UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Enzymatic properties of hepatitis C virus NS3 serine protease and bio-engineering of serine protease… Po, Addy 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2003-0171.pdf [ 5.95MB ]
Metadata
JSON: 831-1.0090883.json
JSON-LD: 831-1.0090883-ld.json
RDF/XML (Pretty): 831-1.0090883-rdf.xml
RDF/JSON: 831-1.0090883-rdf.json
Turtle: 831-1.0090883-turtle.txt
N-Triples: 831-1.0090883-rdf-ntriples.txt
Original Record: 831-1.0090883-source.json
Full Text
831-1.0090883-fulltext.txt
Citation
831-1.0090883.ris

Full Text

ENZYMATIC PROPERTIES OF HEPATITIS C VIRUS NS3 SERINE PROTEASE AND BIO-ENGINEERING OF SERINE PROTEASE INHIBITORS (SERPINS) AGAINST THE NS3 PROTEASE AND ELASTASE by  ADDY PO B . S c , University of British Columbia (Microbiology & Immunology), 2000  A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in  F A C U L T Y OF G R A D U A T E STUDIES (Department of Microbiology and Immunology)  We accept this thesis as confjjrming to the requiredstandard  THE UNIVERSITY OF BRITISH COLUMBIA April, 2003 © Addy Po, 2003  UBC Rare Books and Special C o l l e c t i o n s  - Thesis A u t h o r i s a t i o n Form  Page 1 of 1  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library shall make i t freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.  Department The University of B r i t i s h Columbia Vancouver, Canada Date A W n  , ^  0  Q  3  http://www.library.ubc.ca/spcol1/thesauth.html  3/28/03  Abstract Hepatitis C virus (HCV) has infected millions of people worldwide and emerged as a global health crisis. H C V NS3 serine protease domain (aa:1027-1218) of the H C V polyprotein, is required for the processing and maturation of the H C V viral non-structural (NS) proteins. As an essential protease for viral replication, H C V NS3 protease has been considered as a strategic target for anti-HCV drug development. This thesis investigates the designs and inhibitory properties of novel serine protease inhibitor (serpin) variants against the H C V NS3 protease. First, the substrate specificity of NS3 protease is evaluated for the purpose of generating protein-based inhibitors to target it. Data from kinetic experiments suggested that NS3 protease activity is greatly enhanced with the addition of its NS4A cofactor and that the three key residues in the substrate necessary for efficient recognition and cleavage by NS3 protease are cysteine in P I , serine/alanine in P ' l and aspartic/glutamic acids in P6 position. This information was used to bio-engineer 5 (serpins) variants based on ai-Antitrypsin (ai-AT) scaffold. Variants of a i - A T contained different mutations in the serpin Reactive Site Loop (RSL) involving the P I , P4, P6 & P ' l positions. The biological activities and the newly gained specificities of the a i - A T and its variants were tested mainly by performing SDS-stable complex formation studies and enzyme kinetic experiments directed against NS3 protease and elastase. Results show that NS3 protease was able to form a 72 kDa SDS-stable complex with a i - A T and 4 of its variants. Taken together, the ability of a i - A T and its variants to interact and form SDS-stable complexes with NS3 protease may represent an entry point for a novel class of protein-based H C V NS3 protease inhibitors.  ii  Table of Contents Abstract  ii  Table of Contents  iii  List of Figures  vii  List of Abbreviations  ix  Acknowledgements  xiii  Chapter 1: Introduction  1  1.1 Hepatitis C virus (HCV): Overview  1  1.1.1  The infectious agent  1  1.1.2  H C V genome organization and polyprotein processing  2  1.1.3  Currently available therapeutics for H C V patients and new targets for drug development  4  1.2 H C V NS3/NS4A protease 1.2.1  5  Identification of the NS3 protease and Full-length NS3 protease/helicase  5  1.2.2  The NS4A protease cofactor  7  1.2.3  Substrate specificity of H C V NS3/NS4A protease  9  1.2.3.1 IQFS and M C A substrate  10  1.3 Strategies for inhibition of the NS3/NS4A protease complex  12  1.3.1  H C V NS3 protease; an attractive target for anti-HCV therapy  12  1.3.2  Peptidyl Inhibitors of NS3 protease and their limitations  13  1.3.3  Protein-based Inhibitor: The case of serine protease inhibitors (serpins)  14  ai-antitrypsin (ai-AT) and bio-engineering of serpin reactive site loop (RSL)  16  1.3.4  iii  1.4 Thesis objectives  ;  er 2: Materials and methods  20  21  2.1 Preparation of competent BL21 & D H 5 a E.coli cells  21  2.2 Transformation of E. coli strain BL21 with NS3 recombinant plasmid construct  21  2.3 Expression of recombinant NS3 proteins  22  2.4 Purification of NS3 protease using Fast Protein Liquid Chromatography (FPLC)  23  2.5 Protein analysis with SDS P A G E and Western Blot  23  2.6 Development of H C V NS3 protease assays  25  2.7 Michaelis-Menten steady-state analysis of H C V synthetic peptide substrate... 26 2.8 cti-AT D N A isolation and sub-cloning  27  2.9 Site-directed mutagenesis of cti-AT using Polymerase chain reactions (PCR). 28 2.10  Epression and purification of a i - A T and its variants  31  2.11  Pancreatic elastase assays  32  2.12  Quantification of proteins  32  2.13  Complex formation studies with a i - A T and its mutant variants against elastase, NS3 protease and FLNS3  33  Inhibition assay of elastase and NS3 protease by a i - A T and all its variants  33  Model structure of NS3 protease/NS4A complex, Helicase/NTPase and cti-AT/elastase complex stucture  35  List of buffer solutions  36  2.14  2.15  2.16  iv  Chapter 3: Expressions, purification and characterization of NS3 protease using internally quenched fluorogenic substrates (IQFS)  37  3.1 Expression of recombinant NS3 protease in pLys BL21 E.coli  37  3.2 Purification of recombinant NS3 protease from E.coli supernatant  38  3.3 Optimization of assay conditions for detection of NS3 protease activity  41  3.4 Probing the substrate specificity of NS3 protease using H C V IQFS  43  3.5 Probing the substrate specificity of NS3 protease using G B V IQFS  48  3.6 Summary  51  Chapter 4: Bio-engineering, expression & purification of cti-AT variants (mutants A-E)  53  4.1 Sub-cloning of a i - A T from pDS-56 vector to pET-21 vector  53  4.2  56  Bio-engineering of a i - A T variants  4.3 Purification of a i - A T and its variants  60  4.4  63  Bio-engineering of K D E L mutant A & D  Chapter 5: Characterization of the ability of a i - A T and RSL variants to form complexes with and inhibit the proteolytic activity of elastase  65  5.1 a i - A T , a natural inhibitor of neutrophil and pancreatic elastase  65  5.2 Complex formation between ai-AT, Mutant A & D with elastase  65  5.3 Degradation of Mutant B , C & E by elastase  68  5.4 Inhibition assay of a i - A T and its variants towards elastase  72  5.5 Titration and Progress curve analysis of a i - A T and its variants towards elastase  75  v  Chapter 6: Characterization ai-AT and its variants (Mut A-E) with complex formation studies and inhibition assays against H C V NS3 protease  79  6.1 Serpins as candidates for novel protein therapeutics against H C V NS3 protease  79  6.2 Complex formation between cti-AT and its variants against NS3 protease.... 79 6.3 Studies investigating the effect of enzyme: inhibitor ratio and time of interaction upon complex formation between a i - A T and variants and NS3 protease  83  6.4 Inhibition assay of cti-AT and its variants towards NS3 protease  86  6.5 Complex formation between cti-AT and its variants against FLNS3  88  6.6 Sub-cellular targeting of recombinant serpin-based inhibitors of H C V NS3 protease in human UNS 3/4A cells  91  6.7 Summary  92  Chapter 7: Conclusion  93  7.1 Current view of H C V NS3 serine protease inhibitors 7.2  oci-Antitrypsin and its variants as potential inhibitors against H C V NS3 protease and future directions  93  94  7.3 Animal models and cellular systems for anti-HCV drugs development  95  7.4 Combination therapy for H C V infection  97  Chapter 8: References  99  Chapter 9: Appendix  110  vi  List of Figures Figure 1.1  H C V Genome/Polyprotein processing  3  Figure 1.2  Ribbon representation of FLNS3 & NS3 protease  6  Figure 1.3  NS4A: H C V NS3 protease co-factor  8  Figure 1.4  Serine protease inhibitor (Serpin)  17  Figure 1.5  Conformational polymorphism of inhibitory serpins  18  Figure 3.1  Purification of bacterially expressed histidine-tagged NS3 protease  39  Figure 3.2  NS3 protease activity titration with Pep4A2i-34  42  Figure 3.3  The intramolecularly quenched fluorogenic substrate (IQFS) are suitable for continuous monitoring of H C V NS3/4A protease complex activity  44  Figure 3.4  Substrate specificity of NS3 protease  46  Figure 3.5  NS3/Pep4A protease complex activities on peptide substrates corresponding to trans-cleavage sites NS3/Pep4A protease complex specificities towards G B V - A and G B V - B IQFS  47  Figure 3.6  Figure 3.7  49  Ribbon representation of the catalytic pocket of NS3/Pep4A protease Complex  50  Figure 4.1  Subcloning of a i - A T from pDS-56 to pET-21  55  Figure 4.2  a i - A T and bio-engineered variants  58  Figure 4.3  Ribbon representations of the backbone of the crystal structure of a i - A T wild-type and the reactive site loops of its variants mutant (A-E) Purification of bacterially expressed histidine-tagged cti-AT and its variants  Figure 4.4  59 61  Figure 4.5  SDS-PAGE analysis of a l - A T and all its variants (Mut A-E)  62  Figure 5.1  Serpin complex formation with elastase  67  Figure 5.2  Branched pathway mechanism of serpins as suicide substrate inhibitors.... 68  vii  Figure 5.3  Degradation of Mut B,C & E by elastase  Figure 5.4  Residual activity of elastase with a i - A T after different incubation time.... 73  Figure 5.5  Inhibition of elastase by cti-AT, Mut A & D  74  Figure 5.6  Inhibition constant of ai-AT, Mut A and D vs elastase  76  Figure 5.7  Progress curve of a i - A T , Mut A & D vs elastase  78  Figure 6.1  Complex formation between a i - A T and its variants vs NS3 protease  81  Figure 6.2  Complex formation between a i - A T , Mut A , D , & C with increasing [NS3]  84  Figure 6.3  70  Complex formation between a i - A T , Mut A , D , & C with NS3 protease at different time points  85  Figure 6.4  Inhibition of NS3 protease by a i - A T and its 5 variants  87  Figure 6.5  Complex formation between FLNS3 with a i - A T and all its variants  89  Figure 6.6  Co-localization of oti-AT, Mut A - K D E L & Mut D - K D E L with NS3 in human cells  90  Appendix-1  Tight binding titration of pancreatic elastase by a i - A T  110  Appendix-2  Restriction map of pET-21 (Novagen)  Ill  Appendix-3  Restriction map of p-CMV shuttle vector (Stratagene)  112  Appendix-4  Clustal W alignment of a l - A T and all its variants (amino-acids)  113  Appendix-5  Clustal W alignment of a 1 - A T and all its variants (nucleotides)  Appendix-6  Clustal W alignment of a 1 - A T - K D E L and Mut A - K D E L & Mut D - K D E L (amino-acids) Clustal W alignment of a 1-AT-KDEL and Mut A - K D E L & Mut D - K D E L (nucleotides)  Appendix-7  viii  115  119 120  List of Abbreviations aa  Amino acid  Abz  Anthranillic Acid ( O-Aminobenzoic acid)  Amp  Ampicillin  ai-AT  a 1-antitrypsin  <x2AP  a-2 antiplasmin  bp  Base pair  BSA  Bovine serum almumin  BVDV  Bovine viral diarrhea virus  DNA  Deoxyribonucleic acid  DTT  Dithiothreitol  NS3  Non-Structural 3 protease (protease domain)  eDDnp  Ethylene Diamine 2,4, Dinitrophenyl  E  Enzyme  [E]  Enzyme concentration  EI*  Enzyme/Inhibitor complex  ECL  Electro-chemoluminescence  E.coli  Escherichia coli  EDTA  ethylenediamine tetra-acetic acid  ER  Endoplasmic reticulum  FLNS3  Full length NS3 protease (protease/helicase domains)  FPLC  Fast Protein Liquid Chromnatography  GBV-A  GB virus A  ix  GBV-B  G B virus B  HAART  Highly Active Anti-Retroviral Therapy  HCMV  Human Cytomegalovirus virus  HCV  Hepatitis C virus  HEPES  (N-[2-hydroxyethyl] piperazine-N'-[2-ethane sulphonic acid])  HIV  Human Immunodeficiency virus  HTA  High Throughput Assay  I  Inhibitor  f  Cleaved Inhibitor  IFN-a  Interferon a  IPTG  Isopropyl-p-D-thiogalactosidase  IQFS  Internally Quenched Fluorogenic Substrate  IRES  Internal Ribosome Entry Sites  IV  Intravenous  Ran  Kanamycin  kbp  Kilobase pair  kDa  Kilo Dalton  K  Association constant  a s s  kcat  Catalytic constant  Kd  Dissociation constant  K;  Inhibition constant  K  Affinity constant  m  LB  Luria Bertani medium  x  M  Molar  MCA  Methyl Coumarinamide Substrate  ml  milli litre  mm  milli metre  mM  milli molar  Mut A  Mutant A  Mut B  Mutant B  Mut C  Mutant C  Mut D  Mutant D  Mut E  Mutant E  ul  micro litre  uM  micro molar  nM  nano molar  (NS)  Non-Structural  NRTIs  Nucleoside reverse transcriptase inhibitors  NNRTIs  Non-nucleoside reverse transcriptase inhibitors  PCR  Polymerase Chain Reaction  Pep4A  Peptide H C V NS4A cofactor  pH  Potential of hydrogen  pM  pico Molar  RdRp  RNA-dependent R N A polymerase  RFU  Relative Fluorescence Unit  RNA  Ribonucleic acid  rpm  Revolutions per minute  xi  RSL  Reactive site loop  SDS-PAGE  Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis  Serpin  Serine protease inhibitor  Si  Stochiometry of inhibition  [S]  Substrate concentration  TAE  Tris, acetate, E D T A  TBS  Tris-buffered saline  Triton-X-100  Octyl phenoxy polyethoxy ethanol  UV Vmav  Ultraviolet  [Vo]  Initial Velocity  Maximum velocity of enzyme  Y(N0 )  Nitrotyrosine  2  Amino acids alanine (Ala)  A  leucine (Leu)  L  arginine (Arg)  R  lysine (Lys)  K  asparagines (Asn)  N  methionine (Met)  M  aspartate (Asp)  D  phenylalanine (Phe) F  cysteine (Cys)  C  proline (Pro)  P  glutamate(Glu)  E  serine (Ser)  S  glutamine (Gin)  Q  threonine (Thr)  T  glycine (Gly)  G  Tryptophan (Trp)  W  histidine (His)  H  tyrosine (Tyr)  Y  isoleucine (De)  I  valine (Val)  V  Nucleotide Bases adenine A cytosine  C  guanine  G  thymine  T  uracil  U  xii  Acknowledments I would first like to thank my supervisor Dr. Francois Jean and supervisory committee Dr. Robert Hancock and Dr. Jim Kronstad for their support throughout my graduate school program at U B C . Not only was their guidance instrumental in the completion of this work, but they also provided me with the autonomy to develop my abilities and confidence as a scientist. I would also like to thank Dr. Luiz Juliano, Dr. Nobuko Kakiuchi and Dr. Carl Hashimoto for the IQFS, NS3 protease construct and cti-AT-KDEL construct. I am especially grateful to the anti-viral lab (Pamela, Morgan, Martin, Rebecca, Jen, Laura, and Jessie) for their knowledge and friendship. Special thanks to Pamela Hamill for her constant brilliant suggestions in every aspect of the lab, Martin Richer for aiding in technical advise and the cloning of several mutants and Morgan Martin for input with her unconventional "outside the box" ideas. I would also like to thank Helmut Kae for his technical support and friendship. Financially, I would like to thank CIHR for two years of monetary support and to "Mom and Dad" for their constant supply of delicious food and a comfortable roof to stay in. During my time as a graduate student, many friends including my long time buddies have been supportive and injected much needed fun to balance my life. I would like to thank those friends once again for all their support and friendship over the years; and best of all, guys, it is finally time to join you all in the real world!!!! Last, but certainly not least, I am especially grateful to my family and Sandy whose love, encouragement and unconditional support has been constant in my life. Your support has always been cherished.  xiii  Chapter 1: Introduction  1.1  Hepatitis C virus (HCV): Overview  1.1.1  The Infectious Agent As recently as the late 1980s, few people other than physicians had heard of the  Hepatitis C virus (HCV), a slowly progressing viral infection that over a couple of decades can lead to liver cancer (Di Bisceglie and Bacon, 1999). H C V is the causal agent for a largely chronic liver infection afflicting an estimated 170,000,000 people world-wide and is the primary cause of liver transplantation (Rosenberg, 2001). In general, H C V establishes persistent infections by inducing limited direct cell damage and/or by evading immunologic surveillance. However, the lack of serious direct cell injury allows the virus the opportunity for continued viral replication and establishment of chronic infection. About 20% of the infected individuals develop liver cirrhosis and hepato-cellular carcinoma after 20 years (Alter and Seef, 2000). H C V viral infection leads to cellular damage in vivo through two mechanisms; namely, direct cytopathicity, which is the result of the toxic actions of viral products on infected cells, and immune mediated injury, which is the result of cell lysis of viral-infected cells by either direct lymphocyte cytotoxicity, antibody-mediated damage or viral-induced autoimmunity  (Gonzalez-Peralta et al., 1999). Methods of common  transmission of H C V include: blood transfusion, hemodialysis, organ transplantation, needle sharing in IV drug users and tattooing (Caronia et al., 2001). Although many individuals are chronically infected, many more are newly infected carriers who are unaware of their infection, which could facilitate the spread of the virus (Fanning et al, 2000). Despite the  1  seriousness of the disease, so far there is no vaccine or effective antiviral therapy available (Lechmann and Liang, 2000).  1.1.2  H C V Genome Organization and Polyprotein Processing H C V belongs to the Flaviviridae family, genus hepacivirus. It is an enveloped virus  containing a single-stranded positive-sense R N A genome (9.6 kb) which encodes a single polyprotein of about 3100 amino acids (Bartenschlager., 1997). This polyprotein includes both structural and non-structural (NS) proteins which are cleaved into functional protein products by both cellular and viral proteases (Fig. 1.1 A). Structural proteins incorporated into mature virions are located at the N-terminus of the H C V polyprotein. The first product to be cleaved from the polyprotein is the highly basic core protein which forms the major constituent of the nucleocapsid (Yasui et al., 1998). E l and E2 are highly glycosylated type I transmembrane proteins which have been shown to form a stable heterodimeric complex (Deleersynder et al., 1997). Protein p7, located at the carboxy terminus of E2, is a highly hydrophobic polypeptide of unknown function. A l l H C V structural proteins are cleaved by an unknown host enzyme. The non-structural (NS) proteins are required for viral replication. They include NS2 metallo-protease, which catalyses the cleavage between NS2 and NS3 (Grakoui et al., 1993). NS3 is a bi-functional protein possessing in the amino-terminal -190 residues, a chymotrypsin-like serine-type protease responsible for cleavage at the NS3/4A, NS4A/4B, NS4B/5A and NS5A/5B sites. The carboxyl-terminal of NS3 contains the NTPase/helicase activities essential for the translation and replication of the H C V genome (Bartenschlager et al., 1993). NS4A is an essential cofactor for the activity of NS3 protease and is required for efficient polyprotein processing (Failla et al., 1994). The function of  2  Genome  Proprotein -1  aa  Function  Nucleocapsid Core protein Envelope protein Envelope protein Unknown Metallo-protease Serine Protease Helicase Protease co-factor Unknown  Unknown  RNA-dependent RNA polymerase -3011  aa  B Protease  Helicase  NS3 Domain  NS3 protease - (191 aa) HistidinJ t Protease  Helicase  FL NS3 - (630 aa)  Figure 1.1. HCV Genome/Polyprotein processing. (A) H C V genome organization and virus encoded structural and non-structural proteins. Dotted arrows indicate junctions processed by host proteases, an orange arrow indicates the junction processed by the metallo-protease, solid black arrows indicate junctions processed by the NS3 protease. (B) Histidine-tagged NS3 variants: NS3 protease (191 aa) and FLNS3 (Full length NS3) (630 aa).  3  NS4B is so far unknown. NS5A is a highly phosphorylated protein the function of which is also unknown. NS5B has been identified as the RNA-dependent R N A polymerase (RdRp) required for viral genome replication (Al et al., 1998) (Fig. 1.1 A ) .  1.1.3  Currently Available Therapeutics for H C V Patients and New Targets for Drug Development The development of effective H C V therapeutics has been seriously hampered by the  lack of an efficient in vitro H C V replication system. Moreover, until the recent discovery of the mouse model which permits H C V replication (Mercer et al., 2001), chimpanzees had been the only animal shown to be permissive to H C V infection. At present, Interferon alpha (IFN-a) or pegylated interferon alone or in combination with oral ribavirin, are the only treatments currently approved for chronic Hepatitis C. Combination therapy (IFN + Ribavirin) is rapidly superseding interferon monotheraphy  because dual therapy is  significantly more effective (Foster et al., 2001). The effect of IFN-a may include inhibition of H C V virion production by decreasing the synthesis of viral mRNA and proteins. Ribavirin is a nucleoside analogue which inhibits RNA-dependent R N A polymerase and decreases GTP availability in the cell to cause a reduction in viral protein synthesis (Klaus et al., 2000). However, even the combination therapy for chronic H C V infection is clearly unsatisfactory, as only a small proportion of patients (40%) respond to therapy and the side effects are considerable (Heathcote et al., 2000). Various other compounds have also been tried, including ursodeoxycholic acid, nonsteroidal anti-inflamatory agents, mycophenolate and amantadine. At present, none of these drugs have been shown to have significant advantages over the currently used therapy of IFN + Ribavirin (Craxi et al., 1999). The development of  4  vaccine against H C V also faces a variety of obstacles mainly due to the rapid mutation rate of H C V (Lechmann and Liang., 2000). The importance of NS3 protease in H C V replication (Major et al., 1999) has led to many studies attempting to understand its structure and function, with the ultimate goal being the design of novel anti-HCV therapeutics.  