MICHIGAN STATEU H l I”: if HHHWMWNW(”MINI 3 19 300911 6892 ll! This is to certify that the dissertation entitled THE X-RAY CRYSTALLOGRAPHIC STRUCTURE DETERMINATION OF HUMAN PLASMINOGEN KRINGLE 4 AT 1.9 A RESOLUTION presented by Anne Marie Muiichak has been accepted towards fulfillment of the requirements for Ph . D . degree in Chemistry 6' TEL/@1647 Major proiessor [hue February 12, 1991 MS U i: an Affirmative Adieu/Equal Opportunity Institution 0-12771 ||—___—""'"w _, ‘ LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE MSU Is An Affirmative Action/Equal Opportunity Institution cmva}pt ETERMINATION OF THE X-RAY CRYSTALLOGRAPHIC STRUCTURE RESOLUTION HUMAN PLASMINOGEN KRINGLE 4 AT 1.9 By Anne Marie Mulichak A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1991 65’?— 43/8)? ABSTRACT THE X-RAY CRYSTALLOGRAPHIC STRUCTURE ETERMINATION OF HUMAN PLASMINOGEN KRINGLE 4 AT 1.9 RESOLUTION. by Anne Marie Mulichak Human plasminogen kringle 4 was crystallized from a solution of 30% PEG 8000, 0.12 M ammonium sulfate, 1.2 % dimethyl lormamide. The crystals were orthorhombic, space group P212121, with one molecule per asymmetric unit, and unit cell axes of a = 32.11 A, b=49.09 A, c: 49.39 A. Three-dimensional X-ray intensity data were collected at 1.9 A resolution. The structure was solved by the rotation-translation molecular replacement method, using coordinates of the highly homologous kringle from prothrombin fragment 1 as a model. Struc- ture refinement was performed by restrained least-squares. The final refined K4 structure including 97 solvent molecules has an Fi-factor of 14.2% and an average B value of 18 A2. Most of the K4 three-dimensional structure is well- defined, although the terminal regions of the peptide chain preceding Cyst and beyond CysBO are disordered. The overall peptide folding, internal side chain structure and intricatehydrogen-bonding network observed for K4 is highly conserved from the prothrombin kringle. Interesting features which were not previously observed for the PT structure include the presence of a cis Pro residue and two alternate conformations for the Cys75 side chain. The K4 domain is known to bind lysine and zwitterionic w-amino-carboxylic acids, and the observed binding-site complements this binding specificity, consisting of two oppositely charged Arg71 and Asp55/Asp57 ionic centers separated by a V-shaped hydrophobic trough formed by the indole side chains of Trp62 and Trp72. The orthorhombic crystal packing involves an intimate intermolecular interaction in which a highly ordered sulfate ion is coordinated by Arg and Lys side chains from three separate kringle molecules and in which the Ar932 side chain of a symmetry-related molecule occupies the binding site in a ligand-like manner. The diffraction pattern of the orthorhombic K4 crystals displays repro- ducible relative intensity changes due to X-ray exposure which indicate non- random structural changes occurring within the crystal lattice. A preliminary investigation suggests these changes may involve the disordered terminal regions of the peptide chain and the K4 side chains of the ligand-like intermolecular interaction. ACKNOWLEDGEMENTS Sincere thanks are extended to Dr. Alexander Tulinsky for his guidance and support throughout the course of this work. My stay in his laboratory has been a rewarding experience in many ways. I would also like to express my appreciation to Dr. K.G. Ravichandran for his assistance with the molecular replacement methods and to Dr. Miguel Llinas for providing the plasminogen kringle 4 protein sample. Thanks also to all members of the Tulinsky research group, in particular Dr. Ewa Skrzypczak-Jankun and Dr. K. Padmanabhan for their invaluable help and for sharing so much of their time and experience. Finally, special thanks to my family for their constant support and encourage- ment. in ii illiilmii‘ H __ ‘N‘I‘i ‘ Chapter VII. VIII. XI. XII. XIII. XIV. TABLE OF CONTENTS List of Tables .................... List of Figures .................... Introduction ..................... Crystallization .................... ACA Binding Experiments .............. Data Collection ................... Data Reduction ................... Molecular Replacement ............... Structure Refinement ................. Main Chain and Disulfide Structure .......... Side Chain Structure ................. Intermolecular Interactions and Crystal Packing . . . The Structure of the Lysine-Binding Site ........ Solvent Structure .................. Comparison of Plasminogen K4 and Prothrombin K1 . Comparison of Observed and Modeled Binding Sites . Investigation Into the Cause of Relative Intensity Changes Appendix A .................... Appendix B .................... List of References .................. Page vi viii Table LIST OF TABLES Comparison of kringle sequences aligned to show homolog Abbreviations used. PLG, plasminogen; PT, rothrombin; T A, tissue-plasminogen activator; F Xll, factor XI; UK, urokinase; VB, vampire bat salivary plasminogen activator; HGF, hepatocyte growth factor; APO apolipoprotein(a). ............ Initial results of factorial search for K4 crystallization conditions. Comparison of two plasminogen kringle 4 crystal forms. Intensity measurements of conformation-sensitive reflections taken prior to three-dimenSIonal data collection and between each resolution range of data collection. . . . . . . . Decay slopes determined based on various intensity measure— ments. Slope S calculated as (K~1)/(K-t), K=I,/I, . . . Distribution of structure factor magnitudes (F)' In various resolution shells tor K4 three- dimensional data collection. Highest cross rotation search results . . . . ..... Translation search results for rotation solution 1: (080,40). Highest correlation coefficients (C) are listed. . . . . . Translation search results for rotation solution 2: (85, 80 100) Highest correlation coefficients (C) are listed. . Translation search results for rotation solution 3: (145 125 1,65) Highest correlation coefficients (C) are listed. . . Results of PROFFT least-squares refinement at successive stages of resolution. Comparison of agreement of structure with(+ )andw wxhout (-) solvent molecules for various resolution ranges (8 0 R- factor Is defined by 2(IFOI- chl) IZI Fol. Observed seconda structural elements of K4 main chain. Reverse turns are c assified according to dihedral angles of residues 2 3 [63]. Results oi nmr structure are given for com- parison [37] -------------------- maximum). Page 18 36 44 45 50 52 53 54 58 62 73 14 15 16 17 20 21 22 23 24 25 Hydrogen-bondi interactions of the K4 main chain. Interactions forming )3- -sheet "83) or reverse tums( are indicated. Hydrogen atoms are assigned geometrically idealized positions. . . . . . 74 Geometric parameters of K4 disulfide bonds. . . . . . . . . . . 78 Accessible surface area of K4 main chain and side chains by residue, calculated using programs ACCESS and ACCFMT [65] with a spherical probeo radius 1.4 A. Percent accessible surface values are based on calculated accessible surface for the isolated residue. Asterisks indicate residues having disordered side chains. 85 Hydrogen bonding interactions involving K4 side chains. Hydrogen atoms were assigned geometrically idealized positions. Donor atom is denoted (D), acceptor atom (A). ............... 87 Hydrogen-bonding interactions between sulfate anion and protein side chains. Hydrogen atoms are assigned geometrically idealized positions. Primed residue numbers denote hydrogen donors .pro- vided by symmetry- -related K4 molecules. . . . . . . 124 Protein- solvent hydrogen bonds In which protein atom serves as hydrogen donor ( .................... 125 Protein- solvent hydrogen bonds' In which protein atom serves as hydrogen acceptor. Potential interactions were accepted for protein- solvent distances within 3. 5 A. ............. . 126 Probable solvent-solvent interactions having distances within 3.5 A. 132 Deviations In positions and torsion angles of conserved or highly homologous side chains of PGK4 and PTKI. No torsion angles are given for conserved Ala or Pro residues. . . . . . 142 Final results of PROFFT least-squares refinement at 2. 5 A resolution for second set of orthorhombic data collected after the observation of characteristic relative intensity changes. . . . . . . . . . . . 168 Distribution of magnitude of difference between structure factors calculated from two sets of orthorhombic K4 intensity data measured before and after characteristic relative intensity changes. . . . . . 169 Reflections having largest structure factor discrepancies between first and second K4 data collections ( Ditf2 > 100 where Diffz- - (F1 - F2)? ). ......................... 169 vii Figure 10 LIST OF FIGURES Triple—loop three-disulfide stmcture of the kringle domain. Peptide chain cross-links indicate S- S bonds. . Schematic diagram of plasminogen peptide chain, containing five kringles (K1 K5). Amino (N) and carboxy (C) termini are indicated. Bold chain cross- -links indicate disulfide bridges. Arrows designate 1) activation cleavage point and 2) autolytic cleavage point. ..................... Ligands of kringle Lys-binding site. Abbreviations: ACA, e-amino- caproic acid; AcLys, N-acetyl-L-lysine; AMCHA, trans-4-(amino- methyl)-cyclohexanecarboxylic acid; BASA, p-benzylamine sulfonic acid. ...................... Amino acid sequence and disulfide structure of plasminogen K4. Numbering follows convention of plasminogen K5, with open circles representing deletions with respect to KS. Amino acid types are designated using common one-letter abbreviations. . Schematic diagram of K4 peptide sequence, showing peptide stretches which define the Lys-binding site. Boxes indicate residues which are conserved with prothrombin kringle 1. Darkened circle indicates position of insertion in prothrombin K1 with respect to K4. ................... Stereoview of modeled K4 lysine-binding site [37]. Orthorhombic crystal of human plasminogen kringle 4. Large crystal Is approximately 1. 5 mm in length. . ......... Axial intensity distributions of orthorhombic K4 crystal displaying Type I difraction. X rays at 2000 Watts power (50 kV ,40 mA) for a and c axes, 4000 Watts (50 kV ,80 mA) for b axis. ...... Axial intensity distributions of orthorhombic K4 crystal displaying Type II diffraction X-rays at 2000 Watts power (50 kV, 40 mA). Intensity distributions of b and c axes for orthorhombic K4 crystal 38amAked' In 10mM ACA for four days. X‘rays at 2000 Watts (50kV, ........................... viii Page 13 23 24 11 12 14 15 16 2O 21 22 23 24 25 26 Intensity distributions of b axes for a) native and b) ACA-soaked. monoclInic K4 crystals. X-rays at 2000 Watts (50kV, 40mA).. . 27 Schematic diagram of orthorhombic K4 crystal morphology with respect to crystallographic axes. .............. Intensity distributions of a, b. and c axes for orthorhombic K4 crystal prior to intensity data collection. All axes measured using X-rays at 2000 Watts (50 kV, 40 mA). ............ Schematic diagram of four-circle diffractometer goniostat. The a), III, and x circles are used to orient the crystal. The detector' Is supported by the 26 circle. ................ 32 Intensity peak profile of reflection (1 9 1)versus omega. . . . . 34 Absorption correction curve showing the dependence of the (2 0 0) reflection intensity on III angle. ....... . . . 35 Decay curves for monitor reflections throughout intensity data collection. Resolution range 2- 32° began at 0 hrs, 32- 41° began at 23 hrs, 41 -48° began at 46 hrs, 48- 53° began at 71 hrs. . lntensit distributions of a, b, and c axes for orthorhombic K4 crystal) ollowing data collection. X-rays at 2000 Watts (50 kV, 40 Two- theta dependence of left (squares) and right (circles) back- ground measurements during intensity data collection. . . 41 Two-theta dependence of average negative intensity measure- ments during data collection. ............... 42 Euler angles which relate rotated axes x', y' and z' to their original positions x, y, and z. ................... Dependence of R- factor on scattering angle for final refined K4 structure at 1.9 A resolution. Dashed lines show theoretical coordinate error curves. .................. 64 Ramachandran plot of final refined K4 structure. Open circles denote Gly residues; filled circles denote non- Gly residues. Ener- getically preferred zones are outlined In dashed I.nes . . 66 Distribution of omega angles of refined K4 main chain peptide bonds (single cis peptide bond not included). . . Stereoview of 2IFOI- IFCI electron density for Pr033 which has a cis peptide bond configuration. ................ Stereoview of K4 CA, C, N backbone and disulfide (bold) structure. ........................ 69 27 28 29 3O 31 32 33 34 35 36 37 38 39 40. 41 42 Stereoview of K4 backbone showing Pro side chains and main chain hydrogen—bonding interactions. Hydrogen bonds are indicated wit dashed lines. ................ 7o Avera 9 thermal factors of K4 main chain (solid lines) and side chain dashed lines) structure. Breaks due to deletions with respect to KS numbering convention. ............ 76 Stereoview showing two alternate Cys75 side chain conforma- tions. .......................... 79 Stereoview showing two relative disulfide positions resulting from two Cys75 side chaIn orientations. ............. Stereoview showing a) face-on and b)edge-on views of K4 ructure. ........................ 84 Stereoview showing interactions of adjacent Argt 0 and ArgS2 side chains. Hydrogen bonds are indicated with dashed lines. . . . Stereoview showing hydrogen-bonding interactions (dashed lines) of Gln23 side chain with His31 main chain. ......... Stereoview showing interactions of Asn53 side chain. Potential hydrogen bonds to Asp5 O, Asp57 N, and Asp57 O are indicated with dashed lines ...................... 92 Stereoview of probable hydrogen bond (dashed line) between Lys35 main chain N and lone pair of His33 imidazole N. Hydrogen atoms are shown at geometric ideal positions. . . . . . . 94 Stereoview of K4 backbone and side chains forming internal hydrophobic core. .................... 95 Steroview of perpendicular aromatic stacking interactions between Trp25, Trp62, Ph964, Trp72 and Tyr74. . .......... Stereoviews of perpendicular interactions of a) Pr054 with Tyr41 ring and b) Pro61 with Tyr5 ring. .............. 99 Stereoview showing perpendicular interactions of Pr030 and Prose ........................... 100 Stereoview showing parallel stacking of His31 and His33 rings with Pr030 and Phe64 side chains. ............. 101 Stereoview showing well—defined electron density observed for Tyr9 side chain. ..................... 103 Stereoview showing interactions of sulfate anion with side chains of LysSS and Arg71 of molecule 1, Ly558' of molecule 2, and Arg32" of molecule 3. Additional interactions occur between Arg71/Asp57 and Asp55/Ar932“. . . . . . . . . . . . . . . 106 X 44 45 46 47 48 49 50 51 52 53 54 55 56 57 Stereoview illustrating two-fold screw axis in redirection resulting from trimolecular kringle-kringle interactions at sulfate anion. Sulfate shown in bold .................... 108 Stereoview of intermolecular interaction between Asn43 side chain and Asn76' side chain of symmet mate. Also shown is a solvent-bridged interaction between Gin and Asn76'Hydrogen bonds are indicated with dashed lines. ........... 109 K4 molecules related by two-fold screw axis along y-direction. Residues involved in intermolecular side chain hydrogen bonds are shown in bold ...................... 110 K4 molecules related by two-fold screw axis along z-direction. Solvent molecules which bridge side chains of adjacent symmetry mates are seen at protein-protein interfaces. ......... 111 Solvent-bridged interaction between side chain of Ser14 and Ser69' of neighboring molecule. Hydrogen bonds are indicated with dashed lines ...................... 1 12 Stereoview of K4 lysine-binding site. ............ 114 Stereoview showing ion pair interactions which occur at lysine- bindin site. Sulfate ion and side chains from second symmetry- relate molecule are shown in bold. ............ 118 Distribution of occupancies for ordered solvent molecules. Seven water molecules have an occupancy of 1.0. ......... 121 Distribution of solvent temperature factors. ......... 122 Electron density observed for sulfate anion in orthorhombic K4 crystal stmcture. ..................... 123 Stereoview of internal solvent molecule in K4 structrure. Dashed lines indicate hydrogen bonds which bridge Gln23 and Phe64 main chain atoms. .................... 128 Stereoview showing two solvent molecules found in an empty cavity off the surface of the K4 structure. Dashed lines indicate potential hydrogen bonds. ................. 131 RMS differences between K4 and PTK1 main chain (solid line) and side chain (dashed line) positions. Filled circles indicate conserved residues; open circles indicate highly homologous Tyr/Phe substitutions. ................... 135 Comearison of CA backbone and disulfide structures of K4 (bold) and TK1. ........................ 1 Comearison of CA,C,N,O main chain positions of K4 (bold) and TK1. ........................ 138 xi 59 60 61 62 63 64 65 66 67 68 69 7O 71 72 Stereoview com ring conformations of Pr030, which appears as a'sisomer' In 4 (bold) but as trans' Isomer in PTK1. . 139 Stereoview comparing conformations for Inner disulfides In K4 (bold) and PTK1. Only the more similar of two observed K4 Cys75 side chain orientations Is shown. .............. 141 Comparison of hydrophobic core residues in K4 (bold) and PTK1 structures. ........................ 145 Stereoview showing similar conformations and hydrogen-bonding interactions of conserved residues Glu23 and Asn49 In K4 (bold) and PTK1. ........................ 147 Stereoview comparing(conformations of conserved residues Args2, Asn53, and Asp55 In 4 (bold) and PTK1. Dashed lines indicate conserved hydrogen- bonding interactions ....... . 148 Stereoview comparing conformations of conserved Arg71 and homologous Gln/Glu34 residues in K4 (bold) and PTK1. . . . Stereoviews showing approximately 90° torsional differences in the conformations of a) er14 and b) Ser27' In K4 (bold) and PTK1. Dashed line indicates a hydrogen bonding interaction between Ser27 0G and Thr29 N atoms which is observed only In K4. . . 151 Stereoview comparing conformations of Arg10 in K4 (bold) and PTK1. Dashed lines indicate K4 hydrogen-bonding interactions not observed in PTK1 ...................... 152 Stereoview comparing conformations of Glu73 side chains' In K4 (bold) and PTK1. Both participate in similar hydrogen- bonding interactions with Thr16 side chain. ............. 153 Comparison of K4 lysine- binding site (bold) and corresponding residues of PTK1. .................... 155 Stereoview showing similar solvent positions observed” In K4 (bold) and PTK1. Solvent molecules, W3 and W18, labeled according to K4 quality factor ranking. ................. 156 Stereoview comparing conserved solvent position, W13, in K4 (bold) and PTK1. ..................... 158 Stereoview comparing positions of K4 internal solvent molecule (W1) and nearest solvent site of PTK1 structure. K4 residues are shown In bold. ..................... 159 Cflomparison of observed (bold) and modeled K4 lysine- binding sit es. .......................... 161 Fit of average structure factors from second orthorhombic K4 data collection, applying scale of 1.27 and AB correction of +4 (solid line), to those of original data collection (dashed line). . . . 166 xii 73 74 75 Distribution of scattering angles for set of 334 reflections having very large discrepancies between original and second ortho- rhombic data collections ( (|F1| - lei) > 100). ....... 170 Stereoview showin electron difference density observed near amino terminal of 4 pe tide chain when map is calculated using only 334 reflections havmg very large discrepancies between original and second intenSIty data sets. ........... 172 Stereoview showing electron difference density observed in vicinity of ligand-like intermolecular interaction when difference map is calculated using only 334 reflections having very large discrepancies between original and second intensity data sets. xiii I. INTRODUCTION The kringle stmcture is a type of protein folding domain, characterized by a unique pattern of three peptide loops formed by three disulfide bridges (Figure 1), which appears as a structurally and functionally independent unit of protein architecture [1,2]. Kringles are found repeatedly in the structures of certain proteins, particularly those blood plasma proteins involved in the formation and dissolution of blood clots. Usually preceding a catalytic domain of enzymes, kringles are known in many cases to have a role in protein recognition and enzyme specificity. Among proteins of the blood coagulation cascade, prothrombin includes two kringles [3], one of which mediates the binding of Factor Va [4]; clotting factor Xll carries a single kringle [5]. The fibrinolytic proenzyme plasminogen contains five kringles (Figure 2), which are involved in binding to the fibrin matrix of the blood clot and the inhibitor antiplasmin [6,7]. A number of acti- vators of the fibrinolytic pathway also contain kringle domains: urokinase, an activator of plasminogen found in urine, has one kringle [8,9] and tissue- plasminogen activator contains a pair [10], one of which also binds fibrin [11,12]. A kringle domain has also been identified in the structure of a plas- minogen activator found in vampire bat saliva and is believed to be involved Figure 1. Triple-loo three-disulfide structure of the kringle domain. Peptide chaIn cross-links in icate S-S bonds. 1 Iiilll‘i'.