Insights Into the Serpin Inhibitory Mechanism from Structures of Mutant Serpins in the Mechaelis Complex

INSIGHTS INTO THE SERPIN INHIBITORY MECHANISM FROM STRUCTURES OF MUTANT SERPINS IN THE MECHAELIS COMPLEX

Approved by supervisory committee

Elizabeth J. Goldsmith, Ph.D. Mentor

Jose Rizo-Rey, Ph.D. Chair

Philip J Thomas, Ph.D

Stephen R. Sprang, Ph.D

TO MY FAMILY

INSIGHTS INTO THE SERPIN INHIBITORY MECHANISM FROM STRUCTURES OF MUTANT SERPINS IN THE MECHAELIS COMPLEX

SUL, SOON-HEE

DISSERTATION

Presented to the Faculty of the Graduate School of Biomedical Sciences

The University of Texas Southwestern Medical Center at Dallas

In Partial Fulfillment of the Requirements

For the Degree of

DOCTOR OF PHILOSOPHY

The University of Texas Southwestern Medical Center at Dallas

Dallas, Texas

February 2008

SUL, SOON-HEE, 2008

INSIGHTS INTO THE SERPIN INHIBITORY MECHANISM FROM STRUCTURES OF MUTANT SERPINS IN THE MECHAELIS COMPLEX

Soon-Hee Sul Ph.D.

The University of Texas Southwestern Medical Center at Dallas, 2008

Supervising professor: Elizabeth J. Goldsmith Ph.D.

The serpins belong to a superfamily of protease inhibitors that employ a unique suicide substrate-like inhibitory mechanism. In this mechanism, target protease becomes acylated at the catalytic serine. Deacylation fails to take place, and the serpins undergo dramatic conformational changes in which the acylated protease is translocated 70Å from one pole of the serpin to the other. However, the factors causing suppression of deacylation are not fully understood. The previously solved Michaelis complex of serpin-1B with trypsin/S195A suggests that the P4 (I350) and P1’ (S354) residues on the reactive center loop may be important factors in arresting the hydrolysis reaction, since these residues are involved in interactions with trypsin that are unique to serpins (unlike low molecular weight serine protease inhibitors). Inhibition assay data also show that the

inhibitory activity of serpin was reduced by the introduction of a mutation either at the P4

(I350A) or at the P1’ (S354A) position. The P4 (I350A) mutation nearly completely abolishes inhibitory activity of serpin-1B toward trypsin. In other words, the mutated P4 residue of serpin-1B acts as a substrate, rather than as an inhibitor. A crystallographic approach was used to understand why the serpin-1B (I350A) becomes a substrate; and, hopefully, to gain insight into the serpin inhibitory mechanism. Similarly, the formed

Michaelis complex between trypsin and the serpin-1B (S354A) was also utilized to address the role of this residue in this serpin mechanism. The structural analysis shows that significant conformational changes were observed from the serpin-1B

(I350A)/trypsin complex, but not from the serpin-1B (S354A)/trypsin complex. Due to these conformational changes, the special extensive interactions observed in the wild type

complex were lost in the mutant complexes, the consequence being the destabilization of

the mutant Michaelis complexes and, thus, perhaps destabilizing acylated covalent

intermediate. Also, reduced interactions (mostly P-side of the serpin) induce a conformational change on trypsin Gln192, residue that may be important for arresting hydrolysis. Therefore, the stable and tight interactions in the Michaelis complex may be important in arresting deacylation in the serpin inhibitory mechanism.

TABLE OF CONTENTS

Dedication ...... ii

List of Figures ...... xiii

List of Tables ...... xviii

List of Abbreviations ...... xx

CHAPTER ONE: Introduction ...... 1

1.1 Objectives ...... 6

1.2 Serpin structure ...... 8

1.2.1 Protease subsite nomenclature ...... 14

1.2.2 Native state of Serpin ...... 15

Proximal reactive-loop hinge ...... 16

Functional domain-reactive center ...... 17

Distal reactive-loop hinge ...... 17

Breach ...... 18

Shutter ...... 18

Heparin-binding site ...... 18

1.2.3 Cleaved form of Serpin ...... 20

1.2.4 Latent form of Serpin ...... 20

1.2.5 Partial inserted form of Serpin ...... 21

1.2.6 Serpin polymerization ...... 21

1.3 Serpinopathies ...... 25

1.4 Serpin inhibitory mechanism ...... 26

vii

1.4.1 Serine protease mechanism ...... 31

1.4.2 The mechanism of non-serpin inhibitors ...... 33

1.5 Structural basis of serpin inhibitory mechanism ...... 36

1.5.1 Structure of active serpin1k from manduca sexta ...... 36

1.5.2 The initial noncovalent Michaelis-like complex ...... 38

The unique features of non-covalent serpin-protease complex...... 40

Topological switch point and other conformational change ...... 43

1.5.3 The covalent complex ...... 45

1.6 Description of dissertation research ...... 46

CHAPTER SIX: INTERPRETATION OF KINETIC DATA OF INHIBITION OF

TRYPSIN BY SERPINS-1B (A353K) AND ITS MUTANTS ...... 49

2.1. Introduction ...... 49

2.2 Materials and methods ...... 52

2.2.1 Site-directed mutagenesis of serpin-1b (A353k) ...... 52

2.2.2 Inhibition assay ...... 52

2.3 Results and discussion ...... 53

2.3.1 Mutations at the interface between serpin-1B and trypsin affect serpin

inhibitory activities ...... 53

2.3.2 Effects of the mutations at unique interactions sites (P4 and P1’) on the

serpin activity ...... 56

2.3.3 Effects of the mutation at the P4’site on the activity of the serpin ...... 59

2.3.4 Mutations near the hinge region abrogate serpin activity ...... 60

2.3.5 Effects of the mutations at helix F1 on the serpin activity ...... 61

viii

2.4 Conclusion ...... 61

CHAPTER THREE: Materials and methods of the mutant complex study ...... 62

3.1 introduction ...... 62

3.1.1 Rat anionic trypsin ...... 63

3.1.2 Manduca sexta serpin ...... 64

3.2 Methods and materials ...... 65

3.2.1 Expression and purification of rat anionic trypsin S195A ...... 65

A. Yeast expression ...... 65

preparation of media for trypsin purification ...... 66

B. Purification of trypsinS195A from yeast...... 67

3.2.2 Expression of M. sexta serpin ...... 69

3.2.3 Purification of the mutant serpins ...... 70

A. Preparation of the cell –free extract ...... 70

B. Immobilized metal affinity chromatography ...... 70

C. Anionic exchange chromatography ...... 71

D. Size exclusion chromatography ...... 71

3.2.4 Crystallization ...... 73

A. Seeding ...... 73

B. Streak seeding ...... 74

C. Preparing seed stock ...... 74

D. Micro seeding ...... 74

3.3 Results and Discussion ...... 75

3.3.1 Purification of rat anionic trypsin S195A ...... 75

3.3.2 Purification of serpin-1B/I350A and serpin-1B/S354A ...... 79

3.3.3 Cocrystallization ...... 86

CHAPTER FOUR: Structural determination ...... 91

4.1 Data collection and data reduction ...... 91

4.2 Identifying the space group ...... 92

4.3 Molecular replacement ...... 94

4. 4 Model refinement ...... 95

4.5 Map calculation ...... 96

4.6 Model building ...... 96

4.8 Twinning in complex structures ...... 99

4.8.1 Theoretical background ...... 99

4.8.2 Merohedral twinning ...... 100

4.8.3 Recognizing crystal twinning ...... 102

4.8.4 Estimation of crystal twinning ...... 104

4.8.5 Correcting partially twinned data ...... 106

4.9 Hemihedral twinning in the mutant complexes ...... 107

CHAPTER FIVE: Structural analysis of the Michaelis complexes:Serpin-1B/I350A- trypsinS195A and Serpin-1B/S354A-trypsinS195A ...... 109

5.1 Introduction ...... 109

5.2 Structural analysis of Michaelis complex: Serpin-1B(I350A)/trypsin ...... 111

5.2 1 Overall structures of the mutant Michaelis complex ...... 111

5.2.2 Changes in the reactive center loop ...... 116

5.2.3 Changes at the topological switch point ...... 121

5.2.4 Changes near the reactive center loop...... 123

5.2.5 Other conformational change(in serpin core domain) ...... 125

5.3 Structural analysis of Michaelis complex: Serpin-1B(S354A)/trypsin ...... 130

5.4 Structure of acyl intermediate between chymotrypsin and native peptide as

a potent inhibitor ...... 132

5.5 Discussion ...... 136

CHAPTER SIX: CONCLUSIONAS AND FUTURE DIRECTION ...... 139

CHAPTER SEVEN: Crystallization of Bm-SPN2 ...... 145

7.1 Introduction ...... 145

7.2 Materials and methods ...... 147

7.2.1 Sequence analysis of BmSPN2 ...... 147

7.2.2 Cloning of constructs ...... 147

7.2.3 Expression of bm-spn2 ...... 148

7.2.4 Purification of bm-spn-2 ...... 150

7.2.5 Reductive methylation ...... 152

A. Chemical basis of reductive methylation ...... 152

B. Reagents ...... 153

C. Procedure ...... 154

7.2.6 Crystallization ...... 154

7.3 Results and Discussion ...... 157

7.3.1 Sequence analysis ...... 157

7.3.2 Expression and purification of BmSPN2 (21-426) ...... 162

7.3.3 Expression and purification of BmSPN2 (21-406) and

BmSPN2 (48-406) ...... 166

7.3.4 Reductive methylated BmSPN2 ...... 170

7.3.5 Crystallization and discussion ...... 172

CHAPTER EIGHT: Crystallization of VopA ...... 176

8.1 introduction ...... 176

8.2 Materials and methods ...... 178

8.2.1 Construction of two expression plasmid:Gst-VopA and His-Smt3-

VopA ...... 178

8.2.2 Expression and purification of GST tagged VopA in E.Coli ...... 179

8.2.3 Expression and Purification of His-smt3 tagged VopA in E.Coli ...... 182

8.2.4 Crystallization ...... 186

8.3 Results and discussion ...... 187

8.3.1 Sequence analysis and VopA homology model ...... 187

8.3.2 Expression and purification of GST tagged VopA in E.Coli ...... 190

8.3.3 Expression and Purification of His-smt3 tagged VopA in E.Coli ...... 192

8.3.4 Expression and Purification of VopA (21-201), VopA (21-289) and

VopA (49-289) ...... 196

8.3.5 Crystallization and discussion ...... 199

BIBLIOGRAPHY ...... 201

VITAE

xii

LIST OF FIGURES

FIGURE 1.1 Ribbon diagrams of serpin complex structures ...... 5

FIGURE 1.2 Schematic diagram of the schechter and berger convention of substrate – protease subsite nomenclature ...... 15

FIGURE 1.3 Native conformation of antithrombin ...... 19

FIGURE 1.4 Serpin structures ...... 22~24

FIGURE 1.5 The Branched pathway mechanism of serpins as suicide substrate inhibitors ...... 28

FIGURE 1.6 The kinetic diagram of branched pathway mechanism of serpins ...... 29

FIGURE 1.7 Energy diagram of serpin inhibition ...... 30

FIGURE 1.8 schematic diagram for the chemistry of serine protease reaction ...... 32

FIGURE 1.9 A ribbon representation of rat anionic trypsin bound to bovine pancreatic

trypsin inhibitor ...... 35

FIGURE 1.10 Ribbon diagram of serpin 1K ...... 37

FIGURE 1.11 superposition of active serpin1k with the Michaelis complex of serpin-1B

A353K and trypsin S195A ...... 39

FIGURE 1.12 The interaction between the RCL (P2’- P4) and trypsin in the Michaelis

complex ...... 41

FIGURE 1.13 Superposition of the Michaelis complex ...... 42

FIGURE 1.14 Conformation of the RCL and topological switch point ...... 44

FIGURE 1.15 Ribbon presentation of the final complex form in serpin inhibitory

mechanism ...... 45

xiii

FIGURE 2.1 The mutated residues at the interface of the serpin-trypsin Michaelis complex ...... 50

FIGURE 2.2 Site-directed mutagenesis of serpin residues at the interface of the serpin- trypsin Michaelis complex ...... 51

FIGURE 2.3 Inhibition of trypsin by serpin-1B (A353K) and its mutants ...... 55

FIGURE 2.4 Time course of reaction of P4 -> Ala mutant with serpin ...... 57

FIGURE 2.5 Products of reaction of serpin mutants with trypsin for 30minutes ...... 58

FIGURE 3.1 The SDS-PAGE analysis for the eluted fractions (contained trypsinogen) from Toyopearl column ...... 76

FIGURE 3.2 The activation of trypsinogen by enterokinase ...... 77

FIGURE 3.3 Purification of trypsin by SBTI column ...... 78

FIGURE 3.4 Purification of Serpin-1B/S354A by Ni affinity chromatography ...... 80

FIGURE 3.5 Purification of Serpin-1B/S354A by MonoQ anionic exchange and

Superdex75 chromatography ...... 82

FIGURE 3.6 Purification of Serpin-1B/I350A by Ni affinity chromatography ...... 83

FIGURE 3.7 Purification of Serpin-1B/I350A by MonoQ anion exchange chromatography ...... 84

FIGURE 3.8 Purification of Serpin-1B/I350A by Superdex75 gel filtration chromatography ...... 85

FIGURE 3.9 Checking for complex formation in Native gel ...... 87

FIGURE 3.10 Crystals of Serpin-1B/S354A with trypsin S195A ...... 88

FIGURE 3.11 Crystals of serpin1B/I350A with trypsin S195A ...... 89

FIGURE 3.12 Improved quality crystals of Serpin-1B/I350A with Trypsin S195A ...... 90

xiv

FIGURE 4.1 twinning test for serpin-1B/I350A-trypsinS195A complex ...... 104

FIGURE 5.1 Overall structure of the Michaelis complex between serpin-1B/I350A and rat anionic trypsinS195A ...... 112

FIGURE 5.2 Superposition of the wild type and mutant Michaelis complexes using trypsin as a reference ...... 113

FIGURE 5.3 Molecular surface representation of trypsin with the reactive center loop of serpin showing dots in interface between serpin and trypsin ...... 114

FIGURE 5.4 The comparison of RCL conformation between mutant and wild type complexes ...... 117

FIGURE 5.5 Electron density map of the RCL ( P1’-P17) ...... 118

FIGURE 5.6 Comparison of active sites in RCL (P4-P2’) between the wild type complex and the mutant complex ...... 119

FIGURE 5.7 Stereo diagram showing interactions between the RCL and trypsinS195A ...... 120

FIGURE 5.8 Stereo diagram showing the interactions at the topological switch point 122

FIGURE 5.9 Conformational changes near the reactive center loop ...... 123

FIGURE 5.10 Change near the reactive center loop ...... 124

FIGURE 5.11 Electron density map of the helix E ...... 126

FIGURE 5.12 Changes in the helix E of the serpin ...... 127

FIGURE 5.13 Structural comparison in the helix E and D by superpositions ...... 128

FIGURE 5.14 Conformational change in Helix E of the serpin-1B/I350A ...... 129

FIGURE 5.15 Overall structure of the Michaelis complex between serpin-1B/S354A and rat anionic trypsinS195A ...... 131

FIGURE 5.16 Structure of chymotrypsin (green) in complex with an autolysed peptide inhibitor ...... 134

FIGURE 5.17 Superposition of acylated chymotrypsin-inhibitor peptide with wild-type

Michaelis and serpin-1B(I350A)/trypsin ...... 135

FIGURE 6.1 Michaelis complex of serpin-1B (A353K) with trypsin (S195A) showing location of cysteine mutation used for labeling ...... 142

FIGURE 7.1 Multiple sequence alignment of BmSPN2 with other serpins ...... 158

FIGURE 7.2 Structural disorderness predictions ...... 160

FIGURE 7.3 Construction of His tagged BmSPN2in pET29c ...... 161

FIGURE 7.4 Purification of BmSPN2 by Ni affinity chromatography ...... 163

FIGURE 7.5 Purification of BmSPN2 by MonoQ anionic exchange chromatography 164

FIGURE 7.6 Purification of BmSPN2 (21-426) by Superdex 75 gel filtration chromatography ...... 165

FIGURE 7.7 Purification of BmSPN2 (21-406) by Ni affinity chromatography ...... 167

FIGURE 7.8 Purification of BmSPN2 (21-406) by MonoQ anionic exchange chromatography ...... 168

FIGURE 7.9 Purification of BmSPN2 (21-406) by Superdex 200 gel filtration chromatography ...... 169

FIGURE 7.10 Purification methylated BmSPN2 by monoQ and superdex 75 gel filtration ...... 172

FIGURE 7.11 Crystals of BmSPN2 (21-426) ...... 174

FIGURE 7.12 Crystals of BmSPN2 (21-426) ...... 175

xvi

FIGURE 8.1 Sequence alignment of putative cysteine proteases and adenovirus protease

(AVP) ...... 188

FIGURE 8.2 Differences between VopA and adenovirus protease (AVP) ...... 189

FIGURE 8.3 Purification of GST tagged VopA by glutathione-Sepharose beads ...... 191

FIGURE 8.4 Purification of VopA by superdex75 gel filtration ...... 191

FIGURE 8.5 Purification of VopA by Ni affinity chromatography and Mono Q anionic

exchange chromatography ...... 193-194

FIGURE 8.6 Purification of VopA (1-289) by Superdex 200 gel filtration

chromatography ...... 195

FIGURE 8.7 Purification of VopA(21-289) by Ni affinity column ...... 197

FIGURE 8.8 Purification of VopA (21-289) by MonoQ anionic

chromatography ...... 198

xvii

LIST OF TABLES

TABLE 1.1 X-ray structures of serpins and their complexes ...... 10-12

TABLE 2.1 Association rate constants (ka) of trypsin and serpin-1B (A353K) and its

mutants ...... 58

TABLE 3.1 Summary of reagents used for purification of ssserpin-1B/I330a and serpin-

1B (S354A) ...... 72

TABLE 4.1 Data collection and refinement statistics of serpin-1B (I350A) with Trypsin

S195A ...... 97

TABLE 4.2 Data collection and refinement statistics of serpin-1B/S354A with Trypsin

S195A ...... 98

TABLE 4.3 Hemihedral twin laws ...... 106

TABLE 5.1 Hydrogen bonds involving residues in the contact area of Serpin-1B with trypsin S195A ...... 115

TABLE 5.2 Hydrogen bonds involving residues in the contact area of Serpin-1B/I350A with trypsin S195A ...... 115

TABLE 5.3 Hydrogen bonds involving residues in the contact area of Serpin-1B/S354A with trypsin S195A ...... 131

TABLE 7.1 Summary of reagents used for purification of BmSPN2 ...... 151

TABLE 7.2 Clear strategy screen I ...... 155

TABLE 7.3 Clear strategy screen II ...... 156

TABLE 8.1 Summary of reagents used for purification of GST tagged VopA ...... 181

TABLE 8.2 Summary of reagents used for purification of VopA in smt3-Ulp vector .. 185

xviii

TABLE 8.3 Different constructs of VopA and their properties ...... 188

xix

LIST OF DEFINITIONS

α1AT – α1-Antitrypsin

α1-PI – α1-Protease Inhibitor

Å – Ångström

APS – Advanced Photon Source

B factor – Temperature factor

BPTI – Bovine pancreatic trypsin inhibitor

BmSPN2 – bugia malayi serpin precursor 2

CCD – Charge-Coupled Device

DTT – Dithiothreitol

E.coli – Eschericheria coli

EDTA – Ethylenediamine tetracetic acid

Fc – Calculated structure factor amplitude

Fo – Observed structure factor amplitude

HEPES – N-(2-hydroxyethyl)piperzine-N’2-ethanesulphonic acid

I – Intensity of reflections

IPTG – Isopropyl β-D-thiogalactopyranoside

kDa – Kilo-Dalton

L – Liter

LB – Luria-Broth

MAP – Mitogen Activated Protein

MR – Molecular Replacement

PCR – polymerase chain reaction

PDB – Protein Date Bank rmsd – Root mean square deviation

RCL – reactive center loop

SERPIN – serine protease inhibitor

SDS-PAGE – Sodium dodecyl sulfate polyacrylamide gel electrophoresis

Tris – Tris (hydroxymethyl) aminomethane

VM – Matthews number

VopA – vibrio outer proteinA

xxi

CHAPTER ONE INTRODUCTION

Serpins (serine protease inhibitors or classified inhibitor family I4) are a superfamily of proteins (350-500 amino acids in size) that have a conserved domain consisting of three β-sheets and 8-9 α-helices (Stein and Carrell 1995; Gettins 2002).

Over 1,500 members of this family have been identified up to now in animals, poxviruses, plants, bacteria and archaea (Silverman and Lomas 2004). Most serpins inhibit serine proteases, but there are serpins that inhibit caspases and pepsin-like cysteine proteases.

Also, a few serpins have different functions such as hormone transport, molecular chaperones or tumor suppressors (Potempa, Korzus et al. 1994).

Serpins have been the target of medical research because they regulate processes such as coagulation and inflammation (Silverman, Bird et al. 2001). Although most serpins control proteolytic cascades, some serpins serve as defense against the host immune response in some viruses (Zhu, Wang et al. 2003) and in parasites such as the filarial nematode (Zang, Yazdanbakhsh et al. 1999). In addition to the medical importance of serpins, serpins are also particularly interesting from the viewpoint of their structural biology because they undergo unique and dramatic changes in their conformation when they inhibit target proteases.

In contrast to the small serine protease inhibitors (such as BPTI, Kazal inhibitors, soybean trypsin inhibitors, etc) serpins are relatively large molecules and adopt a unique and extensive conformational change to inhibit serine proteases (Schreuder, de Boer et al.

1994; Ye and Goldsmith 2001). Because the serpin is cleaved during inhibition,

2 inhibitory serpins are called ‘suicide’ or ‘single use’ inhibitors (Huntington, Read et al.

2000). The first step for the ‘suicide’ inhibition mechanism of serpin is the recognition of the cognate protease to form the non-covalent Michaelis-like complex (Huntington J. A.,

Read R. J., et al 2000) (Figure 1.1). Most canonical inhibitors only have this step for their inhibition mechanism (Krowarsch, Cierpicki et al. 2003). However, the serpin adopts a highly unique mechanism, in which the normal proteolytic reaction is arrested.

The serpin acts as a substrate for its cognate protease until the serpin forms covalent acyl enzyme intermediate between the catalytic serine and the reactive center loop

(Whisstock, Skinner et al. 1998). With serpin as substrate, the second half reaction, deacylation, does not occur. Instead, the covalent acyl intermediate is stabilized through massive energetically favorable conformational changes in the serpin and protease

(Bruch, Weiss et al. 1988; Mottonen, Strand et al. 1992; Wang, Mottonen et al. 1996).

After the RCL of serpin is cleaved by the protease, the serpin inserts the cleaved RCL into β-sheet A and translocates the covalently attached protease from top to the bottom of serpin. The active sites residues of translocated protease are disoriented thereby, rendering the protease catalytically incompetent in the covalent acyl-enzyme complex

(Figure 1.1) (Olson, Bock et al. 1995; Shore, Day et al. 1995; Huntington, Read et al.

2000; Lawrence, Olson et al. 2000).

Extensive biochemical studies show that the native state of serpin conformation is not the most stable, rather in a strained metastable state (Wang, Mottonen et al. 1996;

Whisstock, Skinner et al. 1998). So, the strain in the native conformation of the serpin may be the main driving force for structural transition during the complex formation (Im,

Seo et al. 1999; Seo, Im et al. 2000; Silverman, Bird et al. 2001). Also, since the

3 inhibition process of the trypsin by serpin is much slower than the normal turnover rate of trypsin (~70/second) (O'Malley, Nair et al. 1997), distortion of the protease active site and loop insertion occur faster than hydrolysis of the acyl-enzyme intermediate

(Kvassman, Verhamme et al. 1998; Olson, Swanson et al. 2001; Shin and Yu 2002).

Thus, to be an efficient proteinase inhibitor, this insertion process must be much faster than the substrate cleavage reaction (Lawrence, Olson et al. 2000; Shin and Yu 2002).

However, the mechanism of arresting hydrolysis (suppressing deacylation) is not clearly understood.

This unique inhibitory mechanism of serpin is visually proved by X-ray structural studies on several different serpin-protease complexes. For the initial noncovalent Michaelis-like complex, there have been x-ray crystallographic studies on the serpin-1B (A353K) with trypsin (S195A) (Ye, Cech et al. 2001) (Figure 1.1), α1PI-

pittsburgh with anhydrotrypsin (S195A) (Dementiev, Simonovic et al. 2003),

antithrombin with factor Xa (S195A)-heparin (Johnson, Li et al. 2006), and antithrombin

with thrombin (S195A)-heparin complexes (Dementiev, Petitou et al. 2004; Li, Johnson

et al. 2004). For the final kinetically trapped covalent complex, only α1PI-trypsin complex was determined confirming that the protease is translocated and distorted

(Huntington, Read et al. 2000) (Figure 1.1).

The structural analysis of the initial non-covalent Michaelis complexes showed that the conformation of the reactive center loop is different from that of the active serpin alone. Also, the complexed serpins have more contact area with protease than the other small molecular weight serine protease inhibitors (Papamokos, Weber et al. 1982).

Especially, the Michaelis complex structures of antithrombin with factor Xa (S195A)-

4 heparin, antithrombin with thrombin (S195A)-heparin and serpin-1B (A353K) with trypsin demonstrate that they have similar interactions 1) in P4 position (C-terminal to the scissile bond) where the residue is inserted into the hydrophobic groove made by protease and 2) in the P1’ position (directly following the scissile bond) where the residue made close interaction with His 57, one of the catalytic residues in the enzyme. In addition, mutational analysis of these positions (P1’ and P4) verified their significance in complex formation (unpublished data in Chapter 2, Kanost’s lab in Kansas State

University). These structural data and the mutagenesis study can serve as a starting point to understand how deacylation is arrested in serpin inhibitory mechanism.

(A)

(B)

Figure 1.1 Ribbon diagrams of serpin complex structures (A) Michaelis complex of serpin-1B(A353K)-tryspinS195A (PDB code: 1K9O) (B) Final complex form of α-AT with trypsin (PDB code: 1EXZ)

1. 1 Objectives

The Goldsmith lab has had a long-term interest in the mechanism of action of serpins and has made two seminal contributions to the field with the first structure of a

“latent” serpin (Mottonen, Strand et al. 1992) (discussed in section 1.2.4) and the first structure of a serpin-protease Michaelis complex. The lab also contributed the first high- resolution (2.1 Å) structure of an intact serpin (Li, Wang et al. 1999). This research came out of collaboration with Joe Sambrook’s group to develop serpin-resistant protease that might be used clinically (Serpin resistant tissue plasminogen activator) (Madison,

Goldsmith et al. 1989). The success in obtaining the first Michaelis complex and the first high-resolution serpin structure was made possible by going to a model system (serpins from hornworm moth M. Sexta with clones obtained from Michael Kanost’s lab (Kansas

State University), and rat anionic trypsin S195A from Elizabeth Hedstrom’s lab

(Brandeis University)) (Ye, Cech et al. 2001). Since serpin-1B and trypsin are both well expressed and studied (Jiang, Mulnix et al. 1995; Hedstrom 2002), they make an excellent model system for understanding the inhibitory mechanism of serpins by X-ray crystallography (See chapter 3.1 for a detailed discussion). The first Michaelis complex solved by our lab revealed that there was a unique, extensive, and intimate contact surface between the serpin and the protease (Ye, Cech et al. 2001). Among those residues, I350 (P4, counting back from the scissile bond) and S354 (P1’, directly following the scissile bond) residues made unique interactions at the interface between the serpin-1B and the protease (The detailed discussion is given in section 1.5.2) compared to other low molecular weight serine protease inhibitors. In addition, kinetic studies by Kanost’s lab of serpins in which these residues were mutated verified their

7 importance in complex formation (unpublished data). These data offered several clues about the unusual mechanism of action of serpins, which we now propose to investigate

(see the chapter 2 for detailed discussion of kinetic assays).

Although many structures including several Michaelis complexes have been determined, none of the structural studies address how deacylation is arrested. This is the question we are trying to answer in the X-ray crystallographic studies described in this

dissertation. Serpin-1Bs mutated at sites known to contribute to the mechanism are

studied in complex with trypsin: thestructure of trypsin (S195A) in complex with either

serpin-1B (I350A) or serpin-1B (S354). The serpin (I350A) was chosen for the structural

study based upon the result of the inhibition assay by which the Ile350 was revealed to

have an effect on preventing deacylation in the inhibitory mechanism of the serpin (The

detailed discussion is given in Chapter 2). Although the importance of the serpin (S354A)

in the mechanism was not clearly demonstrated, it was selected for further structural

study due to its structural interest. In addition to both Michaelis complexes, we will

discuss a recent publication of an acyl-serine protease complex. This structure shed light

on mechanisms of deacylation arrest.

1.2 Serpin structures

Over eighty crystal structures of serpins have been determined since 1984

(Protein Data Bank), and are listed in Table 1.1 where the serpins are classified with 16 clades, denoted as A through P, based upon its phylogenetic relationships. These serpin structures, together with biochemical data, show that active serpins adopt a metastable state rather than their most stable conformation (Gettins 2002). Serpins consist of three β-sheets (A, B and C) and eight or nine α-helices (A to I), which are highly conserved among family member. In this conformation, the protease recognition site, which is called RCL (the reactive center loop), a 20-25 amino acid flexible loop, is positioned at top on the body of the molecule. Although the number of serpin X-ray structures that have been determined has increased rapidly in recent years, still by far, however, the greatest number of structures have been solved in the noncomplexed state

(Table 1.1). The first structure of a serpin was that of a cleaved serpin (the cleaved human α1-protease inhibitor) (Loebermann, Tokuoka et al. 1984). It was very surprising

that the reactive center loop was not in an exposed location. It was instead integrated into

β-sheet A, the largest β-sheet, as the fourth strand. Then, the structure of the cleaved form of ovalbumin was determined and showed that the reactive center loop of ovalbumin was not inserted into β-sheet A (Stein, Leslie et al. 1990). This demonstrated that a difference in behavior of inhibitory and noninhibitory serpins and suggested how conformational change in serpins lead to protease inhibitory activity. The next structure to be determined was that of the latent form PAI-1 (Mottonen, Strand et al. 1992). It showed that the location of the reactive center loop was the same site of the cleaved form

of the inhibitory α1-PI. However, the uninserted part of the loop was composed of

residues from strand 1 of β-sheet C. This strand was pulled away and provided the linker

role for returning to the top of molecule. In 1994, the first structures of uncleaved serpin

structures of antithrombin (Carrell, Stein et al. 1994; Schreuder, de Boer et al. 1994) and

α1-antichymotrypsin (Wei, Rubin et al. 1994) variants were solved. Since then more than

eighty structures have been solved including complexed serpins (Huntington, Read et al.

2000; Ye 2001; Dementiev, Petitou et al. 2004; Li, Johnson et al. 2004; Dementiev, Dobo

et al. 2006; Johnson, Langdown et al. 2006). Also, the structures of novel types of

serpins have been added to the protein structure data bank. These include an PAI-2 (an

intracellular inhibitory serpin) (Harrop, Jankova et al. 1999), PEDF (a noninhibitory but

physiologically active serpin) (Simonovic, Gettins et al. 2001), crmA (a viral serpin in

cleaved form) (Renatus, Zhou et al. 2000), and a thermophilic prokaryote serpin (Irving,

Cabrita et al. 2003) .