1.2  H C V NS3/NS4A Protease  1.2.1  Identification of the NS3 Protease and Full-length NS3 Protease/Helicase There are two molecular forms of NS3 which have been widely studied so far:  Truncated NS3 protease domain and Full length NS3 (FLNS3) (Fig 1.1 B). H C V NS3 protease consists of approximately 190 amino acids which contain the serine protease domain. Analysis of the X-ray crystal structures of the truncated NS3 protease revealed that the serine protease adopts a chymotrypsin-like fold and the NS3 protease active site is similar to the active site of the other chymotrypsin-like enzymes. The nucleophilic Ser-139, together with the general acid/base catalyst His-57 and Asp-81, form the NS3 catalytic triad (Fig 1.2 B) (Kim et al., 1996). FLNS3 consists of the serine protease domain and the helicase/NTPase domain, an enzyme that can unwind double-stranded regions of R N A or remove regions of secondary structure in an ATP-dependent reaction, allowing the R N A dependent R N A polymerase to copy the positive and negative strands (Shoji et al, 1999; K i m et al, 1998; Yan et al., 1998; Love et al., 1996). The NS3 helicase domain (440 amino acids) is about twice as large as the protease domain (190 amino acids) (Fig. 1.2 A ) (Gallinary et al., 1999). Both the NS3 protease domain itself and the FLNS3 have been shown to be catalytically active in cleaving synthetic peptides mimicking the amino acid sequence of the junctions  5  Figure 1.2. Ribbon representation of FLNS3 & NS3 protease. (A) (FLNS3)NS3/Pep4A protease complex (yellow/purple) and Helicase/NTPase (cyan). (B) Ribbon representation of the backbone of the crystal structure of NS3/Pep4A protease complex; NS3 protease (yellow); NS4A-derived sequence (purple).  6  between the non-structural proteins of H C V (Fattori et al., 2000; Gallinari et al., 1998; Po et al., 2001). The role of NS3 in the processing and maturation of the non-structural junction NS3/4A, NS4A/NS4B, NS4B/NS5A and NS5A/NS5B of the viral polyprotein has been elucidated by transient transfection and cell-free translation^ studies (Bartenschlager et al., 1993; Grakoui et al., 1993; Lin et al., 1994). From these studies it emerged that the NS3/4A site is cleaved in cis and the rest of the junctions are cleaved in trans.  1.2.2  The NS4A Protease Cofactor Although NS3 protease and FLNS3 show some intrinsic proteolytic activity in vitro, a  second viral protein, NS4A, is an essential NS3 protease cofactor for efficient proteolytic processing of the H C V polyprotein (Failla et al., 1994; Bartenschlager et al., 1994; L i n et al., 1994; Tanji et al., 1995; Po et al., 2001). NS4A cofactor is a membrane-bound protein of 54 residues. Region I, residues 1-20, is a highly hydrophobic trans-membrane helix. Region n , residues 21-34, is also hydrophobic, but with a p-strand structure. Region m , residues 35-54, is more hydrophilic with an a-helical conformation (Fig 1.3. A ) (Tomei et al., 1996). NS4A can exert its cofactor function in cis (i.e., when expressed as an NS3-NS4A precursor protein) as well as in trans (i.e., when expressed as a separate molecule to interact with NS3 to cleave other NS polyproteins). The interaction between the NS4A cofactor and NS3 causes the re-arrangement of the NS3 catalytic triad (His-57, Asp-81 and Ser-139) residues, which are important in forming the catalytic pocket of NS3, thereby enhancing the proteolytic activity of NS3 on all cleavage sites (Landro et al., 1997). In addition, the complex formation of the full length NS4A cofactor to NS3 also significantly increases the stability of NS3 in the cytoplasm of cultured mammalian cells as well as targets the NS3  7  Domains of NS4A N  Lumen  t  ER membrane Cytosol  HCV NS4A domain II sequence (Pep 4A):  G - S - V - V - I - V - G - R - I - I - L - S - G - R  B y TML J  IS)" "'  m  Pep 4 A  \  Figure 1.3. N S 4 A : HCV N S 3 protease co-factor (A) R e p r e s e n t a t i o n of N S 3 a n d full length N S 4 A cofactor c o m p l e x attached to the E R of the cell. P r e d i c t e d Structures I: H y d r o p h o b i c trans m e m b r a n e a-helix, II: H y d r o p h o b i c p-sheet, III: Hydrophillic a-helix. Synthetic peptide cofactor only contain d o m a i n II with its a m i n o a c i d s s e q u e n c e s h o w n in red. (B) R i b b o n representation of synthetic peptide cofactor d o c k i n g in N S 3 .  8  protein to the membranes of the endoplasmic reticulum (ER) (Tanji et al., 1995). The production of the full length NS4A cofactor (54 amino acids) for the activation of NS3 in vitro is difficult as full length NS4A is highly hydrophobic and insoluble. However, it has been demonstrated that synthetic 17-mer peptides corresponding to the central domain of NS4A are as efficient as full length NS4A in activating the protease activity of NS3 in vitro. Consequently, such peptides have been utilized in vitro studies of NS3 protease activity (Fig. 1.3. B) (Landro et al., 1997; L i n et al., 1995; Shimizu et al., 1996; Po et al., 2001). Synthetic peptide cofactor has been found to bind to NS3 protease with 1:1 stochiometry (Lin et al., 1995).  1.2.3  Substrate Specificity of HCV NS3/NS4A Protease Substrate recognition by the chymotrypsin-like serine proteases involves the binding  of the substrate on the active surface of the enzyme defined by a specific amino acid sequence. Elucidating the substrate specificity of the H C V NS3 protease is important for the development of high throughput assays for screening of potential H C V protease inhibitors and for the rational design of the H C V protease-specific inhibitors. Sequence comparison among isolated H C V strains has revealed structurally conserved residues flanking the cleavage sites between the NS3, NS4A, NS4B, NS5A and NS5B (Grakoui et al., 1993). Investigations of protease specificity have generally focused on the PI/SI interaction (based on the nomenclature of Schecter and Berger (1967), where P I - P ' l denotes peptide residues on the acyl and leaving group side of the scissile bond, respectively. The adjacent peptide residues are numbered outward, while SI, S ' l , etc. denote the corresponding enzyme binding sites), followed by consideration of the P2-Pn/S2-Sn interactions.  9  The primary specificity of a protease is defined by the side chain of the amino acid that preceeds the scissile bond, i.e. the PI position (De Francesco and Steinkuhler., 1999). In the case of NS3 protease, it was found that a Cys/Thr residue in P I , Asp/Glu at the P6 residue and Ser/Ala at the P ' 1 residue are required for the efficient cleavage by NS3 protease (Fig 3.4.B), with cleavage occurring after cysteine in all trans cleavage sites (i.e. NS4A/4B, NS4B/5A, and NS5A/5B) or after threonine in the intramolecular cleavage site between NS3 and NS4A. However, acidic residue in P6 is not a stringent requirement (Urbani et al., 1997). Preference for cysteine residues in the PI positions of the NS3 substrates can be rationalized on the basis of the peculiar structure of the S1 pocket of the protease. The substrate specificity of the NS3/Pep4A protease complex was studied in detail, using mainly purified recombinant enzyme and synthetic peptide substrate. For efficient activity on peptide substrates, the NS3/Pep4A protease complex requires at least a decamer peptide substrate spanning P6-P'4 (Steinkuhler et al., 1996). Although several research groups have studied the substrate specificity of NS3 protease, none have undertaken a comprehensive study comparing the substrate specificity of both the NS3 protease and the FLNS3 with the same set of substrates (NS4A/4B, NS4B/5A, NS5A/5B) and assay conditions using a high-throughput approach (HTA). Since the exact mechanism of H C V replication in vivo is still largely unknown, this comprehensive substrate specificity study with the same set of substrates is important because potential inhibitors of the H C V NS3 must be able to inhibit both the NS3 protease and the FLNS3.  10  1.2.3.1 IQFS and M C A Substrates Most substrate specificity studies with H C V NS3/Pep4A protease complex were performed and determined with H P L C . Our study used internally quenched fluorogenic substrate (IQFS) and methyl coumarinamide substrate (MCA) containing a cysteine residue in PI position in order to perform continuous monitoring of protease activity with high throughput assays (HTA).screening of the substrate specificity of the purified recombinant NS3 protease. M C A substrates contain only 4-6 amino acids in total and in order to be fluorescently active, the junction between the final amino acid and the fluorophore has to be cleaved.  NH2 C-Xaan-Xaa2-Xaat-X'aan-NH-CH2-CH -Nl 2  N0  2  Abz-Asp-Asp-lle-Val-Pro-Cys-Ser-Met-Ser-Gln-EDDryp  IQFS-1  CH3  Xaan-Xaaj-Xaa^- NH Ac-Glu-Glu-Val-Val-Ala-Cys^MCA  MCA-1  Internally quenched fluorogenic substrates (IQFS), which include 10-12 amino acids including the enzymatically cleaved bond, are characterized by the presence of a donor fluorescent group (Abz) positioned at the N-terminal end of the substrate and a quencher group (EDDnp) positioned at the C-terminal end (Yaron et al., 1979; Jean et al., 1995). The fluorescence energy of an intact substrate is initially quenched by an intra-molecular energy  11  transfer between the donor and acceptor. However, upon cleavage of the substrate, the fluorescence energy is recovered, thereby allowing the increase in fluorescence over time to be measured as cleavage occurs (Yaron et al., 1979). Using a fluorescent spectrofluorometer equipped with a 96 well plate reader, it is therefore possible to measure the amount of increase in fluorescence in many different reactions at once. The results of such a kinetic study can then be analyzed further to obtain the kcat, Vma and k values from the different X  m  reactions that will allow definition of the substrate specificity of the NS3 protease and FLNS3 in vitro.  1.3  Strategies for Inhibition of the NS3/NS4A Protease Complex  1.3.1  H C V NS3 Protease is an Attractive Target for Anti-HCV Therapy There are four major types of protease enzymes (aspartic, serine, cysteine and  metallo) which selectively catalyze the hydrolysis of polypeptide bonds (Leung et al., 2000). H C V NS3 protease, which belongs to the serine protease super family of chymotrypsin like enzymes, mediates the maturation cleavage of viral polyproteins. Virally encoded proteases are attractive targets for drug discovery since they are crucial in the life cycle of many viruses, including H C V (Bianchi and Pessi, 2002). Currently, there are a few designed potent and selective protease inhibitors that slow or halt disease progression. A well known example of a potent viral protease inhibitor is the aspartyl protease inhibitor of the human immunodeficiency virus (HIV). The FflV-1 protease has proven to be an attractive drug target due to its essential role in the replicative cycle of HTV (Campoy et al., 2000). Several low molecular weight inhibitors of HIV-1 protease including saquinavir, ritonavir, indinavir, nelfinavir and amprenavir are now used to treat H I V infections in humans (Leung et al.,  12  2000). These drugs are among the successful examples of drugs developed to target the proteolytic active site of the HIV-1 protease. The concept of inhibiting the viral protease as a means of stopping viral replication can potentially be applied to many other viruses, including H C V . The most intensively studied and therefore best understood target for antiviral therapy against H C V is the NS3 protease. Previous studies have shown that the NS3 protease is responsible for most of the proteolytic maturation events within the non-structural portion of the viral polyproteins (Grakoui A et al., 1993). Studies carried out with HCV and other members of the Flaviviridae family support the hypothesis that inactivation of the homologous serine protease activity in these viruses leads to the production of non-infectious viral particles (Chambers et al., 1990; De Francesco and Steinkuhler., 1999; Major et al., 1999). Thus, the HCV NS3 protease is currently one of the major targets pursued for the discovery of novel anti-HCV drugs. Currently, there are many peptide-based inhibitors directed against the NS3 protease in order to help elucidate the mechanism of substrate recognition by the NS3 and to act as the lead compound for the development of small molecule HCV NS3 protease inhibitors. Such peptides or derivatives thereof may be effective drugs as they may satisfy the criteria necessary for effective small molecule drugs such as; stability, high resistance to proteolytic degradation, good membrane permeability and bioavailability. These properties usually require the compounds to have a low molecular weight (< 1000 Da) (Leung et al., 2000).  1.3.2  Peptidyl Inhibitors of NS3 Protease and Their Limitations Various peptidyl inhibitors including both product inhibitors and substrate inhibitors such  as the a-ketoacid and non-cleavable decapeptides spanning P6-P'4 have been reported to target HCV NS3 protease (Ede et al., 2000; Narjes et al., 2002; Lahm et al., 2002; Llinas-Brunet et al.,  13  1998; Landro et al., 1997; Steinkuhler et al., 2001). The FLNS3 protease cannot efficiently bind small peptide inhibitors, i.e. P6-P1 residues, requiring at least a decapeptide spanning P6-P'4 residues distal to the scissile bond that contribute significantly to binding through hydrophobic and electrostatic interactions (Po A et al., 2001; Steinkuhler et al., 1996). Despite some success with peptidyl inhibitors, significant progress in using peptidyl inhibitors against NS3 protease and FLNS3 protease has been hampered by the very nature of the target: a protein whose binding site is highly charged, solvent-exposed and featureless, and whose minimal decapeptidic substrate relies heavily on charge-charge and hydrogen bonding contacts along an unusually extended surface (Steinkuhler et al, 2001; Bisceglie et al., 1995; Komada et al., 1994; Kolykhalov et al., 1996). It is becoming more evident how the viral polyprotein substrate can compensate for this lack of binding pockets and provide the interactions required for specificity of cleavage during polyprotein processing. Therefore, an alternative strategy to inhibit NS3 protease and FLNS3 is to use a polyprotein substrate that can be recognized by the NS3 protease, such as the use existing serine protease inhibitors (serpins) as an initial scaffold to inhibit the H C V NS3 protease (Carrell., 1986). With the recent finding that the H C V NS3 serine protease can interact with and form an SDS-stable complex with naturally occurring serpins (Drouet et al., 1999), modulation of NS3 enzymatic functions by bio-engineered serpin directed at the H C V NS3 could represent a new therapeutic approach to treating H C V infections.  1.3.3  Protein-Based Inhibitor: The Case of Serine Protease Inhibitors (Serpins) Few protein-based inhibitors have been tested against the H C V NS3 protease so far.  There is some evidence that inhibition of pathogen protease by host cell serpins in the virushost cell relationship can occur in vitro i.e., C l inhibitor (Ci Inh) and a-2 antiplasmin  14  (a2AP). Both of these belong to the serpin family and can interact with the H C V NS3 protease with proteolysis of the serpins and production of a higher molecular weight SDSstable complex (Drouet et al., 1999). In addition, Eglin c, a 70 amino acid reversible potent inhibitor of several serine proteases has been mutated to specifically target the H C V NS3 protease and it has shown some success in inhibiting the NS3 protease (Martin et al., 1998). These observations open the way to new approaches to inhibit H C V replication. Serine Protease Inhibitors (serpins) are a large family of proteins which have been identified from various sources such as viruses and mammals. The primary function of most members of the serpin family is to neutralize over-expressed serine protease activity (Travis et al., 1983). The irreversibility of protease inhibition achieved by the serpins has made them the principal inhibitors controlling both intra and extra-cellular proteolytic pathways. Serpins regulate such diverse physiological processe such as: blood coagulation, fibrinolysis, compartment activation, fetal development and inflammation (Silverman et al., 2001; Gettins et al., 1996; Stein et al., 1995). The structure and mechanistic features of the serpins are typified by ai-antitrypsin (ai-AT) (Carrell et al., 1982). Serpins share a highly ordered structural architecture consisting of three P-sheets, nine a-helices and a reactive site loop (RSL) of -20 amino acids which is solvent-exposed and mimics target protease recognition sites (Fig. 1.4) (Huber et al., 1989; Whisstock et al., 1998). Serpins act as pseudo-substrate and suicidal proteins by forming 1:1 complexes with their target proteases. The resulting irreversible SDS-heat stable complex of enzyme and inhibitor are then removed from circulation for subsequent breakdown (Wright et al., 1995). The mechanism of irreversible inhibition by serpin has three stages: 1) recognition of the serpin and target protease. 2) cleavage at the P I - P ' l bond  15  in the reactive site loop of the serpin that triggers the insertion of the cleaved reactive centre loop through the P-sheet of the serpin molecule, which accompanies translocation of the protease to the opposite pole. 3) formation of SDS-stable complex between the serpin and the protease, which also results in the disruption of the protease active site (Fig. 1.4) (Huntington et al., 2000). The molecular structure and physical properties of serpins also permit these proteins to adopt a number of variant conformations under various physiological conditions, including the native inhibitory forms and several inactive non-inhibitory forms. These include serpins that have already formed irreversible complexes with proteases, ligands or peptides, serpins that were recognized as substrates and cleaved but did not form complexes, or degraded serpins, oxidized serpins and polymerised serpins (Fig 1.5).  1.3.4  ai-antitrypsin (otj-AT) and Bio-engineering of the Serpin Reactive Site Loop (RSL) a i - A T , which is an important component of the serine protease inhibitor (serpin)  system in humans, has been studied since 1955. Although named a i - A T , its target protease is actually elastase (Carrell., 1986). a i - A T is synthesized primarily in the liver and is about -426 amino acids in length (Lobermann., 1984). The reactive site loop (RSL) of the serpin is believed to resemble an ideal substrate for the target enzyme, since the substrate binds but is not hydrolysed and so does not leave the active site of the enzyme. Complex formation involves more than just interaction between the active site of the protease and the RSL of the serpin. However, this interaction is the basis for determining specificity and is therefore the basis of the direct potential for simple engineered changes in function (Travis and Salvesen., 1983). a i - A T is the first protein to have been modified by genetic engineering to display a  16  Figure 1.4. Serine protease inhibitor (Serpin). (A) When a protease reacts with an inhibitor, the inhibitor can either be cleaved while releasing the active protease, or the inhibitor can form a covalent SDS stable complex with the protease. When complex formation occurs, the protease loses its catalytic properties. (B) Ribbon representation of the elastase (E), a i - A T (I), and cci-AT /elastase complex (E/I*). RSL (yellow). (Huntington et al., 2000).  17  F i g u r e 1.5. 2001).  C o n f o r m a t i o n a l p o l y m o r p h i s m of inhibitory s e r p i n s . (Janciauskiene,  18  novel inhibitory activity. a i - A T PI position, methionine 358, was changed to valine to make an ai-AT variant that was resistant to oxidation (Rosenberg et al., 1984; Courtney et al., 1985). B y making changes to the amino acid sequence in the R S L of the ai-AT, it has been possible to target specific proteases thus resulting in an innovation of potential therapeutic value. After the first oxidation-resistant bio-engineered serpin was produced, many more followed, including aiAT-Portland (cti-PDX), a bio-engineered serpin variant of a i - A T highly selective for furin (Tsuji et al., 2002; Jean et al., 1998). a i - P D X was bio-engineered specifically to target furin, an endogenous human serine protease that plays an important role in the proteolytic activation of proteins encoded by many pathogenic agents such as Pseudomonas exotoxin, H I V glycoproteins, Human cytomegalovirus (HCMV) glycoproteins and Influenza A viral hemaglutinin. The inactivation of furin suggests the broad applicability and potential of ai-PDX as a therapeutic agent (Jean et al., 1998). The most recent example of a protein therapeutic for H C M V infections was demonstrated with  ai-PDX. The  production of infectious H C M V is dramatically reduced by the exogenous addition of a i P D X (Jean et al., 2000). A similar strategy could be applied to developing inhibitors of the H C V NS3 protease. a i - A T has a methionine at the PI position of its R S L and inhibits proteases with chymotrypsin-like, trypsin-like and elastase-like specificities (Whisstock et al., 2000). Since the H C V NS3 serine protease domain adopts a chymotrypsin-like fold (Bartenschlager et al., 1993), ai-AT represents a good macromolecular protease inhibitor scaffold to initiate protein engineering studies.  