‘ e \ N 1 72",“ , '9“ “’1I K3 Figure 2 Schematic diagram of plasmin ogen peptide chain, containing five kringles (K1-.K5) Amino (N) )and carboxy ( ) termini are indicated. Bold chain cross-links indicate disulfide bridges. Arrows designate 1) activation cleavage points and 2) autolytic cleavage point. 3 in a molecular interaction with fibrin [13]. Recently the sequence of a platelet protein, hepatocyte growth factor, was found to indicate the presence of a series of four kringle domains [14]. However, most impressive is the series of 38 kringles found in the plasma glycoprotein apolipoprotein(a) [15]. The kringles contain 80-85 residues and have highly conserved amino acid sequences. A comparison of selected kringle sequences is shown in Table 1, from which it can be seen that about 25% of the residues are conserved absolutely. It is evident that the kringle is also an evolutionary unit, which has diverged to produce domains with unique binding characteristics while main- taining a similar overall structure. Based on the observed sequence homo- logies, an evolutionary scheme has been proposed relating a number of kringle-containing proteins [16]. An apparent lack of correspondence between the evolution of the kringle, catalytic, and other domains with which they are associated suggests that the kringle may have been transferred between proteins which had already diverged, and thus, that the kringle may be somewhat genetically autonomous. Plasminogen (Figure 2) is a 91,500 kD glycoprotein found in all body fluids and most tissues but primarily in the blood plasma. As part of the fibrinolytic pathway, plasminogen is the precursor of the enzyme plasmin, which is directly responsible for digestion of the fibrin matrix of blood clots into soluble frag- ments. Plasmin itself is not a highly selective enzyme and is capable of hydrolyzing a number of biological proteins; therefore, the regulation of the enzyme lies in its relationships with additional activator and inhibitor proteins. In the bloodstream, plasminogen is converted to plasmin by tissue-plasminogen activator (t-PA) [17], which is released by the endothelial cells of the blood vessel wall. This activation occurs efficiently only at the fibrin surface as it is enhanced by ternary complex formation, in which plasminogen and t-PA are PLG, UK,urokinase ogy. Abbreviations used aligned to show homol quences, kringle se Table 1. Comparison of plasminogen: PT, F XII, factor XII; touch tactcr; APO, . Relative se apolipo- quence numbers hepatocyte g la is listed peat of APO molecu 24-fold re aspect to PLG K5 given with r S P A T H P S S T S P H R P R E S K T K N G I T C 0 K W S C K T G N G K N Y R G T M PLG K1 uz the. In» 22 Bald 9.01 (4!“ 95K 20 313: (“K :23: IND: PIE-4 0 E at!!! min 333 3:010 000 UUU (4&9: xii-r UUU Inf-IE! >H> BEE-4 >E-‘E-t I‘mfl >I/I 2H 00 mm ><>< 2U) [:10 (90 ("Q (DU Dd: A>< ()0 me xx W N K N M E D L H R H I Y S T T V T G R T C Q A W S S M T P H S H S R T P E Y Y P N C Y H G N G O S Y R G T APO K2 80 I L E C 60 N Y C R N P D N D P O G P N Y C R N P D R E L R D P E K R Y D Y C D W C Y T T G L E E N L K K E PLG K1 N L D E X A PT K1 K HGF K1 x I S R C L T O C N P L I P N D Y C P D P G V R W E Y C N G L I M N Y C R N P D A V A A P Y C Y T R A APO K2 5 both bound to the fibrin substrate [18]. On the molecular level, activation involves cleavage of an Arg-Val bond of the proenzyme to produce a disulfide- stabilized two-chain structure (Figure 2). This is quickly followed by the autolytic cleavage and release of a 76-residue amino terminal activation peptide [19]. The activated plasmin consists of a light chain (MW ~25,000) containing the catalytic site, and a heavy chain (MW ~60,000) composed largely of the five kringle domains (MW ~10,000 each). Fibrinolysis of the blood clot then pro- ceeds as a series of cleavages at exposed Lys-X peptide bonds of fibrin. Once dissolution of the clot is complete, free plasmin is released into the blood plasma, where it is rapidly inactivated by the circulating inhibitor antiplasmin. Inactivation is a two—step process: rapid formation of a reversible complex between plasmin and antiplasmin, followed by a slower conversion to an irreversible complex [20]. Among these interactions, the plasminogen kringles have been implicated in the binding of the fibrin substrate [21] and antiplasmin [22], and, also, in the intramclecular binding of the activation peptide [7]. Specifically, these inter- actions occur via a number of lysine-binding sites found among the kringles, one strong site and 3-4 weaker sites [232425]. The strong binding site, found on kringle 1 (K1), appears to be important in the interactions with fibrin and antiplasmin [21,26]. A weaker site occurs on kringle 4 (K4), which also binds fibrin in the isolated state, although its function in the intact protein is uncertain [21]. In addition, K1 and K4 bind a number of ligands which serve as lysine analogs (Figure 3), including e-aminocaproic acid (ACA), N-acetyl-L-lysine (AcLys), trans-4-(amincmethyl)-cyclohexanecarbcxy|ic acid (AMCHA), and p-benzyl-amine-sulfcnic acid (BASA) [24,27]. Each of these ligands bears a positive and negative ionic group separated at a distance of approximately 6.8 A by a hydrocarbon chain or ring. The binding of these ligands by C0; CO} CO; 505 CH! erg—co-NH-CH err, CH, CH2 CH2 CH2 NH; NH; NH; *NHS sACA AcLys AMCHA BASA CH, i» I": CH, CH I C CH2 9H2 H2N4‘+‘§ NH2 I ’ NH, fNHs Hexylomme Benzomtdine Benzylcmme Figure 3. Ligands of kringle Lys-binding site. Abbreviations: ACA, e-amino- caproic acid; AcLys, N-acetyl-L-lysine; AMCHA, trans-4-(amino-methyl)- cyclohexanecarboxylic acid; BASA, p—benzylamine sulfonic acid. 7 plasminogen is known to have an antifibrinolytic effect in viva. Kringle 5 also carries a weak lysine-binding site with similar binding properties, but which shows no requirement for a negatively charged group, having a high affinity for non-zwitterionic ligands such as benzylamine, benzamidine, and hexylamine (Figure 3), as well as ACA [28]. It has been proposed that the K5 site may be important in the mechanism of fibrinolysis by binding internal lysyl side chains of the fibrin matrix to bring the plasmin catalytic site in contact with the substrate [28]. The K1 site then binds the carboxy terminal lysines which are formed during fibrin degradation. Relatively weak transient binding by the K5 site, coupled with alternation of the K1 site to newly formed terminal lysines, would allow plasmin to move efficiently along the fibrin chain during dissolution. The K5 site probably also corresponds to the "AH-site"[29], which appears to interact with the plasminogen activation peptide and thus to hold the molecule in a closed conformation [30]. Not only is the kringle domain of interest due to its recurrence as a structural unit among this family of proteins, but also the apparent regulatory role of certain kringles in important biological interactions suggests a variety of pharmaceutical opportunities. Thus, a considerable body of research has been concentrated on understanding the structure and behavior of kringles. Prior to this work, however, only the structure of bovine prothrombin kringle 1 (PTK1), as included in PT fragment 1, had been determined crystallographically. This structure was solved in this laboratory at 2.25 A, with an R-factor of 17%.[31,32] The prothrombin kringle, having the overall form of an oblate ellipsoid, displayed an intricate folding pattern with numerous B-turns, but otherwise little organized secondary structure. The structural "nucleus" of this folding appeared to include three elements: 1) a sulfur cluster near the center of gravity of the kringle, formed by the close approach of two disulfides, 2) antiparallel structure 1 [I 8 involving two strands of highly conserved residues, and 3) an aromatic stacked cluster [33]. ‘ Various alternate methods have been employed in the investigation of additional kringle structures, particularly those of plasminogen and t-PA. Probably the most thoroughly studied of these is plasminogen K4. The amino acid sequence and disulfide structure of K4 are shown in Figure 4. The K4 folding has been investigated by circular dichroism experiments [34], which indicated the presence of 64% B-structure, 30% B-turns, and 6% random coil. Specific regions of K4 secondary structure were proposed by Castellino et al. based on Chou-Fasman calculations [34], including B-sheet in the amino acid stretches Ala44-Asn53, Trp62-Thr66, and Val70-Tyr74 and twelve B-turns, with the greatest probabilities calculated at residues His3-Gly6, Asn53-Ala56, Asp67-Val70, and CysBO-Thr83. Seven of the twelve suggested turns coincide with [S-turns reported for the PTK1 folding. Results from aromatic H-nmr, acid-base titration, nuclear Overhauser, and two-dimensional correlated nmr experiments [35,36] have indicated the pre- sence of a buried hydrophobic core stabilizing the kringle fold. Centered on Leu46, this core is composed largely of an aromatic cluster formed by residues Trp25, Tyr41, Tyr50, and Trp62, most of which are strictly conserved, and may also include H1533 and Leu77. The interaction of a Met methyl group with these residues has also been suggested [35]. Models of the K4 backbone folding based on early nmr experiments involved mainly a clustering of the implicated hydrophobic core residues. Based on an observed cross-relaxation between Val(a) and Trp72 resonances, DeMarco et al. [35] also suggested that the amino-terminal peptide is folded back toward the kringle. Recently, a more detailed model of K4 was proposed by Atkinson and Williams [37] from observations of H-nmr spectra, and NOE connectivities. In 10 agreement with the previous work, these authors reported little regular secondary structure other than a short stretch of antiparallel B-sheet between residues Trp62-Ph964 and Tyr74-Trp72 and a possible [Hype interaction between Cys22 and Thr65. Turns in the protein backbone were predicted at Gly19, Ser26, Asn40, Asp55, Ala56, Ly558, Thr66, Ser69, and at each of the six prolines. The close proximity of a number of residues was suggested, including the hydrophobic core previously described, and hydrogen bonds were predicted between amino acid pairs Thr16/Glu73 and His31/Thr65. Only one of the conserved residues involved in PTK1 hyrdcgen~bonding showed any indication of a similar interaction in the K4 spectra. Nevertheless, a three- dimensional model of the K4 structure, generated by employing distance geometry algorithms with NOE-deduced interprcton distances, was found to be highly similar to the crystallographic PTK1 structure. Considerable effort has been directed at understanding the nature of the K4 lysine-binding site, and numerous binding and spectroscopic studies have identified the key residues involved. Chemical modifications of residues Asp57, Arg71 and Trp72 were found to prevent ACA binding [38,39]. Similarly, photo- oxidation of K4 histidine residues greatly reduced affinity for the ligand [40]. From such studies and a knowledge of the ligand-binding specificity, a model of K4 binding developed in which the ligand amino and carboxyl termini interact electrostatically with the complementary charged centers Asp55/Asp57 and HisS1/Arg71 [35,40,41]. In aromatic H-nmr experiments, ligand binding was found to significantly perturb the chemical shifts of Trp62, Ph964, and Trp72, suggesting that these residues are located at the surface of the binding site where they interact strongly with the ligand [41,42]. Also affected to a lesser extent were Trp25, HisSt, and Tyr41. Strong NOE connectivities were observed between these residues and those comprising the hydrophobic core [35], 11 implying that the binding site is contiguous to the latter and is thus primarily supported by the inner kringle loop. This conclusion is consistent with the observation that K4 binding ability is retained after reduction of the Cys1-Cy580 disulfide bridge [43]. Recently, the K4 lysine-binding site, defined by the amino acid stretches 31-35, 54-58, 61-64, and 71-75, was modeled by Tulinsky et al. [44] based on the three-dimensional structure of PTK1 and on information from H-nmr observations. As can be seen in Figure 5, the conservation in amino acid sequence between K4 and PTK1 is concentrated on the inner loop of the kringle, which supports the binding site. However, a potential complication to the modeling is an insertion of a Thr residue in PTK1 relative to K4 at position 59, adjacent to the binding site. According to the modeling procedure used, the basic conformations of the peptide backbone and conserved side chains of PTK1 were retained. Non-conserved K4 side chains for which the corresponding PTK1 residue was sufficiently similar were modeled to follow guide coordinates of the PTK1 side chain. Examples include the modeling of Ar932 (PT Lys), Gln34 (PT Glu), and Asp55 (PT Set). In some cases, the modeling was based primarily on nmr observations. In the cases of side chains for which there were no or poor guide coordinates and no additional information, an extended energy minimized conformation was used, as can be seen for Lys35 (PT lie) and Ly558 (PT Gly) (Figure 6). Finally, the binding site was subjected to energy minimization in both a free and ligand-bound state. The modeled K4 lysine-binding site, shown in Figure 6, is described as a relatively open depression lined by the lipophilic side chains of Trp62, Phe64, Trp72, with anionic Asp55/Asp57 and cationic Arg71 centers at the extremes. In the present work, the X-ray crystal structure determination of human 12 Figure 5. Schematic diagram of K4 peptide sequence, showing peptide stretches which define the Lys-binding site. Boxes indicate residues which are conserved with prothrombin kringle 1. Darkened circle indicates position of insertion in prothrombin K1 with respect to K4. 13 Figure 6. Stereoview of modeled K4 lysine—binding site [37]. 14 plasminogen kringle 4 was undertaken. In addition to obtaining a better understanding of kringle domains and their binding ability, a secondary goal of this research was to assess the validity of the methods used in the previous modeling study. Furthermore, the K4 structure should prove uniquely suited to the further modeling of the apolipoprotein(a) kringles, 37 of which display 75-85% conservation to K4 [15]. II. CRYSTALLIZATION Human plasminogen kringle 4 is isolated from elastase digestion of the enzyme as a heterogenous mixture, 70% having Ala86 and 30% having Val88 at the carboxy terminus [6]. The K4 sample used was provided in the form of a lyophilized powder by Dr. Miguel Llinas, and was stored at freezer tempera- tures. Kringle 4 was first crystallized in this laboratory by Dr. Chang Park [45] using the vapor diffusion method, from a drop consisting of one part 2-5 mg/ml aqueous protein solution to one part reservoir solution of 45% saturated ammonium sulfate in tris saline buffer, pH 7.4. The crystals appeared as small needles, tending to form rosettes, and could only be grown to an appreciable size with repeated seeding. They were characterized to be monoclinic, of space group P21, with four molecules per unit cell, two molecules per asymmetric unit. All crystals observed were found to be twinned along the a' axis. In an effort to find a new set of crystallization conditions which might produce larger, untwinned crystals, a factorial search was made with K4. The factorial search [46] is a method in which a limited set of solutions, representing a random but balanced distribution of common precipitants, salts and buffers, is used to make a rough survey of possible crystallization conditions. Solutions resulting in precipitation of the protein can then be used as a basis for more detailed crystallization experiments. A set of 31 factorial solutions was used, from a suggested list by Dr. Marcos Hatada [47]. Crystallization trials were performed using the vapor diffusion method with 5 III hanging drops consisting of one part 10 mg/ml K4 solution and one part factorial solution. Results of the initial factorial search are shown in Table 2. To test modified crystallization conditions while conserving protein, hanging drops producing amorphous solid or mlcrocrystals were removed from the well solution, allowed to equilibrate 15 r- .co_.m.ccE_.maxo .225. .2 com: 80.2.2.8 28.2.. 9.2.22 .m.m Id .2... .mx Io mode: 68 Id 20822 68 In 2220 6... Id 228... 8.326 282.0. 2.222828. 0mm 6.2.3 EacoEEo .m.<.o.m2omona SecoEEm .n..<..o_oocm.ceo-v.N-_2EoE-N no.2 8:98.853. 16 I I 22.8 .855 2 am I 5 4.8. 888-2528 noun. $8 88 cm... $8 22.5 See... 2 8 I 8. 8.8.88 82.2 ”own. $8 8.. own. $8 22.8 See... 2 8 I 8 .25 maOfiSEm l oEEm... v.82 .2 o; mono... 2 to on I 88 own. $8 22.5 282 2 8 288.2. 2 8 R I 288.8. $8 282 «2 2 8 was: 2 ..o 8 l 9.2 $8 228< See... 2 no a: 2 S 8 l I 8.8... «.2 2 8 2: 2 8 8 028.8. .2222 :2. .8883 $8 8820 8 2 .8 8282. 2 .8 8 28 222.5 2.2. Noun. $8 8.. own. $8 282 m2 2 8 a: 2 8 8 .8 89.228 8.. Own. $8 £520 .2 2 8 88.. 2 8 a l 9.2 $8 I scape. 2 8 8 88.88. 82.. 88 own. $8 22.8 See... 2 8 22.5 2 8 a. .2... 82.28.... 22.82.“. 282 2 am 228.. 85.5. 2 8 a.» 2 8 a. 1 22 $8 2282.2 85.5. 2 8 a: 2 S t I I 228 .3 2 3 228< 2 8 2 I 88 Gun. $8 223 .3 2 8 a: 2 to 2 5.8.8.8 omega ”.w.< 2 md 88 Own. $8 2.23m .coEE< s. —.o 238E. s. to v— ..8 .2208 2.8 88 0."... $8 8.520 No 2 8 8%.. 2 8 n. I § Own. $8 820 m2 2 8 a: 2 3 m. 0.2300: .0 mozomo. mumzamozm v.62 .2 ON 33.3w .coEE< E Nd maoI 2 2.0 «F. I 5.23 I 228<2 3 o. I 8.. own. $8 228< .8522. 2 8 22.6 2 3 m .28.. .88... :2. .o 22220 28 own. $8 2820 no 2 8 2222. 2 to a I I 22.0 «.2 2 N... l A I 8.. GP. $8 2820 m2 2 8 828... 2 .8 a 8:288» 82.. ”9.2 $8 9.2 $8 225 m2 2 8 822. 2 8 m 29950.2... .6 8.88. on .m< 3. ad I 2... 2 —.o v I .828 l I m in 2.1.0:.» 5505 ”2 m; l 2.25... 282 2 ad I N 6.28.8. sea... 2.: 9.2 $8 2820 220.8 2 8 2.82 2 to . 238m 2228.; .8 .23 .oz .mcozficoo COSMNEGES #X 50* cohmmw argue *0 22.60.. EEC. .N 03mg. 17 against a well of the buffer or H20 to redissolve the protein, and returned to a well of a similar crystallization solution having a lowered precipitant concentration. Similarly, drops producing no precipitate were re-equilibrated against a well solution having an incresed precipitant concentration. Based on such experimentation, subsequent efforts were largely concentrated on vari- ations of two sets of conditions: 1) 0.1 M Hepes buffer, 0.2 M ammonium sulfate. 2.0 M sodium potassium phosphate, and 2) 40% polyethylene glycol (PEG) 4000, 0.1 M ammonium sulfate. The K4 crystals produced with these conditions displayed a strong tendency to grow in rosettes or large clusters rather than individually. A number of organic solvents, including dioxane, n-butanol, dimethyl formamide (DMF), t-butyl alcohol and isoproponal, were added in small amounts (1-2%) to the well solutions, some of which appeared to promote the growth of separate independent crystals. After approximate conditions were determined, crystals were grown from 20-30 ul sitting drops. From these experiments, several different K4 crystal forms were grown under similar conditions. Monoclinic crystals isomorphous to those of Park were grown using a reservoir solution of 29-30% PEG 8000 (Fisher Carbowax flakes), 0.12 M ammonium sulfate, with 08-10% n-butanol as an additive, pH 6.2 and an initial protein concentration of 5 mg/ml. These lath-shaped crystals tended to grow to a larger size and were more likely to grow as individual crystals than those monoclinic crystals previously obtained but were crystal- lographicaliy identical to them. They have unit cell parameters of a=32.78(3) A, b=49.17(2) A, c=46.27(3) A, B=100°, with four molecules per unit cell and two molecules per asymmetric unit. They also have a relatively high protein fraction of 66% [45]. A second crystal form of K4 was grown under similar conditions, but using 1.2% dimethyl formamide as an additive, pH=6.0. These crystals are 18 orthorhombic, of space group P212121, with four molecules per unit cell, one molecule per asymmetric unit, and with unit cell parameters a=32.11(1) A, =49.09(2) A, c=49.39(3) A. They also have a high protein content of 62% [45]. The orthorhombic crystals are similar in morphology to the monoclinic form but are generally thicker in the smallest dimension and often show a characteristic end face development (Figure 7). A comparison of the monoclinic and ortho- rhombic crystal forms is given in Table 3. The two additives appeared to promote the growth of different crystal forms, as evidenced by the fact that orthorhombic crystals were grown using DMF, even though the solution was seeded with monoclinic crystals. However, this tendency is not absolute, and both forms have been found to grow in the presence of either organic solvent. On the other hand, the two forms have never been grown simultaneously. Less commonly, a third crystal form of K4 has also been grown under seemingly identical conditions, in the presence of either n-butanol or DMF, and in some cases along with developing orthorhombic crystals. The morphology of these crystals is that of flat plates, which may attain very large surface areas Table 3. Comparison of two plasminogen kringle 4 crystal forms. Crystal form Monoclinic Orthorhombic Space group P21 P212121 Moi/unit cell 4 4 Mol./asym. unit 2 1 Lattice Parameters a ( 32.78 32.11 b (A) 49.17 49.09 c (A) 46.27 49.39 B (deg) 100.67 90.0 Protein fraction (%) 66 62 Vm (A3/dalton) 1.86 1 .99 Figure 7. Orthorhombic crystal of human plasminogen kringle 4. Large crystal is approximately 1.5 mm in length. 