1:(Simonovic, Gettins et al. 2000), 2:(Irving, Cabrita et al. 2003), 3:(Lukacs, Rubin et al. 1998; Li, Wang et al. 1999), 5:(Briand, Kozlov et al. 2001), 6:(Gooptu, Hazes et al. 2000), 7:(Di Giusto, Sutherland et al. 2005), 8:(Di Giusto, Sutherland et al. 2005), 9:(Lukacs, Zhong et al. 1996), 10:(Renatus, Zhou et al. 2000), 11:(Simonovic, Gettins et al. 2000), 12:(Nar, Bauer et al. 2000), 13:(Im, Woo et al. 2002), 14:(Fulton, Buckle et al. 2005), 15:(Beinrohr, Harmat et al. 2007), 16:(Ye and Goldsmith 2001), 17:(Huntington, Read et al. 2000), 18(Law, Irving et al. 2005), 19:(Dementiev, Dobo et al. 2006), 20:(Li, Adams et al. 2007), 21:(Huntington, Pannu et al. 1999), 22:(Al-Ayyoubi, Gettins et al. 2004), 23:(Skinner, Abrahams et al. 1997), 24:(Schreuder, de Boer et al. 1994), 25:(Huntington, McCoy et al. 2000), 26:(Jin 1997), 27:(Skinner, Abrahams et al. 1997), 28:(Stein, Leslie et al. 1991), 29:(Al-Ayyoubi, Gettins et al. 2004), 30:(Law, Irving et al. 2005), 31:(Tucker, Mottonen et al. 1995), 32:(Dunstone, Dai et al. 2000), 33:(Hyun Kyu, Kee Nyung et al. 1995), 34:(Yamasaki, Takahashi et al. 2003), 35:(Jin, Abrahams et al. 1997), 36:(Stout, Graham et al. 2000), 37:(Kim, Woo et al. 2001), 38:(Simonovic, Gettins et al. 2001), 39:(Jankova, Harrop et al. 2001), 40:(Zhou, Huntington et al. 2003), 41:(Dementiev, Simonovic et al. 2003), 42:(Elliott, Lomas et al. 1996), 43:(McGowan, S., Buckle, AM., et al 2006), 44:(Aertgeerts, De Bondt et al. 1995), 45:(Zhang Q, Buckle AM 2007), 46:(Sharp, Stein et al. 1999), 47:(Stout, Graham et al. 2000), 48:(Baglin, Carrell et al. 2002), 49:(Yamasaki, Arii et al. 2002), 50:(Zhou, Stein et al. 2004), 51:(Zhou, Stein et al. 2003), 52:(Huntington, Kjellberg et al. 2003), 53:(Dementiev, Simonovic et al. 2003), 54:(Zhou, Wei et al. 2006), 55:(Li, Adams et al. 2007), 56 (Baglin, Carrell et al. 2002), 57:(Johnson, D.J., Li, W., Adams, T.E., et al 2006), 58:(Engh, Lobermann et al. 1989), 59:(Ryu, Choi et al. 1996), 60:(Baumann, Huber et al. 1991), 61:(Mourey, Samama et al. 1993), 62:(Carrell, Stein et al. 1994), 63:(Skinner, Chang et al. 1998), 64:(Baumann, Bode et al. 1992; Johnson and Huntington 2003), 65:(Johnson and Huntington 2004), 66:(Dementiev, Petitou et al. 2004), 67:(Li, Johnson et al. 2004), 68:(Johnson, Langdown et al. 2006), 69:(Xue, Bjorquist et al. 1998), 70:(Baumann, Bode et al. 1992), 71:(Harrop, Jankova et al. 1999), ** To be published

1.2.1 Protease subsite nomenclature

To have a better understanding about serpin structure and function, the basic nomenclature is explained here (Figure 1.2). A standard system for nomenclature was introduced by Schechter and Berger (Schechter and Berger 1968). This nomenclature allows for accurate description of the cleavage site of protease-substrates and the substrate binding site of protease-inhibitors. According to this nomenclature, the substrate amino acid residue immediately N-terminal to the scissile bond is the P1 residue. Moving from the substrate N-terminus to the P1 residue, residues are indicated as residue P2, P3, P4…. Thus, in serpin structures, the P numbering is applied up to P16 where the topological switch point or residue for strand insertion is located. The residues immediately C-terminal to the scissile bond are labeled P1’, P2’, P3’ … The protease subsites interacting with the P1 residue are denoted as S1, the primary substrate binding sites. Subsites are defined as S1, S2,… proceeding toward the N-terminus from the scissile bond and S1’, S2’ … toward the C-terminus from the scissile bond. The outline of Schechter and Berger nomenclature is described in Figure 1.2.

Figure 1.2 Schematic diagram of the Schechter and Berger convention of substrate- protease subsite nomenclature

1.2.2 Native state of Serpin

The secondary and tertiary structures of all serpins are almost identical within the core domain as represented in Figure 1.3 and Figure 1.4 A. Commonly, serpins have three β-Sheets with β-Sheet A being the largest of these. β-Sheet A has five strands, with the first strand composed of 5-6 residues and all the other strands are similar in length

(12-15 residues). Strands in β-sheet A are antiparallel to another except the central strand. After the insertion of the reactive center loop occurs, β-sheet A becomes completely antiparallel in the cleaved and latent forms of loop-inserted serpins (Figure

1.4 B). The other β-sheets, B and C, are smaller and shorter. They comprise six strands for β-sheet B and four strands for β sheet C. Generally, serpins have eight α-helices named A-H (Figure 1.4 A). Among those helices, only the F-helix is located at the front

16 side of the protein, using the face of β sheet A as a conventional orientation, and all remaining helices are placed on the backside of the protein. In the native conformation, the reactive center loop is located outside of the serpin core domain in a solvent-exposed environment. The N-terminus of the reactive center loop (RCL) is connected to the C- terminus of strand 3 in β-sheet A and the C-terminus of the RCL is connected to the strand 1 in β-sheet C. The RCL binds to the active site of a target protease. For inhibitory serpins the composition and the length of the reactive center loop have been shown to be essential for the inhibition of proteases. Based upon sequence alignment among this family, residues P12-P9 of the reactive center loop show more than 50% conservation of alanine at each position, whereas noninhibitory serpins have few or no alanine at these positions (Irving, Pike et al. 2000).

The length of the reactive center loop is a critical factor that determines whether the serpin has inhibitory activity, since the shortened RCL by mutagenesis (deletion residues within the RCL) is difficult to bring the target protease to the bottom of the serpin (Zhou, Carrell et al. 2001). Thus, the length of the RCL on the N-terminal side of the scissile bond (loop insertion side, P-side) is more restricted than that on C-terminal side (P’-side). The length of N-terminal side is usually 17 residues (only few is 16); however, but that of C-terminal varies from 5 to 9 residues.

Proximal reactive –loop hinge

The residues from P16 to P10 in the reactive center loop are called the proximal

hinge of the loop (Figure 1.3) (Potempa, Korzus et al. 1994; Stein and Carrell 1995). In

this region small aliphatic side chains such as alanine, glycine, and valine are typically

17 conserved. Mutation studies on this region have revealed that the mobility of the reactive center loop is essential to inhibit the target proteinase. The mutated bulky (Such as Trp,

Phe, Ile) or polar side-chains (such as Glu, Asp, Thr) interfere with normal insertion of the N-terminal end of the reactive center loop into β-sheet A, which is critical step in the serpin inhibitory mechanism.

Functional domain-reactive center

Several human mutant serpins carrying mutations of residues in the reactive center show loss of serpin inhibitory activity and even fail to make complex with the target protease. In the mutation of P1 position, it is possible that mutations at this site alter protease specificity rather than loss of inhibitory activity. The Pittsburgh variant of

α-antitrypsin, which has a mutation at the P1 position (Met->Arg), is a potent inhibitor of thrombin rather than an inhibitor of its own target protease, neutrophil elastase (Owen,

Brennan et al. 1983). Conversion of specificity in this serpin variant results in a fatal bleeding disorder.

Distal reactive –loop hinge

The distal hinge of the reactive center loop is comprised of strand 1 of sheet C

(s1C) and the turn connecting this strand to strand 4 of sheet B (s4b) (Figure 1.3) (Stein and Carrell 1995). This region shows dramatic movement when the active antithrombin undergoes conformational changes to the latent form (Whisstock, Skinner et al. 1998).

Breach

At the top of the sheet A, there is the region, called the “breach”, where the reactive center loop first inserts (Figure 1.3) (Whisstock, Skinner et al. 2000). This region is made by the top of s3A, and the top of s5A, s2B, s3B, and s4B. In the serpin- cleaved form, the breach region includes extra strand (s4A) since the cleaved reactive center loop is inserted into a β-sheet A and becomes the strand of β-sheet A. This region

is relatively well conserved and the hydrogen bond network in this place plays a critical

role in support of a native conformation (Harrop, Jankova et al. 1999).

Shutter

In the middle of the serpin there is the region, denoted as the “shutter”, which

controls the opening of the sheet A (Figure 1.3) (Stein and Carrell 1995; Whisstock,

Skinner et al. 2000). The shutter consists of the top of the helix B, the middle of s5A, the

middle of s3A and part of s6B. This region is also well conserved like the breach region

and mutation of residues in the shutter leads to serpin dysfunction (Stein and Carrell

1995).

Heparin-binding site

Several human serpins including antithrombin, heparin cofactor II, protease

nexin, and protein C inhibitor, have a binding site for sulphated polysaccharides (Carrell,

Stein et al. 1994). Upon binding heparin, serpin inhibitory activity is increased

significantly; for example, antithrombin inhibitory activity increases a 1, 000 fold (Olson

and Bjork 1991). Increased activity results from substantial conformational change of

19 the serpin whereby the RCL near the hinge region is expelled from β-sheet A and the P1 residue flips to an exposed enzyme-accessible conformation (Jin, Abrahams et al. 1997;

Whisstock, Pike et al. 2000).

Figure 1.3 Native conformation of Antithrombin (PDB: 2BEH). The reactive center loop is colored as red and distal reactive-loop hinge colored as blue. The residue P1on the RCL shows as stick and P numbers are counted from P1 site to N terminus of the RCL. The residues C-terminal to the scissile bond are labeled as P’ numbers. Pink color shows the domain of heparin binding sites. Two circles indicated to separately Breach and Shutter regions.

1.2.3 Cleaved form of Serpin

Since the first structure of the cleaved RCL (reactive center loop) α-PI was solved (Loebermann, Tokuoka et al. 1984), many other serpins in the cleaved form have been determined (Engh, Lobermann et al. 1989; Baumann, Bode et al. 1992; Mourey,

Samama et al. 1993; Schreuder, de Boer et al. 1994; Huntington, Read et al. 2000;

Simonovic, Gettins et al. 2000; Huntington, Kjellberg et al. 2003; Yamasaki, Takahashi et al. 2003). All of these structures share similar conformation that has, as a remarkable feature, the expansion of β-sheet A as a result of insertion of the cleaved reactive center loop from the fourth strand of the sheet as shown in Figure 1.4 B. Due to this insertion of the residues in the reactive center loop, the environment of these residues changes from solvent-exposed to being mostly buried.

1.2.4 Latent form of Serpin

A second type of uncomplexed serpin structure is the latent form. The latent form of serpin structure was first demonstrated in PAI-1 (Mottonen, Strand et al. 1992).

Later, the heterodimeric structure of antithrombin showed the same latent conformation

(Carrell, Stein et al. 1994). This latent conformation shows the same β-sheet expansion,

similar to the cleaved form of uncomplexed serpin, via the insertion of the reactive center

loop (Figure 1.4 C). However, the major difference between latent and cleaved form of

serpins is that the first strand of β-sheet C of the latent structure becomes C-terminal to the reactive bond, providing a return from the bottom of β-sheet A back to β-sheet B.

1.2.5 Partial inserted form of Serpin

Lastly, the fourth type of uncomplexed serpin, which is designated as δ, has been revealed in a naturally occurring variant of α1-antichymotrypsin (Gooptu, Hazes et al.

2000). The reactive center loop in this structure is partially inserted into β-sheet A up to

P12 (Figure 1.4 D). The lower part of the β-sheet A is occupied by the last turn of F-

helix and the loop connecting it to s3A. This conformation, which shows much more

insertion of the reactive center loop than in active antithrombin, is much less than full

loop insertion of PAI-1. Therefore, this structure demonstrates an intermediate form

involving a different extent of loop insertion.

1.2.6 Serpin polymerization

The first polymeric form of serpin was identified in the Z (Glu342→ Lys) allele

of antitrypsin (Lomas, Li-Evans et al. 1992). This polymeric conformation has been

discovered so far in more than forty variants of five different human serpins

(antichymotrypsin, antitrypsin, antithrombin, C1 inhibitor and neuroserpin) (Stein and

Carrell 1995; Lomas and Carrell 2002; Law, Zhang et al. 2006). Polymeric forms can be

occurred spontaneously both in vivo and in vitro (Figure 1.4 E) and they result in serpin-

related misfolding diseases (serpinopathies) such as emphysema, thrombosis (Stein and

Carrell 1995). Moreover, this phenomenon could be explained as a competition between

‘self’ or cis-reactive center loop insertion process and ‘foreign’ or trans reactive center

loop insertion process (Law, Zhang et al. 2006).

Native α-AT (PDB: 1QLP)

Figure 1.4 Serpin Structures. (A)-(D) Ribbon diagrams of different conformational serpin: reactive center loop are colored in red, β-sheets present as orange color and α- helixes shown in green color. (A) Native state conformation, α-AT (PDB: 1QLP) (B) Cleaved conformation of serpin, cleaved antitrypsin (PDB: 1AS4) (C) Latent conformation of serpin, PAI-1 (PDB: 1C5G) (D) Partial insertion conformation of serpin, partial insertion antichymotrypsin (PDB: 1QMN) (E) polymerized conformation of serpin, cleaved serpin polymer (PDB: 1D5S) Cleavage at the P5/P6 position results in RCL (red) insertion into β-sheet A and the bottom of β-sheet A is filled with an RCL from another molecule.

1.3 Serpinopathies

An essential feature of serpins with respect to their function as protease inhibitors is their ability to undergo dramatic conformational changes. Irreversible inhibition is initiated by cleavage of the reactive center of the inhibitor by the target protease. Since the native state of serpin is a metastable conformation, the cleavage action by the protease triggers the reactive center loop (P-side) to be inserted into the body of the serpin, rendering the irreversible trapping of the proteolytic enzyme by the inhibitor.

Both the complexities of serpin mechanism and structural mobility result in variety of serpin conformation-related human diseases denoted as “serpinopathies” (Stein and

Carrell 1995). Well known serpinophathies include emphysema, cirrhosis, thrombosis and dementia. More than 100 different mutations in variants are associated with inhibitory deficiency or dysfunction (Carrell and Lomas 1997). The majority of these mutations are occur in the domains that control molecular mobility, including two hinges of the RCL and the shutter area, both highly conserved regions (residues) among serpin family.

Misfolded serpin results in two different forms. One is latency and the other is polymerization. The latent state can be promoted from pathogenic mutations in serpins.

This causes disease due to a reduced amount of active inhibitory serpin, such as “wibble and wobble” of the disease-related antithrombin variants (Beauchamp, Pike et al. 1998).

Polymerization of serpin is caused by “domain swap” with one another causing the formation of long-chain polymers (Lomas, Li-Evans et al. 1992; Huntington, Pannu et al.

1999; Dunstone, Dai et al. 2000). Polymeric serpins are inactive and pathogenic. Serpin polymerization leads to diseases in two different ways. Primarily, the lack of active

26 serpin caused by serpin polymerization results in loss of controlling protease activity and tissue destruction, which is found in the case of antitrypsin deficiency. Second, the serpin polymers clog up the endoplasmic reticulum (ER) of cells where serpins are produced. This phenomenon causes cell death and tissue damage such as the case when antitrypsin polymer results in liver damage and cirrhosis (Lomas, Li-Evans et al. 1992).

1.4 Serpin inhibitory mechanism

All serpin-protease inhibition reactions can be described by a general mechanism.

This mechanism can be considered as a branched pathway, which is the suicide substrate inhibition mechanism that is shown in Figure 1.5 and Figure 1.6 (Silverman, Bird et al.

2001).

In this inhibitory serpin mechanism, the proteases recognize the reactive center loop of serpin as ‘bait’ and the sequence of this region is the primary determinant of inhibitory specificity. The serpin acts as a potential substrate. The initial recognition and reaction of serpin by protease is similar to that of serine protease acting on a substrate.

Thus, general serine protease mechanism also will be discussed in section 1.4.1. After initial reaction step of acylation, massive conformational change of serpin results in stabilizing the irreversible trap of the proteolytic enzyme by inhibitor (Figure 1.7).

The general steps of this inhibitory pathway can be described as follows, and as presented in Figure 1.5 and Figure 1.6.

1. Formation of an initial noncovalent Michaelis complex

2. Attack of the reactive site serine on the scissile bond to form covalent ester linkage between serine 195 of the protease and backbone carbonyl of the P1 residue and cleavage of the peptide bond.

3. Only at this point (step 3), with the leaving group departed and the reactive center loop cleaved, can the reactive center loop begin to insert into β-sheet A and transport the covalently bounded protease with it.

4. Upon complete loop insertion the protease is translocated by over 70 Å, and its active site distorted.

5. During the translocation event protease still has a chance to finish the proteolytic reaction and be released from the serpin. This process is the branch point in Figure 2.5. If k3 << k4, the reaction is predominantly an inhibitory process. However, if k3 >> k4, the reaction is predominantly a substrate reaction with the target protease.

Figure 1.5 Branched pathway mechanism of serpins as suicide substrate inhibitors. The serpin (I) inhibition of protease (E) proceeds through an initial noncovalent Michaelis complex (EI) where the protease does not show any conformational change and the serpin only shows minor conformational change in the RCL and some of flexible regions (such as helix D, E). Subsequently an acyl enzyme intermediate (EI*) is formed by the nucleophilic attack of Ser195 in trypsin onto the scissile bond of the serpin. This covalent acyl enzyme intermediate form progresses to either a kinetically trapped loop- inserted covalent complex (EIc) or cleaved serpin (I*) and released protease, which is shown in noninhibitory pathway or substrate pathway. The serpin β-sheet A represents in purple and the RCL as red. Serine protease is in green. Since the structure of the covalent acyl enzyme intermediate form is not known, this form is represented as a circle.

Figure 1.6 Kinetic diagram of branched pathway mechanism of serpins. I represents the serpin and E represents the protease. The noncovalent Michaelis complex is reversibly formed with forward rate constant of k1 and reverse rate constant of k-1. This non-covalent complex progresses through the reaction of peptide bond hydrolysis to form the acyl enzyme intermediate EI* between the catalytically active residue serine and the backbone carbonyl bond of the P1 residue in serpin. This intermediate was achieved with forward rate constant of k2 and back rate constant of k-2 (Although this backward reaction rate constant is almost negligible). At this stage the RCL is cleaved and the leaving group of the P’ side peptide (C-terminus from the scissile bond) is released. The cleaved RCL begins to insert into β-sheet A, and at the same time, the covalently bounded protease is translocated to the bottom of serpin, thereby trapping kinetically with protease as the complex EIc. This insertion process of the RCL (or translocated process of the protease) is achieved with rate constant k4. Since this reaction is competing with the process of deacylation in order for the protease to complete the substrate reaction, there is still room for the protease to release serpin and escape, leaving behind cleaved serpin, I*. This leaving reaction by protease (or deacylation) is occurred with rate constant k3. Shin and Yu (Shin and Yu 2002) showed that the rate of insertion of cleaved RCL is relatively faster than that of deacylation by protease. The kinetically trapped protease (EIc) can be released to free protease and cleaved serpin with rate constant k5, where k5 is much -1 -1 slower than k3. (k5 (more than several days ) <<<

Acyl-enzyme Intermediate

Non-covalent Michaelis complex Final stable Complex

Active protease Cleaved serpin

Figure 1.7 Energy diagram of serpin inhibition Based on the previous studies (Sohl, Jaswal et al. 1998; Cabrita and Bottomley 2004; Shin and Yu 2006), a free energy diagram for the serpin inhibitory mechanism can be presented. The serpin structure of EI* could be closer to a stressed form of I rather than the relaxed forms of EIc or I*. This is because EI* is an intermediate right before serpin undergoes tremendous conformational change. Thus, initially loaded strain energy in the native form would be lasted up to EIc. The pull force for insertion of the reactive center loop and translocation of the protease is derived by the energy differences between EIc and EI*

1.4.1 Serine protease Mechanism

The catalytic triad is located in the active site of the protease, where catalysis occurs. This triad consists of three essential amino acids: histidine (His 57), serine (Ser

195) and aspartic acid (Asp 102). The –OH group of serine acts as a nucleophile attacking the carbonyl carbon of the scissile peptide bond of the substrate. Histidine as a general base removes proton from the serine 195 and then acts as a general acid by donating the proton back to the amine of the cleaved peptide bond. Histidine as a general base also removes proton from water and then acts as a general acid by donating the proton to the serine 195 to regenerate the free enzyme. The carboxyl group of the aspartic acid makes hydrogen bonds with the histidine.

The protease and substrate associate to form a noncovalent complex (ES) known as the Michaelis complex. The –OH group in serine then attacks the carbonyl carbon of the P1 residue. The nucleophilic attack generates a transient tetrahedral intermediate at the carbonyl carbon. An acyl-enzyme intermediate (EP2) forms in which the protease is acylated by the amino terminal product of peptide bond hydrolysis. Then, a water molecule attacks the carbon of the acyl-enzyme intermediate to reform the free serine residue. This final deacylation reaction also occurs through the formation of a tetrahedral intermediate. This mechanism is outlined in Figure 1.8.

Figure 1.8 Schematic diagram for the chemistry of serine protease reaction. The diagram shows the serine195 and histidine 57 of the protease and substrate peptide bond. The movement of electron shows in arrows. Serpin traps the protease in the half of hydrolysis reaction by enzyme. The capturing reaction is circled in pink color. The red dash in acyl-enzyme intermediate form represents covalent bond (Adopted from http://en.wikipedia.org/wiki/serpin).

1.4.2 The mechanism of non-serpin inhibitors

Protein inhibitors of serine protease can be grouped into about ten families based

upon sequence similarity, topological similarity, and mechanism of binding (Laskowski

and Kato 1980). Six of these families are well characterized in terms of the topological

relationships between the disulfide bridges and the location of the reactive sites. They

are Bovine Pancreatic Trypsin Inhibitor (BPTI, Kunitz) family, Pancreatic secretary

trypsin inhibitor (PSTI, Kazal) family, soybean trypsin inhibitor (STI, Kunitz) family,

soybean trypsin inhibitor (Bowman-Birk) family, streptomyces subtilisin inhibitor (SSI)

family, Potato Inhibitor 1 (PI1) family.

Most serine protease inhibitors adopt a substrate like mechanism in which the

inhibitor loop inserts into the substrate-binding groove of the protease (Bode and Huber

1978). These families of inhibitors have been widely termed the ‘canonical’ or ‘lock and

key’ inhibitors. Additionally, these molecules are relatively small (about 10-20 kDa) and

have a “canonical” conformation. The canonical conformation is defined by the highly

similar binding modes observed in most naturally occurring small protein substrates in

which the inhibitor loop inserts into the substrate-binding catalytic triad pocket

(Laskowski and Kato 1980). The interactions of the P1 side chain with the specificity

pocket (S1 subsite) are the primary determinants of specificity for a particular protease

(for example, arginine and lysine at P1 confer trypsin-like specificity). All protease

inhibitors block the entrance of the substrate to the catalytic site through steric hindrance.

This steric hindrance can be produced by binding of an inhibitor peptide segment

directly into the catalytic site (Laskowski and Kato 1980).

In the interfaces of inhibitor:protease complexes, the residues from the protease interact with the loop of inhibitor as tightly as protein interiors (Janin and Chothia 1990).

The interface between protease and inhibitor is complementary and only limited conformational changes are made in complex with a root-mean-square deviations on the order of 3.5Å for flexible region of both protease and inhibitor (McPhalen and James

1987). Since these stabilizing interactions appear to prevent deformation of the reactive site for peptide bond cleavage, the stable complex results in very slow dissociation. This means that deacylation step is delayed due to the tightly bound leaving group, oriented such as to inhibit deacylation step and favoring the regenerating scissile bond (P1-P1’ peptide bond).

One of most well known studies of inhibitor structure is that of the bovine pancreatic trypsin inhibitor (BPTI) (Huber, Kukla et al. 1974). BPTI has a general canonical binding conformation (Figure 1.9), which represents most of protease inhibitors. In this structure, the P1 residue of the inhibitor is located close to the catalytic residues of the protease with the P1 carbonyl carbon interacting with the ser195 γO of the protease. The carbonyl group projects into the oxyanion hole in which it forms hydrogen bonds with Gly193 N and Ser195 N.

Serpins also adopt canonical conformations in which the flexible reactive center loop acts as bait for target protease. However, compared with the small molecule inhibitors as described above, serpins inhibit their cognate protease through irreversible process leading to a kinetically trapped acyl-enzyme intermediate. Once the RCL is cleaved, serpin undergoes massive conformational change before deacylation can take place. At the acyl-enzyme stage, the cleaved RCL is inserted into the β-sheet A and the

35 covalently attached protease is translocated from the top of the serpin to the bottom.

Although numerous studies have been carries out for understanding serpin mechanism, clear explanation for how deacylation is prevented (or how acyl-intermediate is stable) has not been addressed.

Figure 1.9 A ribbon representation of rat anionic trypsin (green) bound to bovine pancreatic trypsin inhibitor (pink). The catalytic triad (orange stick) and lysine (P1, magenta) at the BPTI in its canonical binding conformation are highlighted as stick presentation.

1.5 Structural basis of serpin inhibitory mechanism

In order to address how the deacylation step is suppressed during the translocation of the target protease in serpin inhibitory mechanism, the Michaelis complex solved by the Goldsmith’ lab is an appropriate model system to investigate such an open question (Ye 2001). This Michaelis complex is formed between manduca Sexta serpin-1B (A353K) and rat anionic trypsin. Both proteins have good characteristics for

X-ray crystallographic studies in that they are well expressed and stable and especially, because serpin-1B is not glycosylated (discussed in Chapter 3.1). The structure analyses of serpin-1B and its complex with trypsin S195A (Ye, Cech et al. 2001) serve as a background for my study.

1.5.1 Structure of active serpin 1K from Manduca Sexta

Serpin 1K is a chymotrypsin inhibitor. This serpin gene, which is one of 12 serpins, was found in the hemolymph of the hornworm moth Manduca sexa (Jiang and

Kanost 1997). The same gene encodes all of the Manduca Sexta serpins, and alternative splicing of the final exon produces each serpin. The reactive center loop and two strands of β-sheet B are encoded by the exon. The serpin 1K, like active α1-antitrypsin, has a canonical conformation at the reactive center loop, which is complimentary to P3 through

P’3 of the target protease binding site (Li, Wang et al. 1999). However, the hinge region

(P17-P13) of the reactive center loop in this serpin has tight interactions to a stabilized closed form of β-sheet A similar to ovalbumin and α1-antichymotrypsin.

Tight Interaction

Figure 1.10 Ribbon diagram of serpin 1K (PDB: 1SEK). The RCL is shown in red. Helices are in cyan, and sheets are in magenta.

1.5.2 Initial noncovalent Michaelis-like complex

The first step in the ‘suicide’ inhibition mechanism of serpin is the recognition of the cognate protease to form the non-covalent Michaelis-like complex. Most canonical inhibitors only have this step for their inhibition mechanism. Therefore, the complex formed in this step shows very important features which are the stability and specificity of the complex. Although the inhibition mechanism used by serpins is unlike that of other protease inhibitor families, due to the fact that serpin inhibition involves reaction steps (acylation, loop insertion) beyond formation of the initial noncovalent complex, the

Michaelis complex could still serve as an important aspect of specificity and rate of

serpin inhibition (see chapter 5 for detailed discussions on the importance of the

Michaelis complex). The Michaelis complex in comparison of serpin itself could also

provide important conformational changes for understanding serpin mechanism.

Since the structure of Serpin-1B (A353K) with trypsin S195A was the first solved Michaelis complex in serpin family (Ye, Cech et al. 2001) and Serpin-1B (A353K) is closely related to the chymotrypsin inhibitor serpin 1K (Li, Wang et al. 1999) in sequence and structure, it could be explained the unique feature of the Michaelis complex structure can be compared to active serpin K. Moreover, sequence identity between serpin-1B and serpin-1K is more than 94%. The overall superposition of serpin 1K and the Michaelis complex between serpin-1B with trypsin S195A is presented in Figure

1.11. This superposition shows that there are significant changes in the RCL of the serpin in the Michaelis complex. The detailed analysis is followed in next subsection.

Figure 1.11 Superposition of active serpin 1K (PDB code: 1SEK) (cyan) with the Michaelis complex of serpin-1B A353K and trypsin S195A (Orange) The red square box indicates the important regions to have unique interactions in complex.

The unique features of non-covalent serpin-protease complex

The noncovalent Michaelis serpin1B-trypsin S195A complex (Ye, Cech et al.

2001) revealed that the contact within the reactive center loop is more intimate near the catalytic His 57 than in the well-studied bovine pancreatic trypsin inhibitor (BPTI)- trypsin complex (PDB 2ptc) or the BPTI-trypsin S195A complex (PDB 3tgj) (Figure

1.13) (Ye and Goldsmith 2001).

The P1’ residue of the inhibitor, Ser 354, forms a close contact with His 57 of the protease (Figure 1.12). His 57 also makes a hydrogen bond to the catalytic triad residue

Asp102 in this orientation. The P1’ (Ser 354) residue forms a unique close contact with the catalytic His 57 in this serpin-protease complex. The P1’ Ser residue might provide a hydrogen bonding partner for His 57, an alternative to the normal interaction of the His

57 with Ser 195 in wild type enzyme. Such a contact could release the acylated Ser 195 for the conformational changes associated with strand insertion.

This X-ray structure also revealed that interactions between the serpin P side of the scissile bond and trypsin are much more extensive than those in other inhibitor- protease complexes. Especially, P4 (Ile350) of serpin has the hydrophobic interactions with the side chain of Trp215, Leu99 and the main chains of Lys97 and Thr98 of trypsin

(Figure 1.12).

Also, Gln192 of trypsin forms tight interactions with the main chain of P1’

(Ser354) and P2 (Pro352) on serpin (Figure 1.12).

Figure 1.12 Interactions between the RCL (P2’- P4) and trypsin in the Michaelis complex (PDB code 1K9O). Trypsin is in green and the RCL of serpin is in cyan. The black dash represents hydrogen bond.

Figure 1.13 Superposition of the Michaelis complex (serpin-1B-trypsin S195A) (PDB code: 1K9O) (trypsin shown in orange and serpin shown in magenta) with trypsin bounded to the bovine pancreatic trypsin inhibitor (BPTI) (PDB code: 3TGJ) (trypsin in cyan and BPTI in blue). The catalytic triad and P1residue (Lysine) in inhibitors are presented with stick.

Topological switch point and other conformational change

The position P16 in the reactive center loop of serpin is called “the hinge”, which is the topological switch point for strand insertion. The residues P15 and P14 above the topological switch point commonly have a network of hydrogen bonding interactions in active serpin 1K as well as in the most active serpins. The interactions around the hinge region could stabilize the reactive center loop in active serpin to prevent insertion of the reactive center loop into β-sheet A (Ye, Cech et al. 2001). Although the position of the

P15 (Gly339) residue did not change, the hydrogen bond between P14 (Ala 340) and Trp

184, the body of serpin was lengthened slightly (Figure 1.14). Since this hydrogen bond plays important role on stabilizing the active conformation of the serpin1K, weak hydrogen bond at P15 in the complex may be loosen and make easy for the RCL to be inserted. Moreover, residue P13 (Glu341) including the residues in this hinge region altered their position compared to the active serpin 1K. These conformational changes in the Michaelis complex resulted in loosening “locked” conformation of serpin-1K.

Figure 1.14 Conformation of the RCL and topological switch point (PDB code: 1K9O). Serpin-1B (cyan) is presented in ribbon diagram, and stick residues involves in interactions with the residues (shown in stick) of trypsin (dark green). The dashed lines represent hydrogen bonds.

1.5.3 Covalent complex

Once the acyl-enzyme intermediate is formed (through an ester bond between the catalytic Ser195 of the protease and the P1 residue of the serpin), this cleaved RCL attached to protease is incorporated into β-sheet A and the protease is translocated from the top of the serpin to the bottom. This is shown by the X-ray structure (Huntington,

Read et al. 2000) of…. In this structure, the active sites of the protease are is no longer active. The acyl intermediate is now stable due, in part to a huge structural rearrangement in protease active site resulting in blockage of the deacylation reaction

(Figure 1.15). However, we still cannot understand why deacylation does not take place.

Figure 1.15 Ribbon presentation of the final complex form in serpin inhibitory mechanism (PDB code: 1EZX). The trypsin is shown in cyan. The helices are in green, sheets are in orange and loops are in pink in serpin molecule. The inserted RCL into β- sheet A is highlighted as red.

1.6 Description of dissertation research

This dissertation is composed of three different sections. The first section is main theme of the whole dissertation, where I will show the extent to which we reveal the serpin inhibitory mechanism based upon our structural studies and kinetic data collected in Michael Kanost lab. The first section comprises chapters one through six. These six chapters cover the purpose, background and results of the research. They present the determination and analysis of two three-dimensional structures of a rat anionic trypsin in complex with two different serpin mutants. The atomic structures of these complexes, solved by X-ray crystallography, provide structural insight into the serpin inhibitory mechanism.

The second section is about crystallization of BmSPN-2 (Brugia malayi serpin precursor). This part presents efforts to clone, express, purify and crystallize of BmSPN-

2. Trials on improving crystal quality are presented.

The third section provides the information about VopA (Vibrio outer protein A).

This is a member of a family of type III effectors called YopJ-like proteins. All YopJ-

like proteins function as acetyltransferases. VopA is the extremely potent inhibitor of the

MAPK pathways unlike YopJ which inhibits both MAPK pathways and the NFκB

pathway. This chapter entails efforts to clone, express, purify and crystallize VopA.

Chapter 1 introduces the overall research project of a serpin inhibitory

mechanism and the purpose of performing this project. Also, this chapter provides

background related to main project, encompassing the structural basis on serpin

inhibitory mechanism.