19  1.4  Thesis objectives The main objective of this research was to develop and evaluate bio-engineered serine  protease inhibitors (serpins) directed at the H C V NS3 protease based on the results of substrate specificity studies. To achieve the main objective, several secondary specific aims had to be completed:  i)  Purification of the active recombinant histidine-tagged H C V NS3 protease.  ii)  Development and optimization of an enzymatic assay to monitor H C V NS3 proteolytic activity.  iii)  Determination of the substrate specificity of the H C V NS3/Pep4A protease complex using IQFS ( K , V m  iv)  m a x  , V^/Km).  Bio-engineering of recombinant ai-AT variants to target the H C V NS3 protease.  v)  Purification of the active recombinant histidine-tagged ai-AT and its variants.  vi)  Assessment of the inhibitory activity of bio-engineered oci-AT variants against elastase.  vii)  Assessment of the inhibitory activity of bio-engineered oci-AT variants against the H C V NS3/Pep4A protease complex.  20  Chapter 2: Materials and methods  2.1  Preparation of Competent BL21 & DH5a E.coli Cells A single colony of DH5a or BL21 E.coli cells was used to inoculate 5ml of Luria Bertani  (LB) medium and cells were grown overnight at 37°C in an orbital shaker. The next day, 1 ml of the overnight culture was diluted in 50ml L B and allowed to grow for 2 hours at 37°C until the optical density (Aeoo) of the culture reached 0.4, with shaking. Cells were chilled on ice for 20 minutes and then pelleted at 3000 rpm by centrifugation for 5 minutes at 4°C. The cells were resuspended in 25ml 0.1M CaCl and incubated on ice for 30 minutes. The cells were pelleted 2  again at 3000 rpm by centrifugation for 5 minutes at 4°C and finally resuspended in 3.65 ml CaCl and 0.35 ml of glycerol. Cells were stored at 4°C for at least one hour before use in 2  transformation reactions. The remaining competent cells were snap-frozen in -86°C freezer.  2.2  Transformation of K coli Strain BL21 with the NS3 Recombinant Plasmid Construct. The NS3 pMANS34NSH plasmid construct with a 6X-histidine tag attached to the  minimal protease domain (1027-1218, NS3 protease) was kindly provided by our collaborator (Nobuko Kakiuchi) from Japan. Approximately lOng of NS3 protease plasmid D N A was mixed with 200ul of competent E.coli BL21 and incubated on ice for 30 minutes. Competent cell/DNA mixtures were then subjected to "heat-shock" by incubating at 42°C for 45 seconds. 800 ul of L B was added immediately and samples were incubated at 37°C for 1 hour, with shaking. Cells were then pelleted at 13000 rpm in a microfuge for 30 seconds and  21  800ul of the supernatant discarded. The cell pellet was resuspended in the remaining 200ul of media and then spread onto L B agar plates containing with 100 ^g/ml of ampicillin and 20 ug/ml Chloramphenicol.  2.3  Expression of Recombinant NS3 Proteins. E.coli strain BL21 bearing the NS3 protease expression plasmid was grown in 5 ml of  (LB) containing 100 u.g/ml of ampicillin and 20 |ug/ml chloramphenicol overnight at 31°C. The overnight 5 ml culture was then used to inoculate 500 ml of L B (100 ug/ml of ampicillin) at 37°C for 4 hours until the optical density (A6 o) of the culture reached 0.7. The 0  culture was then supplemented with 700 u M IPTG (Calbiochem), and transferred to a 31°C incubator to grow for another 18 hours to induce expression of the NS3 protease. The induction was performed at 31°C to prevent the formation of inclusion bodies and to produce more soluble protein. Induced bacterial cells were then harvested by centrifugation in 500ml bottles in a GS3 rotor at 4000 rpm at 4°C for 10 minutes. The cell pellets were then frozen in a -86°C freezer and thawed on ice, and resuspended in 20 ml of I X FPLC binding buffer (0.5M NaCl, 20mM sodium phosphate, pH 7.3, 20% glycerol, 0.1% triton-X and 1 capsule of E D T A free protease inhibitor (Roche)). Resuspended/thawed E.coli BL21 cells were then lysed using a sonicator (Misonix) at output level 3 for 3 X 45 seconds intervals. The lysed E.coli cells were then centrifuged at 12000 rpm in SS34 rotor at 4°C for 45 minutes to separate the soluble proteins from the insoluble proteins and cell membranes. The soluble proteins in the supernatant were then filtered using 0.45 u M filter (Millipore) alliquoted and stored at -86°C prior to FPLC purification.  22  2.4  Purification of NS3 Protease Using Fast Protein Liquid Chromatography (FPLC). Protein purification of NS3 protease was done with an automated A K T A purifier  FPLC system (Pharmacia). Frozen E.coli extract supernatant containing histidine tagged NS3 protease was thawed on ice and loaded into the FPLC superloop. It was then allowed to pass through a pre-packed nickel column (Pharmacia) at lml/minute (nickel column was preequilibrated with 15 ml of I X binding buffer—0.5M NaCl, 50 m M sodium phosphate, pH 7.3). After the supernatant had been loaded into the column, proteins with weak interaction to the nickel were removed with 3 X 60 ml sequential washes using I X washing buffer at pH 7.3, 6.0, 4.8 (0.5M NaCl, 50 m M sodium phosphate). NS3 protease was eluted with 20 ml of I X elution buffer (0.5M NaCl, 50 m M sodium phosphate, pH 4.0). Eluted material was collected in 1 ml fractions and stored in aliquots at -86°C. Eluted fractions were then thawed, pooled and dialyzed against neutral pH buffer (50 m M Hepes, 150 m M NaCl, lOmM DTT, pH 7.5) for 3 X 1 hours using the slide-a-lyser dialysis cassette with a 3.5 kD size cut-off membrane (Millipore) to raise the pH of the purified protein solution before analyzing the purity using Sodium Dodecyl Sulfate Poly-Acrylamide Gel Electrophoresis (SDS-PAGE) analysis and Western blots.  2.5  Protein Analysis with SDS PAGE gels and Western Blot. Protein induction levels were determined by comparison of pre-and post IPTG  induced bacterial cell samples run on SDS-PAGE (12% acrylamide) and followed by analysis using anti-histidine specific antibody for a Western blot. The Bio-Rad Mini Protean JJ gel apparatus was used to prepare and run mini gels. A 6% stacking gel was prepared and  23  prior to loading, samples were suspended in reducing sample buffer and were heated to 100°C for 10 minutes. I O U L of protein sample was loaded to each well, and the gel was subjected to electrophoresis at 125V for 75 minutes. For SDS-PAGE analysis, protein were detected with Coomassie blue stain (Biorad) for 1 hour followed by de-staining for 1 hour. For Western blot analysis, proteins form the gels were transferred to 0.45 u M nitrocellulose membrane (Biorad). Protein transfer was done using the semi-dry transfer Trans-blot SD system (Biorad) at 10V for 15 minutes with transfer buffer (see buffer list section 2.16). Unoccupied binding sites within the dried nitrocellulose membrane were blocked with blocking buffer supplied by Qiagen for 18 hours. The membrane was then washed 3 X 10 minutes with TBS-T (TBS, 0.05% Tween), followed by 3 X 10 minutes with TBS before probing with (1 step) HRP-conjugated anti-histidine RGS primary antibody (Qiagen) (anti histidine mouse IgG antibody-1/2000 dilution in blocking buffer) for 1 hour on a shaker at room temperature. For two-step Western blot analysis requiring secondary antibody incubation e.g. anti FLNS3 monoclonal antibody (Novocastra Laboratories Inc), an additional 6 X 10 minutes washes with TBS-T were required between the primary antibody incubation and the secondary antibody incubation (1/10000 secondary antibody dilution in TBS-T) was carried out for 1 hour at room temperature using HRP-conjugated anti-mouse anti-body (Pharmacia). Finally, 8 washes with TBS-T for 10 minutes each were carried out before  proceeding with Chemiluminescence visualization using E C L - H R P  substrate  (Pharmacia) followed by exposure to Hyperfilm (Pharmacia). Hyperfilm was developed using a (Kodak X-omat) x-ray processing machine.  24  2.6  Development of H C V NS3 Protease Assays. To characterize the enzymatic activity of the purified NS3 protease, the ability to  cleave synthetic peptides either with or without the NS4A cofactor was investigated. Commercially available M C A substrate (Ac-E-E-V-V-A-C-AMC) (Calbiochem) was used initially to screen for activity in the FPLC purified NS3 protease. The substrate specificity of NS3 protease was then investigated using synthetic IQFS and peptide NS4A cofactor (AcKKKGSVVIVGPvJILSGR)  designed in our laboratory but synthesized by our collaborator  (Luiz Juliano) from Brazil. The structural representation of M C A , IQFS and the different H C V IQFS used are shown in Fig. 3.4. The cleavage of the IQFS occurs after the cysteine in PI position. P I , P2, etc refers to the amino acids of the substrate prior to the cleavage site. P ' l , P'2, etc refers to the amino acids of the substrate after the cleavage site as described in section 1.2.3. The continuous assays of protease activity were performed using a Gemini Spectromax X S 96 well plate spectrofluorometer. Enzymatic reactions were performed at 31°C. The excitation and emission spectra for M C A and (HCV or G B V ) IQFS substrates were set at 370/460 n M and 320/420 n M , respectively. Each well contained 15uM NS4A cofactor, varying concentrations of (HCV or G B V ) IQFS or M C A substrate, 20 uL of FPLC purified NS3 protease and the final volume in each reaction was adjusted to lOOul using the reaction buffer (50 m M Hepes, 150 m M NaCl, 0.1% triton-X, lOmM DTT, pH 7.5). (HCV or G B V ) NS4A cofactor had to be pre-incubated with NS3 protease for 15 minutes at 31°C prior to any reaction involving NS3 protease to allow proper binding between NS3 protease and NS4A cofactor.  25  2.7  Michaelis-Menten Steady-State Analysis of H C V Synthetic Peptide Substrate. When the purpose of the experiment is to determine the substrate specificity of an  enzyme under specific conditions, V  m a x  and K  m  , assays must be performed at various  concentration of substrate [S], including low concentrations. Otherwise the K  m  cannot be  determined. At low [S], changes in substrate concentration are linearly reflected in the initial rate of reaction, and the slope of this linear part of the reaction allow the sensitive determinations of initial velocity (V ). The (V ) of six different substrate concentrations 0  0  (5,10,20,50,100, and 200 u M of each different IQFS) were determined and the kinetic measurements of NS3 protease cleavage efficiency to the different IQFS were calculated from the least-square fit of initial rates (V ) as a function of the six different substrate 0  concentrations with the help of Sigma plot 2002 software, assuming Michaelis-Menten kinetics. V  m a x  and K value can be extrapolated by Sigma plot software from the Michaelism  Menten curve and the V  m a x  /K  m  values were calculated for comparison of the relative  cleavage efficiency. Kinetic reactions were done in triplicate for each of the 6 different substrate concentrations. Control with no substrate or NS3 protease alone were also done in triplicate.  k, kv Enzyme (E) + Substrate (S) — • Enzyme/Substrate (E/S) —•Enzyme (E) + Product (P) k Catalytic constant (kcat)  =  2  k j , is a catalytic rate constant that measures how much of the  enzyme/substrate (E/S) complex, once formed, is converted to products. Maximum velocity of enzyme (Vma ) = kc (E) , is a rate of reaction that measures how fast X  at  a given amount of the enzyme can catalyze product formation at maximum speed.  26  Affinity constant ( K ) = (k +k3)/ki, is a concentration that measures the intrinsic property 2  m  of an enzyme related to the binding constant for forming (E/S) complex, which also corresponds to the concentration of substrate at which the rate of the reaction is half the maximum velocity ( V  m a x  ).  Specificity constant ( k ^ / K m ) = defines the rate of the overall enzymatic reaction at which the substrate concentration is much lower than K The ratio of V  m a x  /K  m  m  can be used to compare enzymes and measures the efficiency of  the enzyme. Relative substrate specificity of NS3 protease towards the different IQFS junctions can be calculated by comparing V m a / K of NS3 protease against each of the X  m  different IQFS junctions because i f the concentration of the enzyme [NS3 protease] is constant throughout the experiment, the relative substrate specificity of the NS3 protease to the different IQFS will still be comparable. The higher the Vmax/Km value, the higher the specificity of NS3 protease towards the substrate, which means faster turnover and more product formation.  2.8  ai-AT D N A Isolation and Sub-cloning. a i - A T D N A was obtained from Dr. J.L.Christian (Portland, Oregan). oci-AT-KDEL  was obtained from Dr. C. Hashimoto (New haven, Connecticut). cti-AT D N A was transformed into competent DH5-oc E.coli (Novagen) by the standard heat shock (42°C) transformation procedure. cti-AT plasmid D N A propagated in E.coli was pelleted by centrifugation at 13000 rpm (microfuged) for 30 seconds and the supernantant was discarded. Plasmid D N A was then isolated using the D N A mini-prep kit (Qiagen) according to manufacturer's instruction. Plasmid D N A was eluted in a 30ul fraction and stored at -20°C.  27  The a i - A T gene was removed from vector pDS by restriction digest using Xho 1 and Blp I. In a typical restriction digestion, 1.0 ug of pDS-ai-AT plasmid D N A and pET-21 vector D N A were incubated with 10 units of restriction enzymes Xhol and Blp I (NEB) for 3 hours at 37°C. The resulting D N A fragments were resolved by electrophoresis using 0.8% agarose gels (70 volts for 90 minutes). pET 21 Vector (Novagen) was also prepared and purified in the same way as cci-AT. Desired fragments were recovered from agarose gels using the QIA quick Gel Extraction Kit (Qiagen) according to manufacturer's instruction. Ligation of the ai-AT gene into the pET-21 vector (Appendix II) was performed with 50-100 ng (0.015-0.03 pmol) of pET 21 vector and 0.2 pmol a,-AT insert (50 ng of a 500 bp fragment) in a volume of 30 ul (insert to vector ligation ratios of approximately 10:1). l u l of T4 D N A ligase (NEB) was added last and gently mixed by stirring with a pipet tip. Incubation generally lasted 16 hours at 4°C. A l l ligated products were transformed into competent DH5-a E.coli by the standard heat shock (42°C) transformation procedure and plated on L B containing 100 ug/ml of ampicillin (37°C O/N). Grown E.coli colonies were then randomly picked, grown overnight at 37°C, and the plasmid D N A was purified and analyzed for successful sub-cloning by restriction digest analysis. When successful ligation was confirmed, the oti-AT/pET 21 D N A plasmid was transformed to competent BL21 E.coli for expression and protein purification.  2.9  Site-Directed Mutagenesis of ai-AT Using Polymerase Chain Reactions (PCR). A l l the ai-AT variants including the oci-AT-KDEL variants in p C M V shuttle vector  (Appendix m) were mutated by PCR/site-directed mutagenesis using the QuickChange site-  28  directed mutagenesis kit protocol (Stratagene). Oligonucleotide primers were designed using the Vector NTI and Web Cutter 2.0 program according to the QuickChange manual. A l l oligonucleotides primers contained the desired mutation and annealled to the same sequence on opposite strands of the plasmid. They are between 25-45 bases in length, have at least 40% G C content and have a melting temperature of 78°C or higher. A l l oligonucleotide primers were synthesized and purified by Alpha-DNA (Montreal). Each P C R mixture (50ul total volume) contained 5pi of 10X reaction buffer (Stratagene), l u l of (5-50 ng/ul) ds D N A template, 1.25ul (125 ng) of oligonucleotide primer #1, 1.25ul (125 ng) of oligonucleotide primer #2, 5ul (2.5 mM) of dNTP mix (Pharmacia), 35.5ul H 0 , and l u l of PfuTurbo D N A polymerase (2.5 U/ul) (Stratagene). 2  The Cycling parameters for the QuickChange Site-Directed Mutagenesis were as follows:  Segment  Cycles  1  1  2  18  Temperature  Time  95°C 95°C  SO seconds 30 seconds  55°C  1 minutes  68°C  2 minutes/kB (template)  •  The extension time for making all the ai-AT variants (-4.8 kB) was 10 minutes.  •  The extension time for making all the K D E L a A T constructs (-6.8 kB) was 14 r  minutes. •  A l l reactions were allowed to cool down for at least 2 minutes at 4°C before further processing  29  List of Primers Used for the Site-Directed Mutagenesis of pET-21 cci-AT Plasmid Mut A-(use a i - A T plasmid D N A as template) p i (Cys) (AP13)  5' - G - A - G - G - C - C - A - T - A - C - C - C - T - G - C - T - C - T - A - T - C - C - C - C - C - C - T - G - A - G - G - T - C - A - A G-T-T-C-3'  (AP14)  5'  -G-A-A-C-T-T-G-A-C-C-T-C-A-G-G-G-G-G-G-A-T-A-G-A-G-C-A-G-G-G-T-A-T-G-GC-C-T-C-3'  Mut B-(use mut A plasmid D N A as template) p i (Cys),p4 (Ser) & p6 (Asp) (APIS)  5' - G - C - A - G - G - C - G - C - C - A - T - G - T - T - T - G - A - C - G - A - G - T - C - C - A - T - A - C - C - C - T - G - C - T - C T-A-T-C-C-C-C-3'  (AP16)  5' - G - G - G - G - A - T - A - G - A - G - C - A - G - G - G - T - A - T - G - G - A - C - T - C - G - T - C - A - A - A - C - A - T - G G-C-G-C-C-T-G-C-3'  Mut C-(Use mut A plasmid D N A as template) p i (Cys) & p6 (Asp) (API 7)  5'  -G-C-C-A-T-G-T-T-T-G-A-C-G-A-G-G-C-C-A-T-A-C-C-C-T-G-C-T-CT-A-T-C-3'  (AP18)  5' - G - A - T - A - G - A - G - C - A - G - G - G - T - A - T - G - G - C - C - T - C - G - T - C - A - A - A - C - A T-G-G-C-3'  Mut D-(Use mut A plasmid D N A as template) p i (Cys) & p4 (Ser) (AP19)  5' - G - C - C - A - T - G - T - T - T - T - T - A - G - A - G - T - C - C - A - T - A - C - C - C - T - G - C - T - C T-A-T-C-C-C-C-3'  (AP20)  5' - G - G - G - G - A - T - A - G - A - G - C - A - G - G - G - T - A - T - G - G - A - C - T - C - T - A - A - A - A A-C-A-T-G-G-C-3'  Mut E-(Use mut B plasmid D N A as template) (ap 21,22)- p i (Cys),p4 (Ser),p6(Asp)&p'l(Pro). (AP21)  5' - G - A - G - T - C - C - A - T - A - C - C - C - T - G - C - C - C - T - A - T - C - C - C - C - C - C - T - G - A - G  (AP22)  5' - C - T - C - A - G - G - G - G - G - G - A - T - A - G - G - G - C - A - G - G - G - T - A - T - G - G - A - C - T - C - 3 '  30  -3'  PCR reactions were then subjected to Dpn I digestion for 4 hours at 37°C to remove the original plasmid (methylated) D N A . Following the digestion, the D N A contained within each 50ul reaction was precipitated by adding 3 M sodium acetate (5pl): 95% ethanol (lOOul) and incubation at -86°C for 30 minutes. Reaction were then centrifuged at 13000 rpm for 15 mins at room temperature (Eppendorf) to pellet D N A . Excess ethanol was removed and precipitated mutant D N A was resuspended with lOul TE buffer. 3pi of the D N A in TE buffer were then used to transform the mutant D N A into competent DH5-a E.coli using the standard (42°C ) heat-shock method. Transformed D H 5 a cells were grown in overnight cultures and the plasmid D N A was isolated and sent for sequencing. A l l sequencing reactions were performed by the N A P S unit (University of British Columbia). After the mutations for all variants were confirmed, all mutants D N A were transformed to E.coli BL21 for expression.  2.10  Expression and Purification of ai-AT and Its Variants. ai-AT and its variants were expressed and purified using the protocols described for  NS3 protease (section 2.3-2.4) with the following exception: The unbound proteins were removed with 3 sequential washes of 60 ml with buffer of decreasing pH value (0.5M NaCl, 50 m M sodium phosphate, pH 7.3, 5.8, and 5.2). ai-AT was eluted with 20 ml of I X elution buffer (0.5M NaCl, 50 m M sodium phosphate, pH 3.9). Eluted material was collected in different tubes and stored in aliquots at -86°C. Eluted fractions were also dialyzed against (50 m M Hepes, 150 m M NaCl, lOmM DTT, pH 7.5) for 3 X 1 hours using the slide-a-lyser dialysis cassette with a 10 kD size cut-off membrane (Pierce) to raise the pH of the purified  31  protein solution before analyzing the purity using SDS-PAGE analysis and Western blots (section 2.5).  2.11  Pancreatic Elastase Assays. Purified ai-AT and its variants were tested for their inhibitory properties against  elastase. Commercially available M C A substrate (MeoSuc-A-A-P-V-AMC) (Calbiochem) was used to detect the protease activity of elastase (Calbiochem). Continuous assays monitoring elastase activity and the inhibitory activities of ai-AT and variants were performed using a Gemini Spectromax X S 96 well plate spectrofluorometer. Enzymatic reactions were carried out at 31°C. The excitation and emission spectra for M C A were set at 370/460 nM. Each optimized elastase/inhibitor reaction contained the Hepes buffer (50 m M Hepes, 150 m M NaCl, 0.1% triton-X, lOmM DTT, pH 7.5), 9 n M of elastase, 50uM M C A substrate and 1.8 u M of FPLC purified ai-AT or its variants to make a total volume of lOOuL in every reaction well. The concentration of purified cti-AT and its variants were determined using standard Bradford assays (Biorad).  2.12  Quantification of Proteins. Protein concentrations were determined using the Bio-Rad protein assay kit (micro-  assay), which is based on the Bradford dye-binding protein assay, according to the manufacturer's instructions. A standard curve of protein concentration at absorbance 595 nm was produced, using B S A as the standard protein at the following concentrations (ug/ml); 50, 100, 200, 400, 800, 1000. The absorbance of the sample proteins was measured and protein concentrations were determined from the standard curve.  