20 (over 1 A2), but they are extremely thin. As observed with the other two crystal forms, they display a tendency to grow in clusters. Although surprisingly strong in the mother liquor, these crystals splinter when mounted: therefore, this crystal form was not characterized. Both monoclinic and orthorhombic forms of K4 diffract X-rays well, consistent with the high protein fractions, and are quite stable with respect to deterioration due to radiation exposure. However, the monoclinic crystals invariably exhibit the same twinning as was previously observed, with crystal and twin c‘ axes oriented 20 degrees apart. In addition, they are often further split. with secondary "c‘ axes" found within one to two degrees of both crystal and twin. The principal twinning alone greatly complicates the collection and interpretation of intensity data; the secondary splitting, at such small angular differences, renders the crystals useless. Due to the twinning complications, as well as the fact that the monoclinic crystals contain two rather than one kringle in the asymmetric unit, the orthorhombic form was chosen for the present investigation. Some additional points regarding the crystallization of orthorhombic K4 deserve mention. Firstly, as mentioned above, the crystallization conditions have a somewhat acidic pH relative to the physiological pH 7.0. Although an effort was made to incorporate buffer into the crystallization solution and thus ensure a neutral pH, this resulted in inferior crystal growth. Once grown, the crystals were stored in a solution having an increased PEG concentration of 38%. During experimentation with both crystallization and storing solutions it was discovered that an increase in ammonium sulfate concentration increases K4 solubility. As the crystallization appears to be highly sensitive to small changes in ionic strength, it was important not to raise the salt concentration in the storing solution. This sensitivity may also account for the difficulty encountered upon including buffer. Finally, a novel method of seeding was used during this work which proved to be particularly effective and reproducible. Although the original crystal growth generally succeeded in producing orthorhombic crystals of a size suitable for diffraction experiments, macroscopic seeding was used in many cases to further improve crystal size or to ensure growth of orthorhombic rather than monoclinic crystals. A reverse seeding method was used in which addi- tional dissolved protein was slowly introduced to the crystal, in contrast to the commonly used procedure of depositing a seed crystal into a previously prepared protein drop [48]. Crystals which had reached their maximum size in the original mother liquor were first transferred to a fresh 10 ul sitting drop of reservoir solution containing no dissolved protein. The crystal was then allowed to rest for one day to equilibrate with the new conditions, and in so doing, to allow the outermost layer of the crystal to dissolve, exposing a fresh surface for renewed growth. An approximately 10 mg/ml K4 solution was then added in 2-3 ul aliquots, allowing one day between additions, until crystal growth was observed. Generally, the addition of 56 ul protein solution was required to initiate growth, which continued for several days. Once crystal growth had ceased, the entire procedure could be repeated. The advantage of this method in the present case is that it provides a simple and reliable alternative to the time-consuming repetitive rinsing of seed crystals which is othenrvise necessary to expose fresh crystal surface and dissolve any microscopic crystalline material that may serve as a nucleation site for secondary crystal growth. However, it cannot be applied directly for all crystallization conditions but may require some experimentation to find an appropriate drop solution and resting period so that the seed crystal is not immediately dissolved. 22 During experiments with numerous orthorhombic K4 crystals, two signi- ficantly different diffraction patterns have been observed. Although most newly mounted crystals produce a single characteristic pattern (Type 1), in some cases, striking changes in relative intensities occur to give a second pattern (Type II), which appears to reflect some conformational change in the kringle structure. Figures 8 and 9 show the axial intensity distributions from a single orthorhombic K4 crystal as it was first observed (Figure 8) and after it had been exposed for approximately 20 hrs to X-rays and, due to problems with the original mount, had been removed from the capillary and remounted (Figure 9). Most notable are the intensity changes observed for reflections (2 0 0), (8 0 0), (0 014),(0 0 16), (0 0 18), (0 0 22), (O 10 0) and (0 16 0). In addition to these intensity changes, crystals displaying the Type ll diffraction tend to have slightly shorter unit cell dimensions, approximately a=32.15(2) A, b=49.01 (2) A, =49.09(3) A. This phenomenon has also been observed during intensity data collections. In these cases, the changes appear to be gradual and radiation dependent, as mounted crystals do not undergo relative intensity changes with time in the absense of X-ray exposure, even when a partial change has already been induced by previous exposure. 23 - . .26 a .2 258569 mum . ooom 8 3m. x cocomaEU. on? 92:8.qu Rambo 3. gwwmwommwww %%%Hmmmwmmwfimwb>xomv .638 28>) om om ON mF @— Om OF _ . . Jill OF 25 .22 .m 93E l‘ ‘lll ii I (I) 901 it A l m; A ~ h- I - I I m ”X“ x COFOmt—utn = QQ>_ QC_>W_QW_U _N~wao V! O_DEO£._OP:LO ho mCO;__AD_:w_U \xA/thWwLC— _m_x< .m wham mm on om 24 N— Op i (I) 90'] Ill. ACA-BINDING EXPERIMENTS Since ligand-binding by the kringle is a matter of great interest, a logical extension of this work is the structure determination of a kringle-ligand complex. In an effort to obtain isomorphous K4-ligand crystals, diffusion experiments were performed in which native K4 crystals were soaked in a mother liquor containing the ligand ACA. Due to the considerable solvent content of protein crystals, it is possible for small ligands to diffuse into the crystal lattice and form binding interactions with the crystalline protein. Crystals of native K4 in a 40 ul sitting drop of storing solution (38% PEG 8000, 0.12 M ammonium sulfate, 1.4% DMF) were gradually transferred to a soaking solution which was identical to the storing solution but included 10 mM ACA. This was done by slowly ex- changing increasingly large aliquots of the storing solution with the soaking solution (5,10,15,20,25,30,40 ul) at approximately 12 hr intervals. Crystals were soaked for 4 to 7 days. The intensity distribution for axes b and c of an orthorhombic K4 crystal soaked in 10 mM ACA are shown in Figure 10. The intensities agree well with those of the native K4 Type ll distribution, indicating that no binding of the ligand occurred. The same results were obtained for a number of similarly treated K4 crystals. It should be noted that the ACA-soaked orthorhombic crystals were first mounted and their diffraction examined to confirm the crystal type before being returned to a sitting drop for soaking. This prior mounting may well explain the observation of the Type ll pattern. Orthorhombic crystals soaked in higher con- centrations of ACA (25 mM) turned dark and crumbled in the soaking solution. Several monoclinic K4 crystals were also soaked with ACA (25mM). A comparison of the b axial intensity distributions of native and ACA-soaked monoclinic crystals is shown in Figure 11. The differences in relative intensities 25 26 EE or E Umxmom 6330 V! 05 .Axomv 526a m 89:05.5 .8 mmxm o u zm>> ooom am 982* gnu 52 Lo. xomv mzm>> ooom E 29 9.6: m I 3 a Sun 36:95 2 ._o_.a .938 3. oEEoEocto .2 $3 0 new .a .mon w:o%:m_:m%mfiawwwwafixmcwfimfim VN I 3 NF 31 9‘. v or or 9 (I) EJO'I OF 32 Fi ure 14. Schematic dia ram of four-circle diffractometer goniostat. The co, c), an% x circles are used to %rient the crystal. The detector is supported by the 28 circle. 33 A maximum of three extra steps was allowed, if needed, to cross the top of the peak. Background measurements were made on both high- and low-omega sides of the peak From an omega peak profile of an arbitrary reflection (1 9 1) (Figure 15), a total scan range of 0.18° and an offset of 040° from the peak center for background measurements were chosen. A scan speed of 0.40 deg/min was used for peak measurements, with a background to scan time ratio of 0.20. The proper orientation angles for each measured reflection were determined from an orientation matrix which was in turn calculated from the centered positions of twelve strong reflections. The array of reflections chosen had 28 values of 39—43° and were distributed throughout the phi and chi ranges defining the data collection quadrant. Before beginning data collection, the absorption properties resulting from the crystal morphology were determined by measuring the intensity of the (2,0,0) reflection as the crystal was rotated about the phi axis in 10° increments. With the a axis mounted coincident to the phi axis, this set of planes remains in the Bragg condition regardless of the phi angle; however, the intensity of the diffracted beam is affected as the path length of the X-rays through the crystal changes. The resulting absorption correction curve is shown in Figure 16. The purpose of this phi-scan is two-fold; to aid in selecting a quadrant for data collection, and for later use in correcting intensities for crystal absorption. Due to the symmetry of the crystal system, only one eighth of the total reflection sphere was unique and thus only an octant was measured (reflections having positive indices along each axis). Based on the observed absorption profile (Figure 16), the quadrant between phi angles 132° and 222°, which showed the least absorption effects, was chosen for data collection (chi = 0-90°). Data were collected in the ranges 2-32°, 32-41°, 41-48°, 48-52°, each 20 range taking approximately one day to complete. Before beginning 34 INTENSITY (1000 counts/sec) m #- 7.18 ’ 85 752 ' 9.4 ‘ 8.64 8.8 (D Figure 15. Intensity peak profile of reflection (1 9 1) versus omega. 35 .995 e :0 36:95 83093. 8 o my 9: _o 0385qu 9: 9:265 ciao cozomtoo c2983. .9 .2ng e 0 cm ONF omf .0ww. .OOM. _Omm J _ — _ _ 4 I4) H _ .9 -ma m. -3 m .8 w. :3 36 collection of each new resolution range, the integrity of the Type I diffraction pattern was checked by measuring the intensities of the reflections (6.0.0), (8.0.0), (0.0.14), and (0,16,0), which were known to differ markedly between the Type I and Type II patterns. The results of these measurements are shown in Table 4. In addition, the 9 axis intensity distribution was remeasured between each range. The entire data collection required approximately 92 hrs of exposure time. Of 8771 possible unique reflections to 1.75 A resolution, 7647, or 87%, were observed, based on a 20 cutoff. A number of measures of crystal deterioration during data collection were monitored, representing intensities in various resolution ranges. These included a two-dimensional Okl data set from 2° to 15°, collected before and following the data collection, and a short set of selected reflections having 20 values between 45° and 51°, which was measured periodically during data collection. In addition, several monitor reflections were measured throughout the data collection after every 100 reflections. During collection of the 2-32° range, the three reflections (0,14,0), (0,159), and (12,5,1), having 26 values of 25°, 32°, Table 4. Intensity measurements of conformation-sensitive reflections taken prior to three-dimensional data collection and between each resolution range of data collection. Time Intensity (counts/sec) (hrs) 6,0,0 8.0.0 0014 0,16,0 0 3867 246 81 0 1 034 23 4547 220 652 845 46 4390 222 664 81 7 71 4222 235 640 704 92 51 1 2 169 388 554 Average I, 0-71 hrs 4257 231 692 850 37 and 35° respectively, were used as monitors. For successive ranges, the first two of these monitors were substituted for reflections having higher Bragg angles: (0,21,8) and (0,15,18) at 41° and 43° respectively. The intensity decay observed for these reflections during the data collection is shown in Figure 17. Another purpose of these monitors was to track any physical movement of the crystal within the capillary during the experiment. In the event that the intensity of any monitor fell below 80% of the initial measured intensity, the centering array reflections were automatically recentered and a new orientation matrix calculated. 38 .9; K a :88 08-3 .9: 9. a :68 092:. .9; mm a :68 03-9.. .9; o .m Eamon omm-w 096.. 5:20ch €282.00 Emu 5.225 50:98.5 20:8sz SEES .2 $220 >800 .5 9:9“. 3.505 92:. om 04m u 04“ J 040 d 0%” 1 0d? 4' 040 ‘ 04W) (Ii—(II Od— — DOO— 00m— OOON o A M e e o T...“ o 0 c o o o I c 0 o . 000 N o o o S . cow M. .098 m, CON . o o o L O (LIPIIIIPIOBIOI‘ n . . .. . Loomm w. . 00m 030 W 9 mi .) . .. (.489 m.m— .O . M 009 V. DATA REDUCTION The unusually large intensity changes observed in the final measurement of the conformation test reflections (Table 4), as well as the similar abrupt drop in the intensities of the (0.21.8) and (0.15.18) monitor reflections in the final stages of the data collection (Figure 17), suggested that a conformational change may have occurred during measurement of the 48-53° range of reflec- tions. A comparison of pre- and post-data collection axial intensity distributions supported this conclusion (Figures 13,18). Therefore, it was decided that only data at 1.9 A (48°) could reliably be assumed to represent the original Type I diffraction, and thus, the higher resolution data were not used in the K4 structure solution. Before processing the measured data. background measurements were averaged in 29 shells, with approximately 200-250 reflections per shell. Figure 19 shows the 29-dependence of the average background measurements. Similar plots of average background measurements versus phi angle (not shown) indicated no discernible phi dependence. Background-corrected intensities were then calculated as I = (SCAN - (LB + RB)/RATIO) x SCAN RATE where LB and RB designate left and right background measurements and RATIO is the background to scan time ratio. Intensities were considered to be observed according to a cutoff of 2*, where is the magnitude of the average of negative intensity measurements. Based on a plot of dependence on 26 (Figure 20), a minimum acceptable intensity of 13 counts/sec was chosen for reflections with 29 less than 32°, while a value of 4 39 40 .Axomv mum; ooom “m 9.5-x mEm: noSwmcE moxm =< .:o=om=oo Emu 36:95 936:9 .935 3. 03605053 .8 $6. a Ocm .n .c 6 20:33:56 56:25 .3 9:9“. mm _ é}; . .2, a: . as}; was... 9 2 . L I ) 3 mm L I m . 9 ms 3 L om o. I . we 1 v . . I _ m o o. 9 o 9 w L D 9 .( a E m o 1 L 1 o v o a n . m . N: L t m 41 .:o_86__oo Emu 56:25 @556 9:089:68»: 6:29.963 620:3 Em: 6:6 Ammamscmv cm. .0 8:66:88 865-95... .3 9:9”. ON 9 - o... . 0..., . a . a. Li .Nm :3 N ism 4A com, 10 m 4mm. 4W .Qm 42 .:o_8o__oo «60 6556 656593662: 56:25 62606: 696% 3 8:66:33 $65.95... .8. 239m ON or IA 1N I - w. -v 3 N .. 8 I0 1 g .A rm m 4 w. . .2 m .. m. JNFI I 43 count/sec was chosen for reflections beyond 32°. A total of 4869, or 75%, of possible reflections to 1.9 A resolution were accepted as observed. The structure factor modulus, F(hkl), for each reflection was calculated using PDATA [50]. a program written in this laboratory, according to the relation: F(hkl) 2 = l(hkl) x LP x ABS x DEC where l(hkl) is the background corrected intensity, LP is a Lorentz-polarization factor depending on 29 and monochromater characteristics, ABS is an absorption factor which corrects for the differing path lengths of diffracted X-rays through the crystal, and DEC is a correction for intensity decay as a function of X-ray exposure time. An empirical absorption correction was used, as suggested by North at al. [51], derived from the observed phi-dependence of reflections at chi 90°. The maximum absorption correction applied to the data was 1.27. The behavior of the monitor reflections, the 0k! data set, and the high angle hkl data set, as well as the measurements of the array reflections during successive crystal recenterings, were all considered in estimating X-ray deterioration of the crystal. For the latter three, the sums of the individual measured reflection intensities were compared as a function of exposure time. A linear least-squares fit was used to obtain the decay slope for each. The decay factor was then calculated as: DEC = [1/(1 -t - 8)]. where t is exposure time and S is the rate of decay. 44 S = (K-1)/(K*t) K = I(i)/I(g). Based on a comparison of the various decay rates (Table 5), a decay slope of 0.0009 hr'1 was used for data from 2-32°, while a slope of 0.0017 hr‘ was used for data beyond 32°. The maximum decay correction required in 70 hrs of exposure during data collections to 1.9 A was 1.14. The distribution of structure factor magnitudes observed for various resolution ranges is given in Table 6. Table 5. Decay slopes determined based on various intensity measurements. Slope 8 calculated as: (K-1)/(K*t). K=li/ lg. Reflections 29 S Monitors: 0.14.0 25° 0 0.15.9 32° 0 12.5.1 35° 0.00224 0.21.8 41 ° 0.00224 0.15.18 43° 0.00270 Decay file 45-51° 0.00164 Array reflections 40-49° 0.00162 Okl set 2-15° 0.00006 45 Table 6. Distribution of structure factor magnitudes (F) in various resolution shells for K4 three-dimensional data collection. RXS. No. of Number of P3 with magnitude: ( ) 26 max Reflns F<20’ <50 <100 <206 <300' <400 <506 3.5 25.5 1 133 49 4 26 45 31 49 929 3.0 29.8 750 31 15 21 54 30 66 533 2.8 32.0 383 19 16 21 42 13 44 228 2.6 34.5 505 25 1 1 32 75 62 55 245 2.5 35.9 314 22 12 24 50 62 25 1 19 2.4 37.5 372 25 18 32 57 74 27 1 39 2.3 39.2 436 35 30 34 94 92 32 119 2.2 41.0 517 34 44 49 105 127 64 94 2.1 43.1 622 62 73 89 144 1 20 91 43 2.0 45.3 743 90 105 141 189 131 66 21 1 .9 47.9 890 1 36 169 218 236 91 29 1 1 1.8 50.7 1 139 272 272 304 235 43 1 1 2 1 .75 53.9 967 324 286 250 96 10 1 0 Total 8771 1124 1055 1241 1422 886 560 2483 VI. MOLECULAR REPLACEMENT Since the crystallographic structure of the highly homologous prothrombin K1 had been determined previously in this laboratory, the plasminogen K4 structure was solved using the method of molecular replacement [52], in which a highly similar molecule of known three-dimensional structure is used as a model to calculate initial, approximate phases to be used with the experimental data of the unknown structure.This requires that the model stmcture first be positioned correctly in the unit cell of the unknown. The correct position can be determined by making use of the Patterson function P(u.v,w), P(u.v,w) = (1N) 2 IF )2 cos 2n: (hu+kv+lw) which is phase independent and therefore can be calculated both from the observed intensities and from the model coordinates. This positioning can be resolved into two parts: a rotation search and a translation search. The rotation search was carried out using the Search routine of Steigemann's PROTEIN package [53], in which a correlation coefficient of the form c = 2 (Part (xo)) (Patt(R - xc)). was calculated, multiplying the Patterson functions from the observed data and that calculated from the model as it is rotated by a matrix R with respect to the unit cell. The maximum correlation coefficient observed corresponds to the optimum agreement between unknown and model vector sets, and thus to the appropriate model orientation. Rotations were performed with respect to three 46 47 Euler angles 91, 92, 93 (Figure 21), where each complete rotation operation is comprised of the following three components: rotation about the z-axis (91), followed by rotation about the new x-axis (92), and finally rotation about the new z-axis (93). Two independent model structures were used in solving K4. One model employed was the peptide backbone and conserved side chain structure of the highly refined (R-factor of 0.18 at 2.3 A resolution) prothrombin kringle 1 (PTK1), representing 489 atoms or 75% of the K4 structure [32]. The second model was the unrefined monoclinic K4 structure. which had also just been solved at the time of this work [54]. Twenty-six atoms which were not well-defined in the monoclinic structure were omitted, the remaining model accounting for 96% of all possible atoms. Prior to calculations involving the monoclinic K4 model, the coordinates were transformed to an orientation corresponding to PTFl, to facilitate comparison of the rotation search results. Based on the greater diffracting power of the K4 crystals relative to those of prothrombin fragment 1 ( = 40 A2), a thermal parameter of 20 A2 was used in calculating the model Patterson functions. The model Patterson functions were calculated at a 1 A grid spacing. and an oversized triclinic cell of dimensions 75x75x75 A was used. so that the calculated vectors would represent only intramclecular distances. which depend solely on the rotational orientation of the model, rather than intermole- cular distances, which depend also on the translational placement. Both model and unknown K4 Patterson functions were calculated at 12-3.5 A resolution (837 reflections). the limits chosen so as to omit very low resolution terms which include a large solvent contribution, and high resolution terms which are influenced by fine structural details. The rotation search itself was calculated using only a limited set of the largest model Patterson peaks having a specified vector length. Calculations 48 N X Figure 21. Euler angles which relate rotated axes x', y' and z' to thir original positions x, y, and z. 49 involving the PTK1 model were carried out using three different vector sets as a verification of the correctness of the rotation solutions. These included 1) the set of 1500 highest vectors having a length of 7-23 A, 2) the set of 1590 highest vectors having a length of 5-20 A, and 3) the set of 2000 highest vectors having a length of 5-20 A. One search was made using the monoclinic K4 model, using the set of 2000 highest peaks with a 5-20 A vector length. The lower limits on vector length were chosen so as to avoid the Patterson origin, which has an extremely large peak height, but provides no stnlctural information. The upper limits were chosen based on the approximate overall size of the kringle, to exclude vectors having lengths greater than most intramclecular distances. For each of the four model vector sets, an initial search was made in five degree increments over a range of 0-180° through each Euler angle. The orientations resulting in the highest correlation coefficients were then refined using a search in 1 degree increments. ranging 4° on either side of the initial solution. The highest correlations from each rotation search are shown in Table 7. The solution 91:0, 92:75/80. 