Chapter 2 describes the results from serpin inhibition assay. The significance of the residues of serpin at the interface with trypsin will be discussed along with other serpin mutant residues interacting with trypsin and their putative roles in serpin inhibition reaction.

Chapter 3 presents the expression, purification of two mutant serpins (serpin-

1B/I350A and serpin-1B/S354A) and rat anionic trypsin S195A) and co-crystallization of serpin-trypsin; as well as methods of improving crystal quality (high resolution data) are included.

Chapter 4 describes the determination of the crystal structure of two mutant

Michaelis complexes; including collection of X-ray diffraction data, determination of the space group, refining and modeling the structures.

Chapter 5 focuses on analysis of the crystal structures of mutant serpin-protease complexes (serpin-1B/I350A with trypsin S195A, serpin-1B/S354A with trypsin S195A) including an overall description of two mutant complexes, a detailed analysis of the conformation of the reactive center loop, and the other conformational change in these structures. Based upon several studies and the structural data, the structural implication from the change of conformation in the mutant Michaelis complexes is addressed.

Chapter 6 provides conclusions for the X-ay crystallographic project of the

Michaelis complexes between a single trypsin S195A and two mutants serpin-1B/I350A and serpin-1B/S354A. In this chapter, future experiments are proposed to analyze the effect of I350A mutation in greater detail using time resolved fluorescence energy transfer.

Chapter 8 presents the effort to clone, express, purify and crystallize of BmSPN-

2. Several trials on improving crystal quality are also introduced.

Chapter 9 provides the information about VopA and entails efforts to clone, express, purify and crystallize VopA.

CHAPTER TWO INTERPRETATION OF KINETIC DATA OF INHIBITION OF TRYPSIN BY SERPIN-1B (A353K) AND ITS MUTANTS

2.1 Introduction

As discussed in Chapter 1, the Michaelis complex of serpin-1B with trypsin (Ye,

Cech et al. 2001) reveals intimate and extensive interactions between the P-side of the reactive center loop of the serpin and the trypsin S195A. Especially, there is unique hydrophobic interaction at the P4 site of the serpin. This interaction contrasts those formed by the low molecular weight serine protease inhibitors, which only extends through P3-P3’. There is also close interaction between Ser354 at the P1’ position and

His57 of the trypsin.

To investigate the importance of individual residues involved in interactions at the interface, our collaborator, Kanost’s group, introduced several point mutations at interface residues of the serpin-1B (A353K) as shown in Figure 2.1 and Figure 2.2. The construct of the serpin-1B (A353K) was used because its expression is very good and it is a good trypsin inhibitor. Serpin-1B (A353K) will be considered as the wild type serpin throughout this thesis. Effects of the mutations were assessed in inhibition assays against active bovine trypsin. The mutations can be grouped as 1) the unique interaction sites

(P4 and P1’), 2) the P4’ site, 3) near the hinge region, and 4) at helix F1 according to the structural location of the serpin-1B in complex with trypsin as shown in Figure 2.2.

The result of their inhibition assays guided our further crystallographic analysis by directing us to mutant serpins that have altered inhibition characteristics but

50 nevertheless do interact with trypsin. Thus, the kinetic data of Kanost group are presented here.

near Hinge region Helix F1

Figure 2.1 The mutated residues at the interface of the serpin-trypsin Michaelis complex. This figure only shows serpin-1B (helices in cyan, sheets in pink and the RCL in red) at the interface of the Michaelis complex. The mutated residues for inhibitory assay are highlighted in spacefill. The space presentation color follows CPK mode. The P1 residue is presented in stick model.

near hinge at Helix F1 region

Figure 2.2 Site-directed mutagenesis of serpin residues at the interface of the serpin- trypsin Michaelis complex. The entire Michaelis complex (serpin-1B: helices in cyan, sheets in pink and the RCL in red, trypsinS195A in green) is presented in ribbon diagram. The mutated residues for the serpin inhibitory assays are shown in space-filling models and colored in CPK mode. The P1 residue is presented in stick model.

2.2 Materials and Methods

Inhibition assays employing several serpin mutants were performed by Youren

Tang at Michaelis Kanost’s lab in Kansas State University. This section of the Material and Methods was written by Youren Tang and modified by me.

2.2.1 Site-directed mutagenesis of serpin-1B (A353K)

The cDNA encoding serpin-1B (A353K) (Jiang et al., 1995) in H6pQE60 was

available from prior studies. This cDNA contains a mutation of the P1 residue (A353K),

which alters the specificity of the serpin-1B from chymotrypsin to trypsin (Kanost et al.,

1989). Site-directed mutagenesis at the interface of the serpin-1B (A353K) (Figure 3.1)

(I354A, I354E, S354A, S354M E191A, R192A, A345E, V344A, A343E, V342A, E341A,

V340A, G339A) were separately introduced using a QuikChange XL Site-Directed

Mutagenesis Kit (Stratagene) following the manufacturer’s instructions and verified by

sequencing.

2.2.2 Inhibition assays

The concentration of bovine pancreatic trypsin (Sigma) was determined by burst

titration with p-nitrophenyl-p’-guanidinobenzoate (NPGB) according to Chase and Shaw

(1967). For serpin activity assays, 3.5 ng of trypsin freshly diluted (in 5 µl of 50 mM

Tris-HCl, 50 mM NaCl, pH 8.0) were incubated for 10 minutes at room temperature with

varying amounts of serpin (in 20 µl of 50 mM NaH2PO4, pH 8.0, 300 mM NaCl, 20 mM imidazole) and bovine serum albumin (4 µg in 2 µl of 0.9% NaCl). The residual amidase

53 activity of trypsin was assayed by adding 200 µl of 100 µM N-acetyl-Ile-Glu-Ala-Arg-p-

nitroanilide (in 50 mM Tris-HCl, 50 mM NaCl, pH 8.0), and measuring A405 over 10 min.

To measure the inhibition rate constants (kinh) under pseudo-first order conditions, serpin was mixed with 3.5 ng of trypsin at a serpin/trypsin molar ratio of 20 or 10, as described above for inhibition assays, then incubated at room temperature for 15, 30, 45,

60, 75, 90, or 120 sec, and finally, the residual amidase activity of trypsin was measured.

The inhibition rate constants for the pseudo-first order reactions (kinh) were calculated as described by Beatty et al. (1980).

-kinht v= v0e

The equation is describing the loss of activity as function of time, where v0 is the initial activity and v is the activity at the given time. The time dependent loss of activity from v0 to v=0 is therefore a measure of kinh. To maintain pseudo-first conditions during

measurements, a 10-fold excess or more of the inhibitor is typically used.

2.3 Results and discussion

2.3.1 Mutations at the interface between serpin-1B and trypsin affect serpin inhibitory activities

We measured the effects on serpin inhibitory activity of mutations (Figure 2.2)

at the interface by using the inhibition assay in which the residual amidase activity of

trypsin after 10 min incubation with mutated serpin was compared to control reactions

(without added serpin) (Figure 2.3). The result is presented as two groups of mutants (A

and B) in Figure 2.3. The percentage of inhibition, in which the decrease of trypsin

activity is compared to the control reaction of the trypsin in absence of the serpin, is

presented as a function of the molar ratio of inhibitor and trypsin (Figure 2.3). In the

Figure 2.3 most mutations (Figure 3.3A (I354A, I354E, E191A), Figure 3.3B (A345E,

V344A, A343E, V342A, E341A, G339A)) abolish the inhibitory activities toward trypsin.

Two of these mutations, I354A and I354E, are at the unique interaction sites (P4). The

rest of the mutations are located either near the hinge region (A345E, V344A, A343E,

E341A, G339A) or at helix F1 (E191A) of serpin (Figure 2.2). On the other hand, the

mutants (S354A, S354M) at the P1’ site and V342A near the hinge region have the

similar activities to that of the wild type serpin (Figure 2.3). The mutant (R192A) at

helix F1 displays modestly decreased levels of the inhibitory activity compared to the

wild type serpin.

A 100 Serpin-1B (A353K) 75 P4' Leu → Ala P1' Ser → Ala 50 P1' Ser → Met P4 Ile → Ala 25 P4 Ile → Glu Inhibition (%) Arg192 → Ala 0 Glu191 → Ala 0 4 8 12 16 20 Serpin/trypsin molar ratio

B 100

Serpin-1B (A353K) 75 P9 Ala → Glu P10 Val → Ala 50 P11 Ala → Glu P12 Val → Ala P13 Glu → Ala

Inhibition (%) 25 P15 Gly → Ala

0 0 4 8 12 16 20 Serpin/trypsin molar ratio

Figure 2.3 Inhibition of trypsin by serpin-1B (A353K) and its mutants 3.5 ng of trypsin was incubated for 10 min at room temperature with each serpin mutant at different molar ratios. The residual amidase activity of trypsin was measured by using a chromogenic substrate as described in Materials and Methods. The decrease of trypsin activity compared to control reactions without added serpin is indicated as percentage of inhibition. (A) Mutations are introduced at P4, P1’, P4’ and near the RCL region of the serpin-1B. (B) Mutations are located near the topological switch point in the RCL.

2.3.2 Effects of the mutations at unique interactions sites (P4 and P1’) on the serpin activity

1) Mutated serpin (I350A) at the P4 position shows the loss of inhibitory activity perhaps due to enhanced rate of deacylation

The Ile350 at the P4 position directly interacts with the residues of trypsin at the initial stage of the non-covalent Michaelis complex formation. The loss of the hydrophobic interactions of the I350A serpin abolishes the inhibitory activity (Figure

2.3). The inhibition rate constant of the I354A serpin also demonstrates the 22-fold lower activity compared to the wild type serpin (Table 2.1). The lost inhibitory activity of the serpin (I354A) seems to be caused by an increased rate of deacylation since the mutated serpin acts as a substrate for the trypsin. Clearly, it is binding trypsin, since it is being cleaved (Figure 2.4). Every step of the inhibitory mechanism of the serpin is much slower than the normal turnover rate of trypsin (O'Malley, Nair et al. 1997). The rate of acylation of serpin toward trypsin is about 31/second and the insertion rate of the cleaved

RCL is less than 2/second (Shin and Yu 2002). Also, the deacylation rate is about

0.01/second (Shin and Yu 2002). However, the normal turnover rate of the trypsin is about 70/second (O'Malley, Nair et al. 1997; Radisky, Lee et al. 2006). Therefore, even if it might be considered that the rate of loop insertion could be slower than deacylation in this mutated serpin (I350A), it cannot be the main contribution for the mutated serpin being a substrate, due to the fact that the cleaved serpin is detective even in 5 seconds reactions of the time course reaction (Figure 2.4). In contrast to the I354E mutation, we can only observe the native conformation (uncleaved form) of the serpin (I354E) after 30 min reaction with the trypsin (Figure 2.5). This examination of the reaction products

(Figure 2.5) reveals that the serpin (I354E) loses binding affinity toward the trypsin, leading to loss of the inhibitory activity.

In order to understand why deacylation is arrested, visualization of the mutant serpin (I350A) in complex with trypsin S195A by X-ray crystallography would be helpful to understand the mechanistic role of the Il350 residue on the serpin mechanism.

Because this mutant Michaelis complex can be compared with the solved structure of the wild type Michaelis complex to investigate a possible role of the P4 residue on the enhanced deacylation rate. In addition to the mutant Michaelis complex, a recent publication of an acyl-serine protease complex will be helpful to understand the mechanism of deacylation arrest.

seconds minutes 0 5 15 30 45 60 75 2 5 10 20 30 T

80 serpin-trypsin complex cleaved complex 52 intact serpin

cleaved in RSL 35

Figure 2.4 Time course of reaction of P4 Ile → Ala mutant with serpin (molar ratio = 4:1 serpin:trypsin) Time 0 is serpin alone without added trypsin. Lane T is trypsin alone with no serpin, in which the trypsin cannot be detected by the serpin antibody. This is a western blot, with bands detected with rabbit anti-serpin antibody. Black arrows indicate intact serpin band.

Table 2.1 Inhibition rate constants (ka) of trypsin and serpin-1B (A353K) and its mutants. Serpin-1B (A353K) is a wild type serpin for trypsin inhibition assay. P numbering shows the position of the mutation.

-1 -1 Serpin kinh (M s ) 1B(A353K) 1.8 × 105

P1’ S→A 1.1 × 105

P1’ S→M 1.8 × 105

P4 I→A 8.2 × 103

P10 A→V 8.5 × 103

P12 A→V 1.2 × 105

Figure 2.5 Products of reaction of serpin mutants with trypsin for 30 minutes (molar ratio = 4:1 serpin:trypsin). Lane C is serpin alone (P1’Ala mutant) without added trypsin. Lane T is trypsin alone with no serpin. This is a western blot, with bands detected with rabbit anti-serpin antibody. Black arrows indicate intact serpin band.

2) Mutations at the P1’ position have a minor effect on inhibitory activity

As discussed in Chapter 1, the Ser 354 of serpin has the unique interaction with the His 57 of the trypsin. The Ser 354 at P1’ position is well conserved among inhibitory serpin family and a few serpins has met residue in this position. Thus, the inhibition assays on this residue by mutated to Met or Ala was performed to investigate the importance of the Ser 354 in the serpin mechanism. As expected, S354M shows almost same activity as the wild type serpin. The mutated serpin (S354A) also displays slightly reduced inhibitory activity compared to the wild type (Figure 2.3 and Table 2.1).

However, the structural observation on this residue which has a hydrogen bond with the

His57 of the trypsin is interesting to carry on further structural studies. Therefore, it

would be explored how the mutated serpin (S354A) influences the conformation of active

sites (His 57) of the trypsin in the Michaelis complex of the mutant serpin with

trypsinS195A.

2.3.3 Effects of the mutation at the P4’ site on the activity of the serpin

The mutated serpin (L357A) at the P4’ position leads to a dramatic loss of

inhibitory activity (Figure 2.3). A smaller hydrophobic residue seems to disrupt the wild

type van der Waals interactions between Leu 357 of serpin and Tyr 39 of trypsin. The

result would suggest that the interaction of the Leu 357 of the serpin with Tyr 39 of

trypsin may be important in forming the Michaelis complex.

Similarly, interactions between the Glu residue of plasminogen activator

inhibitor-1 (PAI-1) at the P4’ position (as well as P5’ Glu) and tissue-type plasminogen

activator (tPA) are also shown to be important in the formation of Michaelis complex and

60 its inhibitory activity (Madison, Goldsmith et al. 1989; Ibarra, Blouse et al. 2004).

Alanine substitutions at both residues of PAI-1 reduced the rate of PAI-1-tPA Michaelis complex formation by 13.4-fold, and decreased inhibitory activity by 4.7-fold (Ibarra,

Blouse et al. 2004). Therefore, these studies suggest that the interactions around this P4’ residue of the serpin are critical for the recognition of its cognate protease.

2.3.4 Mutations near the hinge region abrogate serpin activity

The hinge is the topological switch point at the P16 position for the strand insertion. In the hinge region of the inhibitory serpins, a more specific sequence restriction is shown that the residues from P12 to P9 have more than 50% conservation of alanine at each position (Gettins 2002). Many kinetic studies elucidated that the mutations of the residues near the hinge region (P16-P10) had great effects on the rate of the RCL insertion, leading to a loss of the overall activity of the serpin (Devraj-Kizuk,

Chui et al. 1988; Perry, Harper et al. 1989; Levy, Ramesh et al. 1990; Aulak, Eldering et al. 1993). Therefore, the mutations of the residues (P16 to P10) near the hinge region become the substrates for their target protease, due to reduced rate of the RCL insertion.

Lost activities of the mutants near the hinge region in our assays are in accordance with results from these earlier mutagenesis studies. All of the mutations (A345E, V344A,

A343E, E341A, G339A), except V340A, almost completely abrogate the ability of the serpin to inhibit the trypsin. The mutation (A342V) of the P12 residue had no effect on the inhibitory activity of the serpin (Figure 2.3), rendering a similar inhibition rate constant to the wild-type serpins-1B (Table 2.1). Since Ala and Val have similar

61 properties, the mutation may not affect RCL insertion or the stability of the final covalent serpin-trypsin complex.

2.3.5 Effects of the mutations at Helix F1 on the serpin activity

E191A had no inhibitory activity, whereas R192A had ~35% activity compared to the wild-type serpin at a serpin/trypsin molar ratio of 5 (Figure 2.3). These results indicate that the interaction between E191 of the serpin and trypsin is very critical. The mutation (E191A) may have an effect on the stabilization of Michaelis complex formation, following inhibition processes such as acylation, cleaved loop insertion. The

R192A mutation might affect the stability of the helix F1 and weaken the interaction.

2.4 Conclusion

Based on results presented here and the published observations of other serpins, we conclude that the P4’, P4, and hinge region residues of the reactive center loop of serpins, as well as some other interface residues (Glu191 and Arg192), are important in the inhibition mechanism of the serpin. The mutations near the hinge region affect the insertion rate of the RCL and the mutants therefore become substrates. The I350A mutant at P4 position also acts as a substrate toward the trypsin, maybe, due to enhanced deacylation rate. Mutations at P4’ (L357A) probably affect Michaelis complex formation.

The mutated serpins (E191A or R192A) at helix F1 have significant effects on the overall inhibitory activity of the serpin. Taken together, the results demonstrated the importance of the individual serpin residues involved in interactions at the interface for inhibiting the trypsin.

CHAPTER THREE MATERIALS AND METHODS FOR THE STUDY OF MUTANT SERPINS AND THEIR PROTEASE COMPLEXES

3.1 Introduction

Since the first structure of serpin (α1-AT) has been solved in 1984 (Loebermann,

Tokuoka et al. 1984) more than 80 structures have been deposited in protein data bank over several decades. Among crystals structures only few complexes between serpin and protease have been determined (Huntington, Read et al. 2000; Ye 2001; Dementiev,

Petitou et al. 2004; Li, Johnson et al. 2004; Dementiev, Dobo et al. 2006; Johnson,

Langdown et al. 2006). This shows how it is difficult to obtain crystals of protease- serpin.

In earlier work the Goldsmith Laboratory, the structure of a Michaelis trypsin- serpin complex was determined. This study made use of serpin-1B, which is normally a chymotrypsin inhibitor. But serpin-1B was mutated at the P1 residue (A353K), which becomes a trypsin inhibitor. Since Serpin-1B is well expressed, stable, and not glycosylated, it is excellent model system for exploring serpin inhibition mechanism by

X-ray crystallographic method. The detailed discussion of serpin-1B and trypsin S195A was made in section 3.1.1 and 3.1.2.

The first structure of the Michaelis complex, initial encounter non-covalent form, of serpin-1B (A353K) with trypsin S195A was presented (Ye, Cech et al. 2001) revealed that the interface made by serpin-protease was extensive and the residues P4, P1’ involved unique interaction with protease unlike other low molecular weight serine protease inhibitor (lock and key type inhibitors). Also, mutational analysis of these

63 residues for serpin activity proved that the importance of these residues in complex

(unpublished data, Kanost’s lab in Kansas State University). These structural and kinetic analyses led me to focus on the project, which could elucidate the role of the residues on the reactive center loop involving unique interaction with protease in serpin inhibitory mechanism. Thus, mutation was introduced into P4 and P1’ positions of serpin. X-ray crystallography method was applied for investigating the mutant Michaelis complex study serpin-1B (A353K, I350A) with trypsinS195A, serpin-1B (A353K, S354A) with trypsin S195A).

In order to determine crystal structures of mutant Michaelis complexes rat anionic trypsin S195A and serpin-1B (A353K, I350A) and serpin-1B (A353K, S354A) were expressed and purified separately. The mutant complex was obtained by cocrystallization method.

3.1.1 Rat Anionic Trypsin S195A

Previous studies showed that the S195A mutant trypsin is inactive showing no detectable peptide or protein substrate hydrolysis during more than 30 days (Carter and

Wells 1988). From the protease binding studies, it also demonstrated that S195A trypsin mutation did not affect the binding determinants of the enzymes (Bergstrom, Coombs et al. 2003). More importantly, the mutant structure of trypsin S195A showed that there are no significant structural changes compared to wild type of rat anionic trypsin (Pasternak,

Ringe et al. 1999). Especially, the catalytic triad residues, including Asp189 residue at the bottom of the specific pocket did not change in comparison to the wild type structure.

Therefore, the S195A trypsin mutant is an excellent mimic of the wild type for structural studies of the Michaelis formation in serpin inhibitory mechanism.

3.1.2 Manduca Sexta serpin-1B

Mandua Sexta serpin-1B is called alaserpin and it inhibits elastase and chymotrypsin.

It is closely related in sequence and structure to chymotrypsin inhibitor serpin 1K.

Compared to other serpins from M. Sexta, serpin-1B shows the broadest selectivity and the mutation of P1 residue to lysine alters Serpin-1B to trypsin inhibitor. This serpin-1B is also well expressed and not glycosylated. Since trypsin has good properties of expression and stability compared to chymotrypsin, mutant serpin-1B (P1, A353K), trypsin inhibitor, would be proper choice for the complex study by X-ray crystallography.

In other words, trypsin S195A and serpin-1B (A353K) were good model system to understand serpin inhibitory mechanism. Therefore, two proteins were utilized for the first structural study of the Michaelis complex.

The rational of my project came from the structural information of the Michaelis complex: serpin-1B (A343K) and rat anionic trypsinS195A. From this study (Ye 2001) I found two important residues (P4 Isoleucine and P1’Serine) deserving worthy of investigation by X-ray crystallography (see chapter 3 for rationale for the choice of these two mutants). Thus, two mutant constructs (P4 or P1’ position) had been made by point mutation and inserted into an expression vector, H6pQE-60.

Two constructs are serpin-1B (A343K, I350A) and serpin-1B (A343K, S354A). For simplicity of the serpin nomenclature, I denote serpin-1B (A343K) as just serpin-1B,

65 serpin-1B (A343K, I350A) as serpin-1B(I350A), serpin-1B (A343K, S354A) as serpin-

1B(S354A), and Michaelis complex of serpin-1B with trypsin as the wild type complex.

3.2 Materials and Methods

3.2.1 Expression and Purification of rat anionic trypsin

The rat anionic trypsinogen was expressed in saccharomyces cerevisiae. The

vector pYT containing the trypsinogen gene was constructed to an α-factor leader

sequence under the control of ADH/GAPDH promoter that directs secretion into the

growth media. This vector was obtained from Dr. L. Hedstrom, Brandeis University.

A. Yeast expression

1. Yeast strain DLM101α with pYT was transformed using EZ transformation kit (Bio-

101). The competent yeast was stored in -80 freezer. (Note: leu = leucine, ura = uracil)

1) 1 tube (~ 50 μl cells) was thawed on ice for each pYT construct

2) 2~1 μg plasmid in less than 5 μl was added in the cells

3) 500 μl of EZ3 solution (bio-101, EZ-yeast transformation) was added and mixed

vigorously

4) The transformation solution was incubated at 30 °C for 45 min and mixed 2~3

times during incubation.

5) 50 to 150 μl of transformed yeast was plated on SC-ura plates and incubated at

30 °C for 2 to 4 days

2. The transformant from the previous procedure was streaked on SC-leu with 8% glucose plates and grow at 30 °C

3. The transformant was inoculated into SC-ura –leu with 8% glucose broth.

Started with a 1ml culture, then diluted to a larger culture

Cells can be frozen at this step by adding sterile glycerol to a final concentration of 15%. Dispense in 1ml aliquots in freezer vials and store at -80.

Preparation of media for trypsin purification

The following SC and YPD media had to be prepared before whole process of purification of trypsin S195A.

(a) Preparation of 1L of SC

6.6 g yeast nitrogen base without amino acids

1.3 g –ura –leu dropout media

20 g agar (for plate)

600 ml water was added and then autoclaved

10 ml Leu (360 mg/ 100 ml stock) for SC –ura medium or 10 ml ura (240 mg/

100 ml stock) for SC –ura medium (Skip this step is medium is SC –leu –ura) had to be added but this step could be skipped for making SC-leu, -ura medium.

400 ml 20% glucose for an 8% final glucose concentration was added for total volume of 1L

(b) Preparation of 1L of YPD

10g yeast extract

20g peptone

20g agar if doing plates

All components was mixed in 930 ml water and then autoclaved.

Finally, 70ml of 20 % glucose was added into medium for 1.5% final glucose

concentration.

Preparation of –ura –leu dropout media: mix following amino acids.

1g adenine sulfate, 1g tryptophan, 1g histidine HCl, 1g arginine-HCl, 1g methionine, 1.5g tyrosine, 1.5g isoleucine, 1.5g lysine-HCl, 2.5g phenylalanine, 5.0g glutamic acid, 5.0g aspartic acid, 7.5g valine, 10g threonine, 18.75g serine

B. Purification of Trypsin from yeast

1. Two 50 ml yeast cultures in SC medium (-leu, -ura) with 8% glucose was

incubated in 250 mL flasks, at 30 °C for about 2 days from a colony or about

1day from 1 mL frozen stock.

2. Fully-grown cultures were diluted 1:20 into 2L of YPD medium with

1.5%glucose (four 500 mL cultures, each in a 2 L flask). These cultures were

incubated for another three days.

3. The cultures were centrifuged in IEC centrifuge at 4000 rpm for 20 minute at 4

°C to remove pellet (cells). The plastic Nalgene-type bottles was used for this

centrifugation step. The supernatants were pooled and 75 mL 1N HCl was added

slowly with stirring in ~30 minute. Another Centrifuge after adding HCl was

performed at 4000 for 20 minute to remove brown precipitate. If not translucent

after 20 minute, centrifuge again.

4. Supernatant was loaded onto an equilibrated Toyopearl 650M column with

peristaltic pump setting 15-20 (this is about 10 ml/min). The column was washed

with Buffer A (100 mm glacial acetic acid, 2mm sodium acetate). The

trypsinogen was eluted with gradient Buffer A to 200 mM Tris, pH8.0, 400 ml

total, at approximately 3 ml/min.

Preparation of Toyopearl 650M column

Column should be stored in 1M NaCl.

Equilibrate in Buffer A (100mM glacial acetic acid, 2 mM sodium acetate

After run, wash with 5 M NaCl, 0.5 M NaOH, then 1m NaCl

5. Trypsinogen-containing fractions were identified by checking absorption at 280

nm, and by using SDS-PAGE. This could done by running every third fraction

or so in a lane of the gel.

6. Fractions that contain trypsin were collected and pooled. The pooled protein

solution was concentrated using a centricon to 15 ml and at the same time, it was

exchanged the buffer with 50 mM Tris pH 6.2, 10mM CaCl, and 0.1% Triton X-

100 (a detergent that stabilizes the enterokinase and keeps it dissolved in aqueous

solution). 1μl enterokinase was added in concentrated solution. In order to

ensure activation of enterokinase, the protein solution was incubated at room

temperature. The level of activation from trypsinogen to trypsin was checked by

SDS-PAGE, and more enterokinase was added as needed.

7. After activation was complete, Activated trypsin was loaded on a SBTI-separose

column (Soybean trypsin Inhibitor) equilibrated in 50 mM Tris, pH 6.5. The

column was washed with 50mm Tris pH 6.5, then 50mm Tris pH 6.5 with 0.5 M

NaCl, then trypsin was eluted with 0.1M formic acid. The trypsin containing

fractions were checked by SDS-PAGE.

8. The trypsin containing fractions were pooled and dialyzed against 1mM HCl

(this protonates the His residue in the catalytic triad and prevents the protease

from reacting with itself).

3.2.2 Expression of M. Sexta serpin

The protein was expressed in BL21 (DE3 pLysS) cell. The six histidine tagged pQE60 plasmid containing different mutants (serpin-1B/I350A andserpin-1B/S354A) were transformed to BL21 (DE3 pLysS) cells. These serpin constructs were obtained from Michael Kanost Lab at Kansas State University). The transformation starts with mixing 1μl of DNA and 50μl of competent BL21cells. Then, the mixed cells with DNA were incubated on ice for 10 min, followed by heat shock at 42°C for 40 seconds and then back on ice for 5 minutes. After adding 50μl of SOC medium to the cells, the cells were incubated for about an hour at 37°C with shaking. The grown cells were plated on antibacterial agar plate containing 50 μg/ml carbenicillin and incubated at 37°C for overnight. A single colony from the plate was inoculated in 100ml of LB medium containing 100 μg/ml ampicillin and then incubated for overnight at 37°C. The overnight culture was inoculated to 1L of LB medium containing 100 μg/ml ampicillin and grown

at 37°C. When OD600 had reached around 0.8, the expression of the protein was induced by adding isopropyl-β-D-thiogalactopyranoside (IPTG) to a final concentration of 1mM

for an additional 4hr. The cells were pelleted by centrifugation at 4000 rpm for 30 min.

Cell pellets were resuspended in pre-cooled lysis buffer. The resuspended cell pellets

were frozen in liquid nitrogen and kept in the freezer at -80°C.

3.2.3 Purification of the mutant serpins

A. Preparation of the Cell-free Extract (from E.coli)

1. The frozen cells were thawed on ice

2. The cells were broken mechanically by the cell suspension twice through a

Franch Press at 15000~18000 psi at 4 °C. When cells were completely broken,

the suspension was changed. This could be indication for level of broken cell.

3. Cell pellet was remove by centrifugation for 30 min at 36000 rpm in a Beckman

Ti-45 rotor under 4 °C.

4. The supernatants were filtered by 0.45μl filtered syringe.

5. The cleared extract (supernatants) was transferred to the prechilled beaker for

next purification step.

B. Immobilized Metal Affinity Chromatography

1. The cleared cell extract was load to a fast-chelating sepharose column

(Amersham Pharmacia) by using a peristaltic pump (Amersham Biociences).

The column was prepared with 0.1 mM NiSO and pre-equilibrated with 5~10

column volume of Ni-binding buffer.

2. After loading the sample (the cleared cell extract), the column was washed by 10

column volume of washing buffer in order to remove non-specific bound proteins.

3. The target protein was eluted with 250 mM imidazole-containing Ni-elution as a

step gradient.

C. Anion Exchange Chromatography

1. The dialyzed protein solution was cleared by 0.22 µm filter syringe.

2. The filtered protein sample was load onto a MonoQ anion-exchange column

(Pharmacia).

3. Following the washing step with the 3 column-volume of MonoQ A buffer, the

protein sample was applied to step gradient for elution of target protein sample.

D. Size-Exclusion Chromatography

1. The eluted protein sample from anion-exchange chromatography was pooled and

concentrated up to a total volume of 2 ml.

2. Following the filtration of sample protein by 0.22 µm filter syringe, the filtrate

was loaded onto Superdex-75 or 200 size-exclusion column. The eluted sample

was pooled and collected. The protein sample for crystallization was usually

concentrated up to ~ 12 mg/ml.

Table 3-1: Summary of reagents used for the purification of

serpin-1B/I350A and serpin-1B/S354A

Ni-NTA Lysis buffer 50mM NaPO4 pH8.0

200 mM NaCl

Ni-NTA wash buffer 50mM NaPO4 pH7.0

200 mM NaCl

2 mM β-mercaptoethanol

Ni-NTA elution buffer 50mM NaPO4 pH7.0

200 mM NaCl

250 mM Imidazole

2 mM β-mercaptoethanol

MonoQ wash buffer 20 mM Tris-HCl Ph 7.6

100 mM NaCl

1mM DTT

MonoQ elution buffer 20 mM Tris-HCl Ph 7.6

1M NaCl

1mM DTT

Gel filtration buffer 20 mM HEPES, pH 7.0

100 mM NaCl

1 mM DTT

3.2.4 Crystallization

The hanging drop of vapor diffusion technique was used for crystallization. In

this setup, 1μl of the protein was mixed with an equal volume of the crystallizing solution

and placed on a siliconized glass cover slip. The cover slip is inverted and sealed into

place over a reservoir containing 200μl of the crystallization solution. The equilibrium

can be reached by diffusion derived from difference in concentration between the drop

and the reservoir.

In order to perform the cocrystallization trypsin S195 A and mutant serpin-1B

were concentrated separately and mixed with 1:1 molar ratio before crystallization setup.

A. Seeding

One method for improving crystal quality is seeding. A seed acts like a template

where more molecules can assemble, and given the proper conditions (time, environment,

etc), the seed can grow into a crystal. Seed-grown crystals can avoid some problems

brought by nucleated crystals. This is because spontaneous nucleation requires a higher

degree of supersaturation, which promotes other aggregation. If supersaturation is

reduced, nucleation cannot occur; thus, crystals may not appear even under good growing

condition. Energetically, seeding is more favorable way to obtain the crystal than to

create a new nucleus. Henceforth, seeding can be used to obtain crystals under

conditions, which may be significantly different from those where the original seed was

obtained.

B. Streak seeding

Before streak seeding was performed, pre-equilibrated drops of crystallization conditions were prepared (usually one day before streak seeding). A crystal was gently touched by the whisker and then, one pre-equilibrated drop was streaked by the whisker.

The seed remain attached to the whisker and are transferred to a new drop. Usually the

whisker is drawn in a straight line across it. Enough seeds remain on the whisker to streak (seed) 3-6 drops.