32  2.13  Complex Formation Studies with ai-AT and Its Mutant Variants against Elastase, NS3 Protease and FLNS3. Elastase complex formation was performed with 1.8 u M of cti-AT and its variants  using the elastase/inhibitor assay buffer (section 2.11) without the addition of M C A substrate. Some reactions were performed with varying concentrations of elastase or different complex incubation times (variations are detailed in the appropriate figure legends in the Result section). NS3 protease complex formation with cti-AT and its variants was performed using the NS3 protease enzymatic assay buffer, NS4A (section 2.6) with the addition of 1.3 u M of serpins. Some reactions were carried out with varying concentration of NS3 protease or different complex incubation times (variations are detailed in the figure legends of the Result section). FLNS3 complex formation with a i - A T and its variants was performed using the NS3 protease enzymatic assay buffer, NS4A (section 2.6) with the addition of 20ul of FPLC purified FLNS3 and 1.3 u M of serpins. The complex formation reaction was allowed to proceed for 4 hours. A l l complex formation reactions were analyzed using SDS-PAGE and Western blot (section 2.5). Anti histidine antibody was used to detect NS3 protease or elastase and a i - A T or all its variants complex formation. Anti FLNS3 antibody was used to detect complex formation between FLNS3 and a i - A T or all its variants.  2.14  Inhibition Assay of Elastase and NS3 Protease by ai-AT and All Its Variants. Inhibition of elastase proteolytic activity by cti-AT and its variants was performed  using elastase/inhibitor assay buffer (section 2.11). Titration assays of elastase activity were performed against cti-AT, Mut A and Mut D with the same buffer condition (section 2.11)  33  except with varying amount of serpins included (variations are detailed in the figure legends of the Result section). In titration assay, the serpins were incubated with elastase for 45 minutes prior to adding the M C A substrate and performing continuous kinetic assay for 3 hours. Progress curve assay of elastase with cti-AT and its variants was also performed with buffer condition (section 2.11) except the amount of elastase used was 9 nM, and the amount of serpins used were also varied (variations are explained in the figure legends of the Result section). In the progress curve assay, elastase, serpins and substrate were added at the same time prior to performing continuous kinetic assay for 3 hours. Inhibition assay of NS3 protease by a i - A T and its variants were performed using the optimized NS3 protease enzymatic assay buffer, (section 2.6) with the addition of 1.3 u M of serpins, however new stock of peptide NS4A cofactor was used for this experiment and was acquired from (Genome B C Proteomic Centre, U.Victoria, BC). NS3 protease and NS4A cofactor were preincubated for 15 minutes first, followed by further 4 hours incubation with the serpins at 31°C. Abz-E-D-V-V-C-C-S-M-S-Y-Q-Y(N0 ) IQFS substrate (Multiple Peptide Sequence, 2  California) was added  immediately before taking continuous kinetic measurements for 2  hours. A l l readings were taken in duplicate.  Calculation  of the Specific Inhibition Constant (Kj) and Stochiometry of  Inhibition (Si). The Specific Inhibition constant (K;) is a concentration that is used to describe how potent a specific inhibitor is against its target. Titration assays with different amount of inhibitors were performed to calculate the amount of residual protease activity. The various residual protease activities were then plotted against the concentration of inhibitor added.  34  From the resulting graph, the value of (Ki) and (S;) can be obtained (for sample calculation see Appendix I).  2.15  Model Structure of NS3 Protease/NS4A Complex, Helicase/NTPase and oci-AT/Elastase Complex Stucture. A l l ribbon representation of the crystal structures was generated by using the Sybyl  version 6.7 software (Tripos, St-Louis).  35  2.16  List of Buffer Solutions  Dialysis buffer-50 m M Hepes, 150 m M NaCl, lOmM DTT, p H 7.5 DNA loading buffer (6X)-4% sucrose, 0.25% bromphenol blue, 0.25% xylene cyanol Enzymatic assay and complex reaction buffer-50 m M Hepes, 150 m M NaCl, 0.1% tritonX, 10mMDTT,pH7.5 Gel destain-5% (v/v) methanol, 7% (v/v) acetic acid, 88% (v/v) water Resolving gel buffer-0.74 M Tris-HCL (pH 8.0), 1% SDS SDS-PAGE sample buffer (6X)-0.28 M Tris-CL, 30% (v/v) glycerol, 1% (w/v) SDS, 0.5 M DTT, 0.0012% (w/v) brompheno blue SDS-PAGE resolving buffer (5X)-0.125 M Tris base, 0.96 M glycine, 0.5% (w/v) SDS Stacking gel buffer-0.122 M Tris-HCL (pH 6.7), 0.1% SDS Sodium phosphate FPLC washing buffer (dibasic)-0.5M NaCl, 50 m M sodium phosphate (dibasic), pH 7.3, 6.0, 5.8, 5.2,4.8 adjusted using sodium phosphate (monobasic) buffer Sodium phosphate FPLC elution buffer (monobasic)-0.5M NaCl, 50 m M sodium phosphate (monobasic), pH 3.9 adjusted using phosphoric acids TBE (lX)-89 m M Tris base, 89 m M boric acid, 2 m M E D T A TBS-100 m M Tris-Cl, pH7.5, 0.9% (w/v) NaCl TBS-Tween-TBS with 0.05% Tween 20 TE buffer-10 m M Tris-Cl, 1 m M E D T A , pH 8.0 Western blot transfer buffer-39 m M Glycine, 48 m M Tris, 0.037% (w/v) SDS, 20% MeOH Western blocking buffer (qiagen)-Qiagen blocking reagent buffer diluted 1/10 in distilled water (500 ml), 5 g of Qiagen blocking reagent, 0.1% Tween 20 Western blocking buffer (milk)-5% milk in TBS-T  36  Chapter 3: Expressions, Purification and Characterization of NS3 Protease Using Internally Quenched Fluorogenic Substrates (IQFS) 3.1  Expression of Recombinant NS3 Protease in pLys BL21 E.coli The serine protease domain of H C V NS3 has previously been expressed in E.coli  HBT01 and purified by affinity chromatography using a nickel-agarose column (Qiagen) (Vishnuvardhan et al., 1997). However, this purification method was performed manually and hence was labor-intensive and time-consuming. This chapter presents work on the development of a novel expression and automated purification method for NS3 protease using Fast Protein Liquid Chromatography (FPLC). Two different strategies have been used successfully to elute histidine-tagged NS3 protease bound to a Ni  2+  column: 1) stepwise  decrease in buffer p H value, 2) increasing imidazole concentration (Sali et al., 1998; Vishnuvardhan et al., 1997). In this study, the strategy used for the purification of the NS3 protease was a stepwise decrease in buffer p H values to elute the NS3 protease. The transformation of pLysS BL21 E.coli with the NS3 protease plasmid construct was successful. pLysS BL21 E.coli was selected as the suitable host for the expression of the NS3 protease because the pLys plasmid encodes the T7 lysozyme that stabilizes the plasmid carrying the NS3 protease gene. BL21 E.coli strain is also protease-deficient. It was chosen in order to eliminate the degradation of the NS3 protease by the bacterial proteases during expression (Grodberg and Dunn., 1988). The expression of soluble NS3 protease was found to be temperature-dependent, because induction of protein expression above 31°C led to the production of insoluble NS3 protease. SDS-PAGE analysis of (+) IPTG and (-) 1PTG  37  bacterial cell cultures after 18 hours of induction showed that a protein band of molecular weight of approximately 23 kDa was induced, and this was consistent with the molecular mass of the NS3 protease calculated from the primary sequence (data not shown). To further confirm the identity of the 23 kDa band, RGS-histidine-tag-specific mouse monoclonal antibody was used to probe Western blots of proteins from induced cell culture. O f the three different primary antibodies tried against the histidine tag of NS3 protease, The RGS mouse anti- histidine antibody seemed to be the best choice, because it specifically bound to the NS3 protease when it was induced in BL21 E.coli cells. The Western blot result shows that there was an intense band at the 23 kDa position only in the IPTG-induced culture, which was an indication that the 23 kDa protein band was the NS3 protease (data not shown).  3.2  Purification of Recombinant NS3 Protease from E.coli Supernatant Prior to proceeding with FPLC purification of the NS3 protease, the enzymatic  activity of the NS3 protease in the crude bacterial cell lysate was tested in an assay using M C A substrate (data not shown). Having established that the expressed NS3 protease was proteoliticly active, FPLC purification of the NS3 protease from the crude supernatant was attempted. A pre-packed 1ml nickel-sepharose column (Pharmacia) suitable for affinity chromatography was used to purify histidine-tagged NS3 protease. Initial attempts at FPLC purification of the NS3 protease crude supernatant using the nickel column did not yield protein of sufficient purity for our assay. Hence, a variety of protocols including changing the pH of the washing and elution buffers, and changing the number and length of washes, were tested in order to optimize the purification conditions. This resulted in the production of much higher purity of NS3 protease and a clear elution peak (Fig. 3.1). The purity of the  38  0  10 20 30 40 50 60 70 80 90 100 110 120 130  Volume (ml) Figure 3.1. Purification of bacterially expressed histidine-tagged NS3 protease. Representative chromatogram of NS3 protease purification using N , binding interaction-chromatography. NS3 protease was eluted at pH -4.0. Inset. SDS/PAGE (lanes 1 and 2) and western blot (lanes 3 and 4). This procedure yielded recombinant NS3 protease (-23 kDa) that was essentially pure and intact as determined by coomassie blue staining (lane 2) and western blot analysis (lane 4). Histidine-tagged NS3 protease was detected by western blot using anti-histidine tag antibody. 2 +  39  eluted NS3 protease was 96%, as determined using an alpha imager spot densitometer. Protein elution was monitored using two wavelengths: 220 nra (adsorption of peptide bonds linking amino acids) and 280 nm (adsorption by aromatic amino acids, tyrosine and tryptophan) (Fig 3.1). In addition, the chromatogram also shows the p H value in the column throughout the purification. Following the binding of the crude bacterial cell lysate onto the N; column, thorough washes using buffers at decreasing pH values (7.3, 6.0, and 4.8) were 2+  carried out to remove any proteins bound non-specifically to the N ;  2 +  column. Histidine-  tagged NS3 protease eluted at pH 4.0 because when the p H was lowered to pH 4.0, the protonation of the 6X-histidine segment disrupted its interaction with the Ni  2+  and caused it  to be repelled from the positively charged nickel ions and thus release the histidine-tagged NS3 protease (Gagnon., 1999). The main advantage of using FPLC to purify recombinant protein is that once the protocols are determined, the purification procedure can be made fully automated to yield consistent results. Moreover, the FPLC machine can be operated at 4°C to decrease protein degradation during the purification process. To analyse the protein content of the different washing and elution peaks, SDS P A G E was performed. As expected, both the injection peak and the washing peaks eluted during the purification process contained many different proteins, including some NS3 proteases (data not shown). However, the peak eluted at p H 4.0 (collected in 20 X 1 ml fractions) contained -96% pure NS3 protease as shown by the SDS-PAGE analysis (Inset of fig 3.1). The purified NS3 protease migrated as a single band with a molecular mass of -23 kDa. The identity of this band was further confirmed by Western blot analysis with the RGS antibody directed against the histidine tag (Inset of Fig 3.1).  40  3.3  Optimization of Assay Conditions for Detection of NS3 Protease Activity Knowing the important role that the NS4A cofactor has on the enzymatic activity of  the H C V NS3 protease, it was important to ascertain the concentration of peptide NS4A cofactor necessary to induce optimal activity of the newly purified H C V NS3 protease. This process was part of the development procedure to find out the best buffer conditions and parameters to use in an enzyme kinetics experiment. A l l enzymatic assays were performed using the Gemini Spectramax X S 96-well plate spectro-fluorometer. Various experiments were designed to determine suitable buffer conditions for the NS3 protease assay. The parameters which were investigated included the buffer type, pH, temperature, effects of detergents and salt concentrations. The result of these experiments indicated that the optimal buffer to use was 50 m M Hepes, 150 m M NaCl, 0.1% Triton-X, lOmM DTT, p H 7.5. It was also found that 15uM of NS4A peptide cofactor was sufficient to induce the desired activity of the H C V NS3 protease under the conditions used (Fig 3.2. A). This result agrees with other published results (Zhang et al., 1997; Steinkuhler et al., 1996) which report using the same concentrations of peptide cofactor to perform NS3 protease assays. The amount of cofactor needed to activate the proteolytic activity of the NS3 protease to half of its maximum level or the affinity constant (Kdo. ) of peptide NS4A cofactor to the NS3 protease was about 4.5 u M . 5  In order to test whether the requirement of cofactor in the activation of the NS3 protease activity was specific to H C V NS4A, cofactors from closely-related Flaviviridae viruses: G B V - A , G B V - B , G B V - C and B V D - V all of which share ~ 30% amino acid sequence homology with the H C V NS4A cofactor were tested using the M C A substrate (Butkiewicz et al., 2000). Results showed that only the H C V NS4A cofactor gave rise to NS3 protease  41  04 0  ,  10  ,  20  ,  30  ,  r  40  SO  [Protease cofactor] (uM)  ,  60  Activation of H C V N S 3 protease activity by H C V , G B V - A , B , C & B V D - V N S 4 A peptide cofactors  E C  60  HCV  GBV-A  GBV-B  GBV-C  BVD-V  HCV NS3 protease with different NS4A cofactors  Figure 3.2. NS3 protease activity titration with Pep4A i-34. (A) Cofactor assay was done with the standardized enzymatic buffer show that 15 u M NS4A cofactor is necessary and sufficient to activate optimal NS3 protease activity. NS4A cofactor was pre-incubated with 20ul of NS3 protease for 15 minutes before performing the enzymatic assay with lOOuM IQFS-1 for 1 hour. (B) 50uM of the peptide cofactors from related Flaviviridae virus ( G B V A , B , C and B V D - V ) failed to activate NS3 protease activity showing that NS3 protease activity is fully dependent and specific for its own H C V - N S 4 A cofactor. Enzymatic assay was done for 1 hour with 15 minutes prior incubation between NS3 protease and NS4 A cofactors. A l l experiments were done in triplicate. Buffers compositions are described in Materials and Methods section 2.6. 2  42  proteolytic activity (Fig. 3.2.B). These results agree with previous findings which showed that NS3 protease activity is highly dependent on and specific for its own cofactor, the H C V NS4A cofactor (Butkiewicz et al., 2000).  3.4  Probing the Substrate Specificity of NS3 Protease Using H C V IQFS Substrate recognition by all serine proteases involves binding of the substrate onto the  active site surface that contains catalytic triads. In order to design a potent protease inhibitor against the H C V NS3 protease, it is important to characterize the substrate specificity of the H C V NS3 protease. In order to address this question, the ability of NS3 protease to cleave IQFS substrates corresponding to the H C V polyprotein junctions NS4A/4B, NS4B/5A and NS5A/5B was tested. IQFS substrates which contain H C V NS junctions P6-P'4 residues are suitable for use in this experiment because they can be cleaved by both the NS3 protease and the FLNS3. In contrast the M C A substrate which contains only the P-side residues can only recognized and cleaved only by NS3 protease (Fig. 3.3). Since the goal of the Dr. Jean's laboratory is to determine the substrate specificity of the NS3 protease and the FLNS3 using the same substrate, only IQFS substrate could be used to perform this experiment. Determining the substrate specificity of the NS3 protease and the FLNS3 also helped to establish which of the NS junctions was the most efficiently cleaved. NS3/4A IQFS junction was not included in this study because it was previously reported that this junction was not cleaved, probably supporting the notion that the site is recognized in an exclusively cisintramolecular fashion (Steinkuhler et al., 1996). In addition, a modified form of the NS5A/5B junction IQFS which was present in high quantity was initially used to test if the NS3 protease proteolytic activity could be  43  A  NH2 C-Xaa -Xaa2-Xaai-X'aa -NH-C^^ n  n  NO  2  Abz-Asp-Asp-lle-Val-Pro-Cys-Ser-Met-Ser-Gln-EDD  IQFS-1 CH3  Ac-Glu-Glu-val-Val-Ala-Cys-MCA  MCA-1  Figure 3.3. The intramolecularly quenched fluorogenic substrates (IQFSs) are suitable for continuous monitoring of HCV NS3/4A protease complex activity. (A) Structural representation of peptidyl MCA-1 and IQFS-1 (modified NS5A/5B). (B) Discrepancy is observed for NS3 variants in the efficiency of cleavage of M C A - 1 and IQFS-1. Only IQFS can be used to measure the activity of FLNS3.  44  monitored using the IQFS substrate. The modified NS5A/5B IQFS has a slightly different sequence from that of the original NS5A/5B B K strain and was cleaved with the highest efficiency by the NS3 protease. Each IQFS substrate was used at varying concentrations (5,10,20,50,100,200 uM) in assay reactions containing constant concentrations of the NS3 protease. The slope of the initial velocity of each reaction was calculated to plot the Michaelis-Menten curves of initial velocity ( V ) against the IQFS concentration. K 0  V  m a x  m  and  values were then obtained from the Michaelis-Menten plots for each different IQFS  junction (Fig. 3.4). K  m  is defined as the Michaelis constant corresponding to the  concentration of substrate at which the rate of the reaction is half of the maximum velocity, (Vmax).  K and Vmax values for each H C V (NS) junction IQFS are in Fig. 3.5. V x / K m values m  ma  are an indication of the relative cleavage efficiency of the NS3 protease at each different junction, since the concentration of the NS3 protease in these kinetic experiments was constant (see Materials and Methods, section 2.6). The V a x / K value for the modified m  m  NS5A/5B junction was the highest (0.46), followed by the B K strain NS5A/5B (0.13), NS4A/4B (0.05) and NS4B/5A (0.02). Based on the values of the Y /K max  m  calculated for the  different H C V (NS) junctions, a putative order of H C V (NS) polyprotein cleavage was proposed. Owing to its high cleavage efficiency, NS5A/5B junction could be cleaved first, followed by the NS4A/4B and NS4B/5A junctions. This result agrees with previous findings (Steinkuhler., 1996). Whether this putative NS cleavage sequence also happens during H C V polyprotein maturation replication in vivo is yet to be determined. Therefore, whether the putative cleavage order is significant in H C V replication in vivo is still unknown. In addition, IQFS 3 & 4 seem to exhibit internal quenching effects at high IQFS substrate concentrations [200 uM]. These phenomena could possibly be the result of product inhibition  45  B  Internally Quenched Fluorogenic Substrate I  HH  /-<^  *  ^-|-Xaa -X»a -X»« -X»« -NH<:H2-CHa-HH n  2  1  n  NOli  Fluorophore  Chromophore  IQFS-1: SA/SB-  *1>, - Asp . Asp - IK -  IQFS-2: 4A/4B  Aft* • Asp - Qlu - Mst - Olu - Glu - Cys • Ala - Ssr - His-  Vs! - Pro- C y s - S s r -  Mo! . Ssr- G i n - EDDnp  IQFS-3:4BISA  A b j . Asp - Cys - Ssr - Thr- P r o - C y s - S s r - Oly- Ssr - Trp - Gin - EDDnp  L s u - G i n - EDDnp  IQFS-J.1: 4B/SA" Ate - Olu - C y s - Thr- Thr- P r o - C y s - S s r - O l y - S s r - Trp - G i n . EDDnp IQFS-4:5A/SB  IQFS-1: 5A/5B*  50  100  A b l - Olu- A s p - V a l -  V s l - Cys - cys - S s r - Mst - Ssr- Tyr-  Gin - EDDnp  IQFS-2: 4A/4B  150  [Substrate] (uM)  200  50  100  150  200  250  200  250  [Substrate] (uM)  IQFS-3:4B/5A  IQFS-4:5A/5B  50  [Substrate] (uM)  100  150  [Substrate] (uM)  Figure 3.4. Substrate specificity of NS3 protease. (A) Natural cleavage sites of the H C V NS polyprotein. (B) Quenched fluorogenic substrates. (C) Representative plots of the initial rate of cleavage (rate) versus substrate concentration [5,10,20,50,100,200 uM], for the selected substrates, IQFS-1 (modified NS5A/5B), IQFS-2, IQFS-3 and IQFS-4, obtained with 20ul of NS3 protease with 15 u M of NS4A. Enzymatic assays were done in triplicates and data analyses were carried out as described in the Materials and Methods section 2.6 and 2.7.  46  r  T NS3  f  I NS4A  NS4B  NS5B  t  t  1 t  2  3  1  i  HCV-Non Structural Junction  K (uM)  NS3/4A NS4A/4B NS4B/5A NS5A/5B NS5A/5B (Modified)  N/A 52 32 22 54  m  1  V  * max  (Rfu/min) N/A  2.7 0.6 2.9 25  V  /K  Relative Cleavage (Rfu/min*^M) Efficiency (%) N/A N/A 0.05 38 2 0.02 15 3 0.13 100 1 0.46 N/A y  max'  i v  m  Figure 3.5. NS3/Pep4A protease complex activities on peptide substrates corresponding to trans-cleavage sites. NS5A/5B (modified) junction was cleaved most efficiently by NS3/Pep4A protease complex followed by natural NS5A/5B, NS4A/4B and NS4B/5A junctions.  