93:40 appeared consistently among the highest solutions of the PTK1 searches and was the only solution common to both PTK1 and MONOK4 searches. The second stage of molecular replacement, the translation search, was performed using the program BRUTE [55). In this procedure. the model structure, which has been correctly orientated in a rotational sense, is translated in increments about the unit cell and at each grid point appropriate symmetry mates are generated and structure factor amplitudes calculated. Agreement with the observed data is determined through calculation of the correlation coefficient: URI 2 - lFol2 ) * ( chl 2 - chI 2l I ( (Fe) 2 - (Fol 2)2 * ( chl 2 - (Fol 2W” Table 7. Highest cross rotation search results. 50 Model PTF1, 7-23 A 1500 pks PTF1. 5-20 A 1590 pks PTFi, 5-20 A 2000 pks MONOK4. 5-20 A 2000 pks 0:01;:de mAQN-A \lmolkwlu-k 05011503104 81 O 35 25 95 25 1 55 30 O 25 1 50 35 95 O 25 1 50 30 35 95 0 85 1 45 65 1 25 1 35 82 75 135 115 70 40 140 135 75 115 140 135 60 75 115 140 135 135 60 80 80 1 25 1 05 60 95 83 40 130 50 40 135 55 60 135 1 00 1 65 85 55 Peak Ht. 1 13.0 1 13.0 1 03.7 99.7 98.6 98.1 97.8 118.9 104.6 103.7 102.2 101.2 118.0 100.3 99.0 96.4 95.6 94.4 238.6 226.2 223.5 218.1 215.1 212.5 Ht/c 999999 999999 99999 9999999 OO-‘QOJV (000de cooocoom QQQhQA-d 51 from which the F(000) term is omitted. BRUTE makes use of a different angular convention than the Euler system which was employed in the rotation search, so the rotation solutions must first be converted to the BRUTE angles a, B, y by the relations a =61-90° [3:62 7 =93+90°. To guard against possible oversight of the correct solution, the translation search was performed on the highest three rotation solutions obtained with the monoclinic K4 model, each of which had been refined in 03° increments. Since translation search results are often ambiguous, with no single convincing solution, the calculations were repeated at four different resolution ranges to aid in interpreting the results. The ranges used were 5.0-4.0 A (276 retls), 5.0-2.8 A (1236 refls), 8.0-4.0 A (525 refls) and 8.0-2.8 A (1435 refls). As a result of the P212121 symmetry, it was only necessary to search an eighth of the unit cell, thus each search was made in 0.5 A increments over ranges in x, y and z of 16, 25 and 25 A respectively. The highest solutions observed for each of the three searches are shown in Tables 8-10. Since a single correct translation solution was not obvious in any case, the possible correctness of a number of highest solutions was determined by inspecting the Ca packing resulting from each, using Evans and Sutherland P8390 stereo graphics with FRODO software [56]. Eleven translation solutions from each of rotation solutions l and II, and eight translation solutions for rotation solution III were examined. Due to the tight K4 crystal packing, all but one were easily rejected based on the presence of interpenetrating symmetry mates or unacceptably short contacts (<5-7 A). Most also showed highly uneven and improbable packing, with very close intermolecular contacts in some regions offset by large solvent cavities. In 52 Table 8. Translation search results for rotation solution 1: correlation coefficients (0) are listed. (0.80.40). Highest Clo X Vector Set 84 A 050555 939 1.. 555005 344099 mean 0: 0.0449 8-2.8 A mean 0: 0.1 599 53 Table 9. Translation search results for rotation solution 2: (85,80,100). Highest correlation coefficients (C) are listed. Vector Set X Y 2 C Clo 5-4A 11.5 2.0 12.0 0.2552 4.8 mean C: 6.0 6.0 24.5 0.2530 4.8 0.0297 6.5 6.5 18.5 0.2258 4.3 6.0 21.0 24.5 0.2235 4.2 52.8 A 7.0 23.5 18.5 0.2969 10.9 mean C= 6.0 6.0 24.5 0.2942 10.8 0.1850 7.0 6.5 18.5 0.2937 10.8 11.0 2.0 12.0 0.2924 10.8 13.0 3.0 21.5 0.2809 10.3 11.0 11.0 7.0 0.2807 10.3 8-4 A 6.5 21.0 19.5 0.1308 3.6 mean 0: 4.0 25.0 21.0 0.1271 3.5 0.0074 7.0 1.5 16.0 0.1242 3.4 1.5 14.0 22.0 0.1229 3.4 82.8 A 7.0 14.0 19.0 0.2090 8.9 mean 0: 7.0 1.5 16.0 0.2068 8.8 0.1193 14.5 21.5 19.5 0.2050 8.8 6.5 21.0 19.5 0.2039 8.7 11.0 2.0 12.5 0.2035 8.7 15.0 19.0 22.0 0.2023 8.6 54 Table 10. Translation search results for rotation solution 3: (145,125,165). Highest correlation coefficients (C) are listed. Vector Set X Y 2 C Clo 5-4A 10.5 4.0 4.5 0.3354 4.8 mean 0: 3.5 7.0 9.5 0.3307 4.8 0.0913 15.0 22.5 8.5 0.3288 4.7 3.0 19.5 10.0 0.3261 4.7 7.5 15.0 17.0 0.3253 4.7 5-2.8 A 3.5 11.5 13.0 0.3367 9.6 mean 0: 10.5 4.0 4.5 0.3363 9.6 0.2149 3.5 7.0 9.5 0.3348 9.6 3.5 19.5 13.0 0.3337 9.5 15.0 4.5 8.0 0.3317 9.5 15.0 22.5 8.5 0.3304 9.4 8-4A 3.5 7.0 9.0 0.1977 4.8 mean 0: 15.0 9.5 8.5 0.1920 4.6 0.0081 15.0 24.5 8.5 0.1864 4.5 8-2.8 A 3.5 7.0 9.0 0.2485 9.1 mean C: 15.0 24.5 8.5 0.2429 8.9 0.1260 15.0 9.5 8.5 0.2394 8.8 3.5 19.5 9.5 0.2329 8.5 55 contrast, translation solution (3.5, 9.5, 10.0) of rotation solution I displayed tight, regular packing with no close contacts except involving one peptide tail, which could be easily accounted for by a slight difference in conformation of this unconstrained, flexible region from that of the monoclinic form. The combined rotation-translation solution was refined using BRUTE. The final solution, with a correlation coefficient of 0.3469, had a rotational orientation of a=267.77, 8:76.68, y=128.67, defined by the matrix: 0.7858 -0.6174 -0.0379 0.1 134 0.2040 -0.9724 0.6080 0.7598 0.2303 and a translation vector of (3.5, 9.5, 10.1 A). initial electron density maps were calculated using phases derived from the monoclinic K4 model, transformed according to the above molecular re- placement solution. The model structure included the K4 backbone from Cyst to Ser82 and those side chains conserved from PTF1, accounting for 506 atoms, or 75% of the complete K4 structure. Two Fo density maps were cal- culated at resolution ranges 83.5 A and 8-2.8 A, using an average thermalparameter of 20 A2. During these and subsequent electron density calculations, extensive use was made of the crystallographic program package PROTEIN [53). The maps were examined using FRODO graphics, during which 18 of the 35 missing side chains were added, to include 567, or 84% of all possible atoms. Based on a Wilson plot from this new expanded model, the average B was corrected to 18.5 A2. This model, having an R-factor of 0.45, was used for the further calculation of a number of density maps: (2|Fol - chl) maps at resolution ranges of 8-2.8 A, 8-3.5 A, 5—2.8 A, and a 5-2.8 A ( [Fol - |Fc| ) map. 56 Inspection of these maps yielded five more side chains (25 atoms) to include 87.6% of the structure. Refinement of the structure began at this point. VII. STRUCTURE REFINEMENT The K4 structure was refined by the method of restrainedleast squares [57], using the fast Fourier transform program PROFFT [58]. This method makes use of idealized geometric parameters which relate atomic positions, such as bond lengths and bond angles, to reduce the number of unknown variables to be found. The deviations of these geometric parameters from their ideal values and the discrepancy between observed and model-calculated structure factors are reduced simultaneously, both properly weighted. The agreement of the model structure and observed intensities is determined by calculation of an R-factor, R = £(IFOI'1FCIVZIFOL which compares observed and calculated structure factors. The K4 structure was refined in three stages, initially including data at 6.0-2.5 A, then extending the resolution to 2.2 A, and finally to 1.9 A. Periodically, the refinement was interrupted for interactive computer graphics sessions using FRODO software, during which the model was manually adjusted to better fit calculated (2|Fo|-|Fc|) and (lFol'chll density maps. During the first stage of refinement, at 6.0-2.5 A resolution, the model structure was assigned an average isotropic thermal parameter (8) of 18 A2. The values applied in weighting the geometrical restraints are given in Table 11. In the early stages of the refinement, a protocol of alternating sets of refine- ment cycles having tight and loose geometric restraints was used, alternately emphasizing the agreement of geometric variables and structure factors. The structure factors were weighted according to a scheme in which 6F was main- tained at approximately half the average discrepancy between Fobs and Fear 57 58 Table 11. Results of PROFFT least—squares refinement at successive stages of resolution. - Resolut'on - RMS Deviations: Target(T/L) 2.5 A 2.2 1.9 A Distances (A): Bond length .020/.030 .023 .020 .018 Bond angle 040/050 059/035 051/035 .044 Planar 14 060/070 070/055 069/055 .047 Planarity Dev from plane (A) .020 .018 .017 .015 Chirality Chiral volume (A3) .150 .208 .233 .234 Nonbonded Contacts (A): Single torsion .550 .241 .217 .178 Mufti le torsion .550 .373 .359 .211 X..Y -bond .550 .334 .382 .260 Torsion Angle (deg): Planar 3.0 2.5 3.0 2.6 Staggered 15.0 31.0 24.7 16.9 Orthonormal 20.0 28.1 23.4 18.9 Thermal Restraints (A2): Main chain bond 2.0/3.0 2.11/2 3.14/3 2.76/3 angle 3.0/4.0 3.20/3 4.23/4 3.48/4 Side chain bond 3.0/4.0 2.64/3 3.98/4 3.61/4 angle 3.0/4.0 3.09/3 4.57/4 4.39/4 R-Factor (%) 19.4 15.1 14.2 (A2) 12.5 16.1 18.0 No. of reflections in calculations 2312 3286 491 9 59 The unrefined K4 structure had an R-factor of 40.7%, but with successive refinement cycles, alternating sets having tight and loose restraints, the R-factor was quickly reduced to 33.4%. At this point, individual B's were introduced and the R-factor was further reduced to 30.3%, the average B falling to 12 A2. Approximately 300 weak reflections having unusually large individual R-factors (>85%) were removed from the calculations, bringing the overall R-factor down to 27.8%. Additionally, the thermal restraints were relaxed somewhat to 2.0, 3.0, 3.0, 3.0 A2 (Table 11), whereas restraints on bond angle and 1-4 distances were tightened to 0.035 A and 0.055 A respectively, to better control the geo- metry of the stmcture. During FRODO graphics interventions, omitted side chains of non- conserved residues were included as suitable density appeared in the electron density maps. In addition, water molecules were added to the structure in the form of oxygen atoms as they became apparent in difference electron density maps. Possible water positions were chosen in the following manner. Difference density (|Fo|-|Fc|) maps were calculated at both 825 A and 6-2.5 A resolution ranges. Difference peaks above a 2.50 threshhold which were common to both maps (within 1 A of each other) were considered. These peaks were evaluated based on computer graphics examination of the difference maps, as well as a 62.5 A (2|Fo|-|Fc|) map. The criteria used for accepting possible solvent positions included that the peak appear in all three maps and be approximately spherical in shape, with the 8-2.5 A peaks being preferably as large or larger than the corresponding 6-2.5 A peaks. In addition, it was required that potential water positions be within hydrogen-bonding distance (2.5-3.5 A) of an appropriate protein atom or another solvent site. New water molecules added to the refinement calculations were assigned an initial B of 18 A2, a slightly higher value than the average protein B, and an occupancy of 0.75 lo sp relate $815 refine updal solver cute 0 as its specie and W; Tl ficanl t that 119 0001an relinen 21.6%. had an Di initially lightly weight 90809 In this riiialive 60 to speed convergence. As the thermal parameter and occupancy are closely related variables, they were not refined simultaneously. Rather, several initial sets of cycles refining solvent coordinates and B values were followed by refinement of solvent coordinates and occupancies, returning occasionally to update refinement of the thermal parameters. During the 2.5 A refinement, one solvent molecule was suspected to be a sulfate ion rather than a water mole- cule based on its refinement to a very low B value and unit occupancy, as well as its position between a number of positively charged K4 sidechains. This species was accordingly reassigned as the central sulfur atom of a sulfate ion and was refined using a constant occupancy of 1.00. The inclusion of solvent structure in the model was found to have a signi- ficant effect on the R-factor. This observation is consistent with the expectation that tightly bound waters of a relatively small molecule should make a sizable contribution to the diffraction pattern. The addition of solvent, along with further refinement cycles alternating tight and loose restraints, reduced the R-factor to 21.6%. The final stmcture at 2.5 A resolution, containing 45 water molecules, had an average B of 12.5 A2 and an R-factor of 19.4%. Data at 2.2 A were then added and similarly refined. The R-factor, which initially increased to nearly 27%, was reduced to 23% using alternating sets of tightly and loosely restrained refinement cycles. At this point the previous weighting of structure factors, using a constant c, was replaced by a e-dependent variable weighting scheme in which 0 = oa + cb(sin 60. - 1/6). In this way, not only the overall weighting on structure factors, but also the relative weights of higher and lower resolution terms could be controlled by 8C thl du bet nor sol 22.t and the addé 61 adjusting the coefficient, ca, and slope, ob. Although the variable weighting scheme did not significantly affect the overall R-factor, agreement was improved for the higher resolution terms. Agreement of low resolution terms, which was sacrificed, was subsequently improved with the continued addition of solvent structure. Approximately 260 low intensity reflections (Fobs<7.5) having unusually large discrepancies were omitted from the calculations. In addition, the thermal restraints were relaxed by increasing the target values to 3.0, 4.0, 4.0 and 4.0 A2 (Table 11). Soon after beginning refinement at 2.2 A, alternation of tight and loose restraints was found to be no longer helpful in the progress of the refine- ment. Instead, further improvement was gained primarily through manual adjustment of the structure based on electron density maps, where necessary corrections were beyond the capability of the least-squares program, and also through the continued addition of solvent. At this stage the criteria used in accepting solvent positions were eased somewhat. Peaks occuring in two of the three maps were considered, and greater flexibility was given to the apparent H-bonding distances, allowing for slight readjustments of the protein structure during subsequent refinement of the altered model. Particularly, distances between solvent positions were evaluated bearing in mind the possibility of non-simultaneous partially occupied sites. The final 2.2 A structure, with 131 solvent molecules, had an R-factor of 15.1% and an average B value of 16 A2 for protein atoms. Addition of 1.9 A resolution data to the refinement increased the R-factor to 22.6%, but the K4 structure was quickly refined to 19.5% using tight restraints and a variable weighting scheme. Difference density maps distinctly showed the positions of oxygens for the suspected sulfate ion and these atoms were added to the model at this time. With further refinement, the R-factor was 62 reduced to 17.9%. Although most of the K4 structure corresponded to the electron density maps exceptionally well, Pr030 was changed from the trans to cis isomer at this point, based on the relatively poor fit and geometric distortions in this region. This substitution further reduced the R-factor and dramatically improved rms deviations of the geometric parameters, allowing the distance restraints to be relaxed to their initial values. During this stage of the refinement. 45 solvent positions, which had refined to very low occupancies and high B values and for which there was poor or no density, were removed from the structure. The final K4 structure at 1.9 A resolution included 97 solvent molecules and had an R-factor of 14.2%. The R-factor after removal of the solvent molecules was 23% (Table 12), indicating a considerable contribution of the solvent structure to the observed diffraction pattern. A summary of the least-squares refinement results at each stage of refine- ment is given in Table 11. During the course of refinement, little or no density was observed for the residues of the tail regions, a-c and 81-87, nor for the side chain atoms of Lys78 and Lys79. These regions of the K4 structure were therefore concluded to be disordered. Furthermore, the density observed for the Table 12. Comparison of agreement of struct re with (+) and without (-) solvent molecules for various resolution ranges (8.0 maximum). R-factor is defined by 2 'F0 ‘ Fc| I z lFol. Resolution Shell Sphere (A) 9+ R- Diff R+ R- Diff 3.5 12.6 13.9 1.3 12.6 13.9 1 .3 2.7 13.7 15.1 1.4 13.1 14.5 1.4 2.4 14.7 16.1 1.4 13.4 14.8 1.4 2.2 13.7 15.0 1.3 13.5 14.8 1.3 2.1 16.0 17.7 1.7 13.7 15.0 1.3 2.0 16.6 18.4 1.8 13.9 15.3 1 .4 1.9 18.8 20.0 1.2 14.2 15.6 1 .4 Cys disu was in Side ( Symb. °XYge E. 2,) Slde ( 63 Cys1-Cy680 disulfide was much less well defined than that for the other two disulfides, suggesting that it also suffers from some disorder. Finally, no density was found for the side chains of Thr12 and Glu39 beyond CB1. The variation in R-factor with scattering angle may be used to estimate the mean coordinate error [59]. As can be seen from Figure 22, the R-factor behavior for K4 agrees well with the theoretical curve corresponding to a coordinate error of 0.15 A. This method of determining error, however, assumes that the discrepancies between observed and calculated structure factors are a result of positional errors only. The actual coordinate error probably exceeds 0.15 A for portions of the structure having high thermal parameters, such as the terminal regions of the peptide chain at and near the Cys1-Cys80 disulfide. The final atomic parameters of the K4 structure are listed in Appendix A. i Abbreviation CA denotes main chain carbon atom from which amino acid side chain branches. Side chain atoms designated b common biochemical symbols in which first character indicates atom type carbon=C, nitro en=N, oxy en=O) and additional character(s) indicates position on side chain B,G,D, E_, ,H proceeding from main chain). Numerals distinguish between equivalent Slde chain positions. 64 35- 33 30" / (I ATIO.25/// 9. ,/ 0 25- x” , E ,/ 8150.20 ,’ I // ’z’ n- / ’ / / J / / / I I / ATIOJS’ 4’ / " ”o / I, /’ 15i- // /’ ’n’ // x’/ ’w’ o / 0 ” 1 g 1 L J 1 0.125 0.150 0.175 0.200 0.225 0.250 S l N (9/ A) Figure 22. Depexdence of R-factor on scattering angle for final refined K4 structure at 1.9 resolution. Dashed lines show theoretical coordinate error curves. VIII. MAIN CHAIN AND DISULFIDE STRUCTURE The observed main chain structure of K4 agrees well with idealized contor- mational parameters. A Ramachandran plot of the refined K4 structure is presented in Figure 23, from which it can be seen that the dihedral angles of nearly all the non-Gly amino acids, with the exception of Met48, fall within the energetically preferred zones. Similarly, the distribution of main chain omega angles (Figure 24) shows the observed values for the trans residues to be concentrated in a narrow range about 180°. A single cis peptide bond is found in the K4 structure, occurring at Pr030 (Figure 25). This bond has an omega angle of 4°, which compares favorably with the ideal value of 0°. The three-dimensional main chain folding and disulfide stmcture of K4 are shown in Figure 26. The backbone structure can be subdivided into four segments which result from the characteristic kringle disulfide bridging pattern and which are defined by amino acid stretches Tyr2-Ly521, Gln23-Tyr50, ArgSZ-Trp62, and Phe64-Lys79. These loops shall be referred to as loops A, B, C, and D, respectively, according to the convention of the prothrombin fragment 1 literature [33]. A striking feature of the K4 structure is the close proximity of the two disulfides Cys22-Cys63 and Cys51-Cys75, which form a four-sulfur cluster and serve as the nucleus of the threesdimensional folding. Due to this arrange- ment, segments B and C, which are bounded by these two disulfides, are each nearly closed three-dimensional loops. The third disulfide, Cys1-Cys80, is approximately 12 A away at the edge of the kringle structure, causing loops A and D to have a more extended conformation. The Pro residues, hydrogen-bonding interactions, and turns (Tables 13,14) which influence the finer details of the K4 folding within each of the four loops are identified in Figure 27. Loop A is a generally S-shaped stretch, with two 65 66 100 1 l I ‘ 1 1 \ I \ ‘I \ .. ' 1 ' l "l , .. ' 1 90 ' ' 2’1 _ i ° ' -’J 1’ l ' . C . ” ' | 1 I 1 .1 l 1 lo -\\ 0 $ l 1 l __ \ ... I ‘\\:o _ \\ .~ .\ CO 0 - a Q \\ . : \\\ \ ‘ x > . :0 _ "‘ I .0 P“./ A 1 l.__/ . _.| _ 1.. _____________ .1 -”—1 .— o I— 'Met48 O 1""? ‘‘‘‘‘‘‘ a "1 "’° 1' l 1 r 1 l I 4” 40 O 90 IOO 1‘0 Fl ure 23. Ramachandran lot of final refined K4.structure. Open circles denote 61?) residues; filled circlespdenote non-Gly resudues. Energetically preferred zones are outlined in dashed lines. WMHJnu. WWE l0 EmmEDZ Fig (sir 67 25 20 20)- 11 10 l CD NUMBER OF RESIDUES ‘ ~ #144 -173 -175 -177 -179 179 177 175 173 171 OMEGA ANGLE (deg) Figure 24. Distribution of omega angles of refined K4 main chain peptide bonds (single cis peptide bond not included). 