C. Preparing seed stock

Seeding stock was made by crushing three or four crystals (depending on the size

of crystal) in a 1ml glass tissue homogenizer with the addition of a stabilizing solution for

preventing the seeds from dissolving. Further solution was added in the sides of

homogenizer for washing the crushed seeds. The seed solution was transferred into a test

tube for storage of seed stock. About 1ul of seed stock was diluted in 1ml of stabilizing

solution and mixed. This seed solution was 10-1 diluted. Further 10-1 to 10-6 dilution was made for micro-seeding method. The temperature was kept constant in order not to dissolve the seeds in solution.

D. Micro seeding

About one fourth to one third volume of one of the serially diluted seeding stock was introduced to pre-equilibrated drops. Usually, the pre-equilibrated drops of crystallization conditions were prepared the day before micro seeding. After adding seed coverslips was resealed in original matrix wells.

3.3 Results and discussion

3.3.1 Purification of rat anionic trypsin

The trypsinogen was expressed in yeast and activated to trypsin later on purification step. Toyopearl SP 650 M and soybean trypsin inhibitor sepharose conjugate

column were utilized for purification of trypsin. Secreted trypsinogen, under control of

α-factor in media, was pooled and clarified by adding acid HCl. The clarified

supernatant was loaded onto Toyopearl SP 650 M column (Toso Haas). Recombinant

trypsinogen was eluted with a gradient of equilibration buffer 200mM Tri, pH 8.0. The

eluted fractions, which contained trypsinogen, were identified by SDS-PAGE analysis

shown in Figure 3.1. The eleven fractions from SDS-PAGE analysis were concentrated

and dialyzed in 50 mM Tris, 10mM CaCl2, 0.1% Triton X-100 to final volume of 10 mL.

The final concentration of 50 nM of enterokinase (Biozyme, CA) was added to activate

the purified and concentrated trypsinogen. The stage of activation was monitored over

time by SDS-PAGE until the reaction was fully completed. Figure 3.2 shows the process

of this activation reaction. Once activation of trypsinogen had finished, the activity of

enterkinase was blocked by the addition of trypsin inhibitor AEBSF (Sigma) to a final

concentration of 500 nM. Trypsin (S195A) was loaded onto soybean trypsin inhibitor

sepharose conjugate column (sigma). After washing, the purified trypsin was eluted with

0.1M formic acid of step gradient. By checking in SDS-PAGE eleven eluted fraction

were pooled (Figure 3.3). The pure trypsin was concentrated for co-crystallized with

serpin.

MWM 37 32 40 41 43 44 46 47 49

kDa 45 25

53 52 50 MWM

Figure 3.1 SDS-PAGE analysis for the eluted fractions (contained trypsinogen) from Toyopearl column. The numbers on top of the gel are fraction numbers and MWM represents molecular weight marker.

MWM 10min O/N kDa

45 25

15 10

Figure 3.2 The activation of trypsinogen by enterokinase

The eluted trypsinogen solution was concentrated and dialyzed in 50 mM Tris, 10mM CaCl2, 0.1% Triton X-100 to final volume of 10 mL. The final concentration of 50 nM of enterokinase (Biozyme, CA) was added to the purified and concentrated trypsinogen for activation. The stage of activation was monitored over time by SDS-PAGE until the reaction was fully completed. MWM represents molecular weight marker. The third lane shows the activation for overnight and second lane shows 10min. activation.

MWM 4 9 10 11 12 13 14 15 16 kDa 45 25 20

15 10

P MWM 26 22 21 20 19 18 17

Figure 3.3 Purification of trypsin by SBTI column The figure of the SDS-PAGE shows the elution factions from soybean trypsin inhibitor sepharose conjugate column. The purified trypsin was eluted with 0.1M formic acid of step gradient from SBTI column. The fractions containing pure trypsin S195A were identified by SDS-PAGE analysis. The numbers on top of the gel represents the eluted fractions numbers and MWM represents molecular weight marker.

3.3.2 Purification of serpin-1B/I350A and serpin-1B/S354A

The serpin-1B/I350A and serpin-1B/S354A were expressed in E.Coli strain

BL21 (DE3) and purified by Ni-NTA agarose (Qiagen) and Mono Q (General Electric), gel filtration (General Electric) columns.

Two mutant serpins show same behavior on each chromatography during purification. Figure 3.4 and Figure 3.6 shows a typical chromatograph from an Ni-NTA column. All peak fractions were pooled and the purity of protein solution checked by

SDS-PhastGel (General Electric). A chromatograph of Mono Q and SDS-PhastGel

(General Electric) shows that the mutant serpin was eluted in flowthrough factions, not in the gradient made by elution buffer. Figure 3.5 and Figure 3.7 clearly prove that most

pure proteins came off from the flowthrough in Mono Q column. For the crystallization

purpose, gel filtration was utilized for checking the homogeneity of protein such as

monomer vs dimmer and removing impurity by size exclusion. The chromatography and

SDS-PhastGels of this step were presented in Figure 3.5 and Figure 3.8.

B 6 5 4 3 2 M M P B pellet M

kDa

31 20 15 6

Figure 3.4 Purification of Serpin-1B/S354A by Ni affinity chromatography Figure A shows a chromatograph of this purification step, which has an elution peak coming from the column as result of a linear gradient from 0 to 250 mM imidazole after a complete wash using 15 column volume wash buffer. Figure B shows an SDS-PAGE of samples taken from the corresponding Ni column fractions. The molecular weight markers are shown in MWM. Pellet represents the solution prepared by mixing water with pellet after centrifugation for removing cell pellet. This was done for checking the solubility of serpin-1B/S354A. ‘B’ indicates the protein solution before loading Ni affinity column whereas ‘P’ indicates the flowthrough solution from the Ni affinity column. The left panel of figure B shows real elution fractions from Ni column. Lane 3 in the bottom of left figure B is matching the highest peak in the Ni column. The numbers after 3 are the fractions after the highest peak fractions. The 6xHis tagged Serpin-1B/S354A is indicated with an arrow.

Figure 3.5 Purification of Serpin-1B/I350A by MonoQ anionic exchange and Superdex75 chromatography. Figure A shows a MonoQ chromatograph for the purification stated above. Most of the pure serpin-1B/S354A was eluted through flowthrough in MonoQ column. The eluted fractions from the linear gradient of 1M NaCl included non-targeted proteins. Figure B shows a Superdex75 chromatogram.

_UV Conc Fractions mAU

3000

2500

2000

1500

1000

500

-500 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Waste 100 50 0 50 ml

Figure 3.6 Purification of Serpin-1B/I350A by Ni affinity chromatography The Chromatograph shows an elution peak coming off from the column as a result of a step gradient from 0 to 250 mM imidazole, after a complete wash using 20 column volume of wash buffer.

611 20 2002:1_UV Concentration _Fractions mA

1500

1000

500

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Waste 0 10 0 20 0 30 0 40 0 ml

M 2 3 4 5 6 7 kDa 198 110 73 45 31 20 15 6

Figure 3.7 Purification of Serpin-1B/I350A by MonoQ anion exchange chromatography. Figure A shows a MonoQ chromatogram. Figure B shows a SDS- PhastGel of samples (lane 2-7) taken from the flowthrough fractions from MonoQ. The 1st lane shows a molecular weight marker. An arrow indicates 6xHis tagged serpin- 1B/I350A.

kDa M 2 3 4 5 6 198 110 73 45 31 20 15 6

Figure 3.8 Purification of Serpin-1B/I350A by Superdex75 gel filtration chromatography. Figure A shows a Superdex75 chromatogram. Figure B shows a SDS-PhastGel of samples (lane 2-5) taken from the corresponding Superdex75 fractions. The molecular weight marker are shown in lane M and the 6x His tagged serpin- 1B/I350A is indicated with an arrow. Lane 6 is concentrated trypsin S195A.

3.3.3 Cocrystallization

Before cocrystallization was performed, a native gel was utilized for checking complex formation (Figure 3.9). The mutant (P1’ S354A) of the Manduca Sexta Serpin

1B (A353K) was co-crystallized with the rat anionic trypsin S195A. Since the crystallization condition for the Michaelis complex of serpin 1B (A353K) and trypsin was 20 % (w/v) PEG 3350, 1.0 M ammonium chloride, 4% (v/v) isopropanol and 0.2 M

Na/K phosphate, pH 6.0–6.5, the first trial was to expand this complex crystallization condition. The crystals were grown at almost every expanded condition. The best crystals were obtained in the condition of 22 % (w/v) PEG 3350, 1.0 M ammonium chloride, 4% (v/v) isopropanol and 0.2 M Na/K phosphate, pH 6.4 and grew to 0.05 mm

0.05 mm 1.0 mm.

The mutant (P4 I350A) of the Manduca Sexta Serpin 1B (A353K) was co- crystallized with the rat anionic trypsin S195A. The expanded crystal conditions of the complex, serpin-trypsin, did not produce any crystals. Thus, the commercially available spare matrix kits were screened for the cocrystallization of the Manduca Sexta Serpin 1B

(A353K) –trypsin. The initial crystal appeared as cluster in the conditions of 30%(w/v)

PEG4000, 0.2M Lithium sulfate monohydrate, 0.1M Tri hydrochloride pH 8.5. Thus, crystallization condition of mutant serpin (P4 I->A)-trypsin was subject to further modification by varying protein concentration, precipitant species, buffer pH, diffusion rate and additives in order to produce single crystal to freeze and collect data. The seeding technique rendered single crystals suitable for X-ray data collection. Among crystals produced by seeding, the best quality crystal was obtained in 25%(w/v)

PEG4000, 0.2M Lithium sulfate monohydrate, 0.1M Tri hydrochloride pH8-9. Single crystals were transferred to 30% (v/v) glycerol, prefrozen in liquid propane and diffracted around 3.5Å at a temperature around 100K on a rotating anode X-ray generator with an

Raxis IV image plate detector (Rigaku/MSC) in the lab. However, this crystal diffracted to 2.7 Å in the advanced Photon Source (APS) at Argonne using synchrotron radiation (

= 0.9794Å) and a charge-coupled device (CCD) detector APS1 on the 19 BM beam.

Figure 3.9 Checking for complex formation in Native gel. S_P1’ represents complex formation between serpin-1B/S354A and trypsinS195A and S_P4 indicates complex formation between serpin-1B/I350A and trypsinS195A. S is a serpin-1B and T is a trypsin S195A. different place in complex, serpin, trypsin proves that the complex between two different molecules has been formed.

Figure 3.10 Crystals of Serpin-1B/S354A with trypsin S195A. The crystals were found in expanded crystal condition of the wild type Michaelis complex (serpin-1B with trypsin S195A). The best expanded condition was 22 % (w/v) PEG 3350, 1.0 M ammonium chloride, 4% (v/v) isopropanol and 0.2 M Na/K phosphate, pH 6.4.

Figure 3.11 Crystals of serpin1B/I350A with trypsin S195A These poor quality crystals (such as crystal bundles and micro crystals) were found in condition #17 from crystal screen I (Hampton Research), containing 30%(w/v) PEG4000, 0.2M Lithium sulfate monohydrate, 0.1M Tri hydrochloride pH8.5.

Figure 3.12 Improved quality crystals of Serpin-1B/I350A with trypsin S195A from micro-seeding method in expanded initial crystal condition. The final improved condition for best diffracted crystals was identified as 25 %(w/v) PEG4000, 4% (v/v) isopropanol, 0.1M Lithium sulfate monohydrate, 0.1M Tri hydrochloride pH8.0. Initially, the cluster crystals were produced in #17 of crystal screen I (Hampton research) condition containing 30%(w/v) PEG4000, 0.2M Lithium sulfate monohydrate, 0.1M Tri hydrochloride pH 8.5. Co-crystallization condition of mutant serpin (P4 I->A)-trypsin was improved by varying protein concentration, precipitant species, buffer pH, diffusion rate and additives in order to produce single crystal to freeze and collect data. The micro- seeding technique produced single crystals having well diffracting quality. The single crystals were transferred to 30% (v/v) glycerol, prefrozen in liquid propane

CHAPTER FOUR STRUCTURE DETERMINATION

X-ray crystallographic techniques have been used in this study to address how the hydrolysis is arrested in serpin inhibitory mechanism by solving two mutant

Michaelis structures; Serpin-1B (I350A)/trypsin and Serpin (S354A)/trypsin complexes.

The crystals of the mutant serpin-protease complexes were improved and optimized using various techniques as discussed in Chapter Three. However, the crystals did not diffract well using rotating anode sources. Therefore, data sets were collected using the Advanced Photon Source (APS) at Argonne, Illinois.

Collected X-ray data was directly solved by using the wild type complex structure published by Sheng Ye and et al (2001). This was accomplished without utilizing molecular replacement, since the wild type and mutant complexes shared the same space group and nearly similar cell dimensions.

This chapter deals with the step-by-step protocol of the procedure used in the structure determination. In addition, both the wild type, and mutant crystals are twinned.

To describe how this challenge was overcome, a brief section on the background of twinning and the strategy of identifying and solving structures of twinned crystals, is included.

4.1 Data Collection and Data Reduction

Diffraction peaks or Bragg reflections of the crystals were screened, with an

20~30 minutes exposure time per 1° rotation, from a Rigaku rotating anode X-ray

92 generator (Rigaku/MSC), which utilizing a RAXIS IV phosphor image plate and monochromatic Cu Kα radiation (λ=1.54 Å). This data was used to determine the crystal symmetry and unit-cell parameters. The resolution of ~3 Å was likely to introduce errors on modeling wrong rotamers of several long, thin side chains (lys, glu, gln, etc) and small side chains (ser, val, thr, etc).

High-resolution data collection was collected using synchrotron radiation ( =

0.9794Å) and a charge-coupled device (CCD) detector APS1 on the 19 BM beam at the

APS (Advanced Photon Source) in Argonne national Laboratory. The exposure time was

30 sec per frame and 1o oscillation rate per frame was used. 180 degrees of rotation (180 frames) were collected on the crystal used for structure determination. HKL2000 was used at the argon synchrotron to examine diffraction quality, perform initial indexing and identify data collection strategy based on the crystal orientation and initial point group determination (Otwinowski, Minor et al. 1997). The DENZO and SCALEPACK components of the HKL2000 program were used for data reduction involving autoindexing of all the Bragg reflections, integration of diffraction intensities, data correction, scaling and post refinement. Primitive hexagonal among 14 possible Bravais lattices was chosen with minimal distortion.

4.2 Identifyingn the Space group

The systematic absences for indicating correct space group were analyzed using the SCALEPACK module within HKL2000. The space group of the previously solved wild type complex was P3121, which represent the Laue groups –31m. However, the best

2 χ statistics (the goodness of fit) and Rmerge for merge data were provided in the space

2 group of P622 although the space group P3121 also gave good χ and Rmerge. In order to

check simple way for confirming the space group and also involving twinning,

Matthew’s coefficient (Vm) was calculated. This is one of ways identifying twinning

crystals if especially the space group appears as the hexagonal, or trigonal (the detail of

crystal twinning will follow this section).

Matthew’s coefficient (Vm) (Matthews 1997) was used to estimate the number of

molecules per unit cell and the solvent content of the crystal. Vm is the ratio of unit cell

volume to protein molecular weight and has units of Å3/Dalton.

unit cell volume Vm = M r ×Z

Where Mr is the molecular weight of the protein and Z is the number of protein monomers in the unit cell. Vm value for most protein in the range of 40-60 kDa is between 2.0 and 3.5 Å3/Dalton (Kantardjieff and Rupp 2003).

However, the Vm for the spacegroup of P622 given the best Rmerge was 1.06

Å3/Dalton. This is physically impossible because half molecule in an asymmetric unit cannot occur. Both complexes of the serpin-1B/I350A and the serpin-1B/S354A with trypsin has space group of p3121 where one molecule per asymmetric unit corresponds to

6 molecules per unit cell. The calculated Vm is 2.97 for serpin-1B/I350A with trypsin and 2.98 for serpin-1B/S354A with trypsin, which are perfectly within the usual range.

Therefore, the space group of P3121 was applied for structural determination and it really

turned out the mutant crystals were twinned like the wild type crystal structure (Ye, Cech

et al. 2001). The details of identification, estimation of twinning fraction and handing

structural studies of twinned crystals will be discussed in section 4.8.

4.3 Molecular Replacement

In order to determine the structure from collected x-ray data set I need to find the phase first after identifying the space group. The electron density is consists of two components comprising of diffraction intensities collected during native data collection and phase associated with the diffraction at each reciprocal lattice point. There are different approaches to determine the phase to solve the structure. They are heavy atom isomorphous replacement (Blundell and Johnson 1976), anomalous diffraction from selenomethionine molecule incorporation (Hendrickson, Horton et al. 1990) and by using known phase information from previously solved structures (Argos and Mathews 1975).

The last method is called molecular replacement. The structure factors are calculated from coordinates of the known structure, thereby leading to an initial solution for the electron density map that is the beginning point of model building and refinement.

Molecular replacement was a definite method to solve the phase for the mutant complex structures because the wild type of complex structure was available. Therefore, phases from the structure factors of the wild type serpin-protease complex (PDB 1K9O) was used as initial estimates of the phases for the mutant complexes as the target molecules.

Three different molecular replacement programs from CNS (Brunger, Adams et al. 1998), AmoRE (Navaza 2001), and PHASER (Storoni, McCoy et al. 2004) were tested to have a better starting model for model building and refinement. All of these programs gave high correlation coefficients for new mutant complexes (Data not shown).

After obtaining the phase information, the density maps were calculated for further refinement and model building.

However, starting model building computed by phase information from molecular replacement was not better than directly using the model of the wild type of the Michaelis complex (PDB code: 1K9O) for refinement and model building. Since the previously solved model of the wild type complex had the same space group with the mutant structures, direct refinement and modeling using the wild type structure became feasible for solving the mutant structures. Therefore, structural determinations of both mutant crystal complexes were directly solved using initial model of wild type complex

(1K9O). The protocol for rigid body refinement in CNS was used to refine this initial model. Then, iterative steps of model building, refinement and map calculation for final structural determination were carried out.

4.4 Model Refinement

Model refinement was performed using the CNS suite (Brunger, Krukowski et al.

1990; Rice and Brunger 1994; Adams, Pannu et al. 1997). Since these crystals were twinned it was essential to utilize the twin refinement programs in the CNS. First, it is very critical step to get an R-free set in which twin related reflections are picked. For this process, “make_CV_twin.inp” was utilized. After getting Rfree set, twin rigid body refinement followed by individual B-factor minimization and Powell minimization in

CNS were carried out. Refinement was stopped at the point where Rfree began to increase.

4.5 Map Calculation

The CNS program was used in the map calculation. Since crystals were twinned, it is very important to use maps generation program for twinned crystals. Hence,

“model_map_twin.inp” inputting the same true data and twin fraction as used in the refinement in order to compute the density map was used. If the maps were still not clear, the new maps were calculated by varying the twin fraction by about 1-2% either way to see if it clears them up a little. The varying twin fractions could be terminated up to the case of perfect twinning (twin fraction=0.5) in which the twin-related reflection intensities were averaged to simulate the case of perfect twinning (Yeates and Charles W.

Carter 1997) which rendered an easily interpretable density map (showing the flexible loops including the RCL).

4.6 Model building

The program O (Jones, Zou et al. 1991) was utilized for model building in structural determination. The process of model building is not the trivial work for twinned structure. It was essential to keep in mind the correct conformation and geometry of secondary or tertiary structure. Also, some of the basic concepts and knowledge about the protein structure (length between Cα and Cα, helix turn, distance of hydrogen bond, residue property) are very useful to conduct the model building. Also,

Molprobity (Davis, Murray et al. 2004) and PROCHECK (Laskowski, Moss et al. 1993) were utilized for checking the model.

Table 4.1 Data collection and refinement statistics

Details of data collection and structure determination for the Michaelis complex: serpin- 1B/I350A with Trypsin S195A

Space group P3121 Unit cell a=112.473 c=97.536 α=β=90° γ=120° Resolution 50-2.7 Observed reflection 340,358 Unique reflection 20,053 Completeness 97.8(91.2) Twinning fraction 0.43

Rmerge 0.124 (0.49)

Rcrystal (%) 15.1

Rfree (%) 23.9 Number of groups Protein atoms 4601 Water atoms 189 r.m.s.d in bond length (Å) 0.007672 r.m.s.d in bond angle (º) 1.88003 Average B factor (Å2) 51.7976

Values between parentheses are for the highest resolution shell.

Rmerge = Σ|(Ihkl) | / Σ(Ihkl), where Ihkl is the integrated intensity of a given reflection Rcryst = Σ|(Fobs) k|Fcalc|| / Σ(Fobs), where Fobs and Fcalc are the observed and calculated structure factors, respectively.

Table 4.2 Data collection and refinement statistics

Details of data collection and structure determination for the Michaelis complex: serpin- 1B/S354A with Trypsin S195A

Space group P3121 Unit cell a=112.952 c=96.982 α=β=90° γ=120 Resolution 500-2.3 Observed reflection 1,308,828 Unique reflection 41,922 Completeness 99%(97) Twinning fraction 0.47

Rmerge 0.12 (0.68)

Rcrystal (%) 0.176

Rfree (%) 0.219 Number of groups Protein atoms 2603 Water atoms 126 r.m.s.d in bond length (Å) 0.008141 r.m.s.d in bond angle (º) 1.51197 Average B factor (Å2) 20.5593

Values between parentheses are for the highest resolution shell.

4.8 Twinning in mutant complex structures

The crystals obtained from two mutant serpins (I350A and S354A) –trypsin complexes appeared as rod shape and diffracted well in synchrotron. The data sets collected to 2.7 and 2.3 Å, respectively for two mutant complexes. Crystallographic analysis and data reduction suggested those crystals belonged to the hexagonal space group P622. However, the spacegroup given the best Rmerge was physically impossible

because the asymmetric unit volume in the crystal is too small to hold the molecules.

The previously solved wild type complex was twinned and its space group was identified

as the trigonal P3121. This information led to test the twinning in crystals using detection_twinning program in CNS indicated that the crystals are twinned. The following section gives the background of twinning and strategy of identifying and solving twinned crystals.

4.8.1 Theoretical Background

Crystal twinning can describe for a special growth irregularity in the crystals and

can be formed when two separate crystals have same crystal lattice points in a

symmetrical manner (Parsons 2003). Macroscopic growth disorders in crystals are

common (e.g crystal splits or multiple crystals), but twinning refers to special cases

where some or all of the lattice directions in the separate fractions are parallel. This

phenomenon results in either partial or complete overlap between the crystal lattices of

the two fractions. Twinning is relatively common phenomenon in crystals of small

inorganic and some organic molecules. This was recognized early by crystallographers

100

(Friedel 1926). In crystallography, macromolecular twinning is not an unfamiliar obstacle for structure solving. There are several examples of well-documented twinning cases (Ban, Nissen et al. 2000).

Several different categories of twinning have been defined and classified according to the three-dimensional coincidence of separate lattice (Donnay and Donnay

1974; Giacovazzo 1992). QTLS (quasi-twin-related symmetry) is the case in which the overlap affects less than three dimensions. This QTLS can also be referred to as nonmerohedral or epitaxial twinning. QTLS crystal are can be distinguished by distinct interpenetrating reciprocal lattices in the three-dimensional diffraction pattern.

Occasionally, these twining crystals can also be identified by visual inspection under the microscope using a polarizing filter. If an indexing process can identify one lattice

exclusively, data on epitaxically twinned crystals can be collected and integrated as usual

(Yeates and Charles W. Carter 1997).

4.8.2 Merohedral Twinning

Twinning symmetry (TLS) can be the case where the lattices overlap in three

dimensions. These type can be further subdivided into pseudomerohedral twinning and

merohedral twinning. Pseudomerohedral twinning can occur at overlapping only approximately in three dimensions. However, merohedral twinning is an case where the lattice overlap exactly in three dimensions. This merohedral twinning either involves two crystal fractions (hemihedral), four factions (tetartohedral) or, for some centric

101 spacegroups upto eight fractions (ogdohedral). In the case of macromolecular crystals only hemihedral TLS has been reported.

The observed intensities cannot be the true crystallographic values even if the diffraction pattern of the hemihedral crystal seem to be normal. Therefore, they are the weighted sum of the intensities of two fractions, h1 and h2 that are related by the twinning operator but not by crystallographic symmetry (Yeates 1997).

Iobs ( h1 ) = αI( h1 ) + (1-α)I( h2 )

Iobs ( h2 ) = (1-α)I( h1 ) + αI( h2 )

Here α, the “twinning fraction,” represents the fractional volume of the specimen

occupied by domains in the second orientation. The case of α = 0 correspond to no

twinning. If the two fragments do not have equal diffracting volume (α < 0.5) we refer to the case as “partial twinning”. If the two fragments have equal or close to equal diffracting volumes (α ≈ 0.5) we call it as “perfect twinning”. If α ranges from one crystal to the next, data sets collected from different crystals could be different and irreconcilable. This situation can be diagnosed as twinning.

Partial twinning does not obscure the true symmetry in the crystal, as the apparent lattice symmetry is the same as the true Laue symmetry. In perfect twinning, however, lattice symmetry appears larger than the true Laue symmetry so the independent twin operator imposes an extra relationship in the crystal, which is not a part of the Laue symmetry. This is only possible for Laue groups capable of supporting a

102 higher-order symmetry (Table 4.3). Therefore, hemihedral or true merohedral twinning

(TLS) occurs usually in the space groups based on point symmetries 3, 4, 6 or cubic systems (Yeates 1997; Chandra, Acharya et al. 1999). However, pseudomerohedral twinning can occur in other point group. In this case, special unit cell geometries produce a lattice with pseudosymmetry that is higher than the actual point group symmetry (Table 4.3).

4.8.3 Recognizing crystal twinning

It is very important to recognize crystal twinning before actually solving the structure. Therefore, it is always important step to check each collected data set for

merohedral twinning. There are several ways for detecting characteristic warning signs

for twinning. Every twinning cannot fit into the cases presented here, but if one finds

several signs of twinning, the possibility of twinning should be given serious

consideration.

1) Packing density: If the asymmetric unit volume in the crystal is too small to hold

the molecules, it is likely that the space group has been wrongly assigned, which

is an indication for perfect merohedral twinning.

2) The lattice symmetry is higher than crystal (Laue) symmetry

3) The space group appears to be trigonal or hexagonal although it itself cannot be a

sign of twinning.

4) Although the data appear to be in order, the structure cannot be solved

103

5) Rmerge value for higher symmetry Laue group is only slightly higher than for the

lower symmetry Laue group.

6) The Patterson function is physically impossible. Without knowing true space

group the translation function could be problematic when the apparent symmetry

is P622. The true point symmetry for apparent P622 could be 6, 321, 312.

Which ones are perpendicular to crystallographic twofold axes cannot be

distinguished from which are perpendicular to the twinning axis by intensity

statistics of the various zones (Yeates and Charles W. Carter 1997).

In order to confirm the crystal twinning we need to perform an analysis of diffraction intensities.

1) Intensity statistics: Intensity statistics from a twinned crystal differs significantly

from untwined ones. Wilson showed that a single crystal the mean and higher

moments of centric and acentric intensities and amplitudes follow a predictable

pattern (Wilson 1949). Pairs of reflections related by the twinning operator have

observed intensities that share more similarity to each other than expected.

2) The second and higher moment of intensities and amplitudes also provide useful

indication. For perfect hemihedral twinning the ratio of /2 for acentric

reflections plotted against resolution should give expected values of 2.0 for

untwined data and 1.5 for perfect hemihedral twinning. Similarly, the ratio

between normalized structure factors /2 is expected to give 0.885 for

untwined and 0.75 for twinned data.

104

2.1

1.9

1.8

1.7 /^2 1.6

1.5

1.4 9 5.67 4.7 4.18 3.83 3.57 3.37 3.21 3.08 2.96 2.86 2.78 2.7 resolution ranges

Figure 4.1 Twinning test for serpin-1B/I350A-trypsinS195A complex The blue line is calculated for data between 9(50) and 2.7 Å from the complex shown in X-axis. The Y-axis shows a ratio of /2, where the values are indicators for twinning. The values of 2.0 is for untwined data and 1.5 is for perfect hemihedral twinning.

4.8.4 Estimation of crystal twinning fraction

In order to estimate the correct twinning fraction the correlation between twin-

related intensities can be used. Then, the data can be corrected by the twinning fraction.

The twinning fraction from intensity data collected from a hemihedrally twinned crystal

can be estimated by a relatively simple method developed by Yeates (Yeates and Charles

W. Carter 1997). Prerequisite however is knowledge about the operator relating the two

crystal fraction (twin law)

105

A parameter H for each pair of reflections related by the twin operator can be defined as :

I ()h − I ()h H = obs I obs 2 I obs ()hI + I obs ()h2

In the absence of errors, H should range between 0 and (1-2α). It is showing that

H is a function of α and the true crystallographic intensities. As the crystallographic intensities should follow Wilson statistics it is possible to derive the expected distribution for H as a function of α. For acentric reflections the cumulative distribution S (H) is linear in H:

H S()H = 1− 2α

α can be easily estimated by calculating the intensity distribution using series of

computer programs that are available, which can be used to identify crystal twinning and

calculate twinning fractions (CNS version1.1twinning program: detect_twinning and

twin_fraction).

106

Table 4.3 Hemihedral twin laws. Adapted from (Yeates and Charles W. Carter 1997)

True point group Space group Twinning operator Reflection related to hkl 3 P3, P31, 2 along a, b h, -k-k, -l P32, R3 a along a*, b* h+k, -k, -l 2 along c -h, -k, l 4 P4, P41, P42, P43, 2 along a, b, a*, b* h, -k, -l I4, I41 6 P6, P61, P65, P62, 2 along a, b, a*, b* h, -h-k, -l P64, P63 321 P321, P3121, 2 along c, a*, b* -h, -k, l P3221 312 P312, P3112, 2 along c, a, b -h, -k, l P3212 23 P23, F23, I23 2 along a, b, c h, -k, l P213, I213 622 P622 2 along a, b, c h,-h -k, l Pseudomerohedral twinning is possible in other point groups as well

4.8.5 Correcting partially twinned data

If the twinning fraction α, can be estimated accurately (not α is 0.5) the true

crystallographic intensities can be recovered from the measured values.

()()1−α I h −αI (h ) I()h = obs I obs 2 1 1− 2α

−αI ()h + (1−α )I (h ) I()h = obs I obs 2 2 1− 2α

This equation can solve the partial twinning problem. However, as α approaches

0.5 the equation becomes degenerate, which makes the crystallographic intensities not be extracted accurately. Although it is impossible to detwin the data when α is closer to 0.5,

107 it can be simulated the case of perfect twinning by simply averaging the twin-related reflection intensities. This leads to the more difficult problem of solving a perfect twinned structure, but at least the uncertainty in α no longer remains.

In the multiple crystals, much further complications arise in datasets collected.

Since the twinning fraction in case of hemihedral twinning may range from crystal to crystal (Valegard, van Scheltinga et al. 1998) it is very important to collect a complete dataset from a single crystal in order to estimate the twinning fraction a correctly (Ito,

Komiyama et al 1995).

4.9 Hemihedral twinning in the mutant complexes: Serpin-1B/I350A with trypsinS195A, Serpin-1B/S354A with trypsinS195A

Detailed analysis of the datasets from both mutant complexes (Serpin-1B/I350A- trypsinS195A, Serpin-1B/S354A-trypsinS195A) using the approaches discussed in section 4.8 revealed that the datasets were hemihedrally twinned with spacegroup P3121.

The twinning was estimated 0.43 for the complex of Serpin-1B/I350A-trypsinS195A and

0.47 for Serpin-1B/S354A-trypsinS195A. Since the mutant complex structures already

have a known phase from the wild type complex of serpin-1B with trypsinS195A, I could

focus on refinement and model building. However, with twinning involved in structure

determination, it was not easy to build and refine model using initial poor quality of

electron density map. It was also essential point that data should not be detwinned using

detwin program in CNS in any steps during solving structure. Therefore, without

detwining data, Rfree set was firstly determined by utilizing “make_CV_twin.inp”. Then, refinement was accomplished from twin rigid body refinement to individual B-factor

108 minimization and Powell minimization in CNS. The calculation of map was done by twin map calculation program in CNS. Because of twinning and low resolution (in case of serpin-1B/I350A-trypsinS195A complex) the process of model building is not trivial work. I had to check what is the correct conformation or geometry of secondary and tertiary structure. Each residue had to be modeled by poor quality map. After only correcting several residues (sometimes 10 residues or more), I checked whether those

residues was correctly placed. Rfree was always used for an indicator for interaction step of modeling and refinement. Also, PROCHECK and Molprobity were utilized for structure quality. More than 50 times iteration was performed to determine the final structure of the complex.