47  by the N-terminal cleavage products of substrate peptides corresponding to the NS4A/4B, NS4B/5A and NS5A/5B cleavage sites, as described in previous work (Steinkuhler et al., 1998). The NS3 protease also displays a high affinity for N-terminal cleavage products of the original substrate, especially with the N-terminal sequence of the original NS5A/5B junction. The modified NS5A/5B substrate did not show any sign of product inhibition, probably due to the nature of its N-terminal sequence. This phenomenon might partly explain the selfregulation and slow propagation of the H C V in infected individuals, as product inhibition by the N-terminal of NS5A/5B junction will inhibit the activity of NS3 protease and stops the processing of other non-structural junctions.  3.5  Probing the Substrate Specificity of NS3 Protease Using G B V IQFS The substrate specificity of NS3 protease against G B V - A & G B V - B virus polyprotein  IQFS substrate was also investigated. G B viruses A & B are closely related to H C V and cause acute hepatitis in tamarins {Sanguinus species), making them attractive surrogate viruses for in vivo testing of anti-HCV inhibitors in a small monkey model. It has been reported that the NS3 protease of G B V - B shares similar substrate specificities with its counterpart in H C V (Scarselli et al., 1997). However, results from our laboratory indicate this is not the case. Figure 3.6. (A) shows the amino acid sequences corresponding to the (NS) cleavage junctions of G B V - A & B , with the putative PI and P6 residues highlighted in red. O f all the G B V IQFS substrates tested using the NS3 protease and the peptide NS4A cofactor, only the NS4A/4B junction of G B V - B was cleaved efficiently under the same assay conditions used to test the NS3 protease against the H C V IQFS substrates. The cleavage  48  A  Internally Quenched Fluorogenic Substrate (IQFS): G B V - A and G B V - B GBV-A NS3-NS4A NS4A-NS4B NS4B-NS5A NS5A-NS5B  Abz-S-L-V-V-V-T-S-W-V-V-Q-EDDnp Abz-T-L-E-T-A-C-G-W-G-P-Q-EDDnp Abz-E-M-E-T-P-A-S-Q-I-V-Q-EDDnp Abz-E-E-E-T-P-T-S-Y-S-Y-Q-EDDnp  not-cleaved not-cleaved not-cleaved not-cleaved  GBV-B NS3-NS4A NS4A-NS4B NS4B-NS5A NS5A-NS5B  Abz-F-T-E-V-N-T-S-G-T-A-Q-EDDnp Abz-E-I-V-E-E-C-A-S-F-I-Q-EDDnp Abz-T-P-T-E-D-D-C-G-L-I-Q-EDDnp Abz-K-S-E-F-S-C-S-M-S-Y-Q-EDDnp  0  20  40  60  80  100  not-cleaved  cleaved  not-cleaved not-cleaved  120  140  160  [Substrate] (uM)  Figure 3.6. NS3/Pep4A protease complex specificities towards G B V - A and G B V - B IQFS. (A) Lists of G B V - A & B IQFS from G B V - A & B non-structural pro-protein junctions that are tested against H C V NS3/Pep4A protease complex. (B) Efficiency of cleavage of internally quenched fluorogenic G B V - B (NS4A/4B) non-structural proprotein related substrate by 20 ul of NS3 protease. Concentration of G B V - B IQFS used were: 5, 10,20, 50, 100, 150 (uM). A l l enzyme kinetics assays were done for 1 hour in duplicates. Buffers compositions are described in Materials and Methods section 2.6.  49  A  Figure 3.7. Ribbon representation of the catalytic pocket of NS3/Pep4A protease complex. (A) Ribbon representation of the NS3/Pep4A protease. (B) Close up of ribbon representation of the binding pocket of the NS3/Pep4A protease catalytic site bound to P6-P'4 decapeptide inhibitors. Color code for NS3/4A: Yellow, hydrophobic, red, negative charge density, blue, positive charge denstiy, magenta, NS4A residue; Color code for the peptide inhibitor: green, carbon, red, oxygen, blue, nitrogen, yellow, sulfur; Hydrogens omitted for clarity (Ingallinella., 2000). 50  efficiency of this junction (Vmax/K = 0.33) (Fig. 3.6 B) was comparable to the cleavage m  efficiency of the H C V (NS) junctions. The reason why only the NS4A/4B junction of the G B V - B was cleaved is probably due to the presence of cysteine residue in the PI position, glutamic acid in the p6 position and alanine in the P ' l position. Having these three amino acids requirements in their respective positions was essential for the efficient cleavage of substrate by NS3 protease, as previously described in the Introduction. None of the other G B V substrates share these sequence requirements for the efficient cleavage by the NS3 protease. Although the NS5A/5B junction of the G B V - B has PI-cysteine and a P'l-serine, it was not cleaved efficiently. It has been suggested that although an acidic residue at P6 is not essential for cleavage by the NS3 protease, the overall negative charge in the P-region of the substrate might be important for improving binding and cleavage efficiency (De Francesco & Steinkuhler., 1999). Figure 3.7 (A&B) highlights the fact that the active site of the H C V NS3 protease is very hydrophobic, with some positive charge around it. Therefore, the presence of a positively charged lysine in P6 of the G B V NS5A/5B junction may have a big impact on the overall charge to the P-region resulting in the disruption of the enzyme/protease interaction. The P-region could be de-stabilized due to the repulsion caused by the introduction of a positively charged area around the S6 position of the NS3 protease induced by the arginine residue. (Fig 3.7 B) (Ingalinella et al., 2000).  3.6  Summary  This chapter describes a successful novel expression and purification protocol for recombinant histidine tagged H C V NS3 protease using a fully automated FPLC system. In  51  addition, enzymatic assay parameters and buffer requirements for characterizing the NS3 protease were determined and used for all subsequent enzymatic experiments involving the NS3 protease. A comparison of the results from the substrate specificity studies using the HCV NS3 protease and the HCV or GBV-A & B IQFS substrates indicates that the presence of cysteine in PI, aspartic/glutamic acid in P6 and serine/alanine in P ' l are required for the efficient cleavage of the substrates by the HCV NS3 protease under the condition used. The Vmax/Km values for modified NS5A/5B junction were the highest (0.46), followed by BK strain NS5A/5B (0.13), NS4A/4B (0.05), and NS4B/5A (0.02). These results will be useful for the design of effective inhibitors against the HCV NS3 protease.  52  Chapter 4: Bio-engineering, Expression & Purification of ai-AT Variants (mutants A-E)  4.1  Sub-cloning of cti-AT from pDS-56 Vector to pET-21 Vector a i - A T was chosen as the first serpin scaffold to initiate protein engineering studies  with the aim of specifically targeting the H C V NS3 protease. It was selected due to its ability to recognize chymotrypsin-like serine proteases and because of previous success with using a i - A T variants as a scaffold for protein-based therapeutics (Jean et al., 1998). In addition, cci-AT is mainly produced in liver cells, the target cells of an H C V infection (Ray et al., 1977; Boskovic et al., 1998). Since H C V primarily infects the liver cells, it was of interest to investigate the possibility of wild-type cti-AT interacting with H C V NS3 protease and to find out whether the bio-engineered cci-AT could improve the interaction and specificity between the NS3 protease and serpins. Prior to commencing sub-cloning of ai-AT, the pDS-56-ai-AT plasmid was sequenced to confirm that there were no mutations from the wild-type sequence. As the sequence obtained was in agreement with the Gen-Bank published sequence, the plasmid was transformed into E.coli BL21 for preliminary expression and purification of the protein. However, a i - A T was inducibly expressed at very low levels and there was not a satisfactory quantity of it produced for purification. In addition, the mutant variants of pDS-56 cti-AT (Mut A & B) were also expressed in such a low quantity that attempts to purify those variants yielded unacceptably low amounts of protein (data not shown).  53  To solve this problem, a different expression vector was used. The pET-21 vector was chosen because it is the most powerful system developed to date for the cloning and expression of recombinant proteins in E.coli. It utilizes the T7 R N A polymerase that is selectively induced and promotes very high protein expression levels. Almost all of the cell's resources are converted to target gene expression; the desired product can comprise more than 50% of the total cell protein only a few hours after induction (Grodberg et al., 1988). The pET-21 vector was selected instead of other pET vectors because it contained the desired components and restriction sites (Xho 1 & BLPI) to allows the entire histidine tagged-cti-AT gene to be sub-cloned from the pDS-56 vector to the new pET-21 vector (Appendix U). The sub-cloning procedure was done according to the manufacturer's instructions (Novagen), pDS-cti-AT and pET-21 were propagated in DH5-ct E.coli and purified with a plasmid mini-prep kit (Qiagen). Approximately 1 ug of the purified cti-AT and pET-21 D N A was digested with Xho 1 & BLP I to drop out the a i - A T insert and to prepare the pET-21 vector for ligation, respectively. After separation on an agarose gel, the a i - A T insert and the pET-21 were recovered using the Qia-Quick Gel Extraction Kit (Qiagen). Ligation was then performed and ligated products were subjected to restriction digest using Xho I and Blp I to confirm successful sub-cloning. Agarose gel analysis showed that the sub-cloned pET-21 a l A T and the original pDS-56 oti-AT both released a i - A T insert (1320 bp) when digested with Xho I and Blp I. The released pDS-56 and pET-21 vectors were of correct size at 3413 bp and 5369 bp, respectively (Fig 4.1 A ) . These results confirmed that pDS-56 a i - A T was successfully sub-cloned in the pET-21 vector.  54  B  —DNA ladder pDS-56/a AT <XhoI/BlD I. r  -Protein ladder pDS-56/ ai-AT  pET-21/^-AT rXhoI/Blp I)  -pET-21/ai-AT I <&~\  5369 bppET-21 3413bppDS-56  - a A T ~ 4 9 kDA r  1320 bp a A T r  Protein ladder  a,-AT ~ 49 k D A  Figure 4.1. Subcloning of ai-AT from pDS-56 to pET-21. (A) ai-AT was successfully sub-cloned from pDS-56 to pET-21as indicated by the Xho I & Blp I digestion of pDS-56 ai-AT and pET-21 ai-AT analysed with D N A agarose gel. (B) SDS gel showing that pET-21 induced ai-AT is present in greater quantity than pDS-56 induced ai-AT. (C) Western blot of uninduced (<D) vs induced (I) pET-21 ai-AT probed with anti-histidine antibody.  55  Following the successful sub-cloning, pET-21 a i - A T and pDS-56 a i - A T expression plasmids transformed into BL21 E.coli were induced under the same conditions and culture samples from each pre-and post induction were analysed by SDS-PAGE to compare the level of expression of a i - A T . As expected, a i - A T (-49 KDa) was expressed at a much higher level from the pET-21 vector as compared to the pDS-56 vector (Fig 4.1 B). The identity of the expressed a i - A T was confirmed by Western blot analysis of un-induced and induced whole cell lysate from BL21, containing pET-21-cti-AT (Fig 4.1 C). The RGS anti-histidine antibody was specific in detection of the induced 6X-histidine tagged recombinant a i - A T (-49 KDa). Following the successful expression of a i - A T , site-directed mutagenesis of the cti-AT gene to create reactive site loop sequence variants was performed to target H C V NS3 protease.  4.2  Bio-engineering of a i - A T Variants Using information regarding the substrate specificity of the NS3 protease from  studies presented in Chapter 3 and others, it was concluded that P I , P6 & P ' l residues determined the cleavage efficiency of the NS3 protease. Previous studies also suggested that an alanine residue in P4 may decrease the cleavage efficiency of the NS3 protease (Zhang et al, 1997). Therefore, in order to better understand how these important residues within an Reactive site loop of a serpin acting as a pseudo-substrate for the NS3 protease would influence its inhibitory properties, 5 different mutants of a i - A T were designed and constructed for testing as potential inhibitors of the H C V NS3 protease. The cti-AT variants generated were:  56  1) Mutant A (Mut A) (PI, M->C) 2) Mutant B (Mut B) (PI, M->C), (P4, A->S), (P6, L->D) 3) Mutant C (Mut C) (PI, M->C),  (P6, L->D)  4) Mutant D (Mut D) (PI, M-»C), (P4, A-»S) 5) Mutant E (Mut E) (PI, M->C), (P4, A-»S), (P6, L->D), (P'l, S->P) Mut A and B were created first, followed by Mut C, D & E . A l l a A T mutants were r  generated by the P C R protocols presented in Section 2.9 and sequenced to ensure that no additional mutations were incorporated into the mutants (NAPS unit, UBC). Clustal W alignment of the nucleic acid and amino acid sequences of wild-type a i A T and all the variants were performed to highlight the detailed changes made to each mutant (Appendix IV & V ) . A l l mutants are identical to the original wild-type a i - A T sequence except for their reactive site loop region (Fig. 4.2 A). A comparison of the reactive site loop residues P6-P'4 of the a i - A T variants are presented (Fig 4.2 B). Ribbon representations of the backbone of the crystal structure of a i - A T wild-type as well as the reactive site loops of its variants, Mut (A-E), are also presented (Fig 4.3). Mutated residues P6, P4, PI & P ' l are shown in ball and stick representation and are solvent-exposed. Mut A was the first to be designed, as it is widely known that PI cysteine is an essential primary requirement for the recognition and cleavage by the H C V NS3 protease. We wanted to test whether the cysteine residue mutation alone in the R S L of a i - A T was sufficient to promote cleavage by the NS3 protease. The combination of RSL mutations introduced into Mut B was predicted to be appropriate to target the serpin to NS3 protease specifically. Mut C & D were generated later to investigate whether any of the P4 or P6 substitutions had a significant impact on improving the specificity of the mutant A  57  RSL (P* ,s—P' 394) 4  3  6Xhis/flag  ai-AT (426 aa)  6Xhis/flag-  Mut A  ( 6 aa)  6Xhis/flag-  Mut B  (426 aa)  6Xhis/flag~ I  42  (426 aa)  6Xhis/flag-  Mut D  (426 aa)  6Xh.s/flag-- M  Mut E  (426 aa)  RSL  B P6 L L D D L D  P5 E E E E E E  P4 A A S A  s s  P3 I I I I I I  P2 P P P P P P  PI M <  C  c c c  P'l  s s s s s p  P'2 I I I I I I  P'3 P P P P P P  P'4 P P P P P P  (Xi-AT  Mut A Mut B Mut D Mut E  Figure 4.2. ai-AT and bio-engineered variants. (A) Structural representation of a i - A T and its 5 different bio-engineered variants. A l l 6 constructs are identical from amino acid (1-384) and (395-426). The reactive site loop (RSL) of a i - A T consists of amino acid (385-394). (B) Detailed representation of the R S L to show the mutated amino acids within the R S L of each cti-AT variants. Red amino acids indicate that it has been mutated from the original a i - A T respective sequence.  58  Figure 4.3. Ribbon representations of the backbone of the crystal structure of the O i - A T wild-type and the reactive site loops of its variants mutant (A-E). Reactive site loops from P6-P'4 residues are shown in red. Mutated residues are shown in ball and stick representation.  59  towards the NS3 protease. Finally, Mut E was included to test i f inhibition by serpin can still be achieved when the reactive site loop can no longer be cleaved by the NS3 protease. The proline in P ' l should inhibit cleavage, as P ' l mutation from serine to proline was previously proven to be recognizable but uncleavable by the NS3 protease (Paolo et al., 2000; Richer et al., unpublished results).  4.3  Purification of cti-AT and Its Variants Plasmids encoding cti-AT and its variants were transformed and expressed in BL21  E.coli using the same protocols as the H C V NS3 protease. 1 L of bacterial cultures expressing a i - A T and the five variants were prepared for purification so that enough materials will be available to perform all related experiments from the same stock. However, some modification to the original protocol of purifying NS3 protease had to be introduced, since the washing and elution buffers used to elute pure NS3 protease did not result in the elution of high purity a i - A T (data not shown). Adjustments of the values of the pH step gradients (2  nd  wash-pH 5.8, 3 wash-pH 5.2 and elution-pH 3.9) were found to be sufficient rd  in order to yield a i - A T and its variants at -96% purity (Fig. 4.4 & 4.5). Despite the success of purifying a i - A T and Mut A , B , D & E , purification of mutant C proved to be difficult. There were 3 lower molecular weight protein bands that were difficult to remove during the purification of mutant C (Fig. 4.5). A l l purified a i - A T and variants were subjected to buffer exchange dialysis to avoid long-term protein instability in low pH environment (pH 3.9). Dialysis buffer consisted of 50mM Hepes, 150 m M NaCL, lOmM DTT and pH 7.5. DTT was included to prevent the formation of disulfide bridges between serpins via the newly introduced cysteine residue at  60  Addy p u r i f i c a t i o n 9 : 1 „ U V J 280nm Addy punfieation49:l_ p H '' '  ' Addy p u r i f i ^ » K m 4 9 : 1 „ l J V 3 _ 2 2 0 u m Addy purificadon49: I^Fractions^  Volume (ml)  Figure 4.4. Purification of bacterially expressed histidine-tagged a i - A T and its variants. Representative chromatogram of a i - A T purification using H binding-interaction chromatography. Inset. SDS/PAGE (lanes 1 and 2) and western blot (lane 3). This procedure yielded recombinant a i - A T and its variants (-49 kDa) that were essentially pure and intact, as determined by the Coomasie blue staining (lane 2) and western blot analysis (lane 3). Histidinetagged cti-AT were detected by western blot using anti-histidine antibody. 2  61  +  •n, 205 116 97 80 66  Protein ! 77 Ladder  ai-AT  Mut A  MutB  MutC  MutD  MutE  *- ftllM* limy"  <*49 kDa  Figure 4.5. S D S - P A G E analysis of a A T and all variants (Mut A - E ) . SDSP A G E result shows that FPLC purified cci-AT and its variants are -96% pure except for mutant C -77% pure as determined by the densitometer. Recombinant histidinetagged cci-AT and all variants are - 49 kda. r  62  the PI site. The 3 X 1 hours buffer exchange dialysis procedure using a lOkD cut off dialysis membrane was carried out according to manufacturer's recommendation and was sufficient to raise the p H of the purified protein solution to pH 7.5. Dialyzed proteins were then aliquoted in 200 pi aliquot, snap-frozen and stored at -86°C.  4.4  Bio-engineering of K D E L Mutant A & D The carboxy terminal tetra-peptide, K D E L , is found in many lumenal E R resident  proteins. When attached onto various molecules, it results in their localization to the endoplasmic reticulum (ER), showing that it is a necessary sequence determinant for this process (Pelham, 1989). E R localization can be achieved in two ways: a) prevention of E R resident proteins from entering newly formed transport vesicles and b) retrieval of those E R residents that escape. Inclusion of a K D E L motif into the amino acid sequence of the desired protein is a useful tool in order to direct it to the E R (Teasdale and Jackson., 1996). A mammalian expression plasmid containing C-terminally K D E L tagged a i - A T sequence was constructed and cloned in the p-CMV vector by our collaborator at Yale University (Appendix JJI). I wished to express KDEL-tagged cti-AT, because it has been shown that the non-structural proteins of H C V are associated with the E R membranes, confirming the hypothesis that the E R is the site of membrane-associated H C V R N A replication (Mottola et al, 2002). In addition to the original a i - A T - K D E L , I have also bio-engineered K D E L Mut A & D according to the P C R protocols described in Section 2.9. Clustal W alignment of all K D E L constructs and translated amino acid alignments are provided in (Appendix V I & VII). The K D E L constructs were sequenced to confirm that no additional mutations were  63  introduced. Preliminary experiments investigating the intracellular localization of oti-ATK D E L variants by immuno-fluorescents were performed by Dr. Pamela Hamill (Chapter 6).  64  Chapter 5: Characterization of the Ability of ai-AT and RSL Variants to Form Complexes with and Inhibit the Proteolytic Activity of Elastase  5.1  <Xi-AT, a Natural Inhibitor of Neutrophil and Pancreatic Elastase In order to determine whether purified recombinant a i - A T and its variants were  biologically active, it was necessary to test their ability to inhibit pancreatic elastase, one of the natural protease targets of cti-AT (Carrell, 1986). There have been various previous attempts to construct variants of cti-AT with improved inhibitory properties against elastase (Kuisetti and Travis., 1996; Matheson et al., 1986). Hence, it was also of interest to determine whether the cti-AT variants created in our laboratory were more effective inhibitors of elastase than the wild-type ai-AT.  5.2  Complex Formation Between ai-AT, Mutant A & D with Elastase A common way to investigate the biological inhibitory function of serpins is to  determine whether they can form SDS-stable complexes with their target protease (see Section 1.3.3). To perform complex formation experiments, suitable reaction buffers and reaction conditions had to be established. After several experiments to optimize the conditions required for ai-AT/elastase complex formation, it was found that a buffer comprised of 50mM Hepes, 150mM NaCl, 0.1% triton-X, 10 m M DTT, p H 7.5 along with 9 n M of elastase and 1.8 u M of purified a i - A T or its variants in a total volume of lOOul  65  reaction mixture was the optimal condition for the detection of complex formations between a i - A T and its variants with elastase. Complex formation experiments were conducted for 1 hourat31°C. cti-AT and its variants have the molecular mass of approximately 49 kDa, whereas pancreatic elastase has the molecular weight of about 21 kDa. Therefore, the molecular weight of the SDS-stable complex between a i - A T and its variants with pancreatic elastase should be ~ 70 kDa. Figure 5.1 A shows that a i - A T and all its variants were stable by themselves. When elastase was added, some of the a i - A T , Mut A and Mut D were cleaved (molecular weight of -45 kDa) and some formed SDS-heat stable complexes with elastase (molecular weight of -70 kDa). These results suggest that a i - A T , Mut A & Mut D produced in the laboratory still retained their biological functions and could represent effective elastase inhibitors. Cleavage of cti-AT and other serpins can occur in 2 different circumstances: 1) cti-AT forms an irreversible inhibitor complex with the target protease, which subsequently gets degraded, or 2) a i - A T is cleaved by target or non-target protease, usually at the R S L region, without the formation of stable inhibitor complexes (Fig. 5.2) (Janciauskiene., 2001). Such cleavages generate a 4 kDa carboxyl-terminal fragment of - 36 residues, which remains noncovalently bound to the cleaved cti-AT. Therefore, upon the addition of denaturing agent (SDS) and heating at 95°C, this 4 kDa fragment is released and the resulting cleaved serpin has a lower molecular weight of -45 kDa as shown in the case of cti-AT, Mut A & D (Fig 5.1 A). In the case of cti-AT, Mut A & D , both pathways are probably occurring simultaneously, as shown by the formation of both SDS stable complexes and cleaved serpins.  