68 ‘ ’ 4!- \‘I' "3‘ o' a. ,1: 1 1.1.2329“. 14's? st . A '. ‘le I ,2? 91’ \‘0 “=. g ~21» .." ‘ ' - 0‘1.» ’ " l‘ ’ ':--‘ .~°.- -'.- i... ' T "‘= -‘:_ ‘5‘ .0 a c “V... {93% . .. ~ ' 09.1, ' 'I' ‘\__"Il.p Br.’ 4 r I ' or '80 . ‘.\7\‘v"l I. I ‘3 .-- C-“ 'i. .1. u ‘ 7". ‘57:», . -91... .8. t~-I."QE’E "l "Ne--24 r 0.. .- .v‘l .1 5n. ‘ '0 55.: = ‘8 ‘ a. ,.4..- Alb ig! r ‘9‘... A'l-e ’ l-Drffi'3“ Figure 25. Stereoview of 2|Fo|-|Fc| electron density for Pr033 which has a cis peptide bond configuration. 69 Figure 26. Stereoview of K4 CA,C,N backbone and disulfide (bold) structure. 70 Fl ure 27. Stereoview of K4 backbone showing Pro sideochains and .main chain hygdrogen-bonding interactions. Hydrogen bonds are indicated wrth dashed lines. 71 B—turns at Hisa-Gly6 and Gly6-Tyr9 combining to form the large upper curve. A sharper hairpin turn at Thr16-Gly19 forms the second curve, which terminates at the central disulfide cluster. Loop B is the longest of the amino acid stretches and its residues comprise nearly half the volume of the kringle. Shortly beyond the disulfide core, residues Ser24-Ser27 form a reverse B-tum, after which the B loop doubles back again in a tight reverse open turn at Pr030, as is commonly observed for cis Pro residues [60]. A short extended stretch of main chain separates these turns from a second bending region, in which a number of con- secutive bends in the chain are caused by Pro38, Pro42, and a [S-type turn involving residues Thr37-Tyr41. The combined effect of these turns results in minimal change in the direction of the main chain; however, an additional B-turn at Tyr41-Ala44 directs the chain back toward the disulfide cluster. One last 8-turn of loop B occurs at Thr47-Tyr50. Loop C is the shortest amino acid stretch and is more tightly constrained by the central disulfide pair. This loop has a single wide turn resulting from a B-turn at Asn53-Ala56 and a 90° bend at Gly60 which does not appear to be stabilized by any hydrogen-bonding inter- actions of the main chain. Two Pro residues occuring in this region have little effect on the path of the main chain: Pr054 results in only a small bend, and Pr061 produces no deviation at all. Finally, loop D has only a single reverse hairpin turn at Pr068, stabilized by a B-type interaction at Asp67-Val70. The remainder of the chain follows a nearly fully extended path, traversing the entire 31 A length of the kringle to the Cysl-Cys80 disulfide. In addition to the B-turn interactions, a number of additional main chain hydrogen-bonding interactions contribute to stabilization of the K4 structure. Extensive hydrogen-bonding is found in regions of loops A and D which cross one another near the disulfide cluster. Short stretches of antiparallel B-sheet occur on both loops involving residues Ser14-Thr16/Ly320-Cy522 and 72 Cys63-Thr65/Arg71-Glu73, respectively. These perpendicular stretches are then stabilized further with respect to each other by hydrogen bonds between Thr15-Asn76, Thr17-Tyr74, and Lys21-Thr66 linking the two loops. A third short stretch of antiparallel B-sheet at ArgSZ-Cys63 stabilizes the ends of loop C. A single main chain hydrogen bond anchors the ends of loop B near the di- sulfides at Trp25-Met48. The remaining interactions stabilize backbone stretches adjacent to disulfide bridges, including a hydrogen bond at Gln23- Phe64 near the Cys22-Cys63 disulfide and two involving Tyr2-Lys78 and Tyr2-Cys80 near the outer Cys1-Cys80 disulfide. Three solvent-mediated hydrogen-bonding interactions also occur in the K4 crystal structure between residue pairs Gln23-Ph864, Tyr41-Leu46, and Ala44-Asn53, the last of which stabilizes the interface between loops B and C. It is evident that the disulfide cluster, adjacent stretches of B-sheet, and inter-loop hydrogen bond links act together to form a highly stabilized zone comprised of the C loop and lower portions of the A and D loops. This is significant in that the K4 Lys-binding site is found in the same region, extending from Asp55/Asp57 on the far edge of the C loop to Arg71 near the lower turn of the D loop. Thus these structural elements appear to serve not only to support the overall kringle folding, but also to rigidly maintain the conformation of the crucial binding region. A summary of the K4 secondary structural elements is given in Table 13, and a complete listing of observed main chain hydrogen—bonding interactions is presented in Table 14. The K4 B-interactions have an average acceptor-t0- hydrogen distance of 2.08 A and average hydrogen-bonding angle of 159°, in good agreement with values observed for other protein structures [61]. Two potential hydrogen bonds having good inter-atomic distances but unusually small bond angles are included in Table 14. The Cys80 N-Tyr2 0 interaction, 73 Table 13. Observed secondary structural elements of K4 main chain. Reverse turns are classified according to dihedral angles of residues 2,3 [63]. Results of nmr structure are given for comparison [37]. X-RAY NMR 8 Structure: 81 Ser14 O - Cys22 N Thr16 N - Ly820 O 82 Ly321 O - Thr66 N Gln23 N - Ph864 O 83 Ar952 N - Trp62 O Ar952 O - Trp62 N 84 Pr061 O- Cys75 N Trp62 -T r74 Cy563 N - Glu73 O Phe64 - rp72 Cy863 O - Glu73 N Thr65 N - Arg71 0 Reverse Turns: T1 His3-GIy4-Asp5-Gly6 ype ll') T2 Gly6-Gln7-Ser8-Tyr9 ype I) T3 Thr1 6-Th r1 7-Th r1 8-Gly19 ype l Gly19 T4 Ser24-Trp25-Ser26-Ser27 ype l Ser26 T5 Thr37-Pr038-Glu39-Asn40 ype lll) Pro,Asn40 T6 Tyr41-Pro42-Asn43-Ala44 ype l) Pro T7 Thr47-Met48-Asn49-Tyr50 ype ll') T8 Asn53-Pr054-Asp55-Ala56 Type I) Pro,Asp55,Ala56 T9 Asp67-Pr068-Ser69-Val70 ype l) Thr66,Pro,Ser69 Other Turns’: T10 Pr030 Pro T1 1 Lys59 Lys60 * Turn T10 is an open, rather than reverse, turn. T11 is an abrupt 90° bend in the main chain which does not meet the criteria for either type of turn [64]. 74 Table 14. H drogen-bonding interactions of the K4 main chain. Interactions forming 8-s eat (8) or reverse turns (1') are indicated. Hydrogen atoms are assigned geometrically idealized positions. Distances (A) Angles degc) DONOR ACCEPTOR H..O N..O NHO CO ON Tyr 2 N L s 78 O 2.15 2.97 144 140 151 Gly 6 N is 3 0 T1 1.91 2.85 149 147 149 T r 9 N Gly 6 O T2 2.13 3.07 162 100 105 T r 16 N lfys20 O 81 1.99 2.93 160 149 155 Thr 17 N I‘ 74 O 2.04 3.02 154 155 150 Gly 19 N T T16 O T3 2.04 3.04 163 118 120 C s 22 N Ser 14 O 81 2.01 2.98 167 149 145 G n 23 N Phe 64 O 82 1.76 2.75 164 130 135 Trp 25 N Met 48 O 2.30 3.18 150 164 161 Ser 27 N Ser 24 0 T4 2.26 3.14 148 103 1 12 Asn 40 N Thr 37 0 T5 2.14 2.85 124 114 128 T l’ 41 N Thr 37 O 1.92 2.80 151 156 165 A a 44 N T r 41 0 T6 2.24 3.25 178 122 123 Tyr 50 N T r 47 0 T7 2.39 3.34 161 150 153 Cys 51 N Ser 13 O 1.90 2.87 163 146 151 Arg 52 N Trp 62 0 $3 2.14 3.07 160 165 1 59 Ala 56 N Asn 53 O 8 2.60 3.55 160 106 106 Tl’p 62 N Arg 52 O 83 1.86 2.86 161 144 145 Cys 63 N GIU 73 O 84 2.26 3.19 156 161 164 Thr 65 N Arg 71 O I34 2.11 3.02 153 128 134 Thr 66 N Lys 21 O 82 1.99 2.86 151 179 170 Val 70 N Asp 67 O T9 2.27 3.25 168 1 19 122 GIU 73 N Cys 63 O 84 2.25 3.17 151 170 172 Cys 75 N Pro 61 O 84 2.09 3.06 167 132 130 Asn 76 N Thr 15 O 1.85 2.86 179 145 145 Cys 80 N Tyr 2 O 1.99 2.74 128 122 137 MAIN CHAIN INTERACTIONS BRIDGED BY SOLVENT (W) Distances (A) A Ies (deg DONOR ACCEPTOR NH..W CO..W NH..O N..O HWO HW c w Leu 46 N T r 41 O 1.97 2.66 4.09 5.09 124 159 150 Asn 53 N A a 44 O 2.12 2.64 4.28 5.14 127 166 127 Phe 64 N Gln 23 O 2.00 2.90 4.35 4.97 124 156 139 75 has a hydrogen bond angle of 128°, which may simply reflect the general dis- order accorded to the structure in the vicinity of the Cys1-Cys80 disulfide. A similar hydrogen bond angle of 124° is measured for the possible Asn40 N- Thr37 0 interaction. In this case, the residues occur at the point of a reverse turn in the main chain (T5), and, although Thr37 carbonyl group actually appears to make a more favorable hydrogen bond with the amide nitrogen of Tyr41 (bond angle 151°), Asn40 is at the proper sequence position for the expected 1-4 8-type interaction. Furthermore, a survey of NH..O hydrogen bonds reported from neutron diffraction studies [62] found there to be a wider distribution of hydrogen bonding angles, with a lower mean value of 132°, when intra- molecular interactions are considered separately from intermolecular inter- actions. Therefore, both potential hydrogen bonds are listed in Table 14 for consideration. Of the 78 residues of the Cys1-Cys80 kringle peptide chain, 47 of the main chain nitrogen atoms and 64 of the carbonyl oxygen atoms participate in some form of hydrogen-bonding. In each case, 25 are accounted for by direct main chain-main chain interactions. Of the remaining hydrogen-bonds, 7 of the amine groups and 13 of the carbonyl groups interact with protein side chains. The balance consists of solvent interactions. A comparison of the observed K4 secondary structure with that proposed by Atkinson and Williams [37] in a recent nmr study shows that many of the 8-turns agree or are immediately adjacent to one another (Table 13). One stretch of antiparallel 8-sheet, Trp62-Tyr74 and Ph664-Trp72, was also correctly identified in the nmr structure, although the additional two observed stretches on loops A and C were not reported. The distribution of average thermal parameters for the main chain as a function of residue number is shown in Figure 28, from which it can be seen that the K4 backbone has a rather low overall average B of approximately 16 A2. 76 .mctoneac mx 9 83mm. 5;) 82.2% 9 mac 9.35 63625 $5. “.9303 Sago 02m new 85. 2.93 59.0 EmE § .o 8269 .959: momam>< .3 9:9“. mum—232 mnoawm HOlOVd EHHLVHBleEi (aw 77 The fluctuations in B faithfully reflect the positions of residues with respect to the three-dimensional folding. The thermal parameters are highest in the terminal regions, c,1-5 and 76-80, which have electron density of marginal quality and are considered to be somewhat disordered. Other notable peaks in B value occur near the ends of three-dimensional loops or bends. These are observed at Ser26, Glu39, and Leu46 and all occur in the generally less stabilized B loop. The peak centered at Pr068 corresponds to the long hairpin turn of loop D. Conversely, residues adjacent to the central disulfide cluster and residues participating in 8-sheet interactions all have B values which are among the lowest observed. The three disulfide bonds, which not only define the characteristic kringle primary loop structure, but also have a fundamental role in the three- dimensional folding, are worthy of some additional discussion. The observed bond distances and bond angles for the three disulfides are presented in Table 15. Interestingly, Cys75 was found to have two alternate side chain positions which differ by a torsional rotation of approximately 90° (Figure 29). Through least-squares refinement of the K4 structure with both Cys75 side chain positions, the difference in orientation was found to have no noticeable effect on the conformation of the adjacent backbone structure or of the Cy851 side chain. The two Cys75 side chain positions result in two alternate Cy551-Cys75 disul- fide bonds having perpendicular orientations (Figure 30). In one case, the CysSt and Cys75 side chains form a left-handed disulfide bond (S-S dihedral angle ~ -90° ) which is parallel to and nearly aligned with the neighboring Cys22-Cy563 bridge. This orientation positions all four sulfur atoms of the cluster within approximately 4.5 A of one another in a box-like arrangement, with a distance of only 3.8 A between the midpoints of the two disulfide bonds. In the second case, the Cys75 sulfur atom is rotated away from the 78 N5 7 mm. vm om .. .Fx 9: 8 mt- 8- 3 .2 88 SN 8856-586 8?. mm- 3- so- 8. 8 :6 3; 896-686 E: - on- .8. E- 8. 9: 56 84 896456 8- mm mm 8 2: z: 5.5 Ba 836-56 .mx «x «x _x maxed 995 5.5 8-5 68.35. 383 8.92. 969 669 8.9. 26 8565 520 85 a O med .Px NX ox Nx wx <0---mo:-0m-:omimo--- . ,4, v w‘dlwehQUESs fun... ;.'I, 17’ i ‘7: .1 . - 95:“. new: Figure 29. Stereoview showing two alternate Cys75 side chain conformations. 80 22 51 Figure 30. Stereoview showing two relative disulfide positions resulting from two Cys75 side chain orientations. 81 Cy522-Cys63 disulfide, to form a right-handed Cys51-Cys75 bridge (S-S dihedral angle ~ 90° ) which is perpendicular to and directed away from its neighbor. The four sulfur atoms now have a T-shaped arrangement, with a distance between disulfide midpoints of 4.9 A, and with the Cys75 sulfur atom approximately 6 A from either sulfur of the Cys22-Cy863 bridge. Based on separate least squares refinements in which one, then the second, side chain position was used, with the other position filled by a free sulfur atom (as solvent), it was estimated that the two alternate Cys75 side chain orientations are each roughly 50% occupied. With the CysSI-Cys75 disulfide in the left-handed orientation, the free sulfur refined to an occupancy of 0.53; with the disulfide in the right-handed orientation, the free sulfur assumed an occupancy of 0.68. In both cases the B factor of the free sulfur refined to the same value of 21 A2. At the same time, the sulfur assigned as $6 of the Cys75 side chain had a fixed occupancy of 1.00, along with all protein atoms, but refined to a lower 8 value in the left-handed disulfide orientation (28 vs 31 A2). These observations may suggest some slight preference for an arrangement in which the two disulfides are parallel. Alternatively, since the K4 structure was refined using only the left-handed disulfide arrangement until the final stages, this difference may merely reflect a bias of the refinement procedure itself. The roughly even distribution in occupancy of the two orientations implies that there is no significant energetic advantage to either arrangement of the four sulfur atoms. Nonetheless, it is surprising that such conformational disorder is permissible at the structural core of the tightly constrained kringle fold. It is probable that the observed Cys75 disorder is static, rather than dynamic. Simple torsional rotation of the Cys75 side chain between the two observed orientations results in the unacceptably close approach of approxi- mately 1.5 A between the Cy551 and Cys75 sulfur atoms. Thus, a 82 conformational change would be extremely hindered without considerable adjustment in position by the neighboring atoms. Moreover, the observed thermal parameters of the Cys51 side chain atoms, as well as of the main chain atoms adjacent to the CysS1-Cys75 disulfide, are quite low, suggesting no such unusual movement. Since the residues following Cys75 in the K4 peptide chain become steadily less ordered, ending in a completely disordered carboxy tail, the presence of two Cys75 conformations may be a result of general flexibility of the terminal peptide regions. The greater conformational freedom of the K4 terminal region in solution may allow flexibility of the CysSl-Cys76 disulfide bridge itself, which is then trapped in one of the two observed orientations upon crystallization. IX. SIDE CHAIN STRUCTURE With the addition of the amino acid side chains, the overall three-dimensional structure of K4 has the appearance of a disc with one convex surface and one flat surface (Figure 31). Of the 76 residues from Cys1 to Cys80 on the K4 peptide chain, 51 (63%) of the side chains are exposed on the surface of the structure while the remaining 27 (37%) are primarily internal. The solvent-accessible surface areas calculated for each residue [65] are listed in Table 16, in which an accessible area of 10% effectively distinguishes buried from external side chains. Several residues, such as Asn49 and Glu73, which appear to be only marginally exposed based on surface accessibility fractions, are observed in the three-dimensional structure to be indeed external but shielded by hydrogen-bonding interactions with neighboring residues. About 75% of the external side chains are hydrophilic. Of these, 21 (55%) are involved in hyrdogen-bonding interactions with adjacent protein side chain or main chain atoms. A complete listing of probable K4 side chain interactions is given in Table 17. Nearly all of the charged side chains are among those found on the surface of the kringle. Although there appears to be no overall pattern to the distribution of charge on the surface, a clustering of like charges is found at either end of the lysine-binding region. The side chain of Asp5 is adjacent to the Asp55/Asp57 pair of the binding site, forming a triply anionic cluster at one end, while Arg71 of the binding site, Argaz, and Lys35 form a similar cationic group at the opposite end. A repeated pairing of similar charged residues is also observed elsewhere on the K4 surface. In addition to the sequential pairs of Ly520/Lys21 and Lys78/Lys79. residues Arng and Ar952 are spatially adja- cent in the three-dimensional structure. These side chains are oriented in an antiparallel manner, with the guanidinium groups of each involved in 83 84 (b) Figure 31. Stereoview showing a) face-on and b) edge-on views of K4 structure. 85 Table 16. Accessible surface area of K4 main chain and side chains by residue, calculated usir? programs ACCESS and ACCFMT [65] with a spherlcal probe of radius 1.4 . Percent accessible surface values are based on calculated accessible surface for the isolated residue. Asterisks indicate residues having disordered side chains. Main Chain Side Chain Residue A essible Surface Accessible Surface Number Type 2 °/. A2 % 1 Cys 11.3 8 42.3 38 2 I?" 1.4 1 34.4 16 3 is 6.5 5 122.6 71 4 Gly 74.6 38 - - 5 Asp 3.7 3 51.0 41 6 Gly 0.2 0 - - 7 Gln 6.7 5 71.0 43 8 Ser 14.7 11 66.2 70 9 Tyr 6.4 5 11.4 5 10 Arg 23.9 18 92.6 41 11 Gly 24.8 13 - - 12 1'hr t i i i 13 Ser 9.9 7 21.8 23 14 Ser 1.7 1 25.0 26 15 Thr 0.0 0 61.2 49 16 Thr 0.5 0 0.0 0 17 Thr 29.8 23 60.7 48 18 Thr 35.3 27 48.8 39 19 Gly 58.6 30 - - 20 Lys 8.2 6 30.7 24 21 Lys 3.8 3 124.8 63 22 cys 0.0 0 0.0 0 23 Gln 1.4 1 2.0 1 24 Ser 2.3 2 36.2 38 25 Trp 0.3 0 0.8 0 26 Ser 20.8 15 65.5 69 27 Ser 8.4 6 27.5 29 28 Met 9.3 39 52.6 37 29 Thr 3.8 3 98.0 78 30 Pro 26.9 24 26.9 19 31 His 0.9 1 12.5 7 32 Arg 3.6 3 183.1 80 33 His 12.4 10 0.5 O 34 Gln 2.3 2 131.8 81 35 #y‘s 0.0 0 57.2 29 37 r 0.0 O 24.3 19 38 Pro 4.7 4 50.9 37 39 Glu t t t i 40 Asn 25.8 20 63.2 45 41 Tyr 4.5 3 59.6 27 42 Pro 28.2 25 75.3 55 43 Asn 16.4 13 117.0 83 Table 16 (cont'd) $3838 9 M 52 8 55 56 57 w m m 62 63 64 65 66 67 68 69 70 N 72 73 74 75 76 m m 9 Ah GU Leu Thr Mm Asn gr s Am Asn 2? p Ala {ASP ys 60 Pro Tm 01s P e Thr Thr Aw Pro Ser Val 80 “m glu yr W5 Asn Leu Lys Lys Cys A N mmpom P9999 mud 99? —.L mmAumAmwbmmUOAOOV3&644454004bhwbb —L N9~9999e9ms -L 9N999998 N 94.9 0') N 86 —L dd—L m —L OOOOONNQOHJICDNOJOOOOVUIUIOV moo -* V1000 _L " ’(DVm-‘wmnbw A \l u—A 338599N9$9899 9 —L mhmbdbbhbkbhbmodm bo;aooaomuymo'm -L -L mmodwmoaw pmpwpmwmmmogpge. a) pm Q Q —L 9 —L 10de VAC» Owomd'mgmmOOGONAmmorm 87 Table 17. Hydrogen bonding interactions involving K4 side chains. Hydrogen atoms were assigned geometrically idealized positions. Donor atom is denoted (D), acceptor atom (A). INTRAMOLECULAR SIDE CHAIN INTERACTIONS Distances (A) Angles Adeg) DONOR ACCEPTOR H..A D.A DHA CA AD 7 Gln NE2 5 Asp OD1 2.07 3.02 157 114 119 10 Arg NH1 7 Gln O 2.02 3.02 161 138 136 10 Arg NH1 9 Tyr O 2.06 3.01 158 117 116 13 Ser N 9 Tyr OH 2.04 2.98 161 149 148 16 Thr OG 73 Glu OE1 1.87 2.55 129 172 162 23 Gln NE2 27 Ser 0 2.19 2.83 119 116 120 23 Gln NE2 31 His 0 2.04 3.02 161 153 148 25 Trp NE 50 Tyr O 1.77 2.71 154 155 158 29 Thr N 27 Ser OG 2.15 3.07 165 126 131 31 His N 23 Gln OE1 1.98 2.82 149 120 129 31 His NE 65 Thr O 2.29 2.92 117 129 145 31 His NE 67 Asp O 2.24 2.97 124 130 135 33 His N 28 Met SD 2.52 3.49 166 146 142 33 His NE2 25 Trp O 2.09 2.83 125 130 134 35 Lys N 31 His ND 1.94 2.92 147 - - 40 Asn ND2 34 Gln O 2.14 2.96 136 135 125 43 Asn ND2 76 Asn OD1 1.86 2.88 169 133 136 49 Asn N02 22 Cys O 1.79 2.72 161 135 134 52 Arg NE 45 Gly O 2.01 3.00 159 130 130 52 Arg NH2 11 Gly O 1.89 2.86 160 121 127 52 Arg NH2 51 Cys O 2.12 2.92 137 157 147 53 Asn ND2 5 Asp O 1.86 2.77 143 145 148 53 Asn ND2 57 Asp O 1.89 2.87 167 111 111 57 Asp N 53 Asn OD1 1.78 2.68 154 126 117 58 Lys N 57 Asp 001 1.87 2.77 153 89 85 62 Trp NE1 55 Asp OD2 1.99 2.79 137 148 136 29 Ser N 67 Asp OD2 1.75 2.67 151 152 146 71 Arg NH2 32 Arg 0 1.99 2.82 137 145 151 74 Tyr OH 57 Asp OD1 1.81 2.74 174 142 143 INTERMOLECULAR INTERACTIONS 32 Arg NH1 55 Asp OD1 1.98 2.96 160 156 149 32 Arg N 57 Asp OD2 1.74 2.60 153 135 136 43 Asn N02 76 Asn OD1 1.79 2.82 174 140 141 71 Arg NH1 57 Asp ODZ 2.28 3.14 140 93 94 88 hydrogen-bonding interactions with main chain carbonyl oxygens, Arg10 with Gln7 and Tyr9, Arg52 with Gly11 and Cy851, thus distancing the similar charges from one another (Figure 32). Additionally, the NE atom of the Ar952 side chain forms a hydrogen bond with the Gly45 carbonyl oxygen. One ion pair is formed between charged side chains on the K4 surface: the positively charged Lys20 residue extends toward the negatively charged Glu73 side chain, bringing the charge centers within 4.0 A of one another. The Lys side chain is also within 4.6 A of a carboxy oxygen of the Asp67 side chain. Among the remaining hydrophilic K4 surface residues, Met28 is particularly interesting as it appears to participate in a hydrogen bond in which the side chain sulfur atom serves as a hydrogen acceptor to the His33 main chain amide group (Table 17). The observed NS and HS distances of 3.49 A and 2.52 A, respectively, (based on idealized hydrogen positions) agree well with the values of similar NH..S hydrogen bonds which have been reported from crystallographic studies of small organic molecules [66,67]. The hydrophobic amino acids which are exposed to the surface include four Pro residues and a number of residues having relatively small side chains, including Gly, Ala, and Val. The bulky His3 side chain is also highly exposed in the K4 crystal structure, but as this residue occurs in the terminal region, it may have a quite different environment in the intact plasminogen molecule. The aromatic rings of residues Tyr41, Tyr50, and Tyr74 are also somewhat exposed, having accessible fractions near 25% (Table 16). In each case, it is the hydro- philic tyrosyl hydroxyl group which is directed outward toward the solvent. The only remaining significantly exposed hydrophobic surface is the indole ring of Trp72. This large hydrophobic patch on the otherwise hydrophilic surface serves as an obvious marker of the lysine-binding site. Since nearly the entire external surface of the kringle is hydrophilic in nature, it would be expected that, 89 Figure 32. Stereoview showing interactions of adjacent Arg10 and Ar952 side chains. Hydrogen bonds are indicated with dashed lines. 