CHAPTER FIVE STRUCTURAL ANALYSIS OF THE MICHAELIS COMPLEXES: SERPIN-1B/I350A-TRYPSIN AND SERPIN1B/S354A-TRYPSIN

5.1 Introduction

In this chapter, two Michaelis complexes of serpin mutants with trypsin (serpin-

1B(I350A)/trypsin, serpin-1B(S354A)/trypsin) are used to analyze the serpin inhibitory mechanism. The structures of both mutant serpin-protease complexes are compared with the wild type of the serpin-protease complex (Ye, Cech et al. 2001).

First, the overall conformations of complexes are discussed in the comparison with wild-type complex. Then, detailed structural comparisons are made in the regions of the Ile350 (P4) and Ser (P1’) in the RCL (Reactive Center Loop), near the N-terminus of the reactive center loop and the topological switch point (Ala342, P12 and Ala 344,

P10). These regions of serpin were previously found to have unique interactions in wild- type complex. Also, the importance of these residues in complex formation was proved by mutational analysis of these positions (unpublished data, Kanost’s lab in Kansas State

University, discussed in Chapter 2).

Second, other conformational changes in the core domain of serpin are discussed in comparison with two structures of serpin-1K and wild type complex.

Third, acyl-enzyme intermediate structure of chymotrypsin-peptide inhibitor complex is analyzed to address how the deacylation is arrested.

Finally, I compare the mutant structures with an acylated chymotrypsin-inhibitor complex in understanding in serpin inhibitory mechanism. Throughout this Chapter 5, I

109

110 mainly focus on the essential features of the Michaelis complexes and the significance of conformational changes in the mutant Michaelis complexes. In addition, I address how the hydrolysis is suppressed at the covalent intermediate acyl enzyme stage and how unique interactions by residues at the protein interface help to arrest the hydrolysis reaction.

For simplicity of the serpin nomenclature, I denote serpin-1B (A343K) as just serpin-1B, serpin-1B (A353K, I350A) as serpin-1B(I350A), serpin-1B (A353K, S354A) as serpin-1B(S354A), and Michaelis complex of serpin-1B (A353K) with trypsin S195A as the wild-type complex. Michaelis complex of serpin-1B (A353K, I350A) with trypsin

S195A as serpin-1B(I350A)/trypsin. Michaelis complex of serpin-1B (A353K, S354A) with trypsin S195A as serpin-1B(S354A)/trypsin. Also, trypsin S195A is denoted as trypsin.

111

5.2 Structural analysis of Michaelis complex: Serpin-1B (I350A)/trypsin

5.2.1 Overall structure of the mutant Michaelis complex

In the mutant Michaelis complex, the trypsin lies at the top of the serpin tertiary core as in the wild type complex (Ye, Cech et al. 2001) close to the β-sheet B and C of the serpin molecule (Figure 5.1). Compared to the wild type complex (PDB entry

1K9O), the binding position of the trypsin is forward and to right as shown in Figure 5.2.

The root mean square (r.m.s) deviation of the two-serpin structures is 0.441 Å when superimposed based upon Cα atoms of all serpin residues. In contrast to the serpins, the overall structure of trypsin is almost identical with the trypsin of the Michaelis complex

(1K9O) (r.m.s.d.= 0.297).

The residues of the reactive center loop (RCL) in the mutant complex make fewer contacts with trypsin, compared to the wild-type complex, especially from residue

P4 through residue P14 (Figure 5.3). There are also few conformational changes in the serpin outside the RCL region including helices A, D, E and connecting loop of secondary structures compared to the wild type structure.

The contact area of a complex between the trypsin and the mutant serpin is 748

Å2 (Solvent-accessible interface). The number of trypsin residues at the interface (also,

interacting with serpin) is 11 and that of serpin residues is 7. Those residues interact with

each other through hydrogen bonding. This contact area of the mutant complex is

significantly reduced, compare to that of the wild-type complex, in which the contact area

is 915 Å. In the wild type complex, the number of residues involved in interface

interactions is 16 for trypsin and 7 for serpin. Therefore, total hydrogen bonds are less in

112 mutant structure (9 hydrogen bonds), compared to the wild type structure (11 hydrogen bonds) as shown in Table 5.1 and Table 5.2.

Figure 5.1 Overall structure of the Michaelis complex between serpin- 1B(I350A)/trypsin: Ribbon representation of the complex (Serpin-1B/I350A (Red reactive center loop and orange elsewhere), TrypsinS195A (dark yellow))

113

Figure 5.2 Superposition of the wild-type and mutant Michaelis complexes using trypsin as a reference: The wild type complex (cyan serpin-1B and dark cyan trypsin), the mutant complex (orange serpin-1B/I350A and dark yellow trypsin). Trypsin is superimposed.

114

Figure 5.3 Molecular surface representation of trypsin with the reactive center loop of serpin showing dots in interface between serpin and trypsin. (A) The wild complex (cyan sticks: reactive center loop) (PDB entry 1K9O) (B) The mutant Complex (Magenta sticks: reactive center loop) with charged surface trypsin. The positive charged surface is in blue and the negative charged surface is in red.

115

Table 5-1 Hydrogen bonds involving residues in the contact area of wild type complex (PDBcode: 1K9O)

Trypsin/S195A Dist.[Å] Serpin-1B E: Ser 147 2.78 I: Glu 191 [OE1] E: Gly 216 [N] 3.36 I: Val 351 [O] E: Gln 192 [NE2] 3.11 I: Pro 352 [O] E: Gly 193 [N] 2.86 I:Lys 353 [O] E: Ala 195 [N] 2.94 I: Lys 353 [O] E: Gln 192 [NE2] 3.05 I: Ser 354 [O] E: Ser 214 [O] 2.88 I: Lys 353 [N] E: Trp 215 [O] 3.33 I: Lys 353 [NZ] E: Ser 190 [OG] 3.22 I: Lys 353 [NZ] E: Phe 41 [O] 3.21 I: Leu 355 [N] E: Tyr 39 [OH] 3.14 I: Leu 357 [N]

Table 5-2 Hydrogen bonds involving residues in the contact area of Serpin- 1B(I350A)/trypsin

Trypsin/S195A Dist.[Å] Serpin-1B E: Gly 216 [N] 3.00 I: Val 351 [O] E: Gln 192 [NE2] 3.02 I: Pro 352 [O] E: Gly 193 [N] 2.56 I:Lys 353 [O] E: Ala 195 [N] 2.88 I: Lys 353 [O] E: Ser 214 [O] 2.67 I: Lys 353 [N] E: Ser 190 [OG] 2.90 I: Lys 353 [NZ] E: Ser190 [O] 2.56 I: Lys 353 [NZ] E: Phe 41 [O] 3.15 I: Leu 355 [N] E: Tyr 39 [OH] 3.14 I: Leu 357 [N]

116

5.2.2 Change in the Reactive center loop (RCL)

The structure of serpin-1B (I350A)/trypsin complex clearly shows that mutation of this residue caused dramatic conformational change within the RCL, especially region between Ile 350 (P4) through Ala 339 (P14) (Figure 5.4). Although the conformation of the active site between P3 and P4’ is similar to that of wild type Michaelis complex, other residues in RCL adopt different conformation. In the RCL region, the residues from P4

(Ile350) to P13 (Glu341) are partially disordered, with weak density for the backbone and side chain, especially at residues Ile 350 (P4) and Gln 346 (P8) (Figure 5.5). However, the electron densities of other residues are clearly visible (Figure 5.5). Compared to wild type Michaelis complex, the region between P4 and P10 appears to be loosened and slips away from the trypsin. The mutant serpin residue Ala350 (P4) lost hydrophobic interactions with the side chain of Trp215, Leu99 and the main chains of Lys97 and

Thr98 on trypsin (Figure 5.6). Different conformation in this region also results in disrupting the van der Waals interactions around Ala 341 (P12) through Ala 344 (P9) of serpin with Gly 219-Cys 220 and Ser 146-Val 149 located on the loop of trypsin (Figure

5.7). Therefore, the trypsin loops at the interface lost stabilizing interactions in the

Michaelis complex of serpin-1B(I350A)/trypsin (Figure 5.7). These changed interactions, in which hydrogen bonds and hydrophobic contact is lost, between the RCL and trypsin in the mutant complex result in less contact area between inhibitor and enzyme (Figure 5.2).

117

Figure 5.4 The comparison of RCL conformation between mutant and wild type complexes. A surface representation shows the trypsin molecule. The RCL of the wild type serpin-1B is shown in cyan and the RCL of the mutant serpin-1B(I350A) is in orange.

118

Figure 5.5 Electron density map (2Fo-Fc) of the RCL( P1’-P17) The residues from P1’ (Serine 354) to P17 (Glutamate 337) was shown as stick models. The contour level (σ) is 1.

119

Figure 5.6 Comparison of active sites in RCL (P4-P2’) between the wild type complex and the mutant complex. The wild type serpin-1B in the Michaelis complex is shown in cyan. The P4 (Ile 350) residue is shown as stick and line elsewhere. Orange ribbon diagram shows trypsin in the mutant Michaelis complex. Orange stick residues involve in interactions with the residues in the (pink sticks) RCL (pink)

120

. cyan represents. ypsin S195A. The mutant serpin-1B/I350A with trypsin mutant serpin-1B/I350A S195A. The ypsin S195A is in orange and the RCL is highlighted as pink. as pink. RCL is highlighted and the orange S195A is in the wild type Michaelis complex:the wild type serpin-1B with tr Figure 5.7 Stereo diagram showing interactions between the Figure 5.7 Stereo RCL and trypsinS195A

121

5.2.3 Change at the topological switch point

The residues from P16 to P10 in the reactive center loop are also called the proximal hinge. The topological switch point for strand insertion is at the P16 position

(Figure 5.8). Active serpin 1K and other active serpins have a network of hydrogen bonding interactions involving P15 (Gly339) and P14 (Ala340) that seem to stabilize the

RCL, preventing insertion of the RCL into sheet A (Li, Wang et al. 1999). These interactions do not change in the wild type complex. There are however noticeable

conformational differences at this hinge region in serpin-1B(I350A)/trypsin complex compared to the wild type complex.

In the Serpin-1B(I350A)/trypsin complex, the hydrogen bond between P14

(Ala340) and Trp184 O of the mutant serpin is lost (Figure 5.8). Another hydrogen bond

made by Ala340 to the body of serpin Lys 187 also no longer appears in the mutant

complex (Figure 5.8). The residue Lys 187NZ also lost hydrogen bond with the residue of Ala342 (P12). Glu 341 (P13) also changes position. On the other hand, a new ion- dipole interaction between Ser183 and Glu235 appears in serpin-1B(I350A). The mutation also gives rise to reduction of the interface area between the mutant serpin-

1B(I350A) and trypsin The contact area involves smaller numbers of both residues and hydrogen bonds, which probably destabilizes Michaelis complex.

122

involve

the RCL is shown in pink sticks. Orange sticks shown in pink the RCL is

ns at the topological switch point.

Figure 5.8 Stereo diagram showing the interactio is represented as ribbon diagram and Serpin-1B/I350A interactions with the RCL residue at the topological switch point. switch point. interactions with the RCL residue at topological

123

5.2.4 Changes near the Reactive center loop

Conformational differences observed in the short α-helix F1 (189-192) in serpin-

1B (I350A) shown in Figure 5.9 and Figure 5.10. This short α-helix F1 is no longer formed and also an ion-dipole interaction between Ser147 of trypsin and Glu191 of serpin-1B in wild type Michaelis complex is lost in serpin-1B (I350A)/trypsin. This fits well with the observation that the mutant E191A has no activity (Discussed in Chapter 2).

In addition, Arg 192 only makes one salt bridge with Glu278 instead of making a pair of salt bridges between them. These results suggest that the interaction between E191 of the serpin and trypsin is important and it might affect the stability of the Michaelis complex.

Figure 5.9 Conformational changes near the reactive center loop Serpin-1B(I350A) is represented in ribbon diagram and the RCL is in pink. Trypsin is shown in olive. The hydrogen bond between Arg 192 and Glu 278 is shown as green dash.

124

Figure 5.10 Changes near the reactive center loop. Important interaction between residue Glu 191 and residue Ser 147 of enzyme is highlighted as stick and magnified. The mutant Michaelis complex: serpin-1B(I350A) (orange)/trypsin(olive) and wild type Michaelis complex: serpin-1B(cyan)/trypsin(dark green) are presented as ribbon diagram. Hydrogen bonds in the wild type complex are shown in blue color and hydrogen bond in serpin-1B(I350A)trypsin is in orange.

125

5.2.5 Other Conformational changes (in serpin core domain)

In comparing our wild type complex with serpin-1B (I350A)/trypsin, we note conformational changes quite distant from the site of interaction. The significance of these changes is not clear. Although the core rigid domain of serpin undergoes almost no conformational change, there are still noticeable rearrangements occurring in helix D, E and strand 1A (s1A). This observation suggests there might be long-distance crosstalk between the RCL and Helix D, E.

In the serpin-1B (I350A)/trypsin complex, the helix E loses secondary structure

(Figure 5.13). This region in the mutant complex shows the most substantial conformational change that could be clearly modeled (Figure 5.11). The positions of the residues in helix E seem to move downward into the bottom of serpin (Figure 5.12). The residues of Ser128 and Ser122 lost hydrogen bonds with Glu56 located in helix B.

However, the strand 1A has a new hydrogen bond between Arg123/NE2 of helix E and the residue Glu129 in strand 1A (Figure 5.14). It is interesting that just one point mutation of P4 site on RCL can trigger conformational changes, ~ 60 Å away from the active site. The conformational linkage is unclear.

126

Figure 5.11 Electron density map (2Fo-Fc) of the helix E Serpin-1B(I350A) is represented in orange ribbon diagram and the residues on helix E are shown as sticks. Green mesh (contour level σ=1) shows the electron density around significantly changed positions of helix E residues compared to the wild type serpin-1B in the Michaelis complex.

127

Figure 5.12 Changes in the helix E of the serpin body. Serpin-1B(I350A) is represented in orange loop diagram and the residues on helix E are shown in stick. Cyan loop shows the wild type serpin from the Michaelis complex (PDB: 1K9O). Green loop represents the active serpin 1K (PDB: 1SEK). Important residues involved in conformational change around this region are highlighted as stick presentation.

128

B 1 2 3

Figure 5.13 Structural comparison in the helix E and D regions by superpositions. The ribbon diagram of figure A shows the superposition with an active serpin- 1K(uncomplexed, green: Figure B, 2) (PDB: 1SEK), Serpin-1B (Cyan: Figure B, 3) (PDB: 1K9O) from the wild type Michaelis complex and serpin-1B/I350A (orange: Figure B, 1) from mutant Michaelis complex. Figure B shows each structures of serpin separately.

129

Figure 5.14 Conformational change in Helix E of the serpin-1B/I350A Overall structural superposition is done with wild type (cyan)(PDB: 1K9O) and mutant Michaelis complexes(orange). Important interaction is shown by stick presentation of the residues in helix E and strand 1A. This region has most significant conformational changes in the body of serpin.

130

5.3 Structural analysis of Michaelis Complex: Serpin-1B (S354A)/trypsin

In contrast to the P4, mutation of P1 residue does not induce large conformational changes in serpin-1B (S354A)/trypsin. The overall structure of serpin-

1B(S354A)/trypsin is very similar to wild type Michaelis complex (Figure 5.15). There is slight rotation in the serpin-1B(S354A) takes place upon its binding to trypsin.

The interface area of the mutant complex is 886 Å. The size of interface area is that between of the wild type (915 Å) and of the serpin-1B(I350A)/trypsin (748 Å). The number of trypsin residues forming contacts in the serpin-1B(S354A) complex is fewer than in the wild type complex. Only 12 residues of trypsin interact with serpin through hydrogen bond interactions as compared to 16 residues in wild type. In three structures, the number of serpin residues at interface involving hydrogen bonds is also 7. Thus, total number of making hydrogen bonds at the interface (10 hydrogen bonds) is also that of between wild type (11 hydrogen bonds) and mutant complex (9 hydrogen bonds). Fewer bonds interact at the interface and there is less contact area. Thus, this complex does not appear to be as stable as the wild type Michaelis complex.

131

Figure 5.15 Overall structure of the Michaelis complex between serpin- 1B(S354A)/trypsin: Ribbon representation of the complex (Serpin-1B(S354A), magenta), Trypsin (cyan) Trypsin is superimposed.

Table 5.3 Hydrogen bonds involving residues in the contact area of Serpin- 1B(S354A)/trypsin

Trypsin/S195A Dist.[Å] Serpin-1B E: Ser 147 [OG] 2.71 I: Glu 191 [OE1] E: Gly 216 [N] 3.11 I: Val 351 [O] E: Gln 192 [NE2] 3.03 I: Pro 352 [O] E: Ala 195 [N] 2.94 I: Lys 353 [O] E: Gly 213 [N] 2.69 I: Lys 353 [O] E: Ser 214 [O] 3.10 I: Lys 353 [N] E: Trp 215 [O] 3.22 I: Lys 353 [NZ] E: Ser 190 [OG] 3.29 I: Lys 353 [NZ] E: Phe 41 [O] 2.91 I: Leu 355 [N] E: Tyr 39 [OH] 3.35 I: Leu 357 [N]

132

5.4 The structure of acyl intermediate between chymotrypsin and native peptide as a potent inhibitor

We looked for a way to directly visualize the acyl enzyme with aldehyde peptide derived from serpin-1B, in order to test the possibility that the I350 residue induce the local conformational change such as active sites of trypsin, leading to preventing deacylation. The sequences of the aldehyde peptides, which possess sequences that correspond to residues on the P-side the scissile bond of serpins, are either Ile residue

(acetyl-NAFIVPR-aldehyde), or Ala residue (acetyNAFAVPR-aldehyde) at the P4 position. It was shown previously that tripeptide aldehyde forms covalent adducts with the catalytic serine which is mimetic form of acyl-intermediate and have been used successfully in crystallographic studies of serine protease (Krishnan, Zhang et al. 1998).

Therefore, it was reasonable to consider utilizing the aldehyde peptide for our crystallographic approach to address how hydrolysis was arrested midstream. Before trypsin was crystallized with the aldehyde peptide, inhibition assay was carried out and it confirmed that the aldehyde peptide really inhibited trypsin (Data not shown). We solved the structure for wild-type trypsin in complex with peptide aldehydes. Unfortunately, although many trials of different crystallizations were made, the electron density of the aldehyde peptide did not appear.

During this study, the structure of an acylated chymotrypsin in complex with an autolysed peptide was published (Singh, Jabeen et al. 2005). The autolysis peptide corresponds to the native segment of chymotrypsin and it is proved to act as a potent inhibitor of chymotrypsin (Dixon and Matthews 1989; Harel, Su et al. 1991; Church,

133

Inglis et al. 2001). In this study (Singh, Jabeen et al. 2005), the structure revealed that the peptide forms an acyl intermediate with covalent bond formation between Ser195 and

Trp14 , the P1 residue of the peptide (Figure 5.16 (A)). Despite the fact that the three catalytic residues Ser195, His57 and Asp102 are in well-ordered conformations; similar to those observed in the native enzyme, it is very surprising that Ser195 is still covalently linked to the peptide. In addition, several water molecules are located around the active site (Figure 5.16(B)) which rules out the possibility that acyl-intermediate is due to inaccessibility of waters. Therefore, the analysis of this structure might provide very important clues to explain how hydrolysis is arrested in the serpin mechanism. We tried to find a possible explanation for why deacylation is arrested (why hydrolytic water cannot position properly for deacylation), based upon the observation of this complex structure. Our possible explanation is that Met192 in the chymotrypsin adopts a unique conformation that prevents water from positioning properly for deacylation.

In wild-type serpin complex Gln192 occupies a very similar space to that of

Met192 in chymotrypsin. Furthermore, Gln192 forms hydrogen bonds with backbone of the serpin (Ser354, P1’ and Pro352, P2). These interactions also appear to block where the water must go. This only became apparent from the comparison of the wild-type serpin-1B complex with the acylated chymotrypsin.

In our complex (serpin-1B (I350A)/trypsin), the corresponding residue Gln192 to chymotrypsin Met192 shows a slight conformational difference (Figure 5.17) such that one hydrogen bond between Gln192 and Ser354 of serpin is lost.

134

Figure 5.16 Structure of chymotrypsin (green) in complex with an autolysed peptide inhibitor (dark green) (A) Ribbon diagram of chymotrypsin-inhibitor (B) stable acyl- conformation of chymotrypsin (Ser195) with the inhibitor

135

Figure 5.17 Superposition of acylated chymotrypsin (green)-inhibitor peptide (dark green) with wild-type Michaelis (serpin-1B (cyan)-trypsin(dark cyan)) and serpin- 1B(I350A)(orange) /trypsin(olive). Hydrogen bonds in wild type complex are shown in blue and orange dashed line is hydrogen bond in serpin-1B(I350A)/trypsin.

136

5.5 Discussion

In Chapter 2, the inhibitory activities of several point mutated serpins are discussed to understand the importance of those mutated residues at the interface between the serpin and the trypsin for the inhibitory mechanism of the serpin. Among those mutations, Ile350 to Ala mutation is particularly interesting since the mutated serpin acts as a substrate. In addition, time course reaction seems to indicate that Ile350 residue of the serpin affect the arrest of deacylation rate. Therefore, X-ray crystallographic structure of the mutated serpin (I350A) and trypsin complex would be helpful to understand the mechanistic role of the Il350 residue on suppressing deacylation.

In contrast to the P4 mutation, the P1’ mutation (S354A) shows ~70% inhibitory activity of the wild type serpin. Also, it cannot be clearly explained which step of serpin mechanism influences the activity of the mutant serpin (S354A). However, since the structural analysis of the Michaelis complex on the P1’ residue is unique (see chapter 1.5 for more detailed discussion), we will explore how the mutated serpin (S354A) affect the conformation of the active sites of the trypsin, which may provide some clues on preventing deacylation in the serpin mechanism.

In this chapter, the Michaelis complexes of both serpin(I350A)/trypsin and serpin(S354A)/trypsin are analyzed to understand how the deacylation is arrested in the inhibitory mechanism of serpin. Before discussed the structural implications on the mechanism, it is explained why the Michaelis complex is chosen for understanding deacylation arrest.

First of all, what is the Michaelis complex? The non-covalent Michaelis complex is formed when the serpin recognizes the cognate protease as the first step of serpin

137 mechanism. In contrast to most canonical inhibitors, which only have this step for their inhibition mechanism, the serpin involves more reaction steps (acylation, loop insertion) beyond formation of the initial noncovalent complex to inhibit target proteases.

Then, how can the Michaelis complex help us to understand arrest of deacylation

in serpin mechanism? It is not possible to determine the acylated serpin-trypsin complex,

in which serpin spontaneously processes dramatic change. Therefore, direct visualization

of the acyl-intermediate cannot be crystallized for addressing the question in serpin

mechanism; how is deacylation suppressed at the stage of acyl-enzyme? However, we

found two structural studies (Basran, Sutcliffe et al. 2001; Jogl, Rozovsky et al. 2003),

which reveal the importance of the Michaelis conformation on enzyme mechanism. The

optimal conformation of the Michaelis complex is very important for the catalysis of the

enzyme. They also suggest that the significance of the Michaelis conformation may be

generally applied for the mechanism of the other enzyme. These studies clearly indicate

the conformation of the Michaelis structure influence the each step of the enzyme

mechanism. In this serpin mechanism, the Michaelis conformation may also have an

effect on each step of serpin mechanism. Therefore, the Michaelis complex of mutated

serpin-trypsin may provide some clues on how the second half reaction is arrested in

comparison with the wild type complex.

To understand how the deacylation is arrested, we analyzed the structures of serpin-1B (I350A)/trypsin, serpin-1B (S354A)/trypsin and acylated chymotrypsin with peptide. Especially, the serpin-1B (I350A)/trypsin reveals that there are the substantial conformation changes mostly in the P-side of the RCL. Due to the conformational

138 changes, the extensive interactions between the serpin and the trypsin are lost. The P4 residue also lost unique hydrophobic interactions with trypsin compared to other low molecular weight protease inhibitors. The loss of hydrophobic contacts at I350A seems to induce dramatic restructuring of the RCL, leading to the losses of interactions at interface and concurrently, reduced contact area between the mutant serpin-1B and the trypsin. These fewer interactions and reduced contact area may, thus, lead to destabilizing the acylated covalent intermediate.

We also observe very important difference between the conformation of the

Gln192 residue of the trypsin in serpin-1B (I350A)/trypsin complex relative to that in the complex with serpin-1B. Residue 192 appears to be responsible for holding the acyl- intermediate, based upon the analysis of the acylated chymotrypsin in complex with an autolysed peptide inhibitor. Another important observation from the acyl-intermediate structure is that the tight interactions in the P-side match the binding site of chymotrypsin with exceptionally high complementarity. There is also a study showing that the tight interactions at P4 position affect serpin activity (Cunningham, Blajchman et al. 1997). In this study, it is revealed that the mutation of P4 Ile to more hydrophobic reside Trp in antithrombin results in increased inhibitor activity for thrombin and elastase.

Thus, the loss of the P-side interactions in Serpin-1B (I350A)/trypsin seem to induce conformational change of the Gln192 residue. These changes perhaps cause

Serpin-1B (I350A) to act as a substrate. In other words, right conformation of Gln192 residue through the tight interactions of P-side may be important in arresting deacylation in serpin inhibitory mechanism.

CHAPTER SIX CONCLUSIONS AND FUTURE DIRECTIONS

The serpins employ a unique suicide substrate-like inhibitory mechanism where serpins are arrested midstream of the normal proteolytic reaction (Whisstock, Skinner et al. 1998). The enzyme accomplishes only first step of acylation, which involves breaking the peptide bond and formation of a covalent intermediate between the catalytic serine and the substrate peptide. The second half reaction of deacylation does not occur.

Instead, a dramatic conformational change is adopted to stabilize the covalently attached serpin trypsin complex (Mottonen, Strand et al. 1992; Wang, Mottonen et al. 1996;

Huntington, Read et al. 2000; Gettins 2002; Silverman and Lomas 2004). The X-ray structure of a serpin-protease complex (Huntington, Read et al. 2000) confirmed that the serpins undergo dramatic conformational changes when cleaved by the protease they inhibit and the protease is translocated 70Å from one pole to the other of serpin.

However, the factors causing suppression of deacylation (or stabilizing acyl-intermediate) during loop insertion of the cleaved reactive center loop of the serpin are not fully understood. Based upon the complex structure of serpin-1B with trypsin/S195A (Ye,

Cech et al. 2001), it was proposed that P4 Ile and P1’ Ser residues on the RCL might be involved in the function of arresting the hydrolysis reaction since those residues involved unique interactions with trypsin. This structural implication was supported by the serpin inhibitory assay data, measured in Michael Kanost’s lab, which showed that the mutation of those residues reduced the inhibitory activity of serpin. The P4 (Serpin-1B/I350A)

139

140 mutation almost abolished the role of serpin toward trypsin. The serpin harboring a mutated P4 residue, rather acted like a substrate, not an inhibitor.

Therefore, the crystallographic visualization of those mutant serpins with trypsin was utilized to understand the inhibitory mechanism. Thus, the structures of the

Michaelis complexes of serpin-1B/S354A with trypsin and serpin-1B/I350A with trypsin have been determined. The structural analyses of those mutant structures might elucidate the consequences resulting from the conformational changes of the Michaelis complex in the serpin inhibitory mechanism. The serpin-1B/I350A with trypsin complex revealed that numerous conformational changes, which reduce interactions between serpin-

1B/I350A and trypsin, led to the reduced contact area. Thus, this complex shows apparent reduced stability of the Michaelis complex and also, perhaps destabilizing acyl- enzyme intermediate. The mutation of P4 of serpin-B/I350A also induced a conformational change of the Glu191 and Arg 192 at the short α helix, where mutation of those residues gave rise to the loss of its activity toward protease. Reduced tight interactions mostly in P side may induce conformational change of the well-conserved

Gln192 residue of trypsin. As discussed in Chapter 5, the Gln 192 in trypsin may very well be part of the mechanism by which deacylation is slowed by not allowing water to position properly for deacylation. Therefore, arrest of deacylation in the serpin mechanism is works through the tight interactions between serpin and trypsin with right orientation of Gln192.

It is clear from the data that has already been obtained (see Chapter 2 for detailed discussion) that the mutation I350A increases the deacylation rate. Further kinetic analysis for the P4 I350A mutation would be valuable to understanding the effects of the

141

I350 mutation on the each step of the serpin mechanism. Here, I propose to investigate individual kinetic step using a time-resolved fluorescence resonance energy transfer

(FRET) and a quench-flow, following the methodology of the previous studies (Shin and

Yu 2002; Kang, Baek et al. 2004; Shin and Yu 2006).

First, the donor-acceptor pair will be chosen based upon previously used ones for serpin system (Stratikos and Gettins 1999; Shin and Yu 2002; Kang, Baek et al. 2004;

Shin and Yu 2006). One of them is tryptophan-dansyl pair, which was successfully used in α1-AT and trypsin (Kang, Baek et al. 2004). We can prepare the dansyl-labeled serpin-

1B and serpin-1B (I350A) in a site-specific manner which engineers a single cysteine residue into the desired location. Several positions such as 314, 121, 360 can be selected at surface location for dansyl labeling (Figure 6.1). The fluorescence properties of the dansyl-labeled α1-AT at those positions were well characterized upon complex formation with trypsin (Stratikos and Gettins 1999). After point mutation on this site dansylation of the resulting serpin-1B variants will be carried out with thiol-selective reagent 1, 5-

IAEDANS, as described elsewhere (Stratikos and Gettins 1999). Dansyl-labeled serpin-

1B is required to verify that kinetic and inhibitory properties of serpin-1B are not affected by dansyl labeling. The SI and association rate constants of dansyl-labeled serpin-1B have to be nearly identical to those of wild type serpin-1B.

A second order rate constant for serpin-1B and dansyl-labeled serpin-1B can be measured by continuous assay procedure (Chaillan-Huntington and Patston 1998). Under pseudo-first order conditions, 0.5nM trypsin is incubated with 5nM serpin-1B in the presence of 200µM D-Pro-Pro-Arg-p-nitroanilide and the reaction is continuously recorded at 405nm (Chaillan-Huntington and Patston 1998). Fitting the reaction curve by

142

the equation kapp = (ka·I0)/(1 + [S0/Km]) provides the rate constant for the reaction.

In the equation Km is the value for trypsin with the substrate, ka is the rate of

acylation and kapp = (ka/Km) × SI. The method to determine the SI value is following next.

Figure 6.1 Michaelis complex of serpin-1B (A353K) with trypsin (S195A) (PDB code 1K9O) showing location of cysteine mutation used for labeling. Sphere models represent the residues for dansyl-labeling. In Serpin-1B (A353K), the reactive center loop is in red. α-helixes are in light green, β-sheets are in orange and loops are in pink. Trypsin is in light cyan.

143

Second, to measure the individual kinetic constants, the stoichiometry of inhibition (SI, the number of moles of serpin inhibitor required to completely inhibit 1

mole of a target protease) will be determined first. This value is defined as 1+kdeac/ktr, which can be determined by the competing rates of deacylation and the protease translocation. SI values can be determined by titration of proteolytic activity (Olson,

Swanson et al. 2001). A series of reactions will be setup containing a constant concentration of [E] but increasing [I]0. After about 30min-1hr incubation, the hydrolytic activity of trypsin of each reaction is measured. A linear decreased hydrolytic activity can be seen with a function of inhibitor concentration. Finally, SI value is calculated by extrapolating the data to zero activity, which is the moles of inhibitor (serpin-1B) required to inhibit a molecule of trypsin (SI).

Third, time-resolved fluorescence resonance energy transfer will be measured using an SFM-4 stopped-flow apparatus (Shin and Yu 2002). Reaction will be carried out at 60nm dansyl-serpin-1B (or serpin-1B(I350A)) and various concentration of trypsin.

Excitation is at 292nm, and emission is measured using a 435nm cut-off filter. The

FRET from tryptophans in trypsin to the dansyl group will be monitored. For example, efficiency of FRET for the 314 dansyl-labeled serpin-1B will increase during the processes from the initial Michaelis formation to the final complex. The progress curve is fitted to a single-exponential equation (y=y0+ exp(-kobst)) and the limiting value

(kobs=kobs,lim[E]/([E]+KM)) can be determined by fitting rates to a hyperbolic function of

trypsin concentration (Kang, Baek et al. 2004). The limiting value of kobs,lim corresponds

144

to partitioning step (ktr + kdeacyl) (Shin and Yu 2002). Therefore, it is possible to obtain

individual rate constants since ktr + kdeacyl = ktr · SI.

Lastly, to directly determine the rate of the RCL cleavage (deacylation), we can

utilize quench-flow experiment with a peptide corresponding to the RCL of serpin-1B

(P1-P14) (Kang, Baek et al. 2004). Quench-flow kinetic measurements can be performed

using the same stopped-flow equipments. After mixing with serpin-1B and trypsin, the

reaction can be quenched with 33.3mM HCl in a predetermined time interval. Since a

histidine residue of trypsin control the rate of both acylation and deacylation quenching

with HCl can protonate the histidine residue, resulting in arresting acylation or

deacylation. Quenched samples will be analyzed by SDS-PAGE and scanned on an

Image Scanner. Gel image will be then processed using the Image Gauge program.