66  B  F i g u r e 5.1. Serpin complex f o r m a t i o n w i t h elastase. (A) 1 hour complex formation  reaction of 1.8 u M of cti-AT and all its variants vs 900 n M elastase. Only ai-AT, Mut A , and D can form a complex with elastase. Mut B , C & E were all degraded during incubation reaction with elastase. (B) Increasing amount of complex formation was observed with the increasing [elastase] (nM) used vs 1.8 u M of ai-AT, Mut A & Mut D (1 hour incubation). (E = elastase ; EI* = enzyme/inhibitor complex ; I = inhibitor ; F = cleaved inhibitor). Buffer compositions for these experiments are described in Materials & Methods sections 2.11 & 2.13. A l l western blots were probed with RGS anti-histidine mouse antibody. 67  E + I <=^EI  EI  Complex inhibitory pathway  E +r  Cleaved substrate pathway  Figure 5.2. Branched pathway mechanism of serpins as suicide substrate inhibitors. I represents the serpin and E represent the protease. EI represent protease/serpin intermediate, EI* represent protease/serpin complex and 1° represent cleaved serpin.  In order to further investigate the complex formation between cti-AT, Mut A & D and elastase, experiments using varying amounts of elastase (90, 120, 180, 300 & 900 nM) reacting with the three serpins were carried out. In general, increasing elastase concentration led to greater amounts of complex formation. However, 180 n M of elastase was sufficient to produce a strong and visible complex between a i - A T , Mut A & D (Fig 5.2 B). In most cases, increasing the concentration of elastase to more than 180 n M did not increase the amount of complex formation. This could be due to the limited amount of active serpins that were present to form complexes with elastase. Nonetheless, results from these experiments showed that the amount of SDS-heat stable complex formation between oti-AT, Mut A & D was dependent upon the concentration of elastase, indicating that complex formation was specific.  5.3  Degradation of Mutant B, C & E by Elastase Muts B , C & E, unlike cti-AT, Mut A & D, failed to form SDS-heat stable complexes  with elastase. Not only did they fail to complex with elastase, they were also completely  68  degraded by elastase by the end of the 1 hour complex formation reaction, as indicated by the failure to detect any protein band of the right molecular weight (Fig 5.3). This phenomenon is very unusual and interesting. The main difference between cti-AT, Mut A & D versus Mut B , C & E was that the former three serpins do not contain any P6 substitution, whereas the latter three serpins all had their P6 residues mutated from leucine to aspartic acid (Section 4.2). There is some evidence that the presence of hydrophobic leucine in P6 is important to allow the conformational change required for a stable complex formation between a i - A T and its target protease (Warshel et al., 1991). Insertion of a bulkier and charged side chain such as arginine or aspartic acid in this case favors the substrate pathway in which the serpin is cleaved into its 45 kDa form and then released from the target protease still in its active form (Dufour et al., 2001). This might explain why Mut B , C & E were all further degraded by active elastase, while the 45 kDa cleaved forms of cti-AT, Mut A & D were not degraded further since all the elastase in the solution was presumably depleted due to complex formation. Upon a more detailed inspection, it was also found that the degradation of Mut B , C & E into their 45 kDa form without complex formation was time-dependent, as after 45 minutes of incubation between these serpins with elastase, all three were completely degraded by elastase (Fig. 5.3). The mechanism of degradation of the three serpins may be different, as Mut B was degraded more readily (complete degradation after 30 minutes) than Mut E and C. The reasons for the degradation of serpins in general are still largely unknown, but it can be speculated that since elastase is a non-specific enzyme able to recognize and cleave different amino acid sequences of different substrates, the excess active elastase in the solution probably managed to cleave newly exposed sites following the cleavage of the P l -  69  Mut B  MutC  MutE  F i g u r e 5.3. Degradation of M u t B , C & E by elastase. 1.8 u M of Mut B , C & E were  degraded degraded inhibitor; described  by 900 n M of elastase as time progresses. A l l of Mut B , C & E were completely by elastase after 45 minutes. (E = elastase; EI* = enzyme/inhibitor complex; I = 1° = cleaved inhibitor). Buffer compositions for these complex experiments are in Materials and Methods section 2.11 & 2.13.  70  P ' 1 in the reactive site loop in a cascade of sequential cleavages. These sequential cleavages eventually lead to a complete degradation of the serpins. Studies of the functional and conformational polymorphism of inhibitory serpins show that under certain physiological conditions, serpins can also undergo conformational change due to mutation, chemical modification or interaction with other molecular species as described in Figure  1.5  (Janciauskiene.,  2001).  Furthermore, one of the most serious and  persistent concerns about affinity chromatography purification of serpins is the potential for product denaturation and conformational change. Elution of the bound protein is often achieved by passing low pH buffer over the column (pH 2 . 5 - 4 . 0 ) . Detailed studies of protein conformation under these conditions have documented permanent conformational changes as the result of such exposure (Gagnon.,  1999).  A n increase tendency toward aggregation or  polymerization and, less commonly, proteolysis of the product are also the result of eluting the proteins at low pH (Gagnon., presented, in December of  2002  1999).  Further evidence of ai-AT polymerization was  (Devlin et al., 2 0 0 2 ) . Devlin et al. showed that the native  serpin architecture is extremely sensitive to mutation and environmental factors. The polymerization of cti-AT at pH 4.0 occurred with initial fast rate, reversible conformational change that resulted in partial loss of secondary structure, followed by subsequent slow rate that is irreversible resulting in the stabilization of the dimmers and subsequent polymer extension. Therefore, it is very likely that a large percentage of the cti-AT and variants purified from the pH 3.9 elution had distorted conformation and formed polymers, which made them inactive and susceptible to protease degradation by the excess active elastase in the solution. This hypothesis is supported by the fact that most of the elastase at low concentration (~ 9 0 0 nM) was able to form SDS-stable complexes with enough active ai-AT.  71  Thus, inactive a i - A T was not degraded or degraded very slowly. However, when high concentrations of elastase (~9 uM) were added to oti-AT, inactive a A T in the solution, r  including any small amounts of elastase/cti-AT complexes that may have been formed, were completely degraded by the excess elastase (data not shown). Further experiments including running a low pH, non-denaturing gel can be performed to investigate the polymerization of a ATatpH3.9. r  5.4  Inhibition Assay of cti-AT and Its Variants Towards Elastase 1.3 u M of a i - A T was pre-incubated with 9 n M of elastase for a range of times prior  to the commencement of enzymatic assays as described in Section 2.14. The results of this investigation are presented in Figure 5.4. They clearly show that the pre-incubation time allowed for the interaction between a,-AT and elastase is crucial in determining the residual elastase activity during the enzymatic assay. As the pre-incubation time between ai-AT and elastase increased, the residual activity of elastase decreased. Forty five minutes of preincubation time between ai-AT and elastase in this case was sufficient to fully inhibit the activity of elastase. Assays assessing the inhibition of elastase by a i - A T and variants were performed using 1.3 u M of each serpin and pre-incubating it with 9 n M of elastase for 45 minutes in the buffer described in the materials and methods, section 2.11. The inhibition assay was performed in duplicate and was allowed to proceed for 3 hours at 31°C, with continuous readings of elastase activity taken every 2 minutes.  Results showed that the control of  elastase alone after 45 minutes of pre-incubation was still active, with residual activity of 100% relative cleavage efficiency against which the activity of the other ai-AT variants was  72  120  100 o  ^ >  80 -  43  O (0  75  ti/2  60  3 min  =  3  "55 9>  40  i  20  s 0 -I 0  ,  ,  ,  ,  ,  ,  ,  ,  *  1  1  5  10  15  20  25  30  35  40  45  50  55  1 60  Time (min)  Figure 5.4. Residual activity of elastase with ai-AT after different incubation time. The length of incubation time that was allowed between 9 n M of elastase and 1.3 u M of a i - A T was directly affecting the residual activity of elastase in vitro. 45 minutes of incubation time between elastase and oci-AT was sufficient to completely inhibit elastase activity in vitro (t\n = 3 min). Enzymatic assay was performed for 3 hours and in duplicate. Enzymatic assays conditions and buffer requirements are described in the Materials and Methods sections 2.11 & 2.14.  73  180 160 H  cti-AT Mut A  Mut B  MutC  Mut D  Mut E  Control  Serpins  Figure 5.5. Inhibition of elastase by ai-AT, Mut A & D. Pre-incubation time between elastase and serpins were 45 minutes and enzymatic assay was done for 3 hours. The 9 n M elastase control is active and the addition of 1.3 u M of a i - A T , Mut A & D completely inhibit elastase activity, whereas 1.3 u M of Mut B , C and E did not inhibit elastase activity. Experiments were done in duplicate. Buffer compositions and enzymatic assays conditions for this experiment are described in the Materials and Methods sections 2.11 & 2.14.  74  compared. cti-AT, Mut A & D completely inhibited elastase activity as expected. However, Mut B,C & E were not able to inhibit the activity of elastase (Fig 5.5). These results agree with the results from complex formation studies and allow for the conclusion that only serpins that can form SDS-heat stable complexes with elastase can disrupt the catalytic active site of elastase, rendering it inactive.  5.5  Titration and Progress Curve Analysis of cti-AT and Its Variants Towards Elastase Titration experiments for ai-AT, Mut A & D against elastase were carried out in the  same way as the inhibition assay, except for the fact that varying amounts of serpin were used to pre-incubate with elastase. The purpose of this experiment was to determine the Stochiometry of Inhibition (SI) and Specific Inhibition Constant (K;) of each of the serpins that inhibit elastase. The amount of each serpin used was (nM): 0, 1.2, 2.5, 4.9, 6.2, 9.3, 12.3, 18.5, 30.9, 61.7, 123.4. A n overlay of a p A T , Mut A and D titration inhibition curves showed that the inhibition constant (Ki) of the three serpins vs elastase were identical (Fig 5.6 A ) . The (Ki) values of a i - A T , Mut A and Mut D against elastase were all 53 p M (Fig 5.6 B,C,D). These (K ) values are all in the p M range, meaning that ai-AT, Mut A & D are all excellent ;  inhibitors of elastase. The Stochiometry of Inhibition (Si) values for oti-AT, Mut A & D against elastase were 2. This means that approximately 2 mol of serpin are required to inactivate 1 mol of elastase, indicating that equal amounts of these serpins went through the complex formation inhibitory and substrate pathways (Figure 5.6 B,C,D). The difference between the progress curves analysis and the titration curve analysis of a i - A T , Mut A & D vs elastase was the lack of pre-incubation time in the progress curve  75  A  B  C  D •  100  80 •  K, = 53 p M  > CO  SI = 2  60 •  ra  T3  'to' QL se &*  40  -  20  i  0 -  0  20  40  60  SO  100  120  20  140  Mut A concentration (nM)  40  60  80  100  120  140  Mut D concentration (nM)  Figure 5.6. Inhibition constant of cti-AT, Mut A and D vs Elastase. Inhibition curves obtained from the titration experiments of ai-AT, Mut A and D vs 9 n M of elastase. The amount of each serpin used were (nM): 0, 1.2, 2.5, 4.9, 6.2, 9.3, 12.3, 18.5, 30.9, 61.7, 123.4. Enzymatic assays were done for 3 hours in triplicates with 45 minutes of preincubation between elastase and serpins. Buffers composition was described in Materials and Methods section 2.11 & 2.14. (A) Over-lay of ai-AT, Mut A and D inhibition curves shows that the (Ki) and (SI) of the three serpin vs elastase are nearly identical. (B) Inhibition curve of a A T vs elastase. Ki = 53 p M . (C) Inhibition curve of Mut A vs elastase. K; = 53 pM. (D) Inhibition curve of Mut D vs elastase. Ki = 53 pM. The (SI) for all three serpins are 2. r  76  analysis. The purpose of performing enzymatic assays without pre-incubation was to find out whether serpins can compete and inactivate elastase in the presence of competitive elastase substrates. This study is important because in vivo it is essential for an effective inhibitor to be able to out-compete the natural substrate of the protease. Figure 5.7 shows that increasing the amount of cti-AT, Mut A & D (nM) decreased the activity of 9 n M of elastase to different extents. Higher amount of serpins inhibit the activity of elastase to a greater extent. Enzymatic assays were done for 3 hours. The inhibition of elastase by a i - A T , Mut A & D obeyed slow-binding inhibition kinetics, as indicated by the biphasic plots, where maximal inhibition was achieved more rapidly with increasing concentrations of ai-AT, Mut A & D . The biphasic plots also showed that the serpins and proteases require time to interact and form complexes which deactivate the protease (illustrated by the linear part of the curve). Once complexes were formed, the process was irreversible and, as a result, all protease had been inactivated (illustrated by the asymptotic component of each curve) (Fig. 5.7).  77  Elastase 150  Increasing  K-AT]  0  1000  2000  3000  4000  5000  6000  7000  8000  9000  10000  Time (sees)  Elastase  0.3 n M  0.7 n M V « -  Increasing [Mut A]  1.4 n M 2.7 n M 13.7 n M  0  laOB  2090  3400  400B  6000  6000  7006  8000  OftflO  108i»  TOTKI I M C S I  Elastase  0.8 n M 1.2 n M  Increasing [Mut D]  1.6 n M 3.1 n M 13.7 n M 1000  2000  3000  4000  5000  6000  7000  8000  9000  10000  Time (sees)  Figure 5.7. Progress curve of a i - A T , Mut A & D vs elastase. Increasing amount of a i A T , Mut A & D decreased 9 n M of elastase activity. The curves also show tight-binding inhibition pattern of serpin/enzyme inhibition. Enzymatic assays were done for 3 hours with no incubation time between elastase and serpins. Experiments were done in duplicates and buffer compositions are described in the Materials and Methods section 2.11. «fe 2.14.  78  Chapter 6: Characterization a A T and Its Variants (Mut A-E) r  with Complex Formation Studies and Inhibition Assays Against HCV NS3 Protease  6.1  Serpins as Candidates for Novel Protein Therapeutics Against H C V NS3 Protease As described in the introduction, there has been evidence of H C V NS3 protease  interaction with host cell serpins in vitro. C l inhibitor and a-2 antiplasmin, both of which belong to the serpin family, can interact with H C V NS3 protease, resulting in proteolysis of the serpins and production of higher molecular weight SDS-stable complexes (Drouet et al., 1999). These observations suggest that the use of serpins engineered to recognize and inhibit the NS3 protease may be a promising new strategy for the development of H C V therapeutics.  6.2  Complex Formation Between ai-AT and Its Variants Against NS3 Protease After characterizing the inhibitory properties of the a i - A T and variant serpins toward  elastase (Chapter 5), the ability of these serpins to inhibit and/or form complexes with NS3 protease was assessed next. The optimization of conditions necessary to conduct complex formation experiments between cti-AT and variants against NS3 protease was necessary before performing in vitro complex experiments. It was determined that 50mM Hepes, 150mM NaCl, 0.1% triton-X, lOmM DTT, p H 7.5, 15uM NS4A cofactor, 20ul of NS3 protease and 1.3 u M of purified a i - A T or its variants in a total volume of lOOul reaction mixture were the optimal conditions under which to detect the formation of complexes  79  between a i - A T and its variants with the NS3 protease. It was also established that 15 minutes of pre-incubation between the NS4A cofactor and the NS3 protease at 31°C was necessary to activate the NS3 protease prior to performing complex experiment with serpins for a period of 4 hours at 31°C. cti-AT and all its variants were mixed with the NS3 protease as described above, and the complex formation reactions were subjected to SDS-PAGE and Western blot analyses. Results show that the NS3 protease which has a molecular weight of -23 kDa can interact with a i - A T and all its variants (-49 kDa) except for Mut E to form SDS-heat stable complexes of molecular weight -72 kDa (Fig 6.1 A & B). Interestingly, the cti-AT that was able to react with the NS3 protease, all formed SDS-stable complexes with NS3 protease and did not go through the substrate pathway as described earlier in Section 5.2. It was expected that a i - A T would not be cleaved as a substrate by the NS3 protease due to its strict cleavage site specificity requiring a cysteine in P I . However, the fact that a i - A T can still form SDS-stable complexes with the NS3 protease without its reactive site loop (RSL) being cleaved first is rather unique. It is widely accepted that in order to form a covalent SDS-stable complex between protease and serpins, the cleavage of the R S L is necessary for the migration of the proteinase to the opposite pole of the serpin (Wilczynska et al., 1995; Janciauskiene., 2001). This observation leads one to believe that the general scaffold, especially the reactive site loop region of ai-AT, had a suitable structure to allow its recognition by the NS3 protease and subsequent complex formation. This supports our initial reasoning behind using a i - A T as a starting scaffold (section 1.3.4) upon which to design specific NS3 inhibitors. The phenomenon of complex formation between a protease and serpin without R S L cleavage is a new concept and the molecular mechanisms underlying this complex formation are yet to be elucidated.  80  Figure 6.1. Complex formation between cti-AT and its variants vs ns3 protease. (A & B) A l l of the 1.3 u M of serpins can form SDS-heat stable complexes with 20ul of NS3 protease except for mutant E . Complex experiments were done for 4 hours with 15 minutes of pre-incubation between NS3 protease & NS4A cofactor. (EI* = enzyme/inhibitor complex ; I = inhibitor ; 1° = cleaved inhibitor). Buffer compositions for these experiments are described in Materials & Methods sections 2.6 & 2.13.  81  Mut A , B , C & D were all proteolytically cleaved by the NS3 protease because some portion of the serpins went through the substrate pathway. These observations agree with our prediction that the introduction of cysteine in the R S L PI position of a i - A T would make it a pseudo-substrate for the NS3 protease. We also found that although the four serpins can form SDS-stable complexes with the NS3 protease, Mut D formed most complexes (Fig 6.1 A & B). We still do not fully understand why Mut D, which has PI and P4 mutations, was able to form the most complexes with NS3 protease. There are three possible hypotheses to explain these results: 1) The pooled stock of Mut D simply contained more biologically active serpin that can form SDS-stable complexes with the NS3 protease as compared to the other stock of serpins. 2) The P I , P4 mutation in Mut D were better recognized by the NS3 protease than the PI mutation in Mut A , thus leading to more complex formations. However, the addition of P6 mutation in Mut B & C improved substrate recognition to such a great extent that most of the Mut B & C that interacted with the NS3 protease went through the substrate pathway instead of the complex inhibition pathway. Therefore, this observation leads to the question of how efficiently a serpin should be recognized by the NS3 to achieve the most complex formation. 