total protein concentration was 20 mg/ml. The precipitant and reservoir solution were 0.2 M sodium phosphate buffer containing 2 M ammonia sulfate (pH 6.0-7.4). The precipitant and protein solution were mixed into the hanging drop in ratios between 1:1 and 1:2 (v/v). After 3 days at 24°C, the precipitant were found in the drops. Prismatic crystals were obtained after about one month. The maximum size of the crystals.is 1.2 × 0.3 × 0.2 mm, and diffracts X-ray to almost 1.6 A. The cell constants are a = 76.9 A, b = 53.4 A, c = 46.6, a = f l = y = 9 0 ° . The space group is P212121, Z = 4 , V = 1.92.105 ,~3, Vm is 1.96 .~3/Da. The crystal contains one molecule per asymmetric unit. X-ray intensity data were collected with a Siemens X-200B area detector to 2.6 .~ and high-resolution intensitoy data were collected with a Weissenberg camera to 1.8 A at a synchrotron radiation beamline. The diffraction data were processed and merged by means of the PROTEIN program package by Steigemann [10] (see Table 1). The coordinates (PDB entry 1MCT) of refined 1.6 ~, resolution crystal structure of PTRY in its complex (BBIT) formed with MCIT-A [4] were used as the search model for molecular replacement calculations. By using the A L M N / C C P 4 program, a finer search for Euler angles space led to the Euler angles [11]: o
o
O1 = 95-0°, ~)2 = 52.5°, 03 = - 147-5°" The translation function was calculated using TFSG E N / C C P 4 program with a correctly oriented model molecule. The highest peak is about 4.40-times as high as the second peak and is proved to be the correct one by following refinement: o
o
o
8 X = 18.30A, B Y = 7.24A, 8Z = 18.64A The initial structural model was obtained with the rotation and translational operation according to the model
parameters for model molecule. Using a SGI 7 0 / G T graphic system and T O M / F R O D O software [12,13], the model was built and improved on the basis of 2 F o - F c and F o - F c Fourier maps. The models of two autolysis sites were based on the following procedure: (i) calculate Fc excluding the atoms of six residues around every terminals, (ii) calculate 2 F o - F~, F o - F c maps using above F~, (iii) fit a o n e - residue model for every terminal to these maps, (iv) calculate F¢ including the atoms of these new residues, and (v) repeat this procedure till whole model has been built. The model was refined with the energy restraint crystallographic refinement procedures in program X-PLOR [14,15]. After the crystallographic R-value had dropped to 0.333, group and restraint B-values were refined on 1.8 ,~ resolution data. Solvent molecules were inserted at stereochemically reasonable positions where the difference density exceeds 3. Finally the individual atomic B-values were also refined. The final R-value for 11691 reflections with 7.0-1.8 .~ resolution is 0.184. The final refinement characteristics is given in Table 2.
3. Results and discussion The overall structure features of EPT are very similar to those of PTRY [4], and of HTNA [6] (PDB entry 1TLD). Comparison of these structures was done using the computer program QUANTA. If these molecules are expressed
Table 2 Final model parameters of the EPT structure Number of active protein atoms Number of solvent atoms Number of metal ions Standard deviation (r.m.s.) from target values Bond length (,~) Bond angle (°) Torsion angle (o) Improper (,~) Resolution range used for refinement (/~) No. of unique reflections used for refinement R-value Estimated mean coordinate error from R-value according to Luzzati [18] (A) a Without omitting reflection data.
1527 72 1 0.010 1.26 27.52 2.08 7.0-1.8 11691 a 0.184 0.25
79
Q. Huang et al. / Biochimica et Biophysica Acta 1209 (1994) 77-82
in equivalent coordinate systems, their main-chain fold shows a high level of agreement, and the r.m.s, deviation between EPT and PTRY is 0.40 ,~ for all main-chain atoms, theolargest deviation which occurs in position Asn25 is 2.7 A except the autolysis positions. Those between EPT and HTNA are 0.47 A, Gly-78, 3.9 A, respectively. Fig. l(a,b) shows the comparisons of Ca backbone atoms of the EPT and PTRY, and EPT and HTNA. There are many solvent molecules in protein crystals, and they have much affect on the function of biological molecules. In an EPT crystal, solvent molecules occupy 39% of the entire crystal volume. Compared with HTNA [6], a low packing density orthogonal bovine fl-trypsin crystal (including 57% solvent molecules), and high packing density orthogonal bovine trypsin crystal [16] (including 44% solvent molecules), the EPT molecule is packed very tightly in the unit cell. This confirms the low Vrn value of the EPT crystal (only 1.96 ,~3/Da). The quantity of solvent molecules involved is almost the same as trigonal bovine trypsin crystal [17]. In our final model, 72 reliable water molecules and one Ca ion were found, with 18 water molecules common to both structures (deviations
less than 1.0 ,~). Every EPT molecule in the unit cell has contacts with 12 neighbour molecules. The number of independent contacts sites is 6. Fig. 2(a,b) shows the autolysis sites of EFT at Lys 6°Ser 61 and Lys145-Ser146. The density maps showed Lys-60 and Lys-145 were not in contact with Ser-61 and Ser-146, respectively. Although at the autolysis site Lys145-Ser146, the electron density map is noisy and broken, considering that this is a disordered region of trypsin, it is reasonable. In EPT and in HTNA the peptide from Ser-146 to Ser-150 has highest B-factors, (see Fig. 3 and ref. [6], Fig. 1), and in PTRY has same high B-factors value (see ref. [4], Fig. 4). Fig. 3 shows the B-values distribution of every aminoacid residue of EPT. The average B-values of EPT are little higher than that of PTRY or HTNA. This illustrate that the EPT has more disordered areas than those of PTRY or HTNA. It is related to the autolysis of peptide bond of Lys145-Ser146 and Lys6°-Ser 61. Peptide msp74-Asn 79 and L e u l l 4 - A r g 117 show high B values, these are also same as PTRY and HTNA. In our final refinement results, His-57, Ser-195 and Asp-102 have smaller temperature factors, so the accuracy of these residues can be reliably
(b) 99
!