By incubating serpin-1B with the peptide (RCL sequence, P1-P14) (Mahadeva,

Dafforn et al. 2002), the peptide forms the binary complex with serpin-1B leading to

generating substrate form of serpin-1B. This substrate form can prevent insertion of the

RCL of serpin-1B (Mahadeva, Dafforn et al. 2002). Therefore, reaction of substrate form

of serpin-1B with trypsin will result in complete cleavage of the RCL without undergoing

loop insertion. The rate of the appearance of cleaved serpin-1B is determined by quench-

flow, SDS-PAGE, and densitometric analyses. This rate can correspond to deacylation if

the step of deacylation is rate limiting in reaction of the RCL cleavage. The value of this

reaction rate can be compared to that from stopped-flow by which the rate of deacylation

indirectly obtained.

CHAPTER SEVEN CRYSTALLIZATION OF BM-SPN2

7.1 Introduction

Serpins have been identified in vertebrates, insects, plants and viruses. In mammals, serpins play an important role on extracellular functions such as complement activation, fibrinolysis, coagulation and inflammation and some serpins have unique roles in the control cell mobility, chromatin folding and Toll signaling (Potempa, Korzus et al.

1994). The naturally occurred mutations of serpin led to human diseases such as thrombosis and emphysema (Stein and Carrell 1995).

Although nematodes are placed a lower part on the invertebrate evolutionary tree, they are very closely related to human disease organisms (Aguinaldo, Turbeville et al.

1997). Nematode worm parasites are among the most prevalent in tropical countries with more than one billion infected people in the world today (Zang and Maizels 2001). The filarial nematode Brugia malayi is the causative agents of lymphatic filariasis. After filarial nematodes are transmitted to their human hosts by mosquitoes, they migrate to the lymphatic vessels where they release millions of micro filasiae into the blood vessel.

Adult worms may survive for long period (about 7 years) and microfilariae may live about 1 year. However, the mechanism of evading the human response by the parasite has not been understood.

BmSPN-2, a serpin, in B. malayi is only expressed in the microfilarial stage and secreted into the bloodstream (Zang, Yazdanbakhsh et al. 1999). Because Bm-SPN-2 expresses in a stage-specific manner it may have very important roles in the parasite life

145

146 cycle. Therefore, it may be the targets for development of novel immuno- or chemoprophylactic agents.

Recent study (Stanley and Stein 2003) shows that BmSPN2 does not have a function of protease inhibition, although it shares many structural features common to other serpin. Thus, BmSPN2 is not an atypical inhibitor, but a new noninhibitory serpin.

X-ray structure of BmSPN2 might provide a starting point for future detailed structure/function analyses into understanding the possible mechanisms of BmSPN2 that could lead to development of therapeutics against lymphatic filariasis.

In this chapter, I described the effort on making the crystal of BmSPN2, which could have the diffraction quality for determination of this new structure. First, expression and purification of BmSPN2 was presented. High level of expression and purity of BmSPN2 led to initial crystallization trials. However, obtaining initial crystals

did not have enough quality (e.g. size, single) for diffraction. Thus, extensive effort on

improving the quality of crystals was carried out and presented here. Another way for

improving crystal quality was made on the construct of BmSPN2 based upon the

sequence alignment with the other well-know serpin family. The different BmSPN2

constructs were also expressed and purified for crystallization. Lastly, methylation in

order to obtain better crystals of BmSPN-2 was performed and described here.

147

7.2 Material and Method

Original BmSPN2 from B. malayi construct was provided from Dr. Rick Maizels in Kansa State University. BmSPN2 cloned into pET29c with 5’ NdeI restriction site and

3’ Xho1 restriction site. This construct encoded a C-terminal 6Xhis tag and a truncated form of BmSPN-2 (21-428) without the putative N-terminal signal peptide (amino acid residues 1-20).

7.2.1 Sequence analysis of BmSPN2

Several sequences of solved serpin structures (Manduca Sexta serpin-1B, α1-AT,

ATIII, ovalbumain, PAI-I) were chosen and multiple sequence alignment generated using

software T-COFFEE (Notredame, Higgins et al. 2000). The sequence was also analyzed

using a program of prediction structural disorder region. This is called IUPs, intrinsically

unstructured proteins, which recognizes intrinsically unstructured/ disordered region of

protein from the amino acid sequence based on the estimated pairwise energy content

(Dosztanyi, Csizmok et al. 2005). This program was built by utilizing assumption that

globular proteins are composed of amino acids that form a large number of favorable

interactions, whereas, intrinsically disorder proteins have less amino acid composition of

making favorable interaction for stable structure (Dosztanyi, Csizmok et al. 2005).

7.2.2 Cloning of Constructs

Each construct was subcloned into the PET 29c vector containing C-terminal 6X

histidine-tag. All the constructs were generated based upon previous sequence analysis.

The PCR reaction condition I used is as follows.

148

2 μl template

1 μl primer 1 (50pmol/μl)

1 µl primer 2 (50pmol/µl)

5 µl Pfu reaction buffer (10X)

1 µl Pfu enzyme

1 µl dNTP (10mM) mixture

39 µl water

50 µl total reaction volume

This mixture was used for PCR reaction in order to amplify the target sequence

1st Cycle: 94 °C for 2 minutes to denature the template

2nd ~ 30th Cycle: 94 °C for 30 seconds to denature the template

58 °C for 40 seconds for primer annealing

72 °C for 90 seconds for primer extension

Once PCR reaction was completed, the PCR product was checked by DNA electrophoresis and purified. And this PCR product was digested by restriction enzymes and applied to ligation reaction by T4 DNA ligase, followed by DNA sequencing.

7.2.3 Expression of Bm-SPN2

A. Transformation

1. 1 µl of DNA plasmid vector was mixed with 20 µl of BL21 (DE3) cells and the

mixture was incubated on ice for 5 minutes

2. Heat-shock was applied to cell-DNA mixture for 40 sec by using 42°C water

bath.

149

3. The cell-DNA sample was transferred onto ice promptly and it kept on ice for 3

minutes.

4. 80 µl of sterilized LB or SOC medium was added and the cells was grown for 1

hour at 37°C in the temperature-controlled shaker

5. Transformed cells were incubated at 37ºC on the LB plate containing 30μg/ml

kanamycin.

B. Cell Growth

1. Streak out E. coli strain BL21 (DE3) transformed with the Pet 29c vector

encoding BmSPN2 on a LB plate containing 30µg/ml of kanamycin.

2. The LB plate was incubated for overnight at 37°C to grow the transformed

colonies.

3. A single colony was inoculated into 100 ml of LB medium containing 30µg/ml

of kanamycin in a 250 ml flask and the cells were grown at 37°C in a shaker

(250~300 rpm) overnight.

4. 20 ml of the overnight-grown culture was used to inoculate 1L of sterilized LB

medium with 30 µg/ml of kanamycin.

5. The cells were grown at 37°C until OD at 595 nm reached about 0.8. (It takes

about 8 ~ 9 hours).

6. 0.5 ~ 1 mM IPTG was utilized to induce BmSPN2 expression and grow the cells

at 30°C for overnight

7. The cells were harvested by centrifugation 30 min at 4000 rpm (at 4°C). for using

a Beckman J6-MI centrifuge with a 6x1 L swing-out rotor.

150

8. The resulting cell pellet was resuspended with about 35 ml of chilled Ni-binding

buffer and treat it with the protease inhibitor cocktail.

9. The resuspended cell pellet was freezed in liquid nitrogen and stored in –80°C

freezer.

7.2.4 Purification of Bm-sPN-2

All the main buffer components used for purification were listed in Table 7-1.

The frozen cells were disrupted by cell disruptor and centrifuged at 16,000rpm for 30min in Beckman Ti45 rotor at 4°C. The supernatant from centrifugation was passed through a

0.45 μm filter to remove particles. The filtered supernatant was load on a Fast-chelating sepharose column (Amersham Pharmacia) charged with 0.1mM NiSO4 and equilibrated

with 10 column volume of nickel binding buffer. Nonspecifically bound proteins were

removed by washing with low concentration imidazole buffer. Bound BmSPN2 was

eluted by elution buffer, which contained a high concentration imidazole. Eluted

fractions containing serpin were identified by SDS-PAGE analysis, following collected

fractions and dialyzed in monoQ equilibration buffer. Following dialysis, the protein was

passed through a 0.22μm filter and loaded onto MonoQ column (Pharmacia). In this

anion exchange chromatography step, the protein was eluted in flowthrough. Bound

protein eluted in a step gradient was found to be impure. Thus, MonoQ purification did

not change serpin concentration and volume from the step of Ni purification. The

purified serpin was concentrated and dialyzed in gel filtration buffer. Then, filtered and

151 loaded onto a Superdex-75 (GE) size-exclusion column, equilibrated with crystallization buffer. The eluted peak of serpin was pooled and concentrated to ~12 mg/ml.

Table 7-1 Summary of reagents used for purification of BmSPN2

Ni-binding buffer 50 mM Tris HCl, pH 8.0

300 mM NaCl

10 mM Imidazole

Ni-elution buffer 50 mM Tris HCl, pH 8.0

300 mM NaCl

250 mM Imidazole

MonoQ buffer A 50 mM Tris HCl, pH 8.0

50 mM NaCl

2 mM DTT

MonoQ buffer B 50 mM Tris HCl, pH 8.0

1 M NaCl

2 mM DTT

Gel filtration buffer 50 mM Tris HCl, pH 8.0

50 mM NaCl

2 mM DTT

152

7.2.5 Reductive Methylation

Obtaining single well ordered crystals is an essential step in determining the three dimensional structure of proteins by X-ray diffraction. Although expression is still a major obstacle to researching numerous numbers of medically important proteins, soluble expression does not guarantee crystals and three dimensional structures. Many proteins are soluble and express of high levels. Even if they are homogeneous, those proteins may not crystallize because of intrinsic flexible nature of their surface residues. The long

chain residues of lysine on the surface of the proteins tend to be flexible. These side chains need to be localized in space in order to improve homogeneous inter molecular interactions necessary for formation of a crystal lattice (Means and Feeney 1968).

Reductive methylation of free amino acids is very efficient method for improving (or obtaining) crystals (Rayment and Charles W. Carter 1997; Walter, Meier et al. 2006).

From the previous study, the methylation of lysine residues resulted only in minor changes in the overall protein conformation where the deviation with the same protein without methylation is in similar range (Rypniewski, Holden et al. 1993). Therefore, the method of reductive methylation of surface-exposed lysine residues was employed to improve BmSPN2 crystals.

A. Chemical Basis for reductive methylation

Reductive methylation of lysine residues is a two step mechanism (Rayment and

Charles W. Carter 1997). First, initial formation of a Schiff base has to occur between the ε-amino group of a lysine residue and either a ketone or aldehyde. The Schiff base is

153 then reduced to a secondary or tertiary amine as indicated for formaldehyde in Equation

1. Formaldehyde reacts with lysine residues rapidly producing a dimethylated product

(Equation 2). This is possible because the pKa of monomethyllysine is slightly lower than that of lysine itself. This method does not change the intrinsic charge on the protein, although it might slightly alter the isoelectric point. Also, the N-terminal amino group of the backbone will be reductively alkylated.

1. R-NH2 + CH2O Æ R-N=CH2 + H2O Æ R-NH-CH3

2 R-NH-CH3 + CH2O Æ R-N(CH3)(CH2OH) Æ R-N(CH3)2

B. Reagents

1. Dimethylamine borane complex (DMAB) (Sigma)

2. 36 % solution of formaldehyde (Sigma)

1 stock solutions for each reagents are needed to be prepared for methylation

A. 1M DMAB

1. 1ml of 1M stock solution of DMAB was prepared by dissolving 59mg of DMAB

in 1ml of deionized water taken in a 1.5ml eppendorf tube.

2. The eppendorf tube was covered with aluminum foil to prevent exposure of

solution to light.

B. 1M Formaldehyde

1ml of 1M stock solution of formaldehyde was prepared by adding 84µl of formaldehyde

(36 % solution) to 1ml of deionized water taken in a 1.5ml eppendorf tube.

154

C. Procedure

1. 1ml of 10mg ml-1 protein sample was transferred in 1.5ml eppendorf tubes.

2. The tubes was covered with aluminum foil.

3. 20μl of 1M DMAB solution was added and followed by 40μL of 1M

formaldehyde solution to the tube.

4. The tube was shaked at 4 °C in dark on a gel shaker maintained at 100 rpm for 2

5. steps 3 and 4 were repeated twice.

6. 10μL of DMAB was added and incubated for 12 h at 4 °C in the dark while

shaking.

7. Methylatedprotein was dialyzed with MonoQ washing buffer.

8. MonoQ and size exclusion chromatography were used for separating methylated

proteins from non-methylated proteins.

9. Analyze proteins using SDS PAGE

7.2.6 Crystallization

Initial extensive crystallization screening was made by using commercially available kits manufactured from Hampton or Emerald. However, the conditions from these kits did not produce any crystals. So, I utilized different conditions for crystallization of BmSPN2. Clear Strategy screens were used for more extensive crystallization screening.

“Clear Strategy Screens” CSS-I and CSS-II were made following the crystallization paper (Brzozowski and Walton 2001). These screens have been developed

155 and tested for previously non-crystallized proteins. Because of the development of high- throughput crystallography with more accumulated information about crystallization properties ‘clear strategy screen’ (CSS) have been developed (Brzozowski and Walton

2001). For crystallization of BmSPN2 these screens were made and tested for different

constructs of BmSPN2. Table 7.2 and Table 7.3 show the conditions of CSS-I and CSS-

II.

Table 7.2 Clear strategy screen I: Adopted from (Brzozowski and Walton 2001) PEG (polyethylene glycol), MME (monomethyl ether), 1.5K, 2K, 4K, 5K, 6K, 8K, and 20K correspond to the molecular weight, MPD (2-methyl-2.4-pentanediol)

156

Table 7.3 Clear strategy screen II: Adopted from (Brzozowski and Walton 2001)

157

7.3 Results and discussion

7.3.1 Sequence Analysis

Construct optimization is often essential to obtain a crystal that diffracts to high resolution. Although full length BmSPN-2 was very soluble and stable protein, it failed to produce any well-diffracting crystals. Several different types of expression and

purification trials were applied to obtain crystals and improve crystal quality. Another approach for improving crystal quality was the reductive methylation of BmSPN2.

Unfortunately, no method provided crystals of sufficient quality to determine BmSPN2 structure. Therefore, sequence analyses were utilized and make the new BmSPN2 constructs for crystallization. The design of new constructs proved difficult. The

sequence alignment program, T-COFFEE and the prediction of disordered region for

protein (IPUD) were utilized for designing new constructs. The sequence alignment was

performed with the other serpins that are well characterized and have crystal structures

including Manduca Sexta serpin-1B, α1-AT, ATIII, ovalbumain, PAI-I. The resulting

alignment is shown in Figure 7.1. Based upon this alignment and orderness prediction in

protein (Figure 7.2) I chose three more constructs: BmSPN2 (21-406), BmSPN2 (48-403)

and N-terminal Stag BmSPN2 (21-406). These three constructs were tested for

expression trials and various purification schemes.

158

Figure 7.1 Multiple sequence alignment of BmSPN2 with other serpins using software T-COFFEE (Notredame, Higgins et al. 2000)

159

160

>bmspn2 globular domains: 1-403

MKLFEVIFIV VAVLGGSSRA NSTLNHCSEN NDDSNKLLIT KAQMNFALEL LRYSSQADET SLLSPFAIAS VMSLLYEGAR GETEREMGRV LSADQPKNAF

GAYIECAIKN IHKQSKRNDY SLYYLTKFFV QQNSLKSDFK DITRRRYPYE

LTQINFASPL QVKSIREIFI KWIKKSMNYG ARNIASITST QSLNLFNGIH FTDDWMYEFL PLNHILPFYS PRLRVTDMPM MERTAKFPYY ENQHMQAVSL PFKDSEMQML IILPKKTFDL AKFEDKLTGE ELFSYISALD SSHEITVTIP

KFKYENQICL LNGLKRMGVQ SMFHESNDFS GIFEQRLFTM DDIRNHAYIK IDEKGINNEN SSMPVRNDMV MDKGEAFRAN HPFLYAIIDN HGTVLWIGRF

TGKnretive nsddsmqcee tpkkrqrl

Figure 7.2 Structural disorderness prediction by IPUD

161

Figure 7.3 Construction of His tagged BmSPN2in pET29c

Figure A shows a 1% agarose gel of PCR product of BmSPN2 (21-406). An arrow shows the PCR product. Figure B shows a 1% agarose gel of the NdeI/XhoI digestion products ligated its BmSPN2 (21-406) fragment of the pET29c vector. Products in lanes 1, 4, 6 contain the expected NdeI/XhoI BmSPN2 fragments.

162

7.3.2 Expression and purification of BmSPN2 (21-426)

The full length BmSPN2 (except the first 20 residues considered as putative

signal peptide) from plasmid PET 29c was expressed in E.Coli strain BL21 (DE3) and

purified by Ni-NTA agarose (Qiagen) followed by Mono Q (General Electric), gel

filtration (General Electric) columns.

Figure 7.4 shows a typical chromatogram and a SDS-PAGE gel analysis for the

purification by Ni-NTA column from 6L culture, using linear gradient. In the

chromatogram, the first peak at about 20% of imdiazole gradient was composed of

mostly non-specifically bound proteins. The second peak from 30% though 45% of

imidazole gradient appeared to be the BmSPN2 as shown in 12% SDS-PAGE gel. A

chromatogram from Mono Q and SDS-PAGE gel of fractions show that BmSPN2 was

eluted in flowthrough factions, not in the gradient. Figure 7.5 demonstrates that most

pure BmSPN2 came off in the MonoQ flowthrough. Flowthrough was collected and

concentrated for gel filtration separation to determine the homogeneity of the protein,

including monomer vs dimmer content and removing impurities by size exclusion. The

chromatograms from a Suerdex75 column (10/10) and SDS-PhastGels of this step are

shown in Figure 7.6. The peak fractions containing pure BmSPN2 were pooled and

concentrated.

163

kDa 198 110 73 45 31 20

Figure 7.4 Purification of BmSPN2 by Ni affinity chromatography

Figure A shows Ni affinity chromatograph of BmSPN2. There are two elution peaks resulting from the linear gradient from 0 to 250 mM imidazole, after a complete wash using 15 column volume wash buffer. The first peak contains non-specifically bounded proteins and the second peak has mostly targeted protein. Figure B shows a 15 % SDS-PAGE of samples taken from the corresponding Ni column fractions. The molecular weight markers are shown as MWM. ‘Before loading’ means the protein solution before loading Ni affinity column. ‘Post loading’ means the flowthrough solution from the Ni affinity column. ‘Pellet’ represents the solution prepared by mixing water with pellet after centrifugation for removing cell pellet. This was done to check the solubility of BmSPN-2. The numbers at top of the gel correspond to elution fractions from the linear gradient. Fractions 7-12 from the first peak mostly contain non-targeted proteins. Fractions 26 -27 shown in the gel from the second peak contain BmSPN2. The 6xHis tagged BmSPN2is indicated with an arrow.

164

kDa 198 110 73 45 31 20 15 6

Figure 7.5 Purification of BmSPN2 by MonoQ anionic exchange chromatography.

Figure A shows MonoQ chromatography of BmSPN2. Most of pure BmSPN2 eluted in the MonoQ flowthrough. The eluted fractions from the linear gradient of 1M NaCl included non-targeted proteins. Figure B shows a 12 % SDS-PAGE separation of samples (numbers at top of the gel 2-6) taken from the flowthrough fractions and the linear gradient from MonoQ. The first lane shows the molecular weight marker and the second lane is the sample before loading the protein solution in MonoQ column. Lane 9 corresponds to the first peak from the MonoQ gradient. Lane 11 is from the second peak from elution gradient in MonoQ column. An arrow indicates 6xHis tagged BmSPN2.

165

M 1 2 3 4 5 kDa 198 110 73 45 31 20 15

Figure 7.6 Purification of BmSPN2 (21-426) by Superdex 75 gel filtration chromatography. Figure A shows a Superdex75 chromatogram. Figure B shows a SDS-PhastGel. The first lane is the sample from peak of Ni column and the lane 2 is the protein solution before gel filtration. The lane 4 is from main peak from gel filtration and the lane 3, 5 are from the side peak of gel filtration. M is the molecular weight marker and the 6x His tagged BmSPN2 is indicated with an arrow.

166

7.3.3 Expression and purification of BmSPN2 (21-406) and BmSPN2 (48-406)

Two different constructs in plasmid PET29c were expressed in E. Coli strain

BL21 (DE3). Unfortunately, the protein, BmSPN2 from the construct of BmSPN2 (49-

406) seems to be like forming inclusion body due to only appearing in cell pellets. It was

not possible to move forward for purification trials of this construct, 48-406 bmSPN2.

The other construct of BmSPN2 was well expressed like full length BmSPN2 and it was

soluble and stable as well as the full-length protein. The BmSPN2 from truncated C-

terminal form also behaved like full length protein in purification procedures of every

chromatography. In linear gradient of imidazole in Ni column, Bmspn2 appeared to be

eluted from more than 45%, which was little bit higher elution gradient than the full-

length protein eluted from30%. The Ni chromatogram and SDS-Phast gel were shown in

Figure 7.7. Analysis from Mono Q chromatogram and SDS-Phast shows that the protein was eluted in flowthrough as shown in Figure 7.8. A Superdex 200/75 column was utilized for size exclusion purification. Nice peak from the gel filtration was appeared as shown in Figure 7.9. The peak fractions were collected and concentrated for crystallization trials

167

B kDa M 1 2 3 4 5 198 110 73 45 31 20 15 6

Figure 7.7 Purification of BmSPN2 (21-406) by Ni affinity chromatography

Figure A shows Ni affinity chromatography. There are two elution peaks coming from the column as result of a linear gradient from 0 to 250 mM imidazole, after a complete wash using 15 column volume wash buffer. The first peak contains non-specifically bounded proteins and the second peak has mostly targeted protein. Figure B shows a SDS-PhastGel of samples taken from the corresponding Ni column fractions. The molecular weight markers are shown in M. The first lane 1 is the sample from the first peak in linear gradient of Ni affinity column. The lanes from 2 to 5 are from the second peak from the imidazole elution gradient in Ni affinity column.

168

B kDa M 1 2 3 4 5

198 110 73 45 31 20 15

Figure 7.8 Purification of BmSPN2 (21-406) by MonoQ anionic exchange chromatography. Figure A shows a MonoQ chromatography. Most of pure BmSPN2 (21-406) was eluted through flowthrough in MonoQ column. The eluted fractions from the linear gradient of 1M NaCl included non-targeted proteins (so called as junk proteins). Figure B shows a SDS-PhastGel of samples (numbers at top of the gel 2-6) taken from the flowthrough fractions and the linear gradient from MonoQ. The molecular weight marker is represented as M. The flowthrough fractions are in lane 1 and lane 2. The lanes from 3 to 5 correspond to peaks from the elution gradient in MonoQ column. An arrow indicates 6xHis tagged BmSPN2.

169

M 1 2 kDa 198 110 73

45 31 20 15 6

Figure 7.9 Purification of BmSPN2 (21-406) by Superdex 200 gel filtration chromatography. Figure A shows a Superdex75 chromatography. Figure B shows a SDS-PhastGel. The 1 lane is the protein solution before loading gel filtration and 2 lane is from peak in gel filtration. The molecular weight marker is shown in lane M and the 6x His tagged BmSPN2 is indicated with an arrow. 8.3.4 Reductive Methylated protein

170

7.3.4 Reductive methylated BmSPN2

BmSPN2 is soluble and expresses in high level. It is also homogeneous from the result of light dynamic scattering. However, they resist being improved the quality of crystals. This could be from intrinsic flexible nature of the surface residues such as lysine residue on the surface of the proteins. To improve homogeneous intermolecular interactions for formation of a crystal lattice the lysine residue can be localized by reductive methylation. This could very efficient method for improving the BmSPN2.

Reductive methylation was performed by following description in method and material section. After complete methylation, Mono Q and gel filtration Columns were utilized for obtaining homogenous methylated BmSPN2. Figure 7.10 shows the chromatogram and SDS-Phast gel from Mono Q. One sharp peak in flowthrough presents methylated BmSPN2. Also, in gel filtration column methylated BmSPN2 was eluted.

171

172

C kDa 198 110 73

45 31 20 15

Figure 7.10 Purification methylated BmSPN2 by Mono Q and superdex 75 Gel filtration Figure A shows MonoQ chromatogram. One sharp peak in flowthrough appears in MonoQ column. Figure B shows a chromatogram from superdex 75 gel filtration. One nice peak is shown in gel filtration. Figure C is SDS-PhastGel, which is showing each step of samples. M represents the molecular weight marker. The lane 1 shows a methylated BmSPN2 from one peak of MonoQ column. The lanes 2 to 4 corresponds to the peak fractions in gel filtration

7.3.5 Crystallization and Discussion

Initial extensive crystallization screening for the construct of full length protein

(21-426 ) was made by using commercially available kits manufactured from Hampton or

Emerald. The first crystal appeared in Hampton screen I, #28. The condition was 0.2M

Na acetate, 0.1M Nacacodylate pH6.5, 30%PEG8K. The crystals in initial condition were neither single nor big enough. The extensive efforts were made for improving crystal quality, especially obtaining single crystals. The initial condition was expanded and streak-seeding method was introduced. When additives in additive screen were

173 added, the cupric chloride additive produced many single crystals. Although these crystals were single and very long, the problem was the thickness of crystal. This crystal was too thin to collect data although X-ray beam test proved this was real protein.

Another problem was this thin crystal was not diffracting well. It only has several diffracted spots showing this was protein. From this point another rigorous effect have been made for improving this thin crystals. Seeding was performed and another additives including detergent types were introduced. Again, none of trials was fruitful.

After completing all possible trials for crystallization for BmSPN2, I made different constructs for crystallization. One of them was not soluble. Only construct (21-

406) was stable and soluble to pursuit the crystallization trials. The protein from this construct was tried to crystallize using commercially available screens. However, any crystals have not appeared in these conditions.

After failing of every trial for producing good crystals, reductive methylation method was introduced to the full length protein. This method was reasonable approach to improve the crystals because the long chain of lysine on the surface of protein could prevent forming well ordered crystalline. After successfully complete of methylation of

BmSPN2, methylated protein was used for initial crystal screening by using Hamptons and Emerald because the crystal condition for methylated protein could be changed. Also, the condition producing crystals of BmSPN2 was expanded. These conditions were utilized for methylated BmSPN2. Unfortunately, none of conditions produced any crystals.

Lastly. New home made crystal screens based upon literature (Brzozowski and

Walton 2001) was made for crystallizing proteins from two different constructs (21-426

174 and 21-406). ‘Clear strategy screen’ CSS-I, CSS-II were utilized for crystallization trials.

However, those condition in screens could not be solution for generating crystals of

BmSPN2.

Figure 7.11 Crystals of BmSPN2 (21-426) The cluster crystals were found in condition Hampton screen I, #28 that contained 0.2M Na acetate, 0.1M Nacacodylate pH6.5, 30%PEG8K. These initial crystals could not be reproduced with any possible effects.

175

Figure 7.12 Crystals of BmSPN2 (21-426) These poorly diffracting crystals were found in the initial crystal condition (0.2M Na acetate, 0.1M Nacacodylate pH6.5, 30%PEG8K) with cupric chloride additive.

CHAPTER EIGHT CRYSTALLIZATION OF VOPA

8.1 Introduction

The Gram-negative bacterial Pathogen Vibrio parahaemolyticus is a major causative agent for gastroenteritis outbreaks associated with the consumption of contaminated food (MacCarter 19990). V. parahaemolyticus uses type III secretion systems to inject effector proteins into the cytoplasm of the host cells (Ghosh 2004). In the host cytosol, the effector proteins manipulate the host signaling machinery to promote pathogen infection (Trosky, Li et al. 2007). The effectors act as mimics of eukaryotic signal transduction machinery, allowing V. parahaemolyticus to deregulate host cell signaling pathways. However, these effectors are in quiescent state with chaperones in the pathogenic bacteria. They only show specificity for host substrates, and sometimes require activator proteins or modification that only can be provided by a eukaryotic source (Viboud and Bliska 2005).

One V. parahaemolyticus effector protein is VopA (Vibrio outer protein A).

VopA belongs to a family of type III effectors called YopJ-like proteins that are present in a wide variety of pathogens such as animal and plant pathogens, as well as the plant symbiont Rhizobium (Orth 2002; Viboud and Bliska 2005). YopJ in Yersinia spp., (the founding member of this family), blocks the MAPK and NFκB signaling pathways in the host cell resulting in inhibition of host innate immune response (Orth, Palmer et al. 1999).

All YopJ-like proteins have a catalytic triad and YopJ itself is functioning as an acetyltransferase (Mukherjee, Keitany et al. 2006). YopJ inhibits MAPK and NFκB

176

177 signaling pathways by acetylating on serine and threonine residues on the activation loop of MAP kinases. Although VopA and YopJ share ~55% sequence similarity, their functions are different. VopA is an extremely potent inhibitor of the evolutionary conserved MAPK pathways, but does not inhibit the NFkB pathway (Trosky, Mukherjee

et al. 2004). VopA acetylates a conserved lysine residue in the catalytic loop of all

kinases. This residue is essential for interaction with the γ-phosphate of ATP. An

acetylation of this lysine residue blocks ATP binding, thus, inactivating the kinase

(Trosky, Li et al. 2007).

Therefore, solving the three-dimensional structure of VopA might provide a great

starting point for the design of drugs, including antibacterial reagents. Also, the analysis

of a VopA X-ray structure will be very valuable in developing much more detail

mechanism for VopA.

In this chapter, different constructs of VopA for expression in E. coli and purification trials are described. Also different expression vectors and purification techniques were utilized for improve the quality of VopA protein for crystallization.

Extensive crystal screens are conducted to generate VopA crystals.

178

8.2 Materials and Methods

8.2.1 Construction of two expression plasmid: GST-vopA and His- smt3-vopA

Two initial pGEX-rTev-3-VopA constructs were provided by Dr. Kim Orth’s lab.

One is wild type without any mutation, and the other is a mutant VopA in which the reactive site, cysteine (residue 167), was mutated to alanine. These expression vectors encoded an N-terminal GST tag and full length VopA (aa 1-289). Since the wild type

VopA and the active site mutant VopA seemed to behave the same, I focused on vopA

(C167A) for crystallization trials because wild type vopA activity could cause problems during expression and purification. Full-length vopA (C167A) in pGEX-rTev-3-VopA appeared to be less stable and it seemed to stick to filters during concentration in the absence of detergents. Therefore, the detergent was necessary to stabilize the protein throughout purification. In addition, the full length VopA was subcloned into the plasmid His- Smt3-VopA, which was suggested (by Dr. Phillip Thomas), to be a good plasmid to rescue insoluble proteins.

The full length VopA subcloned in plasmid His- Smt3-VopA was used for extensive crystallization and purification trials. This plasmid was constructed such that protein could be synthesized as C-terminal fusions with an N-terminal 6Xhistidine Smt3 fusion protein. The resulting His-tagged recombinant protein was purified by metal- affinity chromatography, followed by anion exchange and finally by gel filtration. The protease, Ulp1, was utilized to cleave and liberating the His-tagged Smt3 moiety from the

N-terminus of the recombinant protein during, or after purification. Ulp1 does not have any non-specific activity.

179

The SMT3 coding region was cloned into pET28b upstream of the BamHI site to encode an N-terminal histidine-tagged Smt3 fusion protein. In order to produce native- like protein, VopA after cleavage, I inserted the VopA into BamHI-HindIII sites. The coding region was shown to be in frame with the BamHI site, and the 3’ primer contained a stop codon for each subcloning experiment.

Promoter Start Codon 6-histidine tag Smt3 MCS Stop

Ulp1 cleavage site

T7 Met------HHHHHH------ModSmt3-GlyGlySer-protein (VopA)-stop

Lac ATG------GGATCC-coding sequence-TAA

Any BamHI

The native C-terminal glycine of Smt3 overlaps with the coded glycine of BamHI. Ulp1 cleaves VopA after the Gly-Gly motif, releasing VopA with an extra N-terminal serine.