3) Mut D can form more complexes with NS3 protease than Mut A due to the same reason as above. However, the addition of the P6 mutation in Mut B & C could have disrupted the efficient conformational change as described in Section 5.3. The P6 substitution in the R S L of cti-AT should prevent Mut B & C from forming complexes with the NS3 protease (Dufour et al., 2001). Therefore, these results lead one to question whether the complex formation between all the a i - A T serpin mutants and the NS3 protease follow the traditional pathway requiring R S L cleavage and insertion or follow the possible alternative as seen in cti-AT/NS3 protease complex pathway without R S L cleavage. Small amounts of  82  complex formation can also be seen between the serpins and the NS3 protease without the addition of NS4A cofactor, due to the intrinsic activity of the NS3 protease alone. Mut E was not proteolytically cleaved by the NS3 protease. This is because the proline substitution has been shown to be uncleavable when introduced in the P ' l site of the NS3 protease peptide inhibitor (Paolo et al., 2000). Although a proline in P ' l is a potent peptide inhibitor against the NS3 protease in vitro, it is not effective against the FLNS3 (Richer et al., unpublished result). Moreover, when the substitution was made in the RSL of a i - A T , it failed to form a complex with the NS3 protease (Fig 6.1 B), probably due to the major distortion of the a i - A T RSL from the introduction of proline residue. Finally, none of the serpins were completely degraded by excess NS3 protease because, unlike elastase which can cleave a broad range of substrates, the NS3 protease specifically cleaves only after cysteine residue.  6.3  Studies Investigating the Effect of Enzyme: Inhibitor Ratio and Time of Interaction Upon Complex Formation Between ai-AT and Variants and NS3 Protease Figure 6.2 shows the effect of varying the amount of NS3 protease on the amount of  complex formed in a reaction using 1.3 u M of either ai-AT, Mut A , D & C respectively. The increased amount of SDS-stable complexes formed with the increased amount of the NS3 protease suggests that the complex formation was ratio-dependent. 20 ul of NS3 protease usually resulted with the highest amount of complex formation. Figure 6.3 shows that increasing the incubation time between a i - A T , Mut A , D & C and the NS3 protease also increased the amount of complex formation. Additional incubation times of 6 hours and 18  83  MutC  F i g u r e 6.2. C o m p l e x f o r m a t i o n between a i - A T , M u t A , D, & C w i t h increasing  [NS3J.  SDS-stable complex formation between a A T , Mut A , D, and C with NS3 protease increases as [NS3] increases. Complex experiments were done for 4 hours. All complex formation reactions were done with 15 minutes of pre-incubation between NS3 protease & NS4A cofactor. (EI* = enzyme/inhibitor complex; I = inhibitor). Buffer compositions for these experiments are described in Materials & Methods sections 2.6 & 2.13. r  84  Figure 6.3. Complex formation between ai-AT. Mut A , D, & C with NS3 protease at different time points. Complex formation between cti-AT. Mut A , D , & C with NS3 protease increases as the amount of time allowed for the complex reactions increase. A l l complex formation reactions were done with 15 minutes of pre-incubation between NS3 protease & NS4A cofactor. (EI* = enzyme/inhibitor complex ; I = inhibitor; F = cleaved inhibitor). Buffer compositions for these experiments are described in Materials & Methods sections 2.6 & 2.13. 85  hours were also tried, but did not significantly increase the amount of complexes formed (data not shown). Therefore, 4 hours incubation time between the serpins and the NS3 protease was optimal. This result supports the idea that most of the serpins in the pooled samples were inactive since after 4 hours of incubation, most active serpins had already formed complexes with the NS3 protease and there were still plenty of unreacted serpins and the NS3 protease left in the solution. None of the remaining serpins could contribute to the formation of more complexes after 4 hours. However, there was still active NS3 protease left because as described in the next section, when similar concentrations of serpins and NS3 protease were tested in vitro for inhibition assays, NS3 protease activity was not inhibited.  6.4  Inhibition Assay of cti-AT and Its Variants Towards NS3 Protease Although five of the serpins (ai-AT, Mut A , D , B & C), could form SDS-stable  complexes with the NS3 protease, surprisingly they were not able to inhibit the NS3 protease activity in vitro completely (Fig. 6.4). In the control reaction in which the NS3 protease and the NS4A cofactor were pre-incubated together for 4 hours, the NS3 protease was still active. ai-AT, Mut A , B , C slightly decrease the NS3 protease activity by approximately 20%. Mut D , however, decreased the NS3 protease activity by -50%, while Mut E did not decrease the NS3 protease activity. These results are in agreement with those from the complex experiment in which Mut D formed the most complexes with the NS3 protease compared to the other mutants, while Mut E was not able to form any detectable complexes with the NS3 protease. The reason for the incomplete inhibition of the NS3 protease could be due to the low amount of active serpins present in the pooled samples as discussed before in Chapter 5. Therefore, only a small proportion of active serpin was capable of complex formation with  86  140  cn-AT  Mut A  Mut B  MutC  Mut D  Mut E  NS3 protease  Serpins  Figure 6.4. Inhibition of NS3 protease by cti-AT and its 5 variants. The 10 pi of NS3  protease control was active (-450 R F U = 100% relative cleavage efficiency). 1.3 u M of ctr A T and all its variants cannot completely inhibit NS3 protease activity, although Mut D reduced NS3 protease activity level by a factor 1/2. Inhibition assay were done for 2 hours with 15 minutes of pre-incubation between NS3 protease & NS4A cofactor. Enzymatic assays conditions and buffer compositions for these experiments are described in Materials & Methods sections 2.6 & 2.14.  87  the NS3 protease, and subsequently able to inhibit NS3 protease activity. The inhibition assays using low concentrations of NS3 protease were challenging to perform, because the sensitivity of detection of fluorescence from the IQFS was low. Hence, lower enzyme activity resulting in lower rates of IQFS cleavage was difficult to detect. One of the possible solutions to this problem could be producing higher amounts of biologically active serpins via a different method of purification and re-evaluating the ability to inhibit NS3 protease activity. In addition, Mut A , B & C also tend to go through the substrate pathway more than through the complex pathway (Fig 6.2 & 6.3). A n alternative way to deal with the low concentration of active serpins is to genomically express the recombinant a i - A T - K D E L construct in human cell cultures, so that the serpins are over-expressed in cells without being subjected to processing procedures such as purification and dialysis, which might inactivate some of the serpin.  6.5  Complex Formation Between ai-AT and Its Variants Against F L N S 3 Preliminary studies were performed to investigate if the a i - A T and its variant serpins  can form complexes with recombinant FLNS. The reaction conditions and buffer requirements for performing these FLNS3 complex reactions were identical to those with the NS3 protease (Section 6.2), except 20ul of FLNS3 was used instead of 20ul of NS3 protease. Western blot results from these experiments showed that none of the serpins can form detectable SDS-stable complexes with FLNS3 (Fig. 6.5 A & B). However, Mut A , B , C & D were still recognized by the FLNS3 as a substrate and cleaved. These results suggest that perhaps the overall scaffold of the FLNS3 containing the helicase domain does not permit  88  B •4  Mut  B  •  <  Mut C  •  4  Mut E  •  Figure 6.5. Complex formation between FLNS3 with cci-AT and all its variants. (A) 1.3 u M of a i - A T , Mut A and D are cleaved by 20 ul of FLNS3, however no complex is seen at -116 k D A . FLNS3 were also processed with the addition of NS4A cofactor. (B) 1.3 u M of Mut B and C were cleaved by 20 ul of FLNS3 but Mut E is not cleaved due to the proline in P' 1 position. Complex experiments were done for 4 hours. The lack of complex formation could be attributed to many different factors as discussed in the discussion sections. FLNS3 was detected with anti FLNS3 anti-body. Serpins are detected with anti-histidine antibody. Complex experiments were done for 4 hours with 15 minutes of pre-incubation between FLNS3 & NS4A cofactor. (FLNS3 = Full length NS3; FLNS3* = FLNS3 cleaved ; I = inhibitor ; 1° = cleaved inhibitor). Buffer compositions for these experiments are described in Materials & Methods sections 2.6 & 2.13.  89  Figure 6.6. Co-localization of (Xi-AT-KDEL, Mut A - K D E L & Mut D-KDEL with NS3 in human cells. (A-B) Subcellular localization of transiently expressed a i A T - K D E L motif fusion, Mut A - K D E L , and Mut D - K D E L in UNS3-4A cells, which also express NS3-4A protease/cofactor. Double immuno staining was performed to detect oci-AT-KDEL and variants using a FLAG-epitope tag-specific antibody together with NS3 using NS3 polyclonal antibody. (C) Overlay of the double immuno staining shows that both the serpins-KDEL and H C V NS3 were expressed in the same compartment in the cell, which is presumably the ER. (Work done by Dr. Pamela Hamill)  90  complex formations to occur. The other reason for the absence of complex formation could be that the amount of active F L N S 3 was too low to detect any complex formations. A similar observation has been made with the N S 3 protease (data not shown). Note that F L N S 3 was also degraded with time in the presence of the N S 4 A cofactor after 4 hours of incubation without the serpins. The internal cleavage of F L N S 3 only in the presence of the N S 4 A cofactor has been previously reported (Yang et al., 2 0 0 0 ) .  6.6  Sub-cellular Targeting of Recombinant Serpin-based Inhibitors of H C V N S 3 Protease in Human UNS 3 / 4 A Cells Dr. Pamela Hamill performed transfection of mammalian expression plasmids  encoding cti-AT-KDEL, Mut A - K D E L and Mut D - K D E L in human cells (UNS 3 - 4 A ) which inducibly expressed the H C V N S 3 / 4 A protease complex with the control of tetracycline. Figure 6.6 A & B shows that novel variants of the serpin a i - A T fused to a C-terminal K D E L E R retention motif (cti-AT-KDEL) were transiently expressed in UNS 3 - 4 A which had been induced to expressed the viral N S 3 - 4 A protease in tetracycline-regulated manner. The results also showed that H C V N S 3 - 4 A , cti-AT-KDEL and variants were apparently all co-localized in the E R compartment of the cells (to be confirmed by using E R KDEL-receptor marker). The overlay of cti-AT-KDEL and variants with the N S 3 - 4 A protease expressed in the cells suggests that the co-localization of the serpin/protease in same subcellular-compartment might improve the probability of inhibition of the H C V protease by bio-engineered serpins (Fig. 6.6 C).  91  6 . 7  Summary Chapters 4 , 5 and 6 present results of successful bio-engineering of cti-AT variants  and cti-AT-KDEL variants. The purification of these serpins was largely successful, with some of the serpins retaining their biological activity as shown by the ability to form complexes and be cleaved by elastase and the N S 3 protease. Complex formation studies between cti-AT and variants with elastase and the N S 3 protease showed that only a i - A T , Mut A & D can form SDS-stable complexes with elastase, whereas all the serpins except for Mut E can form a complexes with the N S 3 protease. Inhibition studies showed that a i - A T , Mut A & D can completely inhibit elastase activity in vitro. However, none of the serpins can completely inhibit N S 3 protease activity in vitro. Despite the lack of marked inhibition of the N S 3 protease by the novel cti-AT variants and until the problems of inactive serpins proposed in the discussions are addressed by performing purification of cti-AT and variants using an increasing imidazole gradient, the possibility that cci-AT and its variants might be able to inhibit H C V N S 3 protease activity in vitro and in vivo is still viable and commands further investigation.  92  Chapter 7: Conclusion  7.1  Current View of H C V NS3 Serine Protease Inhibitors In light of the success of protease inhibitors in controlling HTV infections, many  research groups in both industry and academia have chosen to embark on programs aimed at identifying potent and selective inhibitors of the NS3 protease as candidate anti-HCV drugs. In the decade that has elapsed since the discovery of H C V , marked progress has been made towards understanding the novel molecular structure and functions of the H C V NS3 protease. Protease inhibitor discovery began with natural product screenings and substratederived analogue-based drug-design, soon progressing to the incorporation of mechanismbased drug design strategies such as using serpin. Today, the design of protease inhibitors involves a powerful combination of all of these traditional drug discovery approaches and it is supplemented by computer-assisted structure-based inhibitor design using 3-dimensional structures of proteases determined by X-ray crystallography, and rapid robotic assay methods to quickly screen various potential inhibitors. These techniques are currently being refined and expanded to accelerate the discovery process for new protease inhibitors. Coupled with advances in molecular and cellular biology, protein chemistry, microbiology, structural biology and molecular pharmacology, there can be little doubt that the next several decades will see the development of a new class of protease inhibitors. To date, there have been reports of peptide-based inhibitors of the NS3 protease and many more research groups are probably working on improving the potency of these inhibitors (Landro et al., 1997, Paolo et al., 2000). However, the remarkably shallow and solvent-exposed substrate binding region which has resulted in weak interactions that are  93  distributed along an extended contact surface between the proteases and the inhibitors, makes the design of potent small molecule inhibitors a quite challenging task.  7.2  ai-Antitrypsin and Its Variants as Potential Inhibitors Against H C V N S 3 Protease and Future Directions Since the first report of interactions of H C V NS3 protease by C l I N H and oc2-  antiplasmin, that can form SDS-stable complex with NS3 protease (Drouet et al., 1999), there have been no other reports on serpins that can interact with and inhibit NS3 protease. Data in this thesis showed the ability of cci-AT and its variants Mut A , B , C & D to interact and form SDS-stable complexes with the NS3 protease. These results provide further evidence of the possibility of using serpin as an alternative way of inhibiting the H C V NS3 serine protease. Some of the advantages of using serpin and a i - A T or its variants as NS3 protease inhibitors include: 1) the availability of extended contact surface needed for proper interaction with the NS3 protease that was absent when using small peptide based inhibitors. 2) a i - A T is a naturally occurring human serpin that is produced in the liver and there have been no reports of diseases linked to the over-expression of a i - A T . Therefore, a i - A T and its variants should be better tolerated by the immune system and should have limited side effects when applied as therapeutics in humans. 3) The introduction of oti-AT variants, Mut B (P1,P4 & P6) or Mut C (PI & P6) lost their specificity towards elastase, but were still able to form SDS-stable complex with NS3 protease. This phenomenon reinforces the possibility of using Mut B and Mut C as good candidates for the design of specific inhibitors of the NS3 protease. 4) a i - A T and its variants can be used to target the E R of the cells where H C V replication is believed to occur, thereby improving the chance of interaction between the serpins and the NS3 protease  94  (Wu., 2001). 5) Serpin is a suicidal inhibitor because it forms an irreversible interaction with the target. It is therefore effective in removing the protease that is bound to it permanently. 6) SP6, a Drosphila serpin that has been mutated to contain cysteine in P I , serine in P ' l and glutamic acid in P6 to target the NS3 protease and the FLNS3, has been shown not only to form SDS-stable complexes with the NS3 protease and the FLNS3, but it also inhibits the protease activity in vitro (Richer et al, in preparation). Despite the listed advantages of using serpin as an inhibitor of the NS3 protease, there are various stages of research that still need to be performed further to support the hypothesis that serpin can be used as one of the potential therapeutic against the H C V NS3. One of the experiments that is currently being performed in Dr. Franqois Jean's laboratory explores cell culture systems of transfecting various serpins into cultured human osteosarcoma cells (UNS 3-4A) which express the H C V NS3 protease and the NS4A cofactor. These intracellularly expressed recombinant serpins should reduce the problems of dealing with possible inactivation of serpins caused by expression and purification from bacterial cells. Future experiments should also include experiments in Huh-7 cell lines that support the replication of H C V replicon or even using animal models such as chimeric mice both of which will be described later.  7.3  Animal Models and Cellular System for Anti-HCV Drugs Development Although both peptide-based and protein-based inhibitors of the H C V NS3 protease  have been identified, there has been no cell-culture assay available for the propagation of H C V in the laboratory which limits the ability to test whether those protease inhibitors can actually inhibit H C V replication in vivo. For unknown reasons, the propagation of H C V in  95  cell cultures including human hepatocytes has been very challenging, because available system all suffer from low reproducibility and efficiency (Bartenschlager., 2001; Kato., 2000). However, the recent development of the replicon system opens up an encouraging possibility for drug discovery (Bartenschager., 2002). Huh-7 cells were originally isolated from the hepatoma tissue of a man suffering from a well differentiated hepatocellular carcinoma (Nakabayashi et al., 1982). Huh-7 is currently the only efficient cell that allows the proper propagation of the H C V replicon. This replicon system will probably be important for the development of HCV-specific drugs. The advantage of using the H C V replicon system for drug development over screening assays carried out with purified viral protease is that not all compounds that were successful when tested in vitro were effective in blocking viral replication in vivo (Young., 2001). Moreover, H C V proteins form a higher-order multiprotein replicase complex, binding sites for a compound might be accessible in vitro when working with an isolated enzyme but not within an infected cells, where this protein is incorporated into a macromolecular complex (replicase). Therefore, drugs identified by in vitro screening will not necessarily be active at blocking R N A replication in the cell (Bartenschlager., 2002). Although H C V replicons can be propagated in Huh-7 cells, the production of full viral particles is still not possible. In addition, until recently only two H C V isolates of genotype l b could be propagated in the cell culture (Lohmann et al., 1999; Guo et al., 2001; Kishine et al., 2002). The understanding of viral pathogenesis and the development of potential drug candidates have been hindered by slow progress in the development of an animal model that can support replicating viruses (Schinazi et al., 1999). Chimpanzees,  96  which are both expensive and not readily available, have been the only non-human animal model for reproducible propagation of H C V over the past decade, until the recent development of the chimeric mouse model that was suitable for studying the human hepatitis C virus in vivo (Mercer et al., 2001). B y transplanting normal human hepatocytes into SCJD mice carrying the plasminogen activator transgene (Alb-uPA), mice with chimeric human liver were generated. These mice allow the replication of H C V viral proteins and the release of infectious viral particles through three generations of mice. These mice represent the first murine model suitable for studying H C V in vivo. Work is currently in progress to test adenovirus recombinant cti-AT and SP6 variants in these murine model (in collaboration with Dr. Norman Kneteman, University of Alberta).  