99
Fig. 1. Main-chain atoms (Ca, C, N, O) overlay of (a) EPT (thick line) on PTRY in BBIT (thin line), and (b) EPT (thick line) on HTNA (thin line).
Q. Huang et al. / Biochimica et Biophysica Acta 1209 (1994) 77-82
80
accepted. Furthermore, the analysis about the conformation of the active center, especially His-57, is accurate. In 1969, Maroux and Desnulle found the specific peptide bond Arg117-Va1118 can break in bovine trypsin at moderate conditions, and this break did not decrease the activity of bovine trypsin. They also indirectly illustrated that some commercial crystallized bovine trypsins still had activity when peptide bonds tys145-tys 146 and Arg llTVal ~8 were both broken, but once the peptide bond between L y s 6 ° - S e r 61 breaks, the trypsin might lose activity. This is possibly because the peptide bond between Lys 6°-
(a)
is in such close a vicinity to His-57 that it may affect catalysis through the conformational changes. However, they did not obtain trypsins cleaved at positions Arg117-Va1118, t y s 1 4 5 - S e r 146 and/or Lys6°-Ser 61. In 1968, Schroeder and Shaw [19] used ion chromatography to separate commercial bovine trypsin, and got single-chain /3-trypsin and double-chain a-trypsin. They found the peptide bond Lys145-Ser 146 of a-trypsin had broken, and the activity did not have clear changes compared with /3-trypsin. In 1969, Smith and Shaw [20] made a-trypsin autolyze at pH 7.8 in the presence of CaC12, and the S e r 612
J
©
©
(b) f
TI-
Fig. 2. The 2 F o - F c maps based on the final model of two autolysis sites of EFT. (a) Site Lys~-Ser 61, (b) Site Lys145-Ser 146.
Q. Huang et al. /Biochimica et Biophysica Acta 1209 (1994) 77-82 80.0
Table 3 Comparison of geometry around the trypsin active center PTRY
HTNA
2.75 3.42 2.77 2.93 2.80 2.93 2.91
2.68 3.44 2.72 2.91 3.10 2.94 2.79
EFT
Trypsinogen
H-Bond distances (,~)
~-~ 60.0
His 57 ND1-Asp 1°20D2 His 57 ND1-Asp 1°20D1 Asp 1°20D2-Ser 214 OG His 57 N-Asp 1°20D1 His 57 NE2-Ser 195 OG Asp 189 OD1-Gly TM N Ser 19° N-Asp i s 9 0 D 2
< < 40.0
>
81
2.70 3.49 2.81 2.76 3.10 2.86 2.90
2.56 3.24 2.71 2.77 2.60 8.26 3.77
H-Bond angle (°)
20.0
I
o 0
25
75
50
100
RESIDUE
125
150
175
200
225
NUMBER
Fig. 3. Average main-chain atoms B-factor distribution of EFT.
peptide chain broke at the position kys188-Asp 189, the autolysate was named @trypsin which is a three-chain trypsin. Gabel and Kasche [21] showed that the specific activity of @trypsin was clearly lower than that of /3trypsin. Guo et al. [22] had isolated three autolysates from porcine /3-trypsin: a double-chain &trypsin which was
Ser 195 CB-OG-His 57 NE2
96.9
87.8
83.1
95.0
of Leu-99 (~2)
15.2
10.8
2.2
4.5
cleaved at position A.rgl17-vall18, a three-chain "y-trypsin, which breaks at the positions Lys159-Ala160 and Arg 117Va1118, and a three-chain g-trypsin which was cleaved at positions Lys159-Ala160 and Lys145-Ser 146. These three autolysates still had trypsin activities (for BAEE, 15000 U / m g , 12000 U / m g , 12000 U / m g , respectively). But until now, EFT is the first trypsin autolysate which break at positions L y s 6 ° - S e r 61 and LyslgS-Ser 146, and was separated, and the first crystal structure of uninhibited native inactivating trypsin.