8.2.2 Expression and purification of GST-VopA in E.Coli

Escherichia coli cells transformed with pGEX-rTev-3-VopA were used for the expression of the protein. The cells were cultured overnight in 2 × YT medium ( 1% yeast extract, 1.6% tryptone, and 0.5%NaCl) containing 100μg/ml ampicillin at 37°C, diluted 1:100 with fresh pre-warmed 2 × YT medium containing 100μg/ml amphicillin and incubated at 37°C with shaking 200rpm. When OD600 had reached a value of 0.8, the

180

expression of GST-VopA was induced by adding isopropyl-β-D-thiogalactopyranoside

(IPTG) to a final concentration of 4μM for an additional 4hr. The cells were pelleted by centrifugation at 5000 rpm for 15 min. Cell pellets were resuspended in pre-cooled lysis buffer (PBS, 1% Triton, 0.05%BME, 1mM PMSF). The resuspended cell pellets were disrupted by cell disruptor and centrifuged at 10,000 rpm for 40min. The supernatant containing GST-VopA was incubated with glutathione-Sepharose beads for 2hr. Non- specifically bound proteins were removed by washing lysis buffer, and the bound protein with GST tag was eluted with 50mM Tris-HCl (pH8.0) buffer containing 20mM reduced glutathione. Adding GST-Tev protease for overnight at room temperature in shaker digested the eluted GST-VopA from beads. The protein, VopA, was eluted from the

GST beads. Superdex 75 gel exclusion column pre-equilibrated with crystallization buffer (20mM HEPES, 100Mm NaCl, 1mM DTT, 0.1mM EDTA) was performed to remove minor contaminants. The majority of the protein eluted in a single major peak and was concentrated to 15mg/ml for further crystallization steps.

181

Table 8-1 Summary of reagents used for purification of GST tagged VopA

Lysis buffer PBS

1% Triton

0.05%BME,

1mM PMSF

GST-binding buffer 50 mM Tris HCl, pH 8.0

300 mM NaCl

10 mM Imidazole

GST-elution buffer 50 mM Tris HCl, pH 8.0

300 mM NaCl

20mM reduced glutathione

Gel filtration 20 mM HEPES, pH 8.0

100 mM NaCl

1 mM DTT

0.1 mM EDTA

182

8.2.3 Expression and purification of His-Smt3-VopA in E.Coli

A. Expression of VopA

VopA was expressed in BL21 (DE3 pLysS) cell. The six histidine tagged smt3 plasmid containing VopA was transformed to BL21 (DE3 pLysS) cells. The transformation starts with mixing 1μl of DNA and 50μl of BL21competent cells. Then, cells and DNA were incubated on ice for 10 min, followed by heat shock at 42°C for 40 seconds, and finally, back on ice for 5 minutes. After adding 50μl of SOC medium to the cells, the transformation reaction was incubated for about an hour at 37°C while shaking.

The cells were plated on antibacterial agar plates containing 30 μg/ml kanamycin and incubated at 37°C overnight. A single colony from the plate was inoculated in 100ml of

LB medium containing 30 μg/ml kanamycin and then incubated overnight at 37°C. The overnight culture was inoculated into 1L of LB medium containing 100 μg/ml kanamycin

and grown at 37°C. When OD600 had reached more 1.5, the expression of the protein was

induced by adding isopropyl-β-D-thiogalactopyranoside (IPTG) to a final concentration of 0.75mM. VopA were continuously grown at lower temperature of 30°C for an additional 4hrs. The cells were pelleted by centrifugation at 4000 rpm for 30 min, then resuspended in pre-cooled lysis buffer. The resuspended cell pellets were frozen in liquid nitrogen and stored at -80°C.

B. Immobilized Metal Affinity Chromatography

1. The supernant from a 6L culture was applied by using a peristaltic pump

(Amersham Biociences) to a fast-chelating sepharose column (Amersham

183

Pharmacia) loaded with 0.1 mM NiSO and pre-equilibrated with 5~10 column

volume of Ni-binding buffer.

2. After washing the column with at least 10-column volumes of washing buffer,

VopA was eluted using 20-colume volumes of a linear gradient of elution buffer.

The initial washing process allowed nonspecifically bounded protein to be

removed.

3. The peak fractions from the linear gradient were collected and dialyzed.

C. Ulp1-protease Digestion

1. Protein fractions eluted from the Ni-column were mixed with MonoQ A buffer to

reduce the concentration of imidazole.

2. Ulp protease was added to protein solution using a 1000:1 volume ratio (e.g. 10μl

Ulp1 is used for 10ml of protein) and then the solution was well mixed.

3. The protein solution with Ulp1 protease was transferred into dialysis bag for a

buffer change (Mono Q buffer A) at 4°C for overnight.

D. Anion Exchange Chromatography

1. Clear the dialyzed protein solution by 0.22 µm filter. At this stage it was also

possible to load this mixture of his-tagged Smt3 and VopA back over metal-

affinity column, but this turned out to be a bad choice due to the non-specific

association of VopA proteins to this column.

184

2. The filtered protein sample was loaded onto a MonoQ anion-exchange column

(Pharmacia).

3. Following a washing step with the 5 column-volume of MonoQ A buffer, the

protein sample was eluted using a linear gradient for elution of target protein

sample. Smt3 loading buffers binds to MonoQ resin in with reasonable ionic

strength (100mm NaCl) and elutes at NaCl concentrations between 200 and 250

mM.

E. Size-Exclusion Chromatography

1. Protein sample eluted from anion-exchange chromatography was pooled and

concentrated up to the total volume of 2 ml.

2. Following the filtration of sample protein by 0.22µm filter, the protein was

loaded onto a Superdex-75 or 200 size-exclusion column. The eluted sample was

pooled and collected. The protein sample for crystallization was usually

concentrated up to a final concentration of ~ 12 mg/ml.

185

Table 8-2 Summary of reagents used for purification of VopA in smt3-Ulp vector

Ni-binding buffer 50 mM Tris HCl, pH 8.0

300 mM NaCl

10 mM Imidazole

Ni-elution buffer 50 mM Tris HCl, pH 8.0

300 mM NaCl

250 mM Imidazole

MonoQ buffer A 50 mM Tris HCl, pH 8.0

50 mM NaCl

2 mM DTT

MonoQ buffer B 50 mM Tris HCl, pH 8.0

1 M NaCl

2 mM DTT

Gel filtration buffer 50 mM HEPES, pH 7.0

50 mM NaCl

2 mM DTT

0.1 mM EDTA

186

8.2.4 Crystallization

Soluble constructs of VopA were concentrated and screened for crystallization by utilizing commercially available sparse matrix kits (Emerald Biosciences and Hampton

Research). The hanging-drop vapor diffusion method was utilized for grow crystals at

16 °C. CSS-I and CSS-II (“Clear Strategy Screens”-I, II) were prepared following the referenced crystallization paper (Brzozowski and Walton 2001). Such screens have been developed and tested for previously non-crystallized proteins. ‘Clear strategy screen’

(CSS) has been developed previously (Brzozowski and Walton 2001) due to the advance of high-throughput crystallography with more accumulated information about crystallization properties. For the crystallization of VopA, these screens were constructed and tested for different constructs of VopA.

187

8.3 Result and discussion

8.3.1 Sequence analysis and VopA homology model

Recent studies show that all YopJ-like proteins have a catalytic triad and are functioning as acetyltransferases (Mukherjee, Keitany et al. 2006). VopA is an extremely potent inhibitor of the evolutionary conserved MAPK pathways (Trosky, Mukherjee et al.

2004). Solving VopA structure would serve as a tool of investigating the chemistry of

YopJ family acting on MAPK kinase. Unfortunately, the full length of VopA either with

GST-tagged or his-smt3-tagged did not produce any crystals. Many trials of different purification methods were applied to generate VopA crystals; however, none of them was fruitful. Therefore, the modification of the VopA construct was considered. Since none of YopJ family has been determined by X-ray crystallography, it was difficult to decide which region of VopA might be flexible or disordered. T-COFFEE, a sequence alignment program, and IUPred, which simulates the prediction of disordered portion of a protein, were utilized for final decision to make VopA constructs. VopA sequence was aligned with YopJ and ArvA. Also, the sequence of VopA was aligned with adenovirus protease because the closest available crystal structure was from adenovirus protease

(Figure 8.1). From the alignment, the homology model of VopA was predicted as shown in Figure 8.2. An N-terminal truncated VopA (aa 21-289) and an N-terminus with C- terminus truncated form of VopA (aa 21-201) were made4 for purification and crystallization trials (Table 8.3).

188

Table 8.3 Different constructs of vopA and their properties

Length of Plasmid Expression Level of solubility construct system expression 1-289 pGEX-rTev E.Coli good Soluble with (GST tagged) detergent 1-289 (C167A) pGEX-rTev E.Coli good Soluble with (GST tagged) detergent 1-289 PET28b E.Coli good good His-Smt3-tagged 1-289 (C167A) PET28b E.Coli good good His-Smt3-tagged 21-201 PET28b E.Coli ? insoluble His-Smt3-tagged 21-289 PET28b E.Coli bad bad His-Smt3-tagged 49-289 PET28b E.Coli bad bad His-Smt3-tagged

YopJ 98 RFIINMGE..GGIHFSVIDYKHI.NGKTSLILFEPANFNMGPA...M VopA 94 RFVVNMGS..GGIHCIAVDCAIK.NGKCSLIGIEPVTMNLGAS...M AvrA 112 RFLVNMGS..SGIHISVVDFRVM.DGKTSVILFEPAACSFGP....A AvrP 171 RGVFRLNP..PSLHHVAVDVRNHSNGQKTLIVLEPITAYDDVYPPAY XopJ 167 RAVLRLER..DGEHHVAADVRRRPNGEASVIVLEPARLL...... F AVP 41 AIVNTAGRETGGVHWMAFAWNPR...SKTCYLFEPFGFSQRLKQVYQ

YopJ 98 LAIRTKTAIERY..QLPDCHFSMVEMDIQR..SSSECGIFSFALAKK VopA 136 LAIRLQSVCKRE...LPETSLVIMETDMQR..SQGECLMFSLFLVKK AvrA 153 LALRTKAALERE..QLPDCYFAMVELDIQR..SSSECGIFSLALAKK AvrP 195 LPGYPQLREEVNTRLRGNAKMSVIETDAQR..SWHDCVIFSLNFALC XopJ 201 VTGHTQLRRQALSQLGENAKFAFIQVGAQK..SAADCLMFDLHFALH AVP 86 FEYESLLRRSAIASSPDRCITLEKSTQSVQGPNSAACGLFCCMFLHA

Figure 8.1 Sequence alignment of putative cysteine proteases and adenovirus protease (AVP). Residues highlighted in red form the highly conserved catalytic triad; yellow highlighting identifies areas which may represent shifts in the conical cysteine protease active site, represented by human adenovirus 2. Abbreviations are: YopJ from Yersinia pestis, VopA from Vibrio parahaemolyticus, AvrA from Salmonella enterica, AvrP from Pseudomonas syringae, XopJ from Xanthomonas campestris, and AVP from human adenovirus 2.

189

VopA homology model

A homology model of VopA was predicted based on the closest available crystal structure of adenovirus protease (Fig. 9.2). The model-building shows that there are significant changes in the active site of the protease, including a major deletion, which is suggestive of repacking of the C-terminus of VopA relative to AVR.

Figure 8.2 Differences between VopA and adenovirus protease (AVP) Coordinates for AVP (PDB file 1AVP) are highlighted with the sites of major insertions or significant amino acid replacements. Although the catalytic triad (AVR His54, Glu71, and Cys122) is conserved in VopA, bulky residues in AVR in the active site, such as Trp55 and Phe73 have replaced by smaller residues. Also, there is a deletion of 4 residues in a sequence partially buried in the structure suggestive of significantly different folding and packing of the C-terminus. Also there are 50 more residues at the C-terminus of VopA of unknown structure.

190

8.3.2 Expression and purification of GST tagged VopA (1-289)

The full length VopA from plasmid pGex-rTev expressed in E.Coli strain BL21

(DE3) and purified by using glutathione-Sepharose beads and gel filtration (General

Electric) columns.

Figure 8.3 shows SDS-PAGE gel analysis for the purification from glutathione-

Sepharose beads. From the gel in this step of purification it was noticed that a relatively high amount of VopA, almost equal amount of VopA, was still bound to GST beads although sufficient elution volume was applied. This seemed to show the nature of VopA which prefers to binding the beads. VopA protein also became stable in present with detergent and glycerol. These characteristics of VopA might present the possibility of exposing hydrophobic patch of this protein. In other words, the folding of VopA might not be the stable form. Figure 8.4 shows the chromatogram and SDS-PAGE gel for the purification by a superdex 75 gel exclusion column. In this step minor contaminants were removed. The majority of the protein eluted in single major peak, which was concentrated to 15mg/ml for further crystallization step.

191 Pellet E1 B1 E2 B2 Conc

VopA GST

Figure 8.3 Purification of GST tagged VopA by glutathione-Sepharose beads Above figure shows a 15 % SDS-PAGE of samples taken from the elution from glutathione-Sepharose beads. Samples from pellet were prepared for checking a solubility of VopA. Lane E represents the elution from glutathione-Sepharose beads by 20mM reduced glutathione. E1 and E2 were eluted from different incubation with glutathione-Sepharose beads. B1 and B2 are the glutathione-Sepharose beads after eluted VopA protein. Significant amounts of VopA still remain in the beads. The VopA and GST are indicated using arrows. Cleaved GST bound onto the beads.

kDa 198 M 1 2 3 4 5 110 73

45 31

Figure 8.4 Purification of VopA by superdex75 gel filtration This shows a 15 % SDS-PAGE of samples taken from superdex75 gel filtration . The VopA and GST are indicated using arrows. The molecular weight marker is labeled as M. The lane 1 is the concentrated VopA before loading onto Gel filtration. The lanes from 2 to 5 are taken factions from gel filtration peak. An arrow indicates VopA.

192

8.3.3 Expression and purification of His-Smt3 tagged VopA (1-289)

The full length VopA from plasmid pET 28b was expressed in E.Coli strain

BL21 (DE3) and purified by Ni-affinity column (Qiagen), Mono Q (General Electric), and the gel filtration (General Electric) columns.

In the Ni affinity column, Vop-Smt3-Histagged protein was eluted by two different imidazole gradients as shown in Figure 8.5. The first peak in the linear imidazole gradient appeared at around 35%, which mostly contained nonspecifically bound proteins with inappropriately folded VopA-smt3-his tagged. The second peak from 70% imidazole gradient seemed to be the target protein as shown in 12% SDS-

PAGE gel (Figure 8.5). Before loading MonoQ column VopA was produced by the Ulp-

1 cleavage between VopA and His-Smt3 tagged. To ensure complete cleavage, the reaction was checked by SDS-PAGE gel in Figure 8.5. After verifying cleavage of His-

Smt3 tagged VopA was loaded onto MonoQ. The chromatogram of Mono Q and SDS-

PAGE gel show that VopA eluted in flowthrough factions, not in the gradient made by elution buffer. Figure 8.5 clearly proves that most of the purified protein elutes in the flowthrough in Mono Q column. The peaks from MonoQ gradient were comprised of the cleaved Smt3-histagged, non-target proteins and small amount of noncleaved VopA- smt3-histagged protein. The fractions from MonoQ flowthrough were pooled and concentrated for separation on gel filtration. The chromatograms from a Suerdex75 column (10/10) and SDS-PhastGels of this step are presented in Figure 8.6. The peak fractions, which contained pure VopA were pooled and concentrated prior to crystallization.

193

M 1 2 3 4 5 6 7 8 kDa 198

110 73 VopA-smt3 45 31 20 VopA

15 Smt3-His

194

Figure 8.5 Purification of VopA by Ni affinity chromatography and Mono Q anionic exchange chromatography Figure A shows the Ni affinity chromatogram. There are two elution peaks coming from the column as result of a linear gradient from 0 to 250 mM imidazole, after a complete wash using 15 column volume wash buffer. The 1st peak contains non-specifically bounded proteins and the 2nd peak has mostly targeted protein. Figure B shows a 12 % SDS-PAGE of samples taken from the corresponding Ni and Mono Q column fractions. The molecular weight markers are shown in M. The first lane 1 corresponds to the first peak from Ni affinity column. The sample of lane 2 came from the second peak of the linear gradient in Ni column. This second peak in Ni column contained Smt3 tagged VopA as a majority with little contamination. The lane 3 and 4 shows Ulp1 digestion process between Smt3 and VopA. Lane 5 presents the protein solution before performing MonoQ purification. Figure C shows a MonoQ chromatogram. Most of pure VopA was eluted through flowthrough in MonoQ column. The eluted fractions from the linear gradient of 1M NaCl included non-targeted proteins. From figure B, samples of the 12 % SDS-PAGE taken from the flowthrough fractions and the linear gradient from MonoQ, show that the flowthrough fractions (lane 6) contain VopA protein. Lane number 7 corresponds the first peak from the gradient in MonoQ. This peak contained most of the His-Smt3 cleaved from VopA-Smt3-His by Ulp1 protease. Lane 8 is the second peak from the elution gradient in MonoQ column. An arrow indicates VopA protein.

195

GF75 vopa 3 21 06:1_UV GF75 vopa 3 21 06:1_Conc GF75 vopa 3 21 06:1_Fractions

mAU

300

250

200

150

100

0 1 2 3 4 5 6 7 8 9 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 0 20 40 60 80 100 120 140 ml

M 1 2 kDa 198 110 73 45 31 20 15

Figure 8.6 Purification of VopA (1-289) by Superdex 200 gel filtration chromatography. Figure A shows a Superdex75 chromatogram. Figure B shows a SDS-PhastGel. Lane1 is from the peak in the gel filtration column and Lane 2 indicates the shoulder of peak in gel filtration. The molecular weight marker is shown in lane M and VopA is indicated with an arrow.

196

8.3.4 Expression and purification of VopA (21-201) and VopA(21-289)

Two different constructs in the plasmid PET28b- His-Smt3 tagged were expressed in E. Coli strain BL21 (DE3). Unfortunately, the protein, VopA- smt3 - his tagged from the construct of VopA (21-201) seems to be forming inclusion body due to only appearing in cell pellets. Therefore, decision was made not to use this construct for further trials of purification. The alternative construct of VopA was expressed; but, the level of protein expression was not as high as the wild type VopA. The Ni chromatogram and SDS-Phast gel are shown in Figure 8.7. Only one peak from the applied linear gradient seems to contain VopA-smt3-histagged protein. Also, the stability VopA was as robust as the full length VopA due to the fact that this protein tends to precipitate after the elution from Ni column. Analysis of Mono Q chromatogram and SDS-PhastGel shows that the protein was not eluted in the flowthough and there was zone spreading throughout the elution gradient on the MonoQ column (Figure 8.10). There were no fractions to collect, concentrate for the purification step of gel filtration. Further purification trials could be performed for rescuing the target protein. However, considering its stability, level of expression, and solubility of this protein, it is not worthwhile to keep trying different modes of purification for these constructs. It needs to be more thoroughly characterized before proceeding purification and crystallization.

197

B kDa 198 110 73

45 31 20 15 6

Figure 8.7 Purification of VopA(21-289) by Ni affinity column Figure A shows the Ni affinity chromatogram, which has one major peak. Figure B is a SDS-PhastGel of samples from peak of Ni column. Low level of VopA expression is shown as arrow indicated.

198

B kDa 198 110 73

45 31 20 15

Figure 8.8 Purification of VopA (21-289) by MonoQ anionic chromatography Figure A shows the MonoQ chromatogram of VopA purification. Figure B indicates the SDS-PhastGel of the sample taken from each peak of fractions in MonoQ column. Lane 1 and 2 are the flowthrough fractions and the other lanes correspond the peaks of the column. An arrow and M represents VopA and molecular weight marker respectively.

199

8.3.5 Crystallization of VopA and Discussion

Initial extensive crystallization screening for the construct of full length protein

(1-289) in GST tagged vector was carried out using commercially available kits manufactured from Hampton and Emerald. However, the stability and solubility of the protein was not sufficient, such that detergent and glycerol had to be added in purification

steps. VopA tends to precipitate easily and required some organic molecules to stabilize.

Unfortunately, VopA in GST tagged vector was not crystallized. Therefore, it was

decided to follow the suggestion of Dr. Philip Thomas, which utilized the Smt3-Ulp vector for obtaining stable, soluble protein. VopA produced using smt3-ulp vector was first screened by commercially available kits (Hampton research: Crystal Screen I, II, III,

Peg/Ion, Index, SalRx, Light Crystal Screen I and Emerald: Wizard screen I, II, III).

None of the provided conditions produced any crystals. At this stage, ‘Clear strategy screen’ CSS-I, CSS-II was tested for the crystallization of VopA. After no indication of crystals from any conditions of commercially available screens, it was proposed that the full length form of VopA may not be possible to crystallize.

After completing all possible trials for crystallization of the full length VopA, different constructs were engineered for crystallization. Since there was no information on the YopJ family, which VopA belongs to, it was difficult to make a decision concerning the length of the constructs for VopA crystallization trials. Based upon sequence alignments with the closest homologue family that produce crystal structures and data from the structural disorder prediction program, several VopA constructs were made. These different constructs were tested for expression level, solubility, and stability of VopA proteins. However, despite extensive efforts in making new constructs, none of

200

VopA constructs expressed sufficiently and was soluble. Purification of VopA protein from different constructs having different length could not continue due to the stated failure.

Finally, different purification methods were attempted. One of them involved purifying VopA proteins using CoA since activation of VopA requires this common cofactor. VopA co-purified with Mek6 because VopA acts as acetyltransferase to Mek6 effectively blocking Mek6 activity. Various chromatography methods such as PS were used to increase of purity the VopA, thereby raising probability of obtaining VopA

crystals. Nevertheless, upon exhausting different trials for the crystallization of VopA,

VopA crystals were not obtained.

It may be proposed that the failure to generate VopA crystals is indicative that

VopA itself is intrinsically flexible and not stable enough for crystallization. However,

there is still opportunity for other trials for growing crystals of VopA. Since this work

focused on E.coli host strains, different expression systems such as eukaryotic insect cell

cultures may produce more stable protein to generated crystals. Additionally, other

homologues in the YopJ family may be utilized for the investigation of structural

mechanism of the YopJ family. In any case, the quality of the protein is an essential

factor to carry out the extensive and comprehensive crystallization screen.

BIBLIOGRAPHY

Aertgeerts, K., H. L. De Bondt, C. J. De Ranter and P. J. Declerck (1995). "Mechanisms contributing to the conformational and functional flexibility of plasminogen activator inhibitor-1." Nat Struct Mol Biol 2(10): 891-897. Aguinaldo, A. M. A., J. M. Turbeville, L. S. Linford, M. C. Rivera, J. R. Garey, R. A. Raff and J. A. Lake (1997). "Evidence for a clade of nematodes, arthropods and other moulting animals." Nature 387(6632): 489-493. Al-Ayyoubi, M., P. G. W. Gettins and K. Volz (2004). "Crystal Structure of Human Maspin, a Serpin with Antitumor Properties: REACTIVE CENTER LOOP OF MASPIN IS EXPOSED BUT CONSTRAINED." J. Biol. Chem. 279(53): 55540- 55544. Argos, P. and F. S. Mathews (1975). "The structure of ferrocytochrome b5 at 2.8 A resolution." J Biol Chem 250(2): 747-51. Aulak, K. S., E. Eldering, C. E. Hack, Y. P. Lubbers, R. A. Harrison, A. Mast, M. Cicardi and A. E. d. Davis (1993). "A hinge region mutation in C1-inhibitor (Ala436-- >Thr) results in nonsubstrate-like behavior and in polymerization of the molecule." J. Biol. Chem. 268(24): 18088-18094. Baglin, T. P., R. W. Carrell, F. C. Church, C. T. Esmon and J. A. Huntington (2002). "Crystal structures of native and thrombin-complexed heparin cofactor II reveal a multistep allosteric mechanism." PNAS 99(17): 11079-11084. Ban, N., P. Nissen, J. Hansen, P. B. Moore and T. A. Steitz (2000). "The Complete Atomic Structure of the Large Ribosomal Subunit at 2.4 A Resolution." Science 289(5481): 905-920. Basran, J., M. J. Sutcliffe and N. S. Scrutton (2001). "Optimizing the Michaelis Complex of Trimethylamine Dehydrogenase. IDENTIFICATION OF INTERACTIONS THAT PERTURB THE IONIZATION OF SUBSTRATE AND FACILITATE CATALYSIS WITH TRIMETHYLAMINE BASE." J. Biol. Chem. 276(46): 42887-42892. Baumann, U., W. Bode, R. Huber, J. Travis and J. Potempa (1992). "Crystal structure of cleaved equine leucocyte elastase inhibitor determined at 1[middle dot]95A resolution." Journal of Molecular Biology 226(4): 1207-1218. Baumann, U., R. Huber, W. Bode, D. Grosse, M. Lesjak and C. B. Laurell (1991). "Crystal structure of cleaved human [alpha]1-antichymotrypsin at 2.7 a resolution and its comparison with other serpins." Journal of Molecular Biology 218(3): 595-606. Beauchamp, N. J., R. N. Pike, M. Daly, L. Butler, M. Makris, T. R. Dafforn, A. Zhou, H. L. Fitton, F. E. Preston, I. R. Peake and R. W. Carrell (1998). "Antithrombins Wibble and Wobble (T85M/K): Archetypal Conformational Diseases With In Vivo Latent-Transition, Thrombosis, and Heparin Activation." Blood 92(8): 2696-2706. Beinrohr, L., V. Harmat, J. Dobo, Z. Lorincz, P. Gal and P. Zavodszky (2007). "C1 Inhibitor Serpin Domain Structure Reveals the Likely Mechanism of Heparin Potentiation and Conformational Disease." J. Biol. Chem. 282(29): 21100-21109.

201

202

Bergstrom, R. C., G. S. Coombs, S. Ye, E. L. Madison, E. J. Goldsmith and D. R. Corey (2003). "Binding of Nonphysiological Protein and Peptide Substrates to Proteases: Differences between Urokinase-Type Plasminogen Activator and Trypsin and Contributions to the Evolution of Regulated Proteolysis." Biochemistry 42(18): 5395-5402. Bode, W. and R. Huber (1978). "Crystal structure analysis and refinement of two variants of trigonal trypsinogen : Trigonal trypsin and PEG (polyethylene glycol) trypsinogen and their comparison with orthorhombic trypsin and trigonal trypsinogen." FEBS Letters 90(2): 265-269. Briand, C., S. V. Kozlov, P. Sonderegger and M. G. Gr?ter (2001). "Crystal structure of neuroserpin: a neuronal serpin involved in a conformational disease." FEBS Letters 505(1): 18-22. Bruch, M., V. Weiss and J. Engel (1988). "Plasma serine proteinase inhibitors (serpins) exhibit major conformational changes and a large increase in conformational stability upon cleavage at their reactive sites." J. Biol. Chem. 263(32): 16626- 16630. Brunger, A. T., P. D. Adams, G. M. Clore, W. L. DeLano, P. Gros, R. W. Grosse- Kunstleve, J. S. Jiang, J. Kuszewski, M. Nilges, N. S. Pannu, R. J. Read, L. M. Rice, T. Simonson and G. L. Warren (1998). "Crystallography & NMR system: A new software suite for macromolecular structure determination." Acta Crystallogr D Biol Crystallogr 54(Pt 5): 905-21. Brzozowski, A. M. and J. Walton (2001). "Clear strategy screens for macromolecular crystallization." Journal of Applied Crystallography 34(2): 97-101. Cabrita, L. and S. Bottomley (2004). "How do proteins avoid becoming too stable? Biophysical studies into metastable proteins." European Biophysics Journal 33(2): 83-88. Carrell, R. W. and D. A. Lomas (1997). "Conformational disease." The Lancet 350(9071): 134-138. Carrell, R. W., P. E. Stein, G. Fermi and M. R. Wardell (1994). "Biological implications of a 3 a structure of dimeric antithrombin." Structure 2(4): 257-270. Carter, P. and J. A. Wells (1988). "Dissecting the catalytic triad of a serine protease." Nature 332(6164): 564-568. Chaillan-Huntington, C. E. and P. A. Patston (1998). "Influence of the P5 Residue on alpha 1-Proteinase Inhibitor Mechanism." J. Biol. Chem. 273(8): 4569-4573. Chandra, N., K. R. Acharya and P. C. Moody (1999). "Analysis and characterization of data from twinned crystals." Acta Crystallogr D Biol Crystallogr 55(Pt 10): 1750-8. Church, W. B., A. S. Inglis, A. Tseng, R. Duell, P. W. Lei, K. J. Bryant and K. F. Scott (2001). "A novel approach to the design of inhibitors of human secreted phospholipase A2 based on native peptide inhibition." J Biol Chem 276(35): 33156-64. Cunningham, M. A., M. A. Blajchman and W. P. Sheffield (1997). "IMPACT OF MUTATIONS AT THE P4 AND P5 POSITIONS ON THE REACTION OF ANTITHROMBIN WITH THROMBIN AND ELASTASE." Thrombosis Research 88(2): 171-181.

203

Davis, I. W., L. W. Murray, J. S. Richardson and D. C. Richardson (2004). "MOLPROBITY: structure validation and all-atom contact analysis for nucleic acids and their complexes." Nucleic Acids Research 32(suppl_2): W615-619. Dementiev, A., J. Dobo and P. G. W. Gettins (2006). "Active Site Distortion Is Sufficient for Proteinase Inhibition by Serpins: STRUCTURE OF THE COVALENT COMPLEX OF {alpha}1-PROTEINASE INHIBITOR WITH PORCINE PANCREATIC ELASTASE." J. Biol. Chem. 281(6): 3452-3457. Dementiev, A., M. Petitou, J.-M. Herbert and P. G. W. Gettins (2004). "The ternary complex of antithrombin-anhydrothrombin-heparin reveals the basis of inhibitor specificity." Nat Struct Mol Biol 11(9): 863-867. Dementiev, A., M. Petitou, J. M. Herbert and P. G. W. Gettins (2004). "Antithrombin[ndash]anhydrothrombin[ndash]heparin ternary complex structure reveals the basis of inhibitor specificity." Nat. Struct. Mol. Biol. Dementiev, A., M. Simonovic, K. Volz and P. G. Gettins (2003). "Canonical inhibitor- like interactions explain reactivity of [alpha]1-proteinase inhibitor Pittsburgh and antithrombin with proteinases." J. Biol. Chem. 278: 37881-37887. Dementiev, A., M. Simonovic, K. Volz and P. G. W. Gettins (2003). "Canonical Inhibitor-like Interactions Explain Reactivity of {alpha}1-Proteinase Inhibitor Pittsburgh and Antithrombin with Proteinases." J. Biol. Chem. 278(39): 37881- 37887. Devraj-Kizuk, R., D. H. Chui, E. V. Prochownik, C. J. Carter, F. A. Ofosu and M. A. Blajchman (1988). "Antithrombin-III-Hamilton: a gene with a point mutation (guanine to adenine) in codon 382 causing impaired serine protease reactivity." Blood 72(5): 1518-1523. Di Giusto, D. A., A. P. R. Sutherland, L. Jankova, S. J. Harrop, P. M. G. Curmi and G. C. King (2005). "Plasminogen Activator Inhibitor-2 is Highly Tolerant to P8 Residue Substitution--Implications for Serpin Mechanistic Model and Prediction of nsSNP Activities." Journal of Molecular Biology 353(5): 1069-1080. Dixon, M. M. and B. W. Matthews (1989). "Is gamma-chymotrypsin a tetrapeptide acyl- enzyme adduct of alpha-chymotrypsin?" Biochemistry 28(17): 7033-8. Dosztanyi, Z., V. Csizmok, P. Tompa and I. Simon (2005). "IUPred: web server for the prediction of intrinsically unstructured regions of proteins based on estimated energy content." Bioinformatics 21(16): 3433-3434. Dosztanyi, Z., V. Csizmok, P. Tompa and I. Simon (2005). "The Pairwise Energy Content Estimated from Amino Acid Composition Discriminates between Folded and Intrinsically Unstructured Proteins." Journal of Molecular Biology 347(4): 827-839. Dunstone, M. A., W. Dai, J. C. Whisstock, J. Rossjohn, R. N. Pike, S. C. Feil, B. F. Le Bonniec, M. W. Parker and S. P. Bottomley (2000). "Cleaved antitrypsin polymers at atomic resolution." Protein Sci 9(2): 417-420. Elliott, P. R., D. A. Lomas, R. W. Carrell and J. P. Abrahams (1996). "Inhibitory conformation of the reactive loop of [alpha]1-antitrypsin." Nat Struct Mol Biol 3(8): 676-681.