90 in the absense of specific side chain interactions with other plasminogen domains, the K4 domain would tend to assume a solvent-exposed conformation in the intact protein. This observation agrees well with the previous interpreta- tions of the kringle domain as a structurally and functionally autonomous unit. The internal residues of K4 include three hydrophilic amino acids: Gln23, Arg 52, and Asn53. The hydrocarbon chain of the Ar952 side chain, although buried, extends toward the kringle surface where the terminal guanidinium group is slightly accessible to the solvent. Two of the internal side chain nitrogen atoms serve as hydrogen-donors to the carbonyl oxygen atoms of Gly11 and Gly45 (Table 17). In the cases of Gln23 and Asn53, the side chains also appear to be fully involved in hydrogen-bonding interactions with the protein backbone, which may add further stability to the three-dimensional folding. The Gln23 side chain carboxy and amino groups bond with main chain His31 N and O atoms, respectively (Figure 33). The Asn53 carboxy group bonds to Asp57 N and the side chain amino group is within hydrogen-bonding distance of both Asp5 O and Asp57 0 (Figure 34). Nearly all the polar groups provided by less hydrophilic internal side chains are also involved in similar interactions. Thr16 and Thr65, which have side chains bearing both hydrophilic and hydrophobic groups, are internal to the K4 structure. The hydroxyl group of the Thr16 side chain acts as a hydrogen donor to the carboxy group of Glu73; however, the side chain of Thr65 appears to have no similar binding partner. The side chain of Tyr9 also bears a buried hydroxyl group which, in this case, appears to act as a hydrogen acceptor to the Ser13 main chain amide group. Lastly, the indole side chains of Trp25 and Trp62 display hydrogen-bonding interactions to the Tyr50 carbonyl and Asp55 carboxy oxygens, respectively. In many cases these side chain interactions occur between residues of separate peptide loops and thus appear to play a 91 Figure 33. Stereoview showing hydrogen-bonding interactions (dashed lines) of Gln23 side chain with His31 main chain. 92 ' ' ' ' hain. Potential ' ' howm lnteractlons of Asn53 .Sld.9 c . :ygflicraegezifiboitgsre’tgvites‘gsso, A5857 N, and Asp57 O are indicated With dashed lines. 93 role in stabilizing the protein secondary structure. A final observation on the hydrogen-bonding interactions of the internal K4 side chains regards the imidazole group of His33. Although there is ample room in the K4 structure to accommodate this side chain, it is positioned so as to make an unusually close contact of 2.92 A between the ND atom and the main chain amide nitrogen of Ly335 (Figure 35). This suggests that the imidazole ring, which has a pKa of 6.0, is not protonated in this case, allowing for a hydrogen bond between the Lys35 amide hydrogen and the lone pair of the His33 side chain nitrogen. Based on the calculated geometric ideal position of the amide hydrogen, this interaction has a hydrogen-to-acceptor distance of 2.09 A and a hydrogen-bonding angle of 147°, both of which agree well with typical values. Since the orthorhombic crystals were obtained from a mother liquor at approximately pH 6.0, the His residues might easily exist in either the protonated or de-protonated form; however, the His33 side chain is not accessible to the solvent (Table 16) and probably does not recognize this influence. On the other hand, the His31 side chain is slightly accessible to the solvent. This side chain has no available neighboring hydrogen donors, and although several potential oxygen hydrogen acceptors are found in close proximity to His31, the calculated hydrogen-bonding distances and angles are poor. Thus, in addition to its possible protonation, even the rotational orientation of the imidazole ring about the CB-CG bond (1 180° ) is ambiguous in this case. A dominant feature of the internal structure of K4 is a central hydrophobic core formed by a number of aromatic and other bulky hydrophobic side chains (Figure 36). Near the center of this core is Leu46, surrounded by the side chains of Trp25, Pr038, Tyr41, and Pr054. Through Trp25, the core extends to include Trp62, Ph664, and nearby His33. The aromatic rings of Trp62 and Phe64 are also adjacent to the side chains of Trp72 and Tyr74. Tyr50 is near the 94 7 251 .0, it a !. 4 Figure 35. Stereoview of probable hydrogen bond (dashed line) between Lys35 main chain N and lone pair of His33 imidazole N. Hydrogen atoms are shown at geometric ideal positions. 95 Figure 36. Stereoview of K4 backbone and side chains forming internal hydr0phobic core. 96 hydrophobic cluster, and, although the side chain is directed outward toward the solvent, it may communicate with the core via a hydrogen-bonding interaction between the Tyr50 carbonyl group and the nitrogen atom of the Trp25 indole ring. A similar but much smaller hydrophobic cluster occurs at Leu77. which is surrounded by the side chains of Tyr2, Tyr9 and Pr061. The specific interactions observed between the residues of the hydro- phobic core are interesting in themselves. A number of the aromatic side chains participate in stacking interactions in which one aromatic ring is oriented in a perpendicular manner toward a second (Figure 37). For example, the edge of the Trp25 indole ring is directed toward the face of the six-membered ring of the Trp62 indole at an angle of approximately 94° and at a distance of 3.2 A from the the plane of the Trp62 ring. Similarly, the Trp62 indole is oriented at an angle of approximately 92° to the face of the Ph864 ring, and the Trp72 indole lies at an angle of 85° to the Tyr74 ring. In both cases, the calculated geometric ideal position of the nearest radial indole hydrogen is 2.8 A from the midpoint of the corresponding benzene ring. In each case it is the six-membered ring of the Trp indole, rather than the nitrogen-containing five-membered ring, which is directly involved in the interaction. This type of perpendicular aromatic stacking interaction is commonly found in protein structures and is a result of the planar structure of the aromatic side chains [68]. Since the hydrogen atoms, which are positively charged relative to the ring carbon atoms, are oriented radially in the plane of the ring, the edge of the side chain has a positive charge with respect to the face. Thus the perpen- dicular stacking arrangement is the most energetically favorable and in fact has a stabilizing effect on the protein structure. The interplanar stacking angles and distances observed in the K4 structure agree well with those reported for other proteins [68]. In several cases, the planar surfaces of these aromatic K4 side 97 25 25 Figure 37. Stereoview of perpendicular aromatic stacking interactions between Trp25, Trp62, Ph864, Trp72 and Tyr74. 98 chains appear to interact in a similar manner with other groups having a partial positive charge. A Leu46 methyl group is positioned 2.8 A from the plane of the Trp25 indole. Likewise, a Leu77 methyl group is 2.7 A from one face of the Tyr9 ring, while the Gly11 main chain amine hydrogen is directed at a distance of 2.5 A toward the opposite face. Perpendicular stacking interactions very similar to those of the aromatic rings but involving Pro residues are also observed in the K4 structure. The edges of the Pr054 and Pr061 side chains are directed at angles of approxi- mately 80° toward the faces of the tyrosyl rings of Tyr41 and Tyr2 respectively, at separations of about 3 A (Figure 38). In addition, the edge of the Pr030 side chain is oriented similarly about 3.5 A from the Pr068 ring face (Figure 39). This type of clustering between Pro and aromatic side chains has been observed previously for other protein structures [69]. This behavior of the Pro residues may be rationalized using the same principles which influence the contacts of the aromatic residues since the Pro side chains also have a nearly planar conformation. Although the hydrogen atoms of Pro side chains do not lie in the plane of the ring, they are directed away from the center, giving the side chain a charge distribution which is less pronounced but analogous to that of the aromatic residues. In contrast, although the two l-Iis residues of K4 both have neighboring planar ring side chains, neither displays an edgewise interaction. The imidazole groups of H1331 and His33 are adjacent to the Pr030 and Phe64 side chains respectively, and in both cases, the neighboring rings are approximately parallel to one another (Figure 40). In the case of Hi633, which appears to be de-protonated, the radial lone electron pair of the imidazole nitrogen atom would disrupt the edge/face dipolar charge distribution of the ring which encourages the perpendicular stacking arrangements. Similarly, although the 99 43% . 431441 71: a)? .e .e E7 €57 Figure 38. Stereoviews of perpendicular interactions of a) Pr054 with Tyr41 ring and b) Pr061 with Tyr5 ring. 100 1.6 9 6: Pt: Figure 39. Stereoview showing perpendicular interactions of Pr030 and Pr068. 101 64 64 Figure 40. Stereoview showing parallel stacking of His31 and His33 rings with Pr030 and Ph964 side chains. 102 form of the His31 side chain is uncertain, the presence of an equilibrium between protonated and de-protonated states may also alter the nature of its interactions. The observed K4 side chain structure is generally in accordance with previous spectroscOpic results. The central hyrdophobic core and aromatic side chain interactions agree well with nmr and NOE experiments [35.37.42], which indicated connectivity between these same residues. Some more detailed observations of Atkinson and Williams [37] are also confirmed, including the specific hydrogen-bonding interactions between Thr16 OGH and a Glu73 carboxy group and between Thr65 and Hi331 (Table 17). Although NOE connectivity was also reported between Hisaa and Leu46, there is no direct contact observed between the two residues. There may, however, be some indirect communication between the two via the Trp25 side chain. A particular point of interest regards the side chain conformation of Tyr9. An nmr investigation of K4 suggested the presence of two alternate orientations for this residue, one in which the side chain was free to rotate and one in which rotation was severely hindered [70]. However, in the crystal structure of K4, Tyr9 has a single well-defined conformation in which the tyrosyl ring is highly immo- bilized (Figure 41). Stabilizing influences in the surrounding structure include the previously mentioned van der Waals interactions of the ring faces to the Leu77 methyl and Gly11 amide groups and hydrogen-bonding of the side chain hydroxyl oxygen. In the present structure, a torsional rotation of 60° about the CA-CB bond would cause the Tyr9 side chain to collide with the amino terminal region of the peptide chain; however, given adequate flexibility of the terminal regions in solution, a rotation of 60-100° would expose most of the tyrosyl side chain to the solvent where it could rotate freely. As was observed for the main chain atoms, the average thermal factors of 103 Figure 41. Stereoview showing well-defined electron density observed for Tyr9 side chain. 104 the side chains correspond closely to their positions in the three dimensional structure (Figure 28). Large thermal factors are observed for the terminal residues of the peptide chain as expected due to the general disorder in this region. The 8 value of Hisa is especially large because the side chain is completely exposed and free to rotate in the solvent. Other residues throughout the peptide chain having particularly large B values include Ser residues 13, 24, 26, and 27, Lys21, Thr29, Gln 34, Asn40, Thr47 and Met48. All are also highly exposed on the surface of the kringle and have no neighboring side chains with which to form stabilizing interactions (cf. Table 17). The side chains of Thr12 and Glu39, for which no density was observed, have similar exposed environments. In contrast, internal side chains, such as those of the Cys, Pro, His, and aromatic residues, have consistently low thermal factors. In some cases, including Tyr2, Trp25, Tyr41 and Tyr74, the side chain B values are actually lower than the values of the corresponding main chain atoms, emphasizing the structural importance of these side chains. Among the aromatic residues, the thermal factor of Tyr50 is relatively high. This can be accounted for by the less restricted environment of this side chain. The Tyr50 ring has a comparatively high accessible surface area, with no interference to movement in the edgewise direction. Furthermore, the Thr12 and Met48 residues adjacent to the faces of the Tyr50 ring are both, to some extent, disordered themselves. X. INTERMOLECULAR INTERACTIONS AND CRYSTAL PACKING In the orthorhombic crystal structure, a number of side chain interactions occur between neighboring kringle molecules of the lattice. The most significant of these is a very complex binding interaction which involves three symmetry- related molecules and a sulfate ion (Figure 42). The sulfate anion can be consi- dered to be primarily associated with one of the kringle domains, which binds the ion via the positively charged side chains of Lys35 and Arg71. In addition, the Arg71 guanidinium group also particiates in an ion pair interaction with the Asp57' side chain carboxy group of a second kringle. Further linking the two molecles, the adjacent positively charged LysSB‘ side chain of the symmetry mate also binds the same sulfate ion. Finally, the Arga2" side chain of yet a third kringle interacts with both the sulfate ion and Asp55 of the first molecule. Thus, all charges in this region are offset, with the negative charges contributed by the sulfate ion (2-), Asp55 (1 -), and Asp57 (1-) balanced by the four positively charged side chains LysSS, Arg71, LysSB', and Arg32". It is obvious that, due to the presence of numerous interactions within a small region on the kringle surface, there is a very strong association between these symmetry mates. Interestingly, the interaction involves three residues which are crucial to the kringle lysine-binding site: Asp55 and Asp57, which comprise the negative ionic center, and Arg71, the positive ionic center. This is highly significant, and the interaction will be mentioned again in a later detailed discussion of the binding site. From the perspective of crystal packing, this trimolecular interaction may be viewed as the result of crystallographic translation of a bimolecular ligand- like interaction, in which Asp55 and Asp57 from the lysine-binding site of one kringle interact with Arg71 and Arg32, respectively, of a second kringle. 105 106 55 55 58 “I 32.. 58 TI 32. 57' }\7:EO 57' 71 Figure 42. Stereoview showing interactions of sulfate anion with side chains of Ly335 and Arg71 of molecule 1, Ly558' of molecule 2, and Ar932" of molecule 3. Additional interactions occur between Arg71/Asp57' and Asp55/Arg32'. 107 Repeated throughout the crystal lattice, this coupling of negative and positive centers forms a kringle chain which corresponds to a crystallographic two-fold screw axis in the x—direction (Figure 43). On a second face of the kringle, an intermolecular hydrogen-bonding interaction occurs between Asn43 ND and Asn76' OD of a neighboring mole- cule. There is also a solvent-bridged interaction between the Gln7 carboxy group and the side chain amide group of the same Asn76' residue (Figure 44). The symmetry mates linked by these interactions are related crystallogra- phically by a two-fold screw along the y-axis (Figure 45). Finally, the only contact between K4 molecules along the third two-fold screw axis in the z- direction (Figure 46) appears to be a solvent-bridged interaction between the Ser14 OG atom of one molecule and the Ser69' OG atom of a second (Figure 47). A comparison of the contacts between kringle molecules related by each of the three P212121 symmetry elements easily explains the observed crystal face development. The intermolecular interactions in the direction of the x-axis are by far the strongest, and, not surprisingly, the orthorhombic K4 crystals grow most rapidly along the 8 axis. Lastly, it is also important to recognize that the interactions observed between K4 domains in the crystal lattice may perturb the participating residuesfrom the conformations which would be found for the isolated molecule in solution. This is certainly true and especially pertinent for the residues of the lysine-binding site, at which the interactions are unusually intimate. 108 ., c - ' ”2‘ «44:54:; sari-w / (a) if"? . .av “‘ «fat -‘ “'v' . tings: 6’ e .’.\‘"- fi‘ #1 0*! ‘ fl : g‘ps 4., fl“ '1‘ I "flit ,8" '2’ . 12;»: f-s' . i yfgezg r40 f \. -$ ’ v a ‘ ’ \ ‘93-"; .4 A“ \ ‘1‘e’?’ ‘9 ‘5! - ”W131” ., [3.339% ‘ ‘a‘w" ‘ero "c 94.391! ‘7 r p ,. . ’ t$;‘.‘$l :‘ O. .311. \° )‘ c .3“ ' ..~ 2' ‘h‘ ‘37 ' a,‘ : )0 ‘. ‘ .\ gig-‘yu‘ ‘ ‘9‘ ‘1'. .519}: ° .5 ‘5 .5153.” Y, a an O. / o “15.... , Q .i-a 4 "5 fl ' 7'0) ; JQA ”£5. , e“\w$é5“ . ’ ’f‘ ‘ ‘ . - a e \Ul‘l‘z' 6 9 3)" g‘ I .05 v. /. #‘t‘Q’ " é» Figure 43. Stereoview illustrating two-fold bold. screw axis in. x-direction resulting from trimolecular kringle-kringle interactions at sulfate anion. Sulfate shown in 109 44% 48s a; W Figure 44. Stereoview of intermolecular interaction between Asn43 side chain and Asn76‘ side chain of symmetry mate. Also shown is a solvent-bridged interaction between Gln7 and Asn76'. Hydrogen bonds are indicated with dashed lines. 110 Figure 45. K4 molecules related by two-fold screw axis along y-direction. lfile'sjidues involved in intermolecular side chain hydrogen bonds are shown in o . 111 Figure 46. K4 molecules related by two-fold screw axis along z-direction. Solvent molecules which bridge side chains of adjacent symmetry mates are seen at protein-protein interfaces. 112 P re 47. Solvent-brid ed interaction between side chain. of Ser14 and Ser69' oIgnueighboring moleculeg. Hydrogen bonds are indicated With dashed lines. XI. THE STRUCTURE OF THE K4 LYSINE-BINDING SITE The K4 lysine-binding site is found on the flat surface of the overall plano- convex disc, where it is bordered by the peptide segments 31-35, 54-58, 61-64 and 71-75. Thus, the site is supported primarily by the inner kringle loop, extending from the outer rim of loop C to the far end of loop D. The binding site itself has the form of an elongated depression on the kringle surface which is lined by the indole rings of Trp62 and Trp72 (Figure 48). These two side chains are oriented in an anti-parallel manner with an interplanar angle of approximately 80°. This arrangement forms a hydrophobic trough which lines the bottom and one wall of the binding site depression. Although the Trp62 side chain, which comprises the bottom surface, has minimal solvent accessibility (5%), the long edge of the Trp72 side chain projects out into the solvent and is exposed to an unusually high extent for this residue type (38%). The Trp62 and Trp72 side chains display perpendicular stacking interactions with the rings of Phe64 and Tyr74 respectively, resulting in a symmetric, stabilized structural framework for the binding site. In addition, the 31-35 peptide segment, which bounds the site across one end of the aromatic trough, includes two His residues', which lend further structural stability to this region. The charged residues which have been implicated in ligand binding are arranged in a semi-circle along the outer edge of the Trp62 indole, completing the remaining walls of the binding site depression. The two negatively charged Asp55 and Asp57 residues are located at one end of the hydrophobic trough, with the carboxy centers within 5.5 A of one another. The positively charged Arg71 residue is positioned at the opposite end of the trough, with the side chain extending outward away from the site, placing the guanidinium group about 12 A from the Asp55/Asp57 pair. In addition, although not identified as 1 13 35 114 55 55 5 62 62 71 1 35 71 31 3 Figure 48. Stereoview of K4 lysine-binding site. 115 necessary to the binding site, Lys35 is also found along the aromatic trough between the Asp and Arg centers. The Ly535 side chain extends into the site where the amino group is approximately 7.6 A from the Arg71 side chain. Thus, there appears to be a charge pair at the positive as well as negative center. It can be seen that the general three-dimensional structure of the binding site corresponds nicely to the observed K4 binding specificity. The terminal positively and negatively charged groups of zwitterionic ligands such as ACA are attracted electrostatically to the complementary opposed anionic and cationic centers of the binding site, while the aromatic trough provides favorable van der Waals interactions for the hydrocarbon body of the ligand. It is believed that the physiological ligand of this binding site is a carboxy terminal Lys residue of fibrin [20], which also has positively and negatively charged ends and therefore may bind in a similar fashion. Though the basic binding mecha- nism may be the same, it is complicated by the fact that, in this case, the Lys ligand carboxy group is simultaneously the terminus of a fragment of fibrin polypeptide. It appears that certain characteristics of the binding site may help to accommodate the fibrin Lys ligand. Of the two ionic centers, the positively charged end of the binding site is considerably more open. The 54-58 peptide stretch, which serves as the upper boundary of the site, crosses closely to one end of the Trp indole surface and the Asp side chains have limited freedom of movement. However, the lower 31-35 boundary recedes slightly from the aromatic trough and the cationic center is supported by the much more flexible Arg (and possibly Lys) side chain. Thus, although the amino group at the end of the fibrin lysine side chain may fit easily into the tightly held anionic region of the binding site, with a less relaxed topology at the Opposite end, steric inter- ference by the fibrin chain might prevent an efficient interaction of the ligand with the cationic center. 