,8,
Fig. 4. Comparison of the active center of (a) EFT (thick line) and PTRY (thin line), (b) EFT (thick line) and ItTNA (thin line).
82
Q. Huang et al. /Biochimica et Biophysica Acta 1209 (1994) 77-82
Because the refined structure is accurate, the present results enable to describe the conformation of the active center in EPT. From the detailed comparison and analysis of EPT with highly refined PTRY and HTNA structures, the conformation parameters of the active center are found to be similar to those of PTRY and HTNA, and in particular the conformation of His-57 showed little changes in this respect. So we conclude that autolysis at the two sites does not change the conformation of the active center and so cannot be held responsible for EPT lose its activity. We have noticed that residues Leu-99 of EPT have lower average B-factors of main-chain atoms, which are just at the entrance of the active slit of the enzyme, and this also appears in trypsinogen [23] (PDB entry 1TGN), an inactivated zymogen. On the other hand, those in PTRY and in HTNA, inhibited and uninhibited active enzymes are not so low. Probably, some flexibility of residue Leu-99 is important for the substrate to enter the active slit of the enzyme (see Fig. 1 and Table 3).
Acknowledgements We thank Prof. N. Sakabe and Dr. T. Nakagawa at PF, KEK in Tsukube, Japan for their help in collecting our X-ray data.
References [1] Li, Yili (1992) Ph.D. thesis, Peking University, Beijing, China. [2] Bode, W., Chen, Z., Bartels, K.S., Kutzbach, C., Schmidt-Kastner, G. and Huber, R. (1983) J. Mol. Biol. 164, 237-282.
[3] Mayer, E., Cole, G., Radhakrishnan, R. and Epp, O. (1988) Acta Crystallogr., B44, 26-38. [4] Huang, Q., Liu, S. and Tang, Y. (1993) J. Mol. Biol. 229, 10221036. [5] Blow, D.M. (1976) Acc. Chem. Res. 9, 145-152. [6] Bartunik, H.D., Snmmess, L.J. and Bartsh, H.H. (1989) J. Mol. Biol. 210, 813-828. [7] Warshel, A. and Russel, S. (1986) J. Am. Chem. Soc. 108, 65696579. [8] Chi, C.W., Tan, F.L., Chu, H.M. (1982) in Proteins in Biology and Medicine (Bradshaw, R.A., Hill, R.L., Tang, J., Liang, C., Tsao, T. and Tsou, C., eds.), pp. 341-362, Academic Press, New York. [9] Maroux, D. and Desnulle, P. (1969) Biochim. Biophys Acta 181, 59-72. [10] Steigmann, W. (1974) Ph.D. thesis, Technical Univ. Munich, Germany. [11] Crowther, R.A. (1972) in The Molecular Replacement Method (Rossmann, R.G., eds.), Int. Sci. Rev. Ser. No. 13, pp. 173-183, Gordon and Breach, New York. [12] Jones, T.A. (1978) J. Appl. Crystallogr. 11, 262-272. [13] Phigrath, J., Saper, M. and Quiocho, F.A. (1984) Methods and Application in Crystallographic Computing (Hall, S. and Ashida, T., eds.), pp. 403-407, Clarendon Press, London. [14] Brunger, A., Kuriyan, J. and Karplus, M. (1987) Science 235, 458-460. [15] Jack, A. and Levitt, M. (1978) Acta Crystallogr. A34, 931-935. [16] Walter, J., Steigemann, W. and Singh, T.P. (1982) Acta CrystaUogr. B38, 1462-1472. [17] Fehlhammer, H., Bode, W. and Huber, R. (1977) J. Mol. Biol. 111, 415-438. [18] Luzzati, V. (1952) Acta Crystallogr. 5, 802-810. [19] Schroeder, D.D. and Shaw, E. (1968) J. Mol. Chem. 243, 2943-2949. [20] Smith, R.L. and Shaw, E. (1969) J. Mol. Chem. 244, 4704-4707. [21] Gabel, D. and Kasche, V. (1973) Acta Chem. Scand. 27. 1971-1981. [22] Guo, H., Guan, Y. and Zhang, L. (1985) Chinese Biochem. J. 1, 53-59. [23] Kossiakoff, A.A., Chambers, J.L., Kay, L.M. and Stroud, R.M. (1977) Biochemistry 16, 654-660.