204

Engh, R., H. Lobermann, M. Schneider, G. Wiegand, R. Huber and C.-B. Laurell (1989). "The S variant of human {alpha}1-antitrypsin, structure and implications for function and metabolism." Protein Eng. 2(6): 407-415. Fulton, K. F., A. M. Buckle, L. D. Cabrita, J. A. Irving, R. E. Butcher, I. Smith, S. Reeve, A. M. Lesk, S. P. Bottomley, J. Rossjohn and J. C. Whisstock (2005). "The High Resolution Crystal Structure of a Native Thermostable Serpin Reveals the Complex Mechanism Underpinning the Stressed to Relaxed Transition." J. Biol. Chem. 280(9): 8435-8442. Gettins, P. G. W. (2002). "Serpin Structure, Mechanism, and Function." Chem. Rev. 102(12): 4751-4804. Ghosh, P. (2004). "Process of Protein Transport by the Type III Secretion System." Microbiol. Mol. Biol. Rev. 68(4): 771-795. Gooptu, B., B. Hazes, W.-S. W. Chang, T. R. Dafforn, R. W. Carrell, R. J. Read and D. A. Lomas (2000). "Inactive conformation of the serpin alpha 1-antichymotrypsin indicates two-stage insertion of the reactive loop: Implications for inhibitory function and conformational disease." Proceedings of the National Academy of Sciences 97(1): 67-72. Harel, M., C. T. Su, F. Frolow, I. Silman and J. L. Sussman (1991). "Gamma- chymotrypsin is a complex of alpha-chymotrypsin with its own autolysis products." Biochemistry 30(21): 5217-25. Harrop, S. J., L. Jankova, M. Coles, D. Jardine, J. S. Whittaker, A. R. Gould, A. Meister, G. C. King, B. C. Mabbutt and P. M. G. Curmi (1999). "The crystal structure of plasminogen activator inhibitor 2 at 2.0 A resolution: implications for serpin function." Structure 7(1): 43-54. Hedstrom, L. (2002). "Serine Protease Mechanism and Specificity." Chem. Rev. 102(12): 4501-4524. Hendrickson, W. A., J. R. Horton and D. M. LeMaster (1990). "Selenomethionyl proteins produced for analysis by multiwavelength anomalous diffraction (MAD): a vehicle for direct determination of three-dimensional structure." Embo J 9(5): 1665-72. Huber, R., D. Kukla, W. Bode, P. Schwager, K. Bartels, J. Deisenhofer and W. Steigemann (1974). "Structure of the complex formed by bovine trypsin and bovine pancreatic trypsin inhibitor : II. Crystallographic refinement at 1.9 A resolution." Journal of Molecular Biology 89(1): 73-101. Huntington, J. A., M. Kjellberg and J. Stenflo (2003). "Crystal Structure of Protein C Inhibitor Provides Insights into Hormone Binding and Heparin Activation." Structure 11(2): 205-215. Huntington, J. A., A. McCoy, K. J. Belzar, X. Y. Pei, P. G. W. Gettins and R. W. Carrell (2000). "The Conformational Activation of Antithrombin. A 2.85-A STRUCTURE OF A FLUORESCEIN DERIVATIVE REVEALS AN ELECTROSTATIC LINK BETWEEN THE HINGE AND HEPARIN BINDING REGIONS." J. Biol. Chem. 275(20): 15377-15383. Huntington, J. A., N. S. Pannu, B. Hazes, R. J. Read, D. A. Lomas and R. W. Carrell (1999). "A 2.6 a structure of a serpin polymer and implications for conformational disease." Journal of Molecular Biology 293(3): 449-455.

205

Huntington, J. A., R. J. Read and R. W. Carrell (2000). "Structure of a serpin-protease complex shows inhibition by deformation." Nature 407(6806): 923-926. Hyun Kyu, S., L. Kee Nyung, K. Ki-Sun, Y. Myeong-Hee and S. Se Won (1995). "Crystal structure of an uncleaved [alpha]1-antitrypsin reveals the conformation of its inhibitory reactive loop." FEBS Letters 377(2): 150-154. Ibarra, C. A., G. E. Blouse, T. D. Christian and J. D. Shore (2004). "The Contribution of the Exosite Residues of Plasminogen Activator Inhibitor-1 to Proteinase Inhibition." J. Biol. Chem. 279(5): 3643-3650. Im, H., E. J. Seo and M.-H. Yu (1999). "Metastability in the Inhibitory Mechanism of Human alpha 1-Antitrypsin." J. Biol. Chem. 274(16): 11072-11077. Im, H., M.-S. Woo, K. Y. Hwang and M.-H. Yu (2002). "Interactions Causing the Kinetic Trap in Serpin Protein Folding." J. Biol. Chem. 277(48): 46347-46354. Irving, J. A., L. D. Cabrita, J. Rossjohn, R. N. Pike, S. P. Bottomley and J. C. Whisstock (2003). "The 1.5 A Crystal Structure of a Prokaryote Serpin: Controlling Conformational Change in a Heated Environment." Structure 11(4): 387-397. Irving, J. A., R. N. Pike, A. M. Lesk and J. C. Whisstock (2000). "Phylogeny of the Serpin Superfamily: Implications of Patterns of Amino Acid Conservation for Structure and Function." Genome Res. 10(12): 1845-1864. Janin, J. and C. Chothia (1990). "The structure of protein-protein recognition sites." J. Biol. Chem. 265(27): 16027-16030. Jankova, L., S. J. Harrop, D. N. Saunders, J. L. Andrews, K. C. Bertram, A. R. Gould, M. S. Baker and P. M. G. Curmi (2001). "Crystal Structure of the Complex of Plasminogen Activator Inhibitor 2 with a Peptide Mimicking the Reactive Center Loop." J. Biol. Chem. 276(46): 43374-43382. Jiang, H. and M. R. Kanost (1997). "Characterization and Functional Analysis of 12Naturally Occurring Reactive Site Variants of Serpin-1 from Manduca sexta." J. Biol. Chem. 272(2): 1082-1087. Jiang, H., A. B. Mulnix and M. R. Kanost (1995). "Expression and characterization of recombinant Manduca sexta serpin- 1B and site-directed mutants that change its inhibitory selectivity." Insect Biochemistry and Molecular Biology 25(10): 1093- 1100. Jin, L. (1997). "The anticoagulant activation of antithrombin by heparin." Proc. Natl. Acad. Sci. USA 94: 14683-14688. Jin, L., J. P. Abrahams, R. Skinner, M. Petitou, R. N. Pike and R. W. Carrell (1997). "The anticoagulant activation of antithrombin by heparin." Proc Natl Acad Sci U S A 94(26): 14683-8. Jin, L., J. P. Abrahams, R. Skinner, M. Petitou, R. N. Pike and R. W. Carrell (1997). "The anticoagulant activation of antithrombin by 쟦 eparin." PNAS 94(26): 14683- 14688. Jogl, G., S. Rozovsky, A. E. McDermott and L. Tong (2003). "Optimal alignment for enzymatic proton transfer: Structure of the Michaelis complex of triosephosphate isomerase at 1.2-A resolution." PNAS 100(1): 50-55. Johnson, D. J. and J. A. Huntington (2004). "The influence of hinge region residue Glu381 on antithrombin allostery and metastability." J. Biol. Chem. 279: 4913- 4921.

206

Johnson, D. J., W. Li, T. E. Adams and J. A. Huntington (2006). "Antithrombin-S195A factor Xa-heparin structure reveals the allosteric mechanism of antithrombin activation." Embo J 25(9): 2029-37. Johnson, D. J. D. and J. A. Huntington (2003). "Crystal Structure of Antithrombin in a Heparin-Bound Intermediate State." Biochemistry 42(29): 8712-8719. Johnson, D. J. D., J. Langdown, W. Li, S. A. Luis, T. P. Baglin and J. A. Huntington (2006). "Crystal Structure of Monomeric Native Antithrombin Reveals a Novel Reactive Center Loop Conformation." J. Biol. Chem. 281(46): 35478-35486. Jones, T. A., J. Y. Zou, S. W. Cowan and M. Kjeldgaard (1991). "Improved methods for building protein models in electron density maps and the location of errors in these models." Acta Crystallogr A 47 ( Pt 2): 110-9. Kang, U.-B., J.-H. Baek, S.-H. Ryu, J. Kim, M.-H. Yu and C. Lee (2004). "Kinetic mechanism of protease inhibition by [alpha]1-antitrypsin." Biochemical and Biophysical Research Communications 323(2): 409-415. Kantardjieff, K. A. and B. Rupp (2003). "Matthews coefficient probabilities: Improved estimates for unit cell contents of proteins, DNA, and protein-nucleic acid complex crystals." Protein Sci 12(9): 1865-1871. Kim, S.-J., J.-R. Woo, E. J. Seo, M.-H. Yu and S.-E. Ryu (2001). "A 2.1 A resolution structure of an uncleaved [alpha]1-antitrypsin shows variability of the reactive center and other loops." Journal of Molecular Biology 306(1): 109-119. Krishnan, R., E. Zhang, K. Hakansson, R. K. Arni, A. Tulinsky, M. S. L. Lim-Wilby, O. E. Levy, J. E. Semple and T. K. Brunck (1998). "Highly Selective Mechanism- Based Thrombin Inhibitors: Structures of Thrombin and Trypsin Inhibited with Rigid Peptidyl Aldehydes." Biochemistry 37(35): 12094-12103. Krowarsch, D., T. Cierpicki, F. Jelen and J. Otlewski (2003). "Canonical protein inhibitors of serine proteases." Cellular and Molecular Life Sciences (CMLS) 60(11): 2427-2444. Kvassman, J. O., I. Verhamme and J. D. Shore (1998). "Inhibitory Mechanism of Serpins: Loop Insertion Forces Acylation of Plasminogen Activator by Plasminogen Activator Inhibitor-1." Biochemistry 37(44): 15491-15502. Laskowski, M. and I. Kato (1980). "Protein Inhibitors of Proteinases." Annual Review of Biochemistry 49(1): 593-626. Laskowski, R. A., D. S. Moss and J. M. Thornton (1993). "Main-chain Bond Lengths and Bond Angles in Protein Structures." Journal of Molecular Biology 231(4): 1049- 1067. Law, R., Q. Zhang, S. McGowan, A. Buckle, G. Silverman, W. Wong, C. Rosado, C. Langendorf, R. Pike, P. Bird and J. Whisstock (2006). "An overview of the serpin superfamily." Genome Biology 7(5): 216. Law, R. H. P., J. A. Irving, A. M. Buckle, K. Ruzyla, M. Buzza, T. A. Bashtannyk- Puhalovich, T. C. Beddoe, K. Nguyen, D. M. Worrall, S. P. Bottomley, P. I. Bird, J. Rossjohn and J. C. Whisstock (2005). "The High Resolution Crystal Structure of the Human Tumor Suppressor Maspin Reveals a Novel Conformational Switch in the G-helix." J. Biol. Chem. 280(23): 22356-22364. Lawrence, D. A., S. T. Olson, S. Muhammad, D. E. Day, J.-O. Kvassman, D. Ginsburg and J. D. Shore (2000). "Partitioning of Serpin-Proteinase Reactions between

207

Stable Inhibition and Substrate Cleavage Is Regulated by the Rate of Serpin Reactive Center Loop Insertion into beta -Sheet A." J. Biol. Chem. 275(8): 5839- 5844. Levy, N. J., N. Ramesh, M. Cicardi, R. A. Harrison and Davis Ae, III (1990). "Type II Hereditary Angioneurotic Edema That May Result from a Single Nucleotide Change in the Codon for Alanine-436 in the C1 Inhibitor Gene." Proceedings of the National Academy of Sciences 87(1): 265-268. Li, J., Z. Wang, B. Canagarajah, H. Jiang, M. Kanost and E. J. Goldsmith (1999). "The structure of active serpin 1K from Manduca sexta." Structure 7(1): 103-109. Li, J., Z. Wang, B. Canagarajah, H. Jiang, M. Kanost and E. J. Goldsmith (1999). "The structure of active serpin 1K from Manduca sexta." Structure 7(1): 103-9. Li, W., T. E. Adams, M. Kjellberg, J. Stenflo and J. A. Huntington (2007). "Structure of Native Protein C Inhibitor Provides Insight into Its Multiple Functions." J. Biol. Chem. 282(18): 13759-13768. Li, W., D. J. D. Johnson, C. T. Esmon and J. A. Huntington (2004). "Structure of the antithrombin-thrombin-heparin ternary complex reveals the antithrombotic mechanism of heparin." Nat Struct Mol Biol 11(9): 857-862. Loebermann, H., R. Tokuoka, J. Deisenhofer and R. Huber (1984). "Human [alpha]1- proteinase inhibitor : Crystal structure analysis of two crystal modifications, molecular model and preliminary analysis of the implications for function." Journal of Molecular Biology 177(3): 531-557. Lomas, D. A. and R. W. Carrell (2002). "Serpinopathies and the conformational dementias." Nat Rev Genet 3(10): 759-768. Lomas, D. A., D. Li-Evans, J. T. Finch and R. W. Carrell (1992). "The mechanism of Z [alpha]1-antitrypsin accumulation in the liver." Nature 357(6379): 605-607. Lukacs, C. M., H. Rubin and D. W. Christianson (1998). "Engineering an Anion-Binding Cavity in Antichymotrypsin Modulates the "Spring-Loaded" Serpin-Protease Interaction." Biochemistry 37(10): 3297-3304. Lukacs, C. M., J. Q. Zhong, M. I. Plotnick, H. Rubin, B. S. Cooperman and D. W. Christianson (1996). "Arginine substitutions in the hinge region of antichymotrypsin affect serpin [beta]-sheet rearrangement." Nat Struct Mol Biol 3(10): 888-893. Madison, E. L., E. J. Goldsmith, R. D. Gerard, M. J. Gething and J. F. Sambrook (1989). "Serpin-resistant mutants of human tissue-type plasminogen activator." Nature 339(6227): 721-4. Mahadeva, R., T. R. Dafforn, R. W. Carrell and D. A. Lomas (2002). "6-mer peptide selectively anneals to a pathogenic serpin conformation and blocks polymerization. Implications for the prevention of Z alpha(1)-antitrypsin-related cirrhosis." J Biol Chem 277(9): 6771-4. McPhalen, C. A. and M. N. G. James (1987). "Crystal and molecular structure of the serine proteinase inhibitor CI-2 from barley seeds." Biochemistry 26(1): 261-269. Means, G. E. and R. E. Feeney (1968). "Reductive alkylation of amino groups in proteins." Biochemistry 7(6): 2192-2201.

208

Mottonen, J., A. Strand, J. Symersky, R. M. Sweet, D. E. Danley, K. F. Geoghegan, R. D. Gerard and E. J. Goldsmith (1992). "Structural basis of latency in plasminogen activator inhibitor-1." Nature 355(6357): 270-273. Mottonen, J., A. Strand, J. Symersky, R. M. Sweet, D. E. Danley, K. F. Geoghegan, R. D. Gerard and E. J. Goldsmith (1992). "Structural basis of latency in plasminogen activator inhibitor-1." Nature 355(6357): 270-3. Mourey, L., J.-P. Samama, M. Delarue, M. Petitou, J. Choay and D. Moras (1993). "Crystal Structure of Cleaved Bovine Antithrombin III at 3[middle dot]2 A Resolution." Journal of Molecular Biology 232(1): 223-241. Mukherjee, S., G. Keitany, Y. Li, Y. Wang, H. L. Ball, E. J. Goldsmith and K. Orth (2006). "Yersinia YopJ Acetylates and Inhibits Kinase Activation by Blocking Phosphorylation." Science 312(5777): 1211-1214. Nar, H., M. Bauer, J.-M. Stassen, D. Lang, A. Gils and P. J. Declerck (2000). "Plasminogen activator inhibitor 1. Structure of the native serpin, comparison to its other conformers and implications for serpin inactivation." Journal of Molecular Biology 297(3): 683-695. Navaza, J. (2001). "Implementation of molecular replacement in AMoRe." Acta Crystallogr D Biol Crystallogr 57(Pt 10): 1367-72. Notredame, C., D. G. Higgins and J. Heringa (2000). "T-coffee: a novel method for fast and accurate multiple sequence alignment." Journal of Molecular Biology 302(1): 205-217. O'Malley, K. M., S. A. Nair, H. Rubin and B. S. Cooperman (1997). "The Kinetic Mechanism of Serpin-Proteinase Complex Formation. AN INTERMEDIATE BETWEEN THE MICHAELIS COMPLEX AND THE INHIBITED COMPLEX." J. Biol. Chem. 272(8): 5354-5359. Olson, S. T. and I. Bjork (1991). "Predominant contribution of surface approximation to the mechanism of heparin acceleration of the antithrombin-thrombin reaction. Elucidation from salt concentration effects." J. Biol. Chem. 266: 6353-6364. Olson, S. T., P. E. Bock, J. Kvassman, J. D. Shore, D. A. Lawrence, D. Ginsburg and I. Bj?k (1995). "Role of the Catalytic Serine in the Interactions of Serine Proteinases with Protein Inhibitors of the Serpin Family." J. Biol. Chem. 270(50): 30007-30017. Olson, S. T., R. Swanson, D. Day, I. Verhamme, J. Kvassman and J. D. Shore (2001). "Resolution of Michaelis Complex, Acylation, and Conformational Change Steps in the Reactions of the Serpin, Plasminogen Activator Inhibitor-1, with Tissue Plasminogen Activator and Trypsin." Biochemistry 40(39): 11742-11756. Orth, K. (2002). "Function of the Yersinia effector YopJ." Current Opinion in Microbiology 5(1): 38-43. Orth, K., L. E. Palmer, Z. Q. Bao, S. Stewart, A. E. Rudolph, J. B. Bliska and J. E. Dixon (1999). "Inhibition of the Mitogen-Activated Protein Kinase Kinase Superfamily by a Yersinia Effector." Science 285(5435): 1920-1923. Otwinowski, Z., W. Minor and Charles W. Carter, Jr. (1997). [20] Processing of X-ray diffraction data collected in oscillation mode. Methods in Enzymology, Academic Press. Volume 276: 307-326.

209

Owen, M. C., S. O. Brennan, J. H. Lewis and R. W. Carrell (1983). "Mutation of antitrypsin to antithrombin. alpha 1-antitrypsin Pittsburgh (358 Met leads to Arg), a fatal bleeding disorder." N Engl J Med 309(12): 694-698. Papamokos, E., E. Weber, W. Bode, R. Huber, M. W. Empie, I. Kato and M. Laskowski (1982). "Crystallographic refinement of Japanese quail ovomucoid, a Kazal-type inhibitor, and model building studies of complexes with serine proteases." Journal of Molecular Biology 158(3): 515-537. Parsons, S. (2003). "Introduction to twinning." Acta Crystallogr D Biol Crystallogr 59(Pt 11): 1995-2003. Pasternak, A., D. Ringe and L. Hedstrom (1999). "Comparison of anionic and cationic trypsinogens: the anionic activation domain is more flexible in solution and differs in its mode of BPTI binding in the crystal structure." Protein Sci 8(1): 253-258. Perry, D. J., P. L. Harper, S. Fairham, M. Daly and R. W. Carrell (1989). "Antithrombin Cambridge, 384 Ala to Pro: A new variant identified using the polymerase chain reaction." FEBS Letters 254(1-2): 174-176. Potempa, J., E. Korzus and J. Travis (1994). "The serpin superfamily of proteinase inhibitors: structure, function, and regulation." J. Biol. Chem. 269(23): 15957- 15960. Radisky, E. S., J. M. Lee, C.-J. K. Lu and D. E. Koshland, Jr. (2006). "Insights into the serine protease mechanism from atomic resolution structures of trypsin reaction intermediates." PNAS 103(18): 6835-6840. Rayment, I. and Charles W. Carter, Jr. (1997). [12] Reductive alkylation of lysine residues to alter crystallization properties of proteins. Methods in Enzymology, Academic Press. Volume 276: 171-179. Renatus, M., Q. Zhou, H. R. Stennicke, S. J. Snipas, D. Turk, L. A. Bankston, R. C. Liddington and G. S. Salvesen (2000). "Crystal structure of the apoptotic suppressor CrmA in its cleaved form." Structure 8(7): 789-797. Rypniewski, W. R., H. M. Holden and I. Rayment (1993). "Structural consequences of reductive methylation of lysine residues in hen egg white lysozyme: An x-ray analysis at 1.8-.ANG. resolution." Biochemistry 32(37): 9851-9858. Ryu, S.-E., H.-J. Choi, K.-S. Kwon, K. N. Lee and M.-H. Yu (1996). "The native strains in the hydrophobic core and flexible reactive loop of a serine protease inhibitor: crystal structure of an uncleaved [alpha]1-antitrypsin at 2.7 a." Structure 4(10): 1181-1192. Schechter, I. and A. Berger (1968). "On the active site of proteases. III. Mapping the active site of papain; specific peptide inhibitors of papain." Biochemical and Biophysical Research Communications 32(5): 898-902. Schreuder, H. A., B. de Boer, R. Dijkema, J. Mulders, H. J. M. Theunissen, P. D. J. Grootenhuis and W. G. J. Hol (1994). "The intact and cleaved human antithrombin III complex as a model for serpin-proteinase interactions." Nat Struct Mol Biol 1(1): 48-54. Seo, E. J., H. Im, J.-S. Maeng, K. E. Kim and M.-H. Yu (2000). "Distribution of the Native Strain in Human alpha 1-Antitrypsin and Its Association with Protease Inhibitor Function." J. Biol. Chem. 275(22): 16904-16909.

210

Sharp, A. M., P. E. Stein, N. S. Pannu, R. W. Carrell, M. B. Berkenpas, D. Ginsburg, D. A. Lawrence and R. J. Read (1999). "The active conformation of plasminogen activator inhibitor 1, a target for drugs to control fibrinolysis and cell adhesion." Structure 7(2): 111-118. Shin, J.-S. and M.-H. Yu (2002). "Kinetic Dissection of alpha 1-Antitrypsin Inhibition Mechanism." J. Biol. Chem. 277(14): 11629-11635. Shin, J.-S. and M.-H. Yu (2006). "Viscous Drag as the Source of Active Site Perturbation during Protease Translocation: Insights into how Inhibitory Processes are Controlled by Serpin Metastability." Journal of Molecular Biology 359(2): 378- 389. Shore, J. D., D. E. Day, A. M. Francis-Chmura, I. Verhamme, J. Kvassman, D. A. Lawrence and D. Ginsburg (1995). "A Fluorescent Probe Study of Plasminogen Activator Inhibitor-1." J. Biol. Chem. 270(10): 5395-5398. Silverman, G. A., P. I. Bird, R. W. Carrell, F. C. Church, P. B. Coughlin, P. G. W. Gettins, J. A. Irving, D. A. Lomas, C. J. Luke, R. W. Moyer, P. A. Pemberton, E. Remold-O'Donnell, G. S. Salvesen, J. Travis and J. C. Whisstock (2001). "The Serpins Are an Expanding Superfamily of Structurally Similar but Functionally Diverse Proteins. EVOLUTION, MECHANISM OF INHIBITION, NOVEL FUNCTIONS, AND A REVISED NOMENCLATURE." J. Biol. Chem. 276(36): 33293-33296. Silverman, G. A. and D. A. Lomas (2004). "Serpin identification, production, and characterization." Methods 32(2): 71-72. Simonovic, M., P. G. W. Gettins and K. Volz (2000). "Crystal structure of viral serpin crmA provides insights into its mechanism of cysteine proteinase inhibition [In Process Citation]." Protein Sci 9(8): 1423-1427. Simonovic, M., P. G. W. Gettins and K. Volz (2001). "Crystal structure of human PEDF, a potent anti-angiogenic and neurite growth-promoting factor." Proceedings of the National Academy of Sciences 98(20): 11131-11135. Singh, N., T. Jabeen, S. Sharma, I. Roy, M. N. Gupta, S. Bilgrami, R. K. Somvanshi, S. Dey, M. Perbandt, C. Betzel, A. Srinivasan and T. P. Singh (2005). "Detection of native peptides as potent inhibitors of enzymes. Crystal structure of the complex formed between treated bovine α-chymotrypsin and an autocatalytically produced fragment, Ile-Val-Asn-Gly-Glu-Glu-Ala-Val-Pro-Gly-Ser-Trp-Pro-Trp, at 2.2 Å resolution." FEBS Journal 272(2): 562-572. Skinner, R., J.-P. Abrahams, J. C. Whisstock, A. M. Lesk, R. W. Carrell and M. R. Wardell (1997). "The 2.6 A structure of antithrombin indicates a conformational change at the heparin binding site." Journal of Molecular Biology 266(3): 601- 609. Skinner, R., W.-S. W. Chang, L. Jin, X. Pei, J. A. Huntington, J.-P. Abrahams, R. W. Carrell and D. A. Lomas (1998). "Implications for function and therapy of a 2.9 a structure of binary-complexed antithrombin." Journal of Molecular Biology 283(1): 9-14. Sohl, J. L., S. S. Jaswal and D. A. Agard (1998). "Unfolded conformations of alpha-lytic protease are more stable than its native state

211

Unfolded conformations of alpha-lytic protease are more stable than its native state." Nature 395(6704): 817. Stanley, P. and P. E. Stein (2003). "BmSPN2, a Serpin Secreted by the Filarial Nematode Brugia malayi, Does Not Inhibit Human Neutrophil Proteinases but Plays a Noninhibitory Role." Biochemistry 42(20): 6241-6248. Stein, P. E. and R. W. Carrell (1995). "What do dysfunctional serpins tell us about molecular mobility and disease?" Nat Struct Mol Biol 2(2): 96-113. Stein, P. E., A. G. W. Leslie, J. T. Finch and R. W. Carrell (1991). "Crystal structure of uncleaved ovalbumin at 1[middle dot]95 A resolution." Journal of Molecular Biology 221(3): 941-959. Stein, P. E., A. G. W. Leslie, J. T. Finch, W. G. Turnell, P. J. McLaughlin and R. W. Carrell (1990). "Crystal structure of ovalbumin as a model for the reactive centre of serpins." Nature 347(6288): 99-102. Storoni, L. C., A. J. McCoy and R. J. Read (2004). "Likelihood-enhanced fast rotation functions." Acta Crystallogr D Biol Crystallogr 60(Pt 3): 432-8. Stout, T. J., H. Graham, D. I. Buckley and D. J. Matthews (2000). "Structures of Active and Latent PAI-1: A Possible Stabilizing Role for Chloride Ions." Biochemistry 39(29): 8460-8469. Stratikos, E. and P. G. W. Gettins (1999). "Formation of the covalent serpin-proteinase complex involves translocation of the proteinase by more than 70 A and full insertion of the reactive center loop into beta -sheet A." Proceedings of the National Academy of Sciences 96(9): 4808-4813. Trosky, J. E., Y. Li, S. Mukherjee, G. Keitany, H. Ball and K. Orth (2007). "VopA inhibits ATP binding by acetylating the catalytic loop of MKKs." J. Biol. Chem.: M706970200. Trosky, J. E., S. Mukherjee, D. L. Burdette, M. Roberts, L. McCarter, R. M. Siegel and K. Orth (2004). "Inhibition of MAPK Signaling Pathways by VopA from Vibrio parahaemolyticus." J. Biol. Chem. 279(50): 51953-51957. Tucker, H. M., J. Mottonen, E. J. Goldsmith and R. D. Gerard (1995). "Engineering of plasminogen activator inhibitor-1 to reduce the rate of latency transition." Nat Struct Mol Biol 2(6): 442-445. Viboud, G. I. and J. B. Bliska (2005). "Yersinia Outer Proteins: Role in Modulation of Host Cell Signaling Responses and Pathogenesis." Annual Review of Microbiology 59(1): 69-89. Walter, T. S., C. Meier, R. Assenberg, K.-F. Au, J. Ren, A. Verma, J. E. Nettleship, R. J. Owens, D. I. Stuart and J. M. Grimes (2006). "Lysine Methylation as a Routine Rescue Strategy for Protein Crystallization." Structure 14(11): 1617-1622. Wang, Z., J. Mottonen and E. J. Goldsmith (1996). "Kinetically Controlled Folding of the Serpin Plasminogen Activator Inhibitor 1." Biochemistry 35(51): 16443-16448. Wei, A., H. Rubin, B. S. Cooperman and D. W. Christianson (1994). "Crystal structure of an uncleaved serpin reveals the conformation of an inhibitory reactive loop." Nat Struct Mol Biol 1(4): 251-258. Whisstock, J., R. Skinner and A. M. Lesk (1998). "An atlas of serpin conformations." Trends in Biochemical Sciences 23(2): 63-67.

212

Whisstock, J. C., R. N. Pike, L. Jin, R. Skinner, X. Y. Pei, R. W. Carrell and A. M. Lesk (2000). "Conformational changes in serpins: II. The mechanism of activation of antithrombin by heparindagger." J Mol Biol 301(5): 1287-305. Whisstock, J. C., R. Skinner, R. W. Carrell and A. M. Lesk (2000). "Conformational changes in serpins: I. the native and cleaved conformations of [alpha]1- antitrypsin." Journal of Molecular Biology 295(3): 651-665. Xue, Y., P. Bjorquist, T. Inghardt, M. Linschoten, D. Musil, L. Sjolin and J. Deinum (1998). "Interfering with the inhibitory mechanism of serpins: crystal structure of a complex formed between cleaved plasminogen activator inhibitor type 1 and a reactive-centre loop peptide." Structure 6(5): 627-636. Yamasaki, M., Y. Arii, B. Mikami and M. Hirose (2002). "Loop-inserted and thermostabilized structure of P1-p1' cleaved ovalbumin mutant R339T." Journal of Molecular Biology 315(2): 113-120. Yamasaki, M., N. Takahashi and M. Hirose (2003). "Crystal Structure of S-ovalbumin as a Non-loop-inserted Thermostabilized Serpin Form." J. Biol. Chem. 278(37): 35524-35530. Ye, S., A. L. Cech, R. Belmares, R. C. Bergstrom, Y. Tong, D. R. Corey, M. R. Kanost and E. J. Goldsmith (2001). "The structure of a Michaelis serpin[ndash]protease complex." Nat. Struct. Biol. 8: 979-983. Ye, S., A. L. Cech, R. Belmares, R. C. Bergstrom, Y. Tong, D. R. Corey, M. R. Kanost and E. J. Goldsmith (2001). "The structure of a Michaelis serpin-protease complex." Nat Struct Biol 8(11): 979-83. Ye, S. and E. J. Goldsmith (2001). "Serpins and other covalent protease inhibitors." Current Opinion in Structural Biology 11(6): 740-745. Yeates, T. O. (1997). "Detecting and overcoming crystal twinning." Methods Enzymol 276: 344-58. Yeates, T. O. and Charles W. Carter, Jr. (1997). [22] Detecting and overcoming crystal twinning. Methods in Enzymology, Academic Press. Volume 276: 344-358. Zang, X. and R. M. Maizels (2001). "Serine proteinase inhibitors from nematodes and the arms race between host and pathogen." Trends in Biochemical Sciences 26(3): 191-197. Zang, X., M. Yazdanbakhsh, H. Jiang, M. R. Kanost and R. M. Maizels (1999). "A Novel Serpin Expressed by Blood-Borne Microfilariae of the Parasitic Nematode Brugia malayi Inhibits Human Neutrophil Serine Proteinases." Blood 94(4): 1418-1428. Zhou, A., R. W. Carrell and J. A. Huntington (2001). "The Serpin Inhibitory Mechanism Is Critically Dependent on the Length of the Reactive Center Loop." J. Biol. Chem. 276(29): 27541-27547. Zhou, A., J. A. Huntington, N. S. Pannu, R. W. Carrell and R. J. Read (2003). "How vitronectin binds PAI-1 to modulate fibrinolysis and cell migration." Nat Struct Mol Biol 10(7): 541-544. Zhou, A., P. E. Stein, J. A. Huntington and R. W. Carrell (2003). "Serpin Polymerization Is Prevented by a Hydrogen Bond Network That Is Centered on His-334 and Stabilized by Glycerol." J. Biol. Chem. 278(17): 15116-15122.

213

Zhou, A., P. E. Stein, J. A. Huntington, P. Sivasothy, D. A. Lomas and R. W. Carrell (2004). "How Small Peptides Block and Reverse Serpin Polymerisation." Journal of Molecular Biology 342(3): 931-941. Zhou, A., Z. Wei, R. J. Read and R. W. Carrell (2006). "Structural mechanism for the carriage and release of thyroxine in the blood." PNAS 103(36): 13321-13326. Zhu, Y., Y. Wang, M. J. Gorman, H. Jiang and M. R. Kanost (2003). "Manduca sexta Serpin-3 Regulates Prophenoloxidase Activation in Response to Infection by Inhibiting Prophenoloxidase-activating Proteinases." J. Biol. Chem. 278(47): 46556-46564.

VITAE

Soon-Hee Sul was born in Seoul, Korea, on Jun 8, 1972. She is the daughter of Young-

Suk Sul and Myung-Ja Kim, and the sister of Boung-Gyu Sul. After completing her work at Mi-Lim High School, Seoul, Korea in 1991, she entered Han-Yang University at

Seoul, Korea. Following the completion the degree of Bachelor of Science with a major in Clothing and Textiles from Han-Yang University in Feb 2005, she entered Iowa State

University, majoring in Chemical engineering. After completing the degrees of Bachelor and Master of Science with a major in Chemical engineering from Iowa State University, she entered the graduate shool of Biomedical Engineering at the University of Texas

Southwestern Medical Center at Dallas in 2001 and changed the major to Biomedical

Science in 2002. She was admitted to candidacy in the Molecular Biophysics program in

2003. In 1999, she married Seung-Jae Lee. Their son, Joshua A. Lee was born in 2004.

Permanent Address: 3633 W.Walnut Hill Ln Apt #1092

Irving, TX75038