116 The observed binding site also corresponds well to previous experimental results. Aromatic nmr experiments [41,42] found ligand-binding to perturb most strongly the chemical shifts of the side chains of Trp62, Phe64, and Trp72, in accordance with their observed positions at the center of the site. Signals of the residues Trp25, Hi531,Tyr41 and Tyr74 were also shifted, but to a lesser extent. Of these, Trp25 and Tyr74 communicate with the binding site through aromatic stacking interactions with Trp62 and Trp72 respectively. The His31 side chain is adjacent to the base of the Arg71 side chain and may sense the effect of the ligand on this ionic center. The Tyr41 ring, however, is spatially quite remote from the binding site and may only be affected indirectly through the residues of the hydrophobic core. The side chain of His33 was reported to be sensitive to the binding of bulky, but not linear, ligands [42]. It appears that this ring, which lies parallel to to Phe64, is also indirectly affected by perturbations of the phenylanyl side chain caused by the greater contact with bulkier ligands. The projection of the Lys35 side chain toward the binding site positive center agrees well with the results of chemical modification studies [71] which found the blocking of this residue to weaken affinity of K4 for Lys-sepharose. On the other hand, although a role in ligand-binding has been proposed for His31 [41], this side chain appears to be too far removed to participate directly. Ligand-binding has also been reported to change the flip rate of Tyr9 [37]; however, there is no obvious reason for this observation based on the crystal structure, in which Tyr9 appears only to communicate with the binding site very indirectly. The amino terminal of the K4 peptide chain, which partially restricts the Tyr9 side chain conformation and which is not fully determined in the present X-ray structure, may be responsible for this effect. Although the general characteristics of the lysine-binding site fit the predicted model of ligand-binding, the detailed side chain conformations at the 117 ionic centers do not. Both Asp57 and Arg71 side chains are directed away from the binding pocket, with a separation between the charged centers of approxi- mately 12 A, poorly matching the observed optimal ligand length of 6.8 A. This unexpected result stems from the intermolecular interactions mentioned pre- viously between Arg71 and Asp57 side chains of neighboring molecules (Figure 49). Thus, these residues are both directed outward to make contact with a symmetry mate, rather than inward in a ligand-binding fashion. However, torsional rotation of the Arg71 side chain easily brings the guanidinium group within 8.5 A of the anionic center, in good position for ligand binding. The intermolecular interactions occurring at the binding site appear to affect the conformations of other residues as well. For example, the Ly335 side chain is directed toward the cationic end of the binding pocket to form an ion pair with the sulfate anion which is docked there. This observed behavior suggests the possibility that this residue may similarly reinforce the cationic center during ligand-binding. In the crystal structure Lys58 binds to the sulfate associated with the binding site of a neighboring molecule, drawing the side chain away from the binding site of the host molecule in an extended confor- mation. In solution, the conformation of this side chain may be very different, especially considering the availability of the adjacent oppositely charged Asp57 side chain of the binding site. Perhaps the most interesting intermolecular contact is the ligand-like binding of the ArgS2 residue of a second symmetry mate to both the sulfate ion and Asp55 of the binding site. The Argaz side chain projects into the hydrophobic trough, with the aliphatic chain abutted within 2.5 A of the face of the Trp72 indole, mimicking the expected behavior of more typical binding site ligands. Thus, although the kringle-kringle interactions occurring in the crystal structure distort an accurate picture of the free binding site. at the same time they present an interesting view of the kringle ligand 118 Figure 49. Stereoview showing ion pair interactions which occur at lysine- binding site. Sulfate ion and side chains from second symmetry-related molecule are shown in bold. 119 binding ability. The binding site interactions explain the observed inability to bind ACA in the orthorhombic K4 crystals during soaking experiments. The close inter. actions, which occupy both the ionic centers and the hydrophobic pocket, preclude the access of ACA to the binding site. Should this small ligand even- tually gain access to the site and displace the protein-protein interactions, the disruption of these important intermolecular contacts would naturally lead to destruction of the crystal stmcture itself. This possibility may account for the cracking and crumbling of orthorhombic K4 crystals soaked with high concen- trations of the ligand. XII. SOLVENT STRUCTURE The final K4 crystal structure includes 97 ordered solvent molecules. These are listed, along with their crystallographic parameters, in Appendix B. The solvent occupancy and thermal factor distributions are shown in Figures 50 and 51. There is a normal distribution of thermal factors about the mean value of 25 A2, with nearly 75% of the values falling between 20 and 30 A2. This type of distribution does not occur for solvent occupancies; in this case there is a roughly even spread between values of 0.5 and 1.0. Among the solvent molecules, one sulfate ion was found to be associated with each kringle molecule (Figure 52). As was previously described, the sulfate anion is located at the cationic center of the K4 lysine-binding site, where it is coordinated by four Arg and Lys side chains from three separate symmetry related molecules. The geometric parameters of these interactions are given in Table 18. As might be expected from the number of stabilizing hydrogen- bonding interactions, the sulfate has a relatively low average thermal factor of 19 A2, a value which is comparable to the average value for protein atoms. A comparison of the individual atomic B values also agrees well with the observed structure. The central sulfur atom has a B value of 16 A2, which is significantly lower than the average value of nearly 20 A2 related to the surrounding oxygen atoms. This difference can easily be explained by a certain degree of rotational freedom about the center of mass. Furthermore, the fact that atoms O2 and 03 each have B values of 18 A2 whereas O1 and O4 have slightly higher values of 21 and 22 A2 corresponds nicely with the observed binding interactions. The former two oxygens each have contacts with two protein donors whereas the latter two oxygens each have only one. The remaining solvent sites are all assigned as water molecules and are 120 121 d; 6 6:838 cm 3m... $38.08 $83 co>ow 82:00.08 2528 .8820 22 865380 8 8:33:65 .8 2:9“. >OZ (A2) 20.4 169 Table 24. Distribution of magnitude of difference between stmcture factors calculated from two sets of orthorhombic K4 intensity data measured before and after characteristic relative intensity changes. e Numberof (Di 2 Measurements 0-1 1130 1-5 901 5-10 430 10-15 161 15-20 105 20-25 71 25-30 39 30-35 39 35-40 22 40-45 16 45-50 8 50- 100 23 100+ 11 Difi2=(|F1| - "‘202 Table 25. Reflections having largest stmcture factor discrepancies between first and second K4 data collections ( Diff2 > 100, where Diff2 = ( |F1| - [F2] )2). h,k,l 26 (deg) Resolution (A) 01112 1 7 0 14 32.0 2.6 636 2 5 15 5 32.0 2.6 506 3 7 0 15 33.5 2.7 474 4 6 12 11 34.1 2.6 463 5 1 11 0 20.1 4.4 207 6 7 5 0 21.4 4.2 148 7 2 4 6 17.0 5.2 130 8 0 10 0 18.1 4.9 117 9 6 2 0 22.4 4.0 111 10 11 6 5 35.3 2.5 109 11 2 3 9 17.9 5.0 106 170 100L 90F 70 i- 50) 40L 30F NUMBER OF REFLECTIONS 20r 1OF F: 0 5 10 15 202530 3540 29 Figure 73. Distribution of scattering angles for set of 334 reflections having very large discrepancies between original and second orthorhombic K4 data collections ( (|F1|-|F2|)2 > 100). 171 The observed map is generally featureless, showing primarily fragments of positive difference density in the empty solvent regions between K4 molecules. However, two larger regions of difference density are observed. Some unusually shaped density is found near the last ordered residue of the amino terminal (Figure 74). A second region occurs at the lysine-binding site between the Trp72 indole ring and the side chains of symmetry mate Ar932 and Met28 residues (Figure 75). The latter region is the only difference density observed in the immediate vicinity of well-ordered protein structure. A similar difference density map calculated using the remaining 2622 reflections at the same reso- lution range displays density only in the intermolecular solvent cavities, none of which coincides with that observed using the set of 334 highly discrepant reflec- tions. Thus, these observations suggest that the intensity changes of orthorhombic diffraction pattern may be related to structural changes occurring at the poorly ordered amino terminal tail region, which was not observed at all in the K4 structure determination, and at the site of the ligand-like kringle-kringle interaction. A fuller understanding of these changes will require further investi- gation of these preliminary results. 172 Figure 74. Stereoview showing electron difference density observed near amino terminal of K4 peptide chain when map is calculated usmg only 334 reflections having very large discrepancies between original and second intensity data sets. 173 Figure 75. Stereoview showing electron difference density observed in vicinity of ligand-like intermolecular interaction when difference map IS calculated using only 334 reflections having very large discrepancies between original and second intensity data sets. APPENDICES APPENDIX A Atomic coordinates and crystallographic parameters of orthorhombic plasminogen K4 crystal stmcture. Seq Res hb.TypeAtcm C C C C 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 4 4 4 4 5 5 5 5 5 S 5 5 ASP ASP ASP ASP CYS CYS CYS CYS CYS CYS TYR TYR TYR TYR TYR TYR TYR TYR TYR TYR TYR TYR HIS HIS HIS HIS HIS HIS HIS HIS HIS HIS GLY GLY GLY GLY ASP ASP ASP ASP ASP ASP ASP ASP N CA C O N CA C 0 CB SG N CA C 0 CB CG CD1 CE1 CZ C32 CD2 OH CA CB CG NDl CEl N32 CDZ OD1 OD2 X .60960 .96030 -4.82330 .25320 -5.08620 -5.91320 -5.43360 -4.64700 -7.40010 -7.67380 -5.99120 -5.61910 -6.90010 -8.00040 -4.84870 -5.60630 .62590 .27680 .94110 .94680 .25320 .59830 .74660 .82330 .66680 .61830 .92600 .89930 .01980 .65640 .34770 .85250 .77480 .91950 .22760 .61200 .22990 .48320 .18020 .36220 .34690 .53110 .35060 .06590 CoolantesiA) \DQO‘OQQUIUIUIU'I Y .00590 .36460 .87650 .13260 .17910 .80140 .23300 .81020 .84170 .69730 .78500 .21650 .05200 .56600 .37270 .01830 .68930 .41780 .38900 .69640 .97870 .20400 .32200 .32210 .18520 .46040 .27850 .52780 .03100 .45440 .56810 .24940 .64590 .44790 .86660 .79150 .58520 .08770 .41530 .05680 .19250 .34360 .29220 .26610 174 Z 20.70820 21.21420 20.06620 19.16120 20.13850 19.08820 18.82540 19.58930 19.46210 21.04680 17.78000 17.53920 17.49130 17.10570 16.22390 14.95920 14.46600 13.26630 12.56130 13.05260 14.22090 11.38000 17.84770 17.75970 16.49710 15.91720 18.94160 20.23270 20.83400 21.96050 22.11580 21.02990 15.97010 14.77520 13.55340 13.05650 13.06130 11.87160 12.32650 11.49310 10.81850 11.37960 10.62430 12.51850 B (A2) 39.44000 37.01870 34.01270 36.64960 31.06120 28.42870 21.04750 23.79590 29.37410 35.36780 23.51860 20.95490 20.02810 19.37630 17.85780 17.59420 15.64240 16.37740 13.97010 13.93040 12.43000 20.51060 22.59660 22.58150 22.74330 25.77080 26.44550 26.07700 30.09910 31.88160 34.49220 28.36130 23.57470 25.73750 22.86140 25.79960 20.29500 16.47700 11.05580 11.62980 15.78720 14.09670 17.20790 11.68460 W 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 1.00000 H H H~H H'H'H H H‘H'HIH'H'HIH'H H H»H H H HiH mmcmomwmommommmmquqqqqqqq~tma~iam Hiuiaididha hauiuih) h‘h‘k‘h‘h‘k‘k‘h‘H‘H ADK)R>A)F'H|H HWDfD¢3<3:;2;2;E;EIgLS GLY N GLY GLY C GLY O GLN N GLN GLN C GLN O GLN GLN GLN GLN GLN SER N SER CA SER C SER O SER CB SER OG CA CB CG CD1 CD2 C81 CE2 CZ OH CA CB CG CD NE CZ NH1 NH2 GLY N GLY GLY C GLY O ALA N ALA SER N SER SER C 081 N82 —4.97920 -3.85030 -2.52720 -1.49240 -2.56260 -1.31060 -0.77790 0.40880 -1.47280 -1.66130 -1.68430 -0.58130 -2.83030 -1.52910 -1.08980 -0.71250 -0.27530 -2.15240 -3.35650 -0.90010 -0.58040 0.89080 1.68400 -1.08640 -0.74540 -1.33780 0.26290 -0.94410 0.69770 0.06980 0.56160 1.21900 2.61620 2.92500 3.92950 3.07740 3.13910 4.31360 4.40640 3.56790 2.64610 3.56350 2.12010 2.24010 3.42050 4.29120 3.41360 4.44380 4.01560 4.80650 4.73270 2.80330 2.38350 3.33380 bammhmmqmcoto 175 .28400 .68320 .42980 .77870 .71620 .49870 .60790 .87790 .86710 .90940 .39950 .90570 .94970 .39710 .48110 .12800 .06000 .14180 .37720 .09080 .69120 .51600 .74570 .64900 .22510 .56100 .59190 .24750 .30270 .64010 .35330 .13370 .93940 .53460 .32640 .07800 .43530 .63070 .95350 .39030 .58320 .69200 .57350 .17960 .43660 .90500 .15670 .18340 .28560 .39030 .43530 .41530 .54810 .68110 13.61490 14.24510 14.09460 14.28100 13.77930 13.67680 15.09670 15.33020 13.02560 11.51060 11.07570 10.97350 10.87470 16.13290 17.50620 18.07200 19.21700 18.42020 18.39800 17.27270 17.74840 18.06720 17.15860 16.73930 17.16510 18.24660 16.44050 18.54310 16.75630 17.80070 18.05710 19.28340 19.63000 20.13260 20.81450 20.55170 19.84390 18.96750 18.33200 17.42680 16.92400 17.12930 19.76190 20.13000 19.55790 18.82570 19.92040 19.52080 18.36590 17.95910 20.85370 17.81110 16.69510 15.50750 9.63040 13.04590 17.67000 13.50620 16.18800 15.03470 10.33340 14.28420 17.30170 13.37330 11.04240 13.02340 12.24860 10.24800 12.23250 13.75170 17.66110 17.88670 20.54300 17.86730 16.46040 15.88040 18.05230 17.85800 14.92670 13.33350 16.02040 15.79380 13.17220 18.22460 19.25690 13.51460 15.54950 18.63550 20.13970 21.05560 16.58900 18.37720 16.96580 16.02260 14.93750 17.93590 19.02930 22.09100 21.22440 20.69700 20.00350 19.02900 20.22260 17.73360 21.35670 19.92990 17.49400 20.28460 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 13 13 13 14 14 14 14 14 14 15 15 15 15 15 15 15 16 16 16 16 16 16 16 17 17 17 17 17 17 17 18 18 18 18 18 18 18 19 19 19 19 20 20 20 20 20 20 20 20 20 21 21 21 21 SER SER SER SER SER SER SER SER SER THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR THR GLY GLY GLY GLY LYS LYS LYS LYS LYS LYS LYS LYS LYS LYS LYS LYS LYS C62 CA CB 061 CGZ CA CB 0G1 CG2 CA CB 061 C62 3.82310 0.94220 0.10620 3.62160 4.54620 4.28750 5.30250 5.98470 6.29630 3.00050 2.64990 1.53290 0.66540 2.52920 1.47700 2.61050 1.64480 0.65150 -0.59950 -0.68790 1.24450 1.24030 2.67740 -1.64120 -2.96040 -2.88430 -3.82650 -3.58490 -4.40220 -3.95420 -1.85640 -1.61580 -0.57990 ~0.11900 -1.21270 0.17490 ~2.13480 -0.17870 0.79280 2.26050 3.10340 2.60210 4.00130 4.64240 3.91860 4.13430 3.60580 3.73870 3.39400 1.95320 5.96710 6.61300 7.00120 7.47320 III HOHONHHNNwNNwwaLfl HwNHOOHNONOHN 176 .79620 .76740 .99570 .53230 .62090 .52960 .12180 .52030 .34600 .19440 .18640 .77760 .45050 .29740 .12060 .42360 .53750 .05230 .14370 .07210 .08830 .69170 .61020 .59650 .98540 .43120 .23400 .15200 .00640 .80340 .79910 .14080 .98530 .02620 .06220 .56230 .18840 .56890 .21740 .96830 .55780 .08980 .75550 .95210 .21900 .03070 .83390 .01390 .59770 .96090 .00720 .25240 1. 1. 13610 14560 15 16 17 13 12. .09770 12 14. .89080 14 12. .53870 .70820 .26580 .05350 11 10 11 12 11. 13. 9. 8.46060 8.52270 9.13770 7.00470 6.49950 6.91460 7.84050 7.65100 7.07720 7.34200 6.79140 7.71960 5.41970 6.37130 5.77060 6.51320 6. 4 4 3 7 8 8 8 7 7 8 8 5 4 3 1 1 8 9 8 7 .26560 .22760 .08900 14. .73170 84380 69110 18880 56890 44590 54420 41810 01580 .26900 .27810 .40980 .67620 .57680 .27770 .99110 .32390 .00600 .11720 .83880 .67200 .51080 .24470 .92170 .91520 .25240 .33430 .78460 .63250 15.59020 24.84850 33.31560 14.57160 11.76980 14.15020 13.59220 14.67690 18.10470 13.00840 13.11610 13.42020 16.43490 17.55060 17.93550 13.77540 12.36260 12.16110 18.41360 15.97260 13.53530 14.65210 11.48240 14.69090 16.14960 18.34280 19.12300 17.88530 18.58360 12.51520 15.80590 16.37400 13.25580 15.35810 15.14690 16.72810 17.49040 14.92410 16.00880 17.26360 16.00320 14.54220 13.73360 13.72090 13.86710 15.13420 15.16760 16.99810 19.57120 20.21440 10.35960 15.02520 11.76510 14.34990 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 21 LYS 21 LYS 21 L28 21 LYS 21 LYS 22 023 22 cxs 22 Cys 22 CYS 22 CYS 22 cxs 23 GLN 23 GLN 23 GLN 23 GLN 23 GLN 23 GLN 23 GLN 23 GLN 23 GLN 24 SER 24 sea 24 SER 24 see 24 sea 24 SER 25 TRP 25 TR? 25 TRP 25 TRP 25 TRP 25 TRP 25 TRP 25 TRP 25 TRP 25 TRP 25 TRP 25 TRP 25 TRP 25 TRP 26 SER 26 SER 26 SER 26 SER 26 SER 26 sea 27 sea 27 SER 27 SER 27 569 27 sea 27 SER 28 MET 28 MET CB CG CD CE NZ CA CB SG CA CB CG CD 081 NE2 CA CB OG CA CB CG C01 C02 NE1 C32 C83 C22 C23 CH2 CA CB OG CA CB OG N CA 7.86730 8.95030 9.96620 10.70210 11.69700 6.83520 7.22440 8.75090 9.48850 6.72680 4.91820 9.16400 10.55560 10.81360 9.89620 10.69580 12.05890 12.23500 11.85790 12.93300 12.04130 12.46070 12.47300 12.90360 13.87340 13.85240 11.95840 11.86040 13.23740 13.37880 11.06470 9.67080 9.21180 8.55270 7.87100 7.42680 8.39030 6.16150 7.15010 6.04540 14.22060 15.58880 16.32270 17.32820 16.41450 16.64830 15.86770 16.54770 16.03480 14.85120 16.34270 16.81890 16.90700 16.58050 177 -1.03400 -0.19150 -1.03120 -0.17810 0.62550 2.16520 3.51680 3.71250 3.25230 4.56460 4.75180 4.43170 4.88550 6.05320 6.85760 5.31360 5.85240 6.43150 5.96910 7.54940 6.23460 7.35360 8.66280 8.66890 7.27300 6.05900 9.69830 11.03730 11.57990 12.25520 11.94310 11.50390 10.97900 11.62180 10.71130 11.14850 12.11390 11.09590 12.09360 11.56950 11.24800 11.74640 10.87060 11.32380 11.98840 10.63720 9.68580 8.89560 9.18880 9.52130 7.40960 6.61620 9.06570 9.26340 9.73530 .42160 .16430 .18860 .40960 9.54010 9.07830 9.07030 9.96490 .14050 10.16230 .04720 .82790 .77220 .09670 .36750 .02830 .66510 .60690 .67160 .25680 .12850 .30230 .15540 .73270 .49610 .93990 .25020 .90300 .85130 .17620 .47260 .65800 .60170 .50980 .29010 .29970 .71080 .72460 .39770 .71190 .39310 .36820 .83050 .64260 .08030 .00290 .98400 .58230 .33660 7.38800 6.33290 4.57340 3.16960 H H H -q G1m>¢>FJa>~Jaa H H H H H 131410 H Cic> H H UIU165G>FJCD~JCD\D\Dc>w101010 H H h»UIUIa\~J~JCDKOPOFJ\D‘D‘4‘” .22230 .60770 .44070 .57360 .73650 .23270 .40470 .65790 .82310 .56550 .64940 .99850 .48250 .98220 .51120 .01450 .08280 .18340 .93470 .42160 .98710 .14630 .35420 .52640 .22940 .78970 .65520 .13580 .67430 .29400 .95720 .18970 .48010 .30090 .26820 .50070 .05110 .17640 .09920 .43610 .15880 .82170 .10020 .87080 .34400 .04080 .48980 .07600 .75710 .06360 .47060 .74440 .45000 .11480 15.17810 10.92750 19.47760 25.53290 28.34660 30.03060 16.52970 12.54610 14.83030 17.39640 14.68480 16.33100 14.49650 13.60380 17.71470 17.90770 16.60080 16.50850 18.19550 20.73270 18.58920 13.75040 22.33080 23.37130 26.90280 28.01710 24.05430 22.95290 20.31520 28.14320 26.31530 28.44780 31.70050 31.25860 24.63710 33.42000 35.03130 33.01360 24.70070 22.77740 19.54140 18.81350 17.49650 14.75820 17.54330 17.41090 13.59450 11.07120 14.09380 13.29680 12.32900 16.32410 17.47020 18.25490 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 42 PRO N 42 PRO CA 42 PRO C 42 PRO 0 42 PRO CB 42 PRO CG 42 PRO CD 43 ASN N 43 ASN CA 43 ASN C 43 ASN O 43 ASN CB 43 ASN CG 43 ASN 001 43 ASN N02 44 ALA N 44 ALA CA 44 ALA c 44 ALA o 4 4 ALA CB 45 CL! N 45 GLY CA 45 GLY C 45 GLY O 46 LED N 46 L30 CA 46 L00 C 4 6 LED 0 4 6 LED CB 4 6 LEU CG 46 LEU C01 46 LED C02 47 THR N 47 THR CA 47 THR C 47 THR O 47 THR CB 47 THR 0G1 47 THR C02 48 MET N 48 NET CA 48 NET C 48 MET O 48 MET CB 48 NET CC 48 NET SD 48 NET CE 49 ASN N 49 ASN CA 49 ASN C 49 ASN 0 49 ASN CB 49 ASN CG 49 ASN 001 9.99550 10.35920 9.30070 9.72600 10.94780 10.83020 10.27660 8.05620 7.01170 6.22120 5.27560 6.01080 6.73430 7.56440 6.38530 6.51760 5.73320 5.94750 5.22910 5.98870 6.87960 7.35240 7.96550 7.72220 8.69810 9.29970 10.44650 11.60840 9.65820 8.62810 7.14040 8.84290 10.14840 11.18460 11.03630 9.89240 11.19170 10.33460 10.65500 12.15320 12.17570 11.36810 11.58870 11.60730 12.20780 13.88970 13.97890 10.46410 9.63040 8.13300 7.33640 9.94960 9.60270 9.61100 180 22.20180 22.24930 22.05930 21.67350 23.67550 24.35700 23.42640 22.33690 22.18080 20.88290 20.81360 23.34680 24.70360 24.84290 25.68170 19.96560 18.77780 17.57850 16.58300 18.28820 17.61120 16.56410 15.43220 14.26410 15.63400 14.58160 13.80450 14.13850 15.16400 15.16490 15.06640 16.32560 12.81140 11.99760 10.51580 10.06210 12.28590 11.35620 13.71330 9.79030 8.34870 8.07340 8.76990 7.55830 8.01900 7.31380 7.30560 7.09720 6.74450 6.80430 6.02760 5.33030 4.22370 4.47200 11.89290 13.26790 14.32040 15.44100 13.47810 12.17620 11.14880 14.01770 15.02930 14.83500 15.63240 14.88540 15.04400 15.91440 14.23930 13.94140 13.69890 14.59410 14.42700 12.24850 15.46740 16.36330 15.49950 15.86620 14.44430 13.61780 14.28280 13.97510 12.24920 11.12730 11.42500 10.16460 15.09870 15.78450 15.44300 15.16950 17.32040 18.06100 17.56880 15.50650 15.18780 13.91850 12.95090 16.34090 17.63980 17.75740 19.58690 13.97560 12.81920 13.18740 12.63930 12.27570 13.27470 14.49290 17.50050 16.41940 15.72270 21.95940 18.25610 22.09740 19.15520 15.43760 16.17630 17.39920 19.25880 17.51300 18.46060 22.15520 19.97530 14.94220 10.87260 12.50860 11.66240 9.42520 14.82520 16.33960 17.35930 20.09010 13.47900 20.47230 22.66040 27.84500 20.41880 26.24330 19.05700 17.72640 17.76700 21.31720 19.60880 18.74460 21.37420 32.20110 28.70860 19.05280 18.09020 14.83100 17.25400 20.45360 28.96210 36.42290 36.59890 15.09220 17.23290 14.00790 15.07070 15.89860 17.05890 14.89480 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 49 50 50 50 50 50 50 50 50 50 50 50 50 51 51 51 51 51 51 52 52 52 52 52 52 52 52 52 52 52 53 53 53 53 53 53 53 53 54 54 54 54 ASN ND2 TYR N TYR CA TYR C TYR O TYR CB TYR CG TYR CD1 TYR CD2 TYR CEl TYR CEZ TYR CZ TYR OH CYS N CYS CA CYS C CYS O CYS CB CYS SG ARG N ARG CA ARG C ARG O ARG CB ARG CG ARG CD ARG NE ARG CZ ARG NH1 ARG NH2 ASN N ASN CA ASN C ASN O ASN CB ASN CG ASN ODl ASN ND2 PRO N PRO CA PRO C PRO 0 54 PRO CB 54 PRO CG 54 PRO CD 55 ASP N 55 ASP CA 55 ASP C 55 ASP 0 55 ASP CB 55 ASP CG 55 ASP 001 55 ASP OD2 56 ALA ll 9.31190 7.80310 6.41880 5.50480 5.97660 6.40750 7.27020 7.92970 7.44680 8.72190 8.23080 8.84990 9.64940 4.24460 3.24830 3.07440 2.85070 1.91310 2.13590 3.12760 2.94330 2.18680 1.84090 4.26410 5.07210 4.42270 5.14270 4.99410 5.84200 3.99840 1.94480 1.20130 1.84500 1.22000 0.27380 -1.16340 -O.73390 -2.43330 3.03220 3.77260 3.12400 3.27420 5.11970 5.06050 3.76180 2.47790 1.85190 0.36670 -0.27120 2.25170 1.78290 1.93660 1.33500 -0.09510 181 3.04950 7.63850 7.74740 8.42420 9.32440 8.56690 8.00850 8.92110 6.63260 8.44760 6.14880 7.07230 6.65900 8.00130 8.61340 10.12910 10.40680 7.89790 6.17270 10.96780 12.41620 13.02910 12.39410 13.18490 12.70890 13.15490 12.83770 11.79510 11.66020 10.92360 14.36250 15.01120 16.40010 17.39870 15.10480 15.59220 15.73050 15.85210 16.39120 17.62760 18.63230 19.87570 17.21840 15.73600 15.23840 18.10710 18.90760 19.19950 19.49250 18.26080 16.79670 16.34080 16.15850 19.03170 12.66360 14.15870 14.58040 13.55800 12.86520 15.86620 16.95830 17.77930 17.14120 18.82920 18.17370 19.01650 20.04090 13.50320 12.62740 13.01180 14.20250 12.91910 12.27710 12.00240 12.24200 11.06990 10.05410 12.53270 13.74850 15.01500 16.22870 17.03110 18.06680 16.82790 11.16370 10.05840 9.83820 10.27180 10.49160 .37530 .23020 .63660 .25140 .06450 .11810 .33390 8.42900 8.11620 8.72940 7.10450 6.01840 6.18400 5.14230 4.69520 4.51000 3.38570 5.48790 7.40710 mmmmooaxo 12.85170 10.88760 13.21310 13.88100 17.20920 17.57610 20.75600 20.21410 22.34790 25.89360 24.76730 29.00820 34.34030 11.06220 13.24870 9.97620 9.20030 11.43400 16.28430 10.30190 14.25930 14.11390 13.05940 14.27760 16.32140 12.76630 14.77490 16.01850 18.28030 14.50100 11.31110 11.52120 13.58050 11.78300 11.69980 13.01740 13.06600 14.07390 11.22050 11.41650 13.38730 12.82520 14.23850 15.33780 12.84220 15.18710 13.12600 11.38800 12.83300 10.51120 13.72180 16.01240 13.42990 11.68160 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 56 ALA CA 56 ALA C 56 ALA 0 56 ALA CB 57 ASP N 57 ASP CA 57 ASP C 57 ASP 0 57 ASP CB 57 ASP CG 57 ASP 001 57 ASP 002 58 LYS N 58 LYS CA 58 LYS C 58 LYS O 58 LYS CB 58 LYS CG 58 LYS C0 58 LYS CE 58 LYS N2 60 GLY N 60 GLY CA 60 GLY C 60 GLY o 61 PRO N 61 PRO CA 61 PRO C 61 PRO 0 61 PRO CB 61 PRO CG 61 PRO CD 62 TRP N 62 TRP CA 62 TR? C 62 TRP 0 62 TRP CB 62 TRP CG 62 TRP C01 62 TRP 002 62 TRP N01 62 TRP 082 62 TRP CB3 62 TRP C22 62 TRP 023 62 TRP CH2 63 CYS N 63 CYS CA 63 CYS C 63 CYS 0 63 CYS CB 63 CYS SG 64 P30 N 64 PHE CA ~1.52680 -2.53680 -3.48530 -1.72540 -2.29230 -3.20920 ~4.37000 -4.55990 -2.46380 -3.29880 -3.88510 -3.37700 -5.14900 -6.33440 -6.02510 -6.89670 -7.24700 -6.60270 -7.78010 -7.27410 -8.43300 -4.83080 -4.50110 -3.02510 -2.44650 -2.54470 -1.11450 -0.76200 -1.63590 -0.89510 -2.17070 -3.24540 0.44710 0.99790 2.49780 2.97300 0.62400 1.13910 0.58400 2.40340 1.42970 2.53580 3.36580 3.63970 4.46480 4.60740 3.11270 4.59030 5.01420 4.20360 5.09360 4.47160 6.28700 6.97230 182 19.26970 18.59680 19.23950 20.79710 17.26190 16.47740 15.99790 16.46640 15.38650 14.95530 13.85770 15.64080 15.03270 14.49300 13.69210 13.58110 13.80230 12.54490 11.59490 10.33060 9.53360 13.15130 12.39190 11.98520 12.17970 11.42840 11.04430 10.00240 9.19070 10.77530 11.05730 11.35590 10.08850 9.29140 9.14510 9.76130 9.87900 11.27550 12.49440 11.56470 13.50730 12.98390 10.71590 13.57860 11.33860 12.72320 8.30620 8.11700 7.54610 6.96420 7.22430 5.48430 7.77370 7.19470 7.73360 6.82150 6.25720 7.70580 6.63120 5.81770 6.69290 7.80220 5.05450 3.86160 3.96810 2.83390 6.17140 6.91380 8.15870 9.03700 5.91130 5.34480 5.05460 4.38680 3.94800 8.36370 9.58350 9.50600 8.41140 10.59100 10.67540 9.61480 9.24300 12.14780 12.87210 11.86540 9.08750 8.00230 8.20740 9.16940 6.63800 6.41910 6.80070 5.77620 6.35920 5.73660 5.18210 5.18380 4.59310 4.59410 7.39480 7.45290 6.11410 5.37120 8.58240 8.37390 5.81350 4.64040 13.28000 17.27430 12.94670 16.64670 10.90280 11.40680 11.25650 12.53130 12.56140 9.31180 14.59780 16.40770 14.63630 13.11570 12.95630 13.77270 12.41050 12.21690 17.28150 17.21310 20.36770 10.73520 15.39730 8.97370 13.38480 8.42290 9.25970 11.05180 8.59910 14.73470 10.04330 11.26080 8.44130 9.56880 8.63000 9.67660 8.02550 11.91790 10.45270 12.97060 11.94570 10.56000 9.65610 9.06270 14.13680 13.64050 9.74700 8.41670 7.00000 8.16420 10.76770 11.80500 7.40510 9.44350 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 64 64 64 64 64 64 64 64 64 65 65 65 65 65 65 65 66 66 66 66 66 66 66 67 67 67 67 67 67 67 67 68 68 68 68 68 68 68 69 69 69 69 69 69 70 70 70 70 70 70 70 71 71 71 PHE P88 P38 PHE PHE PHE PBS P88 P33 THR THR THR THR THR THR THR THR THR THR THR THR THR THR ASP ASP ASP ASP ASP ASP ASP ASP PRO PRO P R0 P R0 P R0 PRO PRO S ER SER SER SER S ER 8 ER VAL VAL VAL VAL VAL CB CG CDl CD2 C31 C82 0092:} CB 061 C62 CA CB 0G1 CG2 CA CB CG OD1 OD2 CA CB CG CD CA CB OG CA CB CG1 CG2 CA HHH HH H HHHH mmmwwbmmmmdamquONv-lmxoooqmqqmommomoooommbmomqdqmqmqmmqq .12070 .23670 .27570 .08250 .62660 .34280 .37470 .11870 .60770 .03430 .08000 .43370 .15630 .03500 .73350 .45630 .74740 .07620 .02540 .99890 .59280 .56080 .85340 .96770 .85750 .31210 .69810 .47050 .09350 .14490 .63530 .27130 .76920 .70650 .80390 .90870 .02780 .99660 .67240 .58750 .46200 .70290 .95190 .80710 .38310 .25910 .87470 .33700 .33830 .12090 .85930 .77270 .40020 .84120 183 5.72940 5.52850 7.93940 9.29730 9.41570 10.40710 10.67020 11.69610 11.78620 4.78620 3.33800 2.69370 3.23460 2.74360 2.85920 1.52950 1.62370 1.00810 0.06370 -O.65820 0.25710 -0.69130 1.29380 -0.03210 -0.90970 -0.04200 1.01290 -1.49460 -2.21430 -3.00850 -2.02750 -0.51610 0.25180 0.55670 1.62950 -0.54390 -1.76280 -1.76800 -0.25110 -0.03980 0.91820 1.25750 -1.42750 -2.14700 1.31410 2.17980 3.32670 3.14550 1.26140 1.87950 0.09580 4.45820 5.64920 5.99440 OOHOOHNQmmUJWothwaubbowaHbNU-bO‘UI .17240 .41420 .32560 .76070 .44820 .57920 .87700 .98590 .67770 .26490 .64660 .41440 .55410 .60450 .22630 .83140 .13180 .85220 .67160 .44310 .13760 .50010 .23430 .91930 .74640 .58380 .41400 .56480 .30620 .39750 .80170 .18560 .33380 .34960 .97510 .93010 .09030 .00500 .48050 .43650 .08420 .01130 .68010 .56320 .84310 .37890 .60750 .52500 .53250 .09810 .42140 .30290 .68620 .67040 .00000 .32230 .59600 .65530 .87570 .63800 .47980 .15630 7.00000 10.78030 15.78060 16.06150 14.08520 20.23390 27.28010 10.67040 12.61690 15.44580 16.20060 18.35680 14.56670 14.56790 16.90410 17.10900 19.04450 18.46930 19.72200 20.19260 21.93500 24.41730 27.35660 20.58290 23.19560 21.70190 25.71440 24.56820 25.68550 23.26330 20.89680 21.80540 18.75410 24.43830 23.49460 32.42190 18.44730 18.10820 13.95070 14.74790 22.17210 18.44520 23.54400 13.76180 10.19370 10.33660 H H \Ot‘H HIHI~94140404 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 .00000 76 76 76 77 77 77 77 77 77 77 77 78 78 78 78 79 79 79 79 79 80 80 80 80 80 80 ASN ASN L80 L80 L80 L80 L80 L80 L80 GLY GLY GLY GLY ALA ALA ALA ALA ALA CYS CYS CYS CYS CYS CYS CG 001 CD1 CB SG -3.32660 -4.53100 -2.83000 -4.25840 -4.66140 -6.13440 -6.58280 -3.84030 -2.68130 -2.33700 -2.98300 -6.81320 -8.25020 -8.27620 -7.39270 -9.28850 -9.62190 10.32850 11.49240 10.55680 -9.73000 10.32220 11.83050 12.34510 -9.67570 -7.97390 185 .27210 .51170 .63690 .16370 .90640 .29930 .55610 .21140 .48220 .99840 .18180 .27390 .63690 .11340 .52670 .75750 .15490 .18210 .76000 .73740 .61180 .69670 .99200 10.45460 10.81710 10.67690 \ommmmoqqqmmasqmasmmbbooo O 12.64950 12.49430 13.47310 13.08770 14.26450 14.10220 12.96100 14.35470 15.24970 15.10700 16.70860 15.22600 15.21020 15.69360 16.46000 15.18440 15.48200 16.83870 16.80220 14.41930 17.91530 19.23990 19.20470 18.15770 20.04750 20.53690 15.71670 18.70470 18.31940 15.68120 22.32530 22.46190 27.33470 23.34450 24.91510 27.05950 23.24050 23.63340 26.18590 22.96030 23.76350 31.45840 33.29060 34.60390 42.25590 29.76180 32.58660 32.84410 35.98390 35.42010 33.88050 33.55440 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 * Cys75 36 position for left-handed CysSt ~Cys75 disulfide is listed. Alternate position: 75 CYS SG 0.05650 5.74360 12.58280 APPENDIX 8 Crystallographic parameters of solvent molecules included in orthorhombic plasminogen K4 structure. SULFATE ION “0'“ Coorrhales (A) 3 (AZ) 00am x v SUL S 6.04600 13.79300 -0.71820 16.11150 1.00000 SUL 01 7.54750 13.85560 -0.94280 21.15120 1.00000 SUL 02 5.63250 12.32430 —0.66800 18.73570 1.00000 SUL 03 5.68230 14.45160 0.53690 18.19410 1.00000 SUL 04 5.49010 14.35350 -1.96820 22.26630 1.00000 WATER MOLECULES Molecule Coordinates (A) BM?) Ooapancy OocZB x v z WAT 1 8.14400 8.80200 7.85500 11.52090 1.00000 0.0868 WAT 2 2.93120 16.49340 13.10580 15.33130 1.00000 0.0652 WAT 3 9.87130 18.26770 13.82180 16.01820 1.00000 0.0624 WAT 4 6.55930 22.40100 11.55240 16.09920 0.97110 0.0586 WAT 5 9.17060 15.57890 0.25740 17.58870 1.00000 0.0569 WAT 6 -0.51150 -1.37260 14.76610 18.06600 1.00000 0.0553 WAT 7 2.91490 0.91610 15.77660 18.50280 0.98150 0.0521 WAT 8 1.31900 14.45220 14.36750 19.61940 1.00000 0.0510 WAT 9 4.08680 21.20420 10.83830 20.77920 1.00000 0.0481 WAT10 14.03830 4.04510 8.83350 21.24800 0.98060 0.0452 WAT11 1.98940 19.52340 12.12240 20.95460 0.94420 0.0426 WAT12 12.09470 2.31630 10.36810 20.52660 0.90380 0.0398 WAT13 -4.42460 8.60070 9.87010 23.66090 0.96730 0.0395 WAT14 13.09140 20.17470 -3.84530 21.95910 0.92770 0.0392 WAT15 —5.09810 20.50150 13.93880 20.42830 0.87610 0.0376 WAT16 -0.95950 5.96410 19.76940 23.48920 0.93240 0.0370 WAT17 8.76720 19.66080 16.92080 16.54330 0.78040 0.0368 WAT18 1.17260 18.44640 20.30720 21.93220 0.89850 0.0368 WAT19 -11.34030 7.62520 2.05880 23.58300 0.92060 0.0359 WAT20 -7.81180 19.48780 13.86410 24.03110 0.92290 0.0354 WAT21 -6.32700 9.44520 7.65280 23.54150 0.88310 0.0331 WAT22 —4.64140 17.59170 16.19520 24.76660 0.90440 0.0330 WAT23 15.63340 16.55410 2.76790 18.52510 0.76320 0.0314 WAT24 2.30020 22.15740 7.74780 20.68870 0.80380 0.0312 WAT25 4.10970 21.52670 3.58500 27.77960 0.92640 0.0309 WAT26 8.09450 19.18900 1.60580 24.71720 0.86380 0.0302 WAT27 13.87010 10.61930 12.70080 24.42410 0.84400 0.0292 WAT28 5.63950 -1.50940 12.39980 23.12480 0.80820 0.0282 WAT29 17.32760 3.82280 -1.56320 21.48180 0.77200 0.0277 186 WAT30 WAT31 WAT32 WAT33 WAT34 WAT35 WAT36 WAT37 WAT38 WAT39 WAT40 WAT41 WAT42 WAT43 WAT44 WAT45 WAT46 WAT47 WAT48 WAT49 WATSO WAT51 WATSZ WAT53 WAT54 WATSS WAT56 WAT57 WAT58 WAT59 WATGO WAT61 WAT62 WAT63 WAT64 WAT65 WAT66 WAT67 WAT68 WAT69 WAT70 WAT71 WAT72 WAT73 WAT74 WAT75 WAT76 WAT77 WAT78 WAT79 WAT80 WAT81 WAT82 WAT83 1.07520 7.37460 8.59300 -7.01080 -5.41130 -9.14960 2.54900 14.22730 -3.98990 19.71730 -1.51440 9.28310 9.21980 16.55120 19.79080 -6.05990 -6.95920 18.58560 -10.37590 -5.94720 3.01100 0.20930 13.74920 -2.97660 1.60560 16.66230 5.43820 6.57880 -14.87130 11.73540 17.67050 2.11940 11.83690 -2.47280 17.07320 15.71770 -0.47660 14.51460 -5.51850 -0.53460 -6.00600 10.14880 11.21850 3.48610 0.40760 -8.10570 7.84460 -6.95080 9.11830 18.89580 11.89480 -11.50910 5.76950 -11.21510 18.02580 13.71020 -3.94450 7.05680 6.33840 15.44410 5.92680 6.06660 11.82920 6.59780 18.03670 2.67000 29.13880 0.92310 8.52060 19.33930 3.98320 15.13360 6.44610 4.67950 1.77990 1.20000 -1.03770 2.19530 -4.48170 16.78990 -4.37160 2.61340 18.55570 5.74080 14.16970 19.67240 18.57560 -2.21180 15.00490 0.40070 12.57540 2.98170 19.19510 6.14900 17.34550 21.14330 0.46650 18.85100 15.87420 7.27740 3.89870 6.48910 23.86170 25.60080 4.40620 11.68690 22.00980 13.23320 187 1.26310 18.94010 -2.47020 1.41020 8.97950 9.31830 -1.30420 14.53030 20.96460 -1.11220 3.15820 16.71740 3.97850 3.03780 4.86650 8.18540 10.77360 3.42730 19.31160 4.52860 19.72750 22.63510 3.79360 16.31060 3.47160 17.42910 8.63590 18.08440 11.54520 -3.10090 7.84700 15.19930 15.50610 10.18750 9.82540 -0.53480 21.23530 6.70790 18.26410 -0.19010 3.28200 2.98850 -6.82690 21.51430 22.49770 10.73090 -4.20870 6.84190 5.90840 4.21400 23.15630 23.04300 18.90260 7.02330 23.82720 22.53120 22.25090 24.99070 28.15760 22.77290 22.50030 24.27070 30.63600 25.24530 25.34530 27.92460 25.07610 27.15210 28.67870 21.00730 27.98210 28.22940 26.98050 28.11790 27.46170 28.72050 26.72870 28.84340 25.77120 21.81740 27.60390 28.15420 25.07690 21.18020 26.86530 22.47940 28.86080 28.34880 30.86680 23.42530 19.05330 22.87120 26.94980 33.75030 23.84430 24.37450 24.91860 28.46260 23.96800 25.00740 27.20930 30.01110 35.84140 34.74710 27.38410 29.76500 21.67910 23.78250 OOOOOOOOOOOOOOOOOOOOOOOOCOCOCOOOOOOOOOOOOOOOOOOOOOOOOO .80780 .75150 .73510 .77610 .81300 .72780 .72240 .72720 .81660 .73400 .71900 .75350 .70990 .73770 .74510 .62630 .70870 .70870 .69210 .70400 .68990 .70440 .67520 .69510 .65100 .59470 .66770 .67280 .63110 .56450 .63430 .57090 .63830 .63210 .64200 .55530 .50070 .54810 .59460 .66120 .55060 .55200 .55550 .59000 .52680 .53350 .55310 .57540 .62820 .61240 .53920 .55910 .47580 .49630 0. 0. 0. 0. .0235 .0233 .0232 .0218 .0218 .0213 .0204 .0203 .0201 .0200 .0194 .0187 .0180 .0178 .0178 .0176 .0173 .0173 .0171 .0168 .0164 .0162 .0161 .0161 .0159 .0150 .0150 .0145 .0141 .0141 .0133 .0132 .0132 .0131 .0131 .0130 .0127 .0125 .0124 .0122 .0116 .0114 .0112 .0110 .0110 .0108 .0106 .0105 .0104 .0104 OOOOOOOOOOOOOOOOOOOOOOOOCOOOOOOOOOOOOOOOOOOOOOOOOO 0274 0251 0243 0241 WAT84 WATBS WAT86 WAT87 WAT88 WAT89 WAT90 WAT91 WAT92 WAT93 WAT94 WAT95 WAT96 06020 65770 93810 50570 .15570 .89400 .87550 .91670 .98050 .58860 .57400 .45270 .36070 .72120 .04250 .92780 .92780 .63800 .41370 .82060 .80590 .85690 .69120 .87990 .69900 .63420 22. .08140 12 13. .29600 17 15. .57550 21 15. 19. -7. —0. .03230 7 19. 21. 188 52770 21270 75910 74930 81600 63720 09150 60070 80340 31. 30. .96070 18 33. 23. .39240 27 29. 31. .61100 24 33. 33. 26. 22. 14770 07720 59600 78800 60170 93040 15100 90540 79520 14050 0000000000000 .55710 .54720 .42710 .56470 .47050 .49660 .49620 .48380 .41880 .48350 .47950 .39450 .34580 0000000000000 .0100 .0100 .0096 .0095 .0093 .0090 .0083 .0073 .0071 .0071 .0068 .0058 .0054 LIST OF REFERENCES 1O 11 12 13 14 15 LIST OF REFERENCES Castellino,F.M., Ploplis,V.A., Powell,J.R., Strickland,D.K. (1981) J.Biol. Chem. 256. 4778-4782. Trexler,M., Patthy,L. (1983) Proc.NatI.Acad.Sci. USA 80, 2457-2461. Magnusson,S., Petersen,T.E., Sottrup—Jensen,L., Claeys,H. (1975) in Proteases and Biological Control (Reich,E., Rifkin,D.B., Shaw,E., eds) pp 123-149, Cold Spring Harbor Laboratories, Cold Spring Harbor,NY. Esmon,C.T., Jackson,C.M. (1974) J.Biol. Chem. 249, 7791-7797. McMullen,B.A., Fujikawa,K. (1985) J.Biol.Chem. 260, 5328-5340. Sottrup-Jensen,L., Claeys,H., Zaidal,M., Petersen,T.E.,Magnusson,S. (1978) Frog. Chem. Fibrinolysis hrombolysis 3,191-209. Lerch,P.G., Rickli,E.E., Lergier,W., Gillessen,0. (1980) Eur.J.Biochem. 107, 7-13. Steffens,G.J., Gunzler,W.A., Otting,F., Frankus,E., Flohe,L. (1982) Hoppe- Sey/er's Z. Physio]. Chem. 363, 1043-1058. Gunzler,W.A., Steffens,G.J., Otting,F., Kim,S-M.A., Frankus,E., Flohe,L. (1982) Hoppe-Seyler's Z. Physio]. Chem. 363, 1155-1165. Pennica,D., Holmes,W.E., Kohr,W.J., Harkins,R.N., Vehar,G.A., Ward,C.A., Bennett,W.F., Yelverton,E., Seeberg,P.H., Heynober,H.L., Goeddel,D.V., Collen,D. (1983) Nature 201, 214-221. Verheijen,J.H., Caspers, M.P.M., Chang,G.T.G., de Munk,G.A.W., Pouwels, P.H.,Enger-Valk,B.E. (1986) EMBO J. 5, 3525-3530. van Zonneveld,A.-J., Veerman,H., Pannekoek,H. (1986) J.BioI.Chem. 261, 14214-14218. Gardell,S.J., Duong,L.T., Diehl,RE., York,J.D., Hare,T.R.,Register,R.B., Jacobs,J.W., Dixon,R.A.F., Friedman,P.A. (1989) J.BioI.Chem. 264, 1 7947-1 7952. Nakamura,T., Nishizawa,T., Hagiya,M., Seki,T., Shimonishi,M.,Sugimura, A., Tashiro,K., Shimizu,S. (1989) Nature 342, 440-443. McLean,J.W., Tomlinson,J.E., Wun-Jin ,K., Eaton,D.L., Chen,E.Y., Fless, G.M., Scanu,A.M., Lawn,R.M. (1987) ature, 330, 132-137. 189 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 190 Patthy,L. (1985) Cell 41, 657-663. Collen,D. (1980) Thromb.Haemostasis 43, 77-89. Wiman,B., Collen,D. (1978) Nature 272, 549-550. Violand,B.N., Castellino,F.J. (1976) J.Biol.Chem. 251, 3906-3912. Christensen,U., Clemmensen,l. (1977) Biochem J. 163, 389-391. Lucas,M., Fretto,L, McKee,P. (1983) J.Biol. Chem. 258, 4249-4256. Hochschwender,S., Laursen,R. (1981) J.Biol. Chem. 256, 1 1 166-1 1171. %a12rkus,G., DePasquale,J.L., Vlfissler,F.C. (1978) J.Biol.Chem. 253, 727- Markus,G., Priore,Ft.L., Vlrlssler,F.C. (1979) J.BioI.Chem. 254, 1211-1216. Markus,G., Evers,J.L., Hobika,G.H. (1978) J.Biol.Chem. 253, 733-739. Thorsen,S., Clemmensen,l., Sottrup-Jensen,L., Magnusson,S.(1981) Biochim. Biophys.Acta 668, 377-387. Winn,E., Hu,S., Hochschwender,S., Laursen,R. (1980) Eur.J.Biochem. 104, 579-586. Thewes,T., Constantine,K., Byeon,l-J.L, Llinas,M. (1990) J.BioI.Chem. 265, 3906-3915. Christensen,U. (1984) Biochem J. 223, 413-421. Mangel,W.F., Lin,B., RamakrishnanN. (1990) Science 248,69-73. gggnsclgAq Park,C.H., Skrzypczak-Jankun,E. (1988) J.MoI.BioI. 202, -9 . Seshadri,T.P. unpublished results of this laboratory. Park,C.H., Tulinsky,A. (1986) Biochemistry 25, 3977-3982. Castellino,F.J., De Serrano,V.S., Powell,J.R., Johnson,W.R.,Beals,J.M. (1986) Arch. Biochem. Biophys. 247, 312-320. 3099M3a8rco,A., Laursen,R.A., Llinas,M. (1985) Biochim.Biophys.Acta 827, 6 - 0. Williams,R.J.P., Esnout,P., Lawrence,M., Cederholm-Williams,S.A. (1986) FEBS Lett. 209, 111-116. Atkinson,R.A., Williams,R.J.P. (1990) J.MoI.BioI. 212, 541-552. Hochschwender,S.M., Laursen,Ft.A. (1981) J.BioI.Chem. 256, 11172- 1 1 176. 39 41 42 43 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 1 91 Trexler,M., Vali,Z., Patthy,L. (1982) J.Biol.Chem. 257, 7401 -7406. Lerch,P.G., Rickli,E.E. (1980) Biochem.Biophys.Acta 625, 374-378. Ramesh,V., Petros,A.M., Llinas,M., Tulinsky,A., Park,C.H. (1987) J.Mol.Bio/. 198, 481 -498. DgygfioA" Laursen,R.A., Llinas,M. (1986) Arch.Biochem.Biophys. 244, Trexler,M., Patthy,L. (1983) Proc.Nat/.Acad. Sci. USA 80. 2457-2461. Tulinsky,A., Park,C.H., Mao,B., Llinas,M. (1988) Proteins 3, 85-96. MulichakA.M., Park,C.H., Tulinsky,A., Petros,A.M., Llinas,M. (1989) J.Biol. Chem. 264, 1922-1923. Carter,Jr.,C.W., Carter,C.W. (1979) J.Biol.Chem. 254, 12219-12223. Hatada,M. Hoffmann-LaRoche Inc., (1987) personal comm. Thaller,C., Eichele,G., Weaver,LH., \Mlson,E., Karlsson,R.,Jansonius,J.N. (1985) Methods Enzymol. 1 14, 132-135. Wyckoff,H.W., Doscher,M., Tsemoglou,D., lnagami,T., Johnson,L.N., Hardman,K.D., Allewell,N.M., Kelly,D.M., Richards,F.M. (1967) J.Mol.Bio/. 27, 563-578. Buck,D., Tulinsky,A., Gaier,J. (1982) unpublished results. North,A.C.T., Philips,D.C.,Mathews,F.A. (1968) Acta Cryst. 24A, 351-359. Rossman,M.G., Blow,D.M. (1962) Acta Cryst. 15, 24-31. Steigemann,w. (1974) Dissertation, Technische Universitat, Munchen. Ravichandran,K.G. (1989) Dissertation, Michigan State Univ., East Lansing, Michigan. Fujinaga,M., Read,R.J. (1987) J.Appl.Cryst. 20, 517-521. Jones,T.A. ( 1982) In: Computational Crystallography, Sayre 0. ed. Oxford: Clarendon Press, pp 303-317. Hendrickson,W.A., Konnert,J.H. (1980) In Biomolecular Structure, Func- tion, Conformation and Evolution, R. Srinivasan ed. 1, 4357, Oxford: Pergamon. Finzel,B.C. ( 1987) J.Appl.CIyst. 20, 53-55. Luzzati,P.V. (1952) Acta Cryst. 5, 802-810. Ramachandran,G.N., Mitra,A.K. (1976) J.MoI.Bio/. 107, 85-92. 61 62 63 64 65 66 67 68 69 70 71 72 192 Gorbitz.C.H. (1989) Acta Cryst 845, 390-395. Taylor,R., Kennard,0., Versichel,W. (1984) Acta Cryst. 840, 280-288. Venkatachalam,C.M. (1968) Biopolymers 6, 1425-1436. Crawford,J.L., Ugscomb,W.N., Schellman,C.G. (1973) Proc.Nat.Acad.Sci. USA 70, 538-54 . Lee,B., Richards,F.M. (1971 ) J.MoI.BioI. 55, 379-400. Downie,T.C., Harrison,W., Raper,E.S., Hepworth, MA. (1972) Acta Cryst. 828. 283-290. Form,G.F1., Raper,E.S., Downie, TC. (1973) Acta Cryst. B29, 776-782. Burley,S.K., Petsko,G.A. (1985) Science 229, 23-28. Tulinsky,A., Vandlen,R.L., Morimoto,C.N., Mani,N.V., Wright,L.H. (1973) Biochemistry 1 2, 4185-4192. DeMarco,A., Pluck,N.0., Banyai,L., Trexler,M., Laursen,R.A., Patthy,L., Llinas,M., Williams,R.J.P. (1985) Biochemistry 24, 748-753. Trexler,M., Banyai,L., Patthy,L., Pluck,N.0., Williams,R.J.P. (1985) Eur.J. Biochem. 152, 439-446. James,M.N.G., Sielecki,A.R. (1983) J.Mol.BioI. 163, 299-361. HICHIGRN STATE UNIV. 1 ll lllllllllill“ 09 1 1 6892 {Ill llllllllll 312930