7.4  Combination Therapy for HCV Infection HTV protease inhibitors have proven to be valuable therapeutics in combination with  nucleoside reverse transcriptase inhibitors (NRTIs) and non-nucleoside reverse transcriptase inhibitors (NNRTIs) (a drug combination known as "highly active antiretroviral therapy" (HAART)) in the treatment of HTV infections (De Clercq., 2002). Likewise, a successful therapy for H C V will probably have to involve a combination of drugs targeting different stages of viral replication cycles to reduce the chance of developing viral resistance. Possible targets for anti-HCV drug development in the future may include the inhibition of viral entry, the inhibition of IRES, the inhibition of capping, the inhibition NS3/4A serine protease, the inhibition of ribonucleic acid (RNA) helicase and the inhibition of RNA-dependent R N A polymerase. In addition, antisense oligonucleotides or ribozymes may also become part of the cocktail inhibitors to minimize H C V replication. Finally, immunotherapies to enhance  97  H C V - s p e c i f i c i m m u n e r e s p o n s e s are a l s o attractive strategies to c o n t r o l H C V i n f e c t i o n s a n d t o p r e v e n t c h r o n i c l i v e r d i s e a s e ( C o r n b e r g et a l . ,  2001;  L e y s s e n et a l . ,  2000;  D y m o c k et a l . ,  2000). In  c o n c l u s i o n , the  intense  search for  inhibitors  against  various  targets  of  H C V  r e p l i c a t i o n w i t h the h e l p o f r a p i d l y a d v a n c i n g t e c h n o l o g i e s a n d the d e v e l o p m e n t o f n e w c e l l l i n e s a n d a n i m a l m o d e l s t o test t h e s e c o m p o u n d s w i l l p r o b a b l y r e s u l t i n t h e d i s c o v e r y potent c o c k t a i l inhibitors o f H C V replication i n the near future.  98  of  Chapter 8: References Al, R. H., Y. Xie, Y. Wang, and C. H. Hagedorn. 1998. E x p r e s s i o n of recombinant hepatitis C virus non-structural protein 5 B in E s c h e r i c h i a coli. V i r u s R e s 53:141-9. Alter, H. J . , and L. B. Seeff. 2 0 0 0 . R e c o v e r y , persistence, a n d s e q u e l a e in hepatitis C virus infection: a perspective o n long-term o u t c o m e . S e m i n Liver Dis 20:17-35. Attwood, M. R., J . M. Bennett, A. D. Campbell, G. G. Canning, M. G. Carr, E. Conway, R. M. Dunsdon, J . R. Greening, P. S. Jones, P. B. Kay, B. K. Handa, D. N. Hurst, N. S. Jennings, S. Jordan, E. Keech, M. A. O'Brien, H. A. Overton, J . King-Underwood, T. M. Raynham, K. P. Stenson, C. S. Wilkinson, T. C. Wilkinson, and F. X. Wilson. 1999. T h e d e s i g n a n d synthesis of potent inhibitors of hepatitis C virus N S 3 - 4 A proteinase. Antivir C h e m C h e m o t h e r 10:259-73. Banda, M. J . , A. G. Rice, G. L. Griffin, and R. M. Senior. 1988. A l p h a 1-proteinase inhibitor is a neutrophil chemoattractant after proteolytic inactivation by m a c r o p h a g e e l a s t a s e . J Biol C h e m 263:4481-4. Barbato, G., D. O. Cicero, F. Cordier, F. Narjes, B. Gerlach, S. Sambucini, S. Grzesiek, V. G. Matassa, R. De Francesco, and R. Bazzo. 2 0 0 0 . Inhibitor binding i n d u c e s active site stabilization of the H C V N S 3 protein s e r i n e p r o t e a s e d o m a i n . E m b o J 19:1195-206. Bartenschlager, R., L. Ahlborn-Laake, J . Mous, and H. Jacobsen. 1 9 9 3 . Nonstructural protein 3 of the hepatitis C virus e n c o d e s a serine-type proteinase required for c l e a v a g e at the N S 3 / 4 a n d N S 4 / 5 junctions. J Virol 67:3835-44. Bartenschlager, R., L. Ahlborn-Laake, J . Mous, and H. Jacobsen. 1994. Kinetic a n d structural a n a l y s e s of hepatitis C virus polyprotein p r o c e s s i n g . J Virol 68:504555. Bartenschlager, R. 1997. C a n d i d a t e targets for hepatitis C virus-specific antiviral therapy. Intervirology 40:378-93. Bartenschlager, R., and V. Lohmann. 2 0 0 0 . Replication of hepatitis C virus. J G e n Virol 81:1631-48. Bartenschlager, R., and V. Lohmann. 2 0 0 1 . N o v e l cell culture s y s t e m s for the hepatitis C virus. Antiviral R e s 52:1-17.  99  Bartenschlager, R. 2 0 0 2 . Hepatitis C virus replicons: potential development. Nat R e v Drug D i s c o v 1:911-6.  role for drug  Bianchi, E., S. Orru, F. Dal Piaz, R. Ingenito, A. Casbarra, G. Biasiol, U. Koch, P. Pucci, and A. Pessi. 1 9 9 9 . Conformational c h a n g e s in h u m a n hepatitis C virus N S 3 protease upon binding of p r o d u c t - b a s e d inhibitors. Biochemistry 38:13844-52. Bianchi, E., and A. Pessi. 2 0 0 2 . Inhibiting viral opportunities. B i o p o l y m e r s 66:101-14.  proteases: challenges and  Boskovic, G., and S. S. Twining. 1998. L o c a l control of a l p h a l -proteinase inhibitor levels: regulation of a l p h a l - p r o t e i n a s e inhibitor in the h u m a n c o r n e a by growth factors a n d cytokines. B i o c h i m B i o p h y s A c t a 1403:37-46. Butkiewicz, N., N. Yao, W. Zhong, J . Wright-Minogue, P . Ingravallo, R. Zhang, J . Durkin, D. N. Standring, B. M. Baroudy, D. V. Sangar, S. M. Lemon, J . Y. Lau, and Z. Hong. 2 0 0 0 . V i r u s - s p e c i f i c cofactor requirement a n d chimeric hepatitis C v i r u s / G B virus B nonstructural protein 3. J Virol 74:4291-301. Caronia, S., M. F. Bassendine, R. Barry, P. Mills, N. V. Naoumov, R. Fox, J . Lowes, D. Hollanders, L. Murray-Lyon, W. L. Irving, R. D. Goldin, and G. R. Foster. 2 0 0 1 . Interferon plus a m a n t a d i n e v e r s u s interferon a l o n e in the treatment of naive patients with chronic hepatitis C : a U K multicentre study. J Hepatol 35:512-6. Carrell, R. W., J . O. Jeppsson, C. B. Laurell, S. O. Brennan, M. C. Owen, L. Vaughan, and D. R. Boswell. 1 9 8 2 . Structure a n d variation of h u m a n a l p h a 1antitrypsin. Nature 298:329-34. Carrell, R. W. 1986. R e a c t i v e - c e n t r e variants of a l p h a 1-antitrypsin. A n e w range of anti-inflammatory agents. Biotechnol G e n e t E n g R e v 4:291-309. Chambers, T. J . , R. C. Fletterick, and C. M. nonstructural protein N S 3 site-specific c l e a v a g e s in 902.  Weir, A. Grakoui, D. W. McCourt, J . F. Bazan, R. J . Rice. 1 9 9 0 . E v i d e n c e that the N-terminal d o m a i n of from yellow fever virus is a s e r i n e protease r e s p o n s i b l e for the viral polyprotein. P r o c Natl A c a d S c i U S A 87:8898-  Cornberg, M., H. Wedemeyer, and M. P. Manns. 2 0 0 1 . Hepatitis C : therapeutic perspectives. F o r u m ( G e n o v a ) 11:154-62. Courtney, M., S. Jallat, L. H. Tessier, A. Benavente, R. G. Crystal, and J . P. Lecocq. 1 9 8 5 . S y n t h e s i s in E. coli of a l p h a 1-antitrypsin variants of therapeutic potential for e m p h y s e m a a n d thrombosis. Nature 313:149-51.  100  Craxi, A., C. Camma, and M. Giunta. 1 9 9 9 . Definition of r e s p o n s e to antiviral therapy in chronic hepatitis C . J Hepatol 31 Suppl 1:160-7. De Clercq, E. 2 0 0 2 . Strategies in the d e s i g n of antiviral drugs. Nat R e v Drug D i s c o v 1:13-25. De Francesco, R. and C. Steinkuhler. 1 9 9 9 . Structure a n d function of hepatitis C virus N S 3 - N S 4 A S e r i n e proteinase. J . Viral H e p a t 6: 149-167. Deleersnyder, V., A. Pillez, C. Wychowski, K. Blight, J . Xu, Y. S. Hahn, C. M. Rice, and J . Dubuisson. 1 9 9 7 . Formation of native hepatitis C virus glycoprotein c o m p l e x e s . J Virol 71:697-704. Devlin, G. L , M. K. Chow, G. J . Howlett, and S. P. Bottomley. 2 0 0 2 . A c i d Denaturation of alpha(1)-Antitrypsin: Characterization of a N o v e l M e c h a n i s m of S e r p i n Polymerization. J Mol Biol 324:859-70. Di Bisceglie, A. M. 1995. Hepatitis C a n d hepatocellular c a r c i n o m a . S e m i n Liver D i s 15:64-9. Dimasi, N., F. Martin, C. Volpari, M. Brunetti, G. Biasiol, S. Altamura, R. Cortese, R. De Francesco, C. Steinkuhler, and M. Sollazzo. 1997. Characterization of e n g i n e e r e d hepatitis C virus N S 3 protease inhibitors affinity s e l e c t e d from h u m a n pancreatic secretory trypsin inhibitor a n d minibody repertoires. J Virol 71:7461-9. Drouet, C , L. Bouillet, F. Csopaki, and M. G. Colomb. 1 9 9 9 . Hepatitis C virus N S 3 serine protease interacts with the serpin C 1 inhibitor. F E B S Lett 458:415-8. Dufour, E. K., J . B. Denault, L. Bissonnette, P. C. Hopkins, P. Lavigne, and R. Leduc. 2 0 0 1 . T h e contribution of arginine r e s i d u e s within the P 6 - P 1 region of a l p h a 1-antitrypsin to its reaction with furin. J Biol C h e m 276:38971-9. Dymock, B. W., P. S. Jones, and F. X. Wilson. 2 0 0 0 . N o v e l a p p r o a c h e s to the treatment of hepatitis C virus infection. Antivir C h e m C h e m o t h e r 11:79-96. Ede, N. J . , S. N. Eagle, G. Wickham, A. M. Bray, B. Warne, K. Shoemaker, and S. Rosenberg. 2 0 0 0 . S o l i d p h a s e s y n t h e s i s of peptide a l d e h y d e p r o t e a s e inhibitors. Probing the proteolytic sites of hepatitis C virus polyprotein. J P e p t S c i 6:11-8. Failla, C , L. Tomei, and R. De Francesco. 1994. Both N S 3 a n d N S 4 A a r e required for proteolytic p r o c e s s i n g of hepatitis C virus nonstructural proteins. J Virol 68:375360.  101  Fanning, L , E . Kenny-Walsh, J . Levis, K. R. Choudhury, B. Cannon, M. Sheehan, M. Whelton, and F. Shanahan. 2 0 0 0 . Natural fluctuations of hepatitis C viral load in a h o m o g e n e o u s patient population: a prospective study. Hepatology 31:225-9. Fattori, D., A. Urbani, M. Brunetti, R. Ingenito, A. Pessi, K. Prendergast, F. Narjes, V. G. Matassa, R. De Francesco, and C. Steinkuhler. 2 0 0 0 . P r o b i n g the active site of the hepatitis C virus s e r i n e protease by f l u o r e s c e n c e r e s o n a n c e energy transfer. J Biol C h e m 275:15106-13. Feldmann, H., V. E . Volchkov, V. A. Volchkova, U. Stroher, and H. D. Klenk. 2 0 0 1 . Biosynthesis a n d role of filoviral glycoproteins. J G e n Virol 82:2839-48. Gagnon, P. 1999. B i o s y s t e m s P r o c e s s i n g , www.validated.com  Quarterly  Resource  Guide  for  Downstream  Gallinari, P., D. Brennan, C. Nardi, M. Brunetti, L. Tomei, C. Steinkuhler, and R. De Francesco. 1998. Multiple e n z y m a t i c activities a s s o c i a t e d with recombinant N S 3 protein of hepatitis C virus. J Virol 72:6758-69. Gallinari, P., C. Paolini, D. Brennan, C. Nardi, C. Steinkuhler, and R. De Francesco. 1 9 9 9 . Modulation of hepatitis C virus N S 3 protease a n d h e l i c a s e activities through the interaction with N S 4 A . Biochemistry 38:5620-32. Gettins, P. G., P. A. Patston and J . Griffin. 1996. S e r p i n s ; Structure, function a n d biology. ( L a n d e s , A u s t i n , T e x a s ) . Gettins, P. G. 2 0 0 2 . T h e F-helix of s e r p i n s plays a n e s s e n t i a l , active role in the proteinase inhibition m e c h a n i s m . F E B S Lett 523:2-6. Gonzalez-Peralta, R. P., G. J . Galasso, T. Poynard, S. Schalm, H. C. Thomas, and T. L. Wright. 1999. S u m m a r y of the first international s y m p o s i u m o n viral hepatitis. Antiviral R e s 42:77-96. Grakoui, A., D. W. McCourt, C. Wychowski, S. M. Feinstone, and C. M. Rice. 1993. Characterization of the hepatitis C v i r u s - e n c o d e d serine proteinase: determination of proteinase-dependent polyprotein c l e a v a g e sites. J Virol 67:283243. Grakoui, A., D. W. McCourt, C. Wychowski, S. M. Feinstone, and C. M. Rice. 1993. A s e c o n d hepatitis C v i r u s - e n c o d e d proteinase. P r o c Natl A c a d S c i U S A 90:10583-7. Grodberg, J . , and J . J . Dunn. 1988. o m p T e n c o d e s the E s c h e r i c h i a coli outer m e m b r a n e protease that c l e a v e s T 7 R N A p o l y m e r a s e during purification. J B a c t e r i d 170:1245-53.  102  Guo, J . T., V. V. Bichko, and C. Seeger. 2 0 0 1 . Effect of a l p h a interferon o n the hepatitis C virus replicon. J Virol 75:8516-23. Heathcote, J . 2 0 0 0 . Antiviral therapy for patients with chronic hepatitis C . S e m i n Liver D i s 20:185-99. Hedstrom, L., L. Szilagyi, and W. J . Rutter. 1992. Converting chymotrypsin: the role of s u r f a c e loops. S c i e n c e 255:1249-53.  trypsin  to  Huber, R., and R. W. Carrell. 1 9 8 9 . Implications of the three-dimensional structure of a l p h a 1-antitrypsin for structure a n d function of serpins. Biochemistry 28:8951-66. Huntington, J . A., R. J . Read, and R. W. Carrell. 2 0 0 0 . Structure of a s e r p i n protease c o m p l e x s h o w s inhibition by deformation. Nature 6806:923-6 Ingallinella, P., E. Bianchi, R. Ingenito, U. Koch, C. Steinkuhler, S. Altamura, and A. Pessi. 2 0 0 0 . Optimization of the P'-region of peptide inhibitors of hepatitis C virus N S 3 / 4 A protease. Biochemistry 39:12898-906. Janciauskiene, S. 2 0 0 1 . C o n f o r m a t i o n a l properties of serine proteinase inhibitors (serpins) confer multiple pathophysiological roles. B i o c h i m B i o p h y s A c t a 1535:22135. Jean, F., A. Basak, J . DiMaio, N. G. Seidah, and C. Lazure. 1 9 9 5 . A n internally q u e n c h e d fluorogenic substrate of p r o h o r m o n e c o n v e r t a s e 1 a n d furin l e a d s to a potent p r o h o r m o n e c o n v e r t a s e inhibitor. B i o c h e m J 307 ( Pt 3):689-95. Jean, F., K. Stella, L. Thomas, G. Liu, Y. Xiang, A. J . Reason, and G. Thomas. 1998. a l p h a l - A n t i t r y p s i n Portland, a bioengineered serpin highly selective for furin: application a s a n antipathogenic agent. P r o c Natl A c a d S c i U S A 95:7293-8. Jean, F., L. Thomas, S. S. Molloy, G. Liu, M. A. Jarvis, J . A. Nelson, and G. Thomas. 2 0 0 0 . A protein-based therapeutic for h u m a n cytomegalovirus infection. P r o c Natl A c a d S c i U S A 97:2864-9. Kakiuchi, N., S. Nishikawa, M. Hattori, and K. Shimotohno. 1 9 9 9 . A high throughput a s s a y of the hepatitis C virus nonstructural protein 3 s e r i n e proteinase. J Virol M e t h o d s 80:77-84. Kato, N., and K. Shimotohno. 2 0 0 0 . S y s t e m s to culture hepatitis C virus. C u r r T o p Microbiol Immunol 242:261-78  103  Kim, J . L , K. A. Morgenstern, C. Lin, T. Fox, M. D. Dwyer, J . A. Landro, S. P. Chambers, W. Markland, C. A. Lepre, E. T. O'Malley, S. L . Harbeson, C. M. Rice, M. A. Murcko, P. R. Caron, and J . A. Thomson. 1 9 9 6 . Crystal structure of the hepatitis C virus N S 3 protease d o m a i n c o m p l e x e d with a synthetic N S 4 A cofactor peptide. C e l l 87:343-55.  Kishine, H., K. Sugiyama, M. Hijikata, N. Kato, H. Takahashi, T. Noshi, Y. Nio, M. Hosaka, Y. Miyanari, and K. Shimotohno. 2 0 0 2 . S u b g e n o m i c replicon derived from a cell line infected with the hepatitis C virus. B i o c h e m B i o p h y s R e s C o m m u n 293:993-9. Kolykhalov, A. A., S. M. Feinstone, and C. M. Rice. 1996. Identification of a highly c o n s e r v e d s e q u e n c e element at the 3' terminus of hepatitis C virus g e n o m e R N A . J Virol 70:3363-71. Komoda, Y., M. Hijikata, Y. Tanji, Y. Hirowatari, H. Mizushima, K. Kimura, and K. Shimotohno. 1994. P r o c e s s i n g of hepatitis C viral polyprotein in E s c h e r i c h i a coli. G e n e 145:221-6. Kwong, A. D., J . L. Kim, G. Rao, D. Lipovsek, and S. A. Raybuck. 1999. Hepatitis C virus N S 3 / 4 A protease. Antiviral R e s 41:67-84. Lahm, A., A. Yagnik, A. Tramontano, and U. Koch. 2 0 0 2 . Hepatitis C virus proteins a s targets for drug development: the role of bioinformatics a n d modelling. C u r r Drug Targets 3:281-96. Landro, J . A., S. A. Raybuck, Y. P. Luong, E. T. O'Malley, S. L. Harbeson, K. A. Morgenstern, G. Rao, and D. J . Livingston. 1997. M e c h a n i s t i c role of a n N S 4 A peptide cofactor with the truncated N S 3 protease of hepatitis C virus: elucidation of the N S 4 A stimulatory effect v i a kinetic a n a l y s i s a n d inhibitor m a p p i n g . Biochemistry 36:9340-8. LaPlante, S. R., D. R. Cameron, N. Aubry, S. Lefebvre, G. Kukolj, R. Maurice, D. Thibeault, D. Lamarre, and M. Llinas-Brunet. 1 9 9 9 . Solution structure of s u b s t r a t e - b a s e d ligands w h e n b o u n d to hepatitis C virus N S 3 protease d o m a i n . J Biol C h e m 274:18618-24. Lechmann, M., and T. J . Liang. 2 0 0 0 . V a c c i n e d e v e l o p m e n t for hepatitis C . S e m i n Liver Dis 20:211-26. Leyssen, P., E. De Clercq, and J . Neyts. 2 0 0 0 . P e r s p e c t i v e s for the treatment of infections with Flaviviridae. C l i n Microbiol R e v 13:67-82, table of contents.  104  Lin, C , B. M. Pragai, A. Grakoui, J . Xu, and C. M. Rice. 1994. Hepatitis C virus N S 3 serine proteinase: t r a n s - c l e a v a g e requirements a n d p r o c e s s i n g kinetics. J Virol 68:8147-57. Lin, C , and C. M. Rice. 1 9 9 5 . T h e hepatitis C virus N S 3 serine proteinase a n d N S 4 A cofactor: establishment of a cell-free t r a n s - p r o c e s s i n g a s s a y . P r o c Natl A c a d S c i U S A 92:7622-6. Llinas-Brunet, M., M. Bailey, G. Fazal, S. Goulet, T. Halmos, S. Laplante, R. Maurice, M. Poirier, M. A. Poupart, D. Thibeault, D. Wernic, and D. Lamarre. 1998. P e p t i d e - b a s e d inhibitors of the hepatitis C virus serine protease. Bioorg M e d C h e m Lett 8:1713-8. Loebermann, H., R. Tokuoka, J . Deisenhofer, and R. Huber. 1984. H u m a n a l p h a 1-proteinase inhibitor. Crystal structure analysis of two crystal modifications, molecular m o d e l a n d preliminary a n a l y s i s of the implications for function. J M o l Biol 177:531-57. Lohmann, V., F. Korner, J . Koch, U. Herian, L. Theilmann, and R. Bartenschlager. 1 9 9 9 . Replication of s u b g e n o m i c hepatitis C virus R N A s in a h e p a t o m a cell line. S c i e n c e 285:110-3. Luisetti, M., and J . Travis. 1996. Bioengineering: a l p h a 1-proteinase inhibitor sitespecific m u t a g e n e s i s . T h e prospect for improving the inhibitor. C h e s t 110:278S283S. Martin, F., N. Dimasi, C. Volpari, C. Perrera, S. Di Marco, M. Brunetti, C. Steinkuhler, R. De Francesco, and M. Sollazzo. 1998. D e s i g n of selective eglin inhibitors of H C V N S 3 proteinase. Biochemistry 37:11459-68. Matheson, N. R., H. L. Gibson, R. A. Hallewell, P. J . Barr, and J . Travis. 1 9 8 6 . R e c o m b i n a n t D N A - d e r i v e d forms of h u m a n a l p h a 1-proteinase inhibitor. S t u d i e s o n the alanine 3 5 8 a n d cysteine 3 5 8 substituted mutants. J Biol C h e m 261:10404-9. Mercer, D. F., D. E. Schiller, J . F. Elliott, D. N. Douglas, C. Hao, A. Rinfret, W. R. Addison, K. P. Fischer, T. A. Churchill, J . R. Lakey, D. L. Tyrrell, and N. M. Kneteman. 2 0 0 1 . Hepatitis C virus replication in m i c e with chimeric h u m a n livers. Nat M e d 7:927-33. Mottola, G., G. Cardinali, A. Ceccacci, C. Trozzi, L. Bartholomew, M. R. Torrisi, E. Pedrazzini, S. Bonatti, and G. Migliaccio. 2 0 0 2 . Hepatitis C virus nonstructural proteins a r e localized in a modified e n d o p l a s m i c reticulum of cells e x p r e s s i n g viral s u b g e n o m i c replicons. Virology 293:31-43.  105  Nakabayashi, H., K. Taketa, K. Miyano, T. Yamane, and J . Sato. 1982. Growth of h u m a n h e p a t o m a cells lines with differentiated functions in chemically defined m e d i u m . C a n c e r R e s 42:3858-63. Narjes, F., K. F. Koehler, U. Koch, B. Gerlach, S. Colarusso, C. Steinkuhler, M. Brunetti, S. Altamura, R. De Francesco, and V. G. Matassa. 2 0 0 2 . A d e s i g n e d P1 cysteine mimetic for covalent a n d non-covalent inhibitors of H C V N S 3 protease. Bioorg M e d C h e m Lett 12:701-4. Pelham, H. R. 1 9 8 9 . Control of protein exit from the e n d o p l a s m i c reticulum. A n n u R e v Cell Biol 5:1-23. Pessi, A. 2 0 0 1 . A personal a c c o u n t of the role of peptide r e s e a r c h in drug discovery: the c a s e of hepatitis C . J P e p t S c i 7:2-14. Pietschmann, T., V. Lohmann, G. Rutter, K. Kurpanek, and R. Bartenschlager. 2 0 0 1 . Characterization of cell lines carrying self-replicating hepatitis C virus R N A s . J Virol 75:1252-64. Plotnick, M. I., M. Samakur, Z. M. Wang, X. Liu, H. Rubin, N. M. Schechter, and T. Selwood. 2 0 0 2 . Heterogeneity in serpin-protease c o m p l e x e s a s d e m o n s t r a t e d by differences in the m e c h a n i s m of c o m p l e x b r e a k d o w n . Biochemistry 41:334-42. Po, A., M. Martin, M. Richer, M. Juliano, L. Juliano, and F. Jean. 2 0 0 1 . Probing the substrate specificity of Hepatitis C virus non-structural 3 serine p r o t e a s e by intramolecularly q u e n c h e d fluorogenic peptide substrate. P e p t i d e s (The W a v e O f T h e Future) 17 A m e r i c a n Peptide S y m p o s i u m , 5 6 3 - 5 6 4 . th  Ray, M. B., V. J . Desmet, and W. Gepts. 1 9 7 7 . alpha-1-Antitrypsin immunoreactivity in islet cells of adult h u m a n p a n c r e a s . C e l l T i s s u e R e s 185:63-8. Rosenberg, S., P. J . Barr, R. C. Najarian, and R. A. Hallewell. 1984. S y n t h e s i s in yeast of a functional oxidation-resistant mutant of h u m a n alpha-antitrypsin. Nature 312:77-80. Rosenberg, S. 2 0 0 1 . R e c e n t a d v a n c e s in the molecular biology of hepatitis C virus. J M o l Biol 313:451-64. Ryan, M. D., S. Monaghan, and M. Flint. 1998. V i r u s - e n c o d e d proteinases of the Flaviviridae. J G e n Virol 79 ( Pt 5):947-59. Sali, D. L., R. Ingram, M. Wendel, D. Gupta, C. McNemar, A. Tsarbopoulos, J . W. Chen, Z. Hong, R. Chase, C. Risano, R. Zhang, N. Yao, A. D. Kwong, L. Ramanathan, H. V. Le, and P. C. Weber. 1998. S e r i n e protease of hepatitis C virus e x p r e s s e d in insect cells a s the N S 3 / 4 A complex. Biochemistry 37:3392-401.  106  Scarselli, E., A. Urbani, A. Sbardellati, L. Tomei, R. De Francesco, and C. Traboni. 1997. G B virus B a n d hepatitis C virus N S 3 s e r i n e p r o t e a s e s s h a r e substrate specificity. J Virol 71:4985-9. Schechter, I., and A. Berger. 1 9 6 7 . O n the s i z e of the active site in p r o t e a s e s . I. P a p a i n . B i o c h e m B i o p h y s R e s C o m m u n 27:157-62. Schinazi, R. F., E. Han, P. L. Black, X. Yao, and S. Dagan. 1999. C e l l - b a s e d a n d a n i m a l m o d e l s for hepatitis B a n d C v i r u s e s . Antivir C h e m C h e m o t h e r 10:99-114. Shimizu, Y., K. Yamaji, Y. Masuho, T. Yokota, H. Inoue, K. Sudo, S. Satoh, and K. Shimotohno. 1996. Identification of the s e q u e n c e o n N S 4 A required for e n h a n c e d c l e a v a g e of the N S 5 A / 5 B site by hepatitis C virus N S 3 protease. J Virol 70:127-32. Silverman, G. A., P. I. Bird, R. W. Carrell, F. C. Church, P. B. Coughlin, P. G. 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. 2 0 0 1 . T h e s e r p i n s a r e a n e x p a n d i n g superfamily of structurally similar but functionally d i v e r s e proteins. Evolution, m e c h a n i s m of inhibition, novel functions, a n d a r e v i s e d nomenclature. J Biol C h e m 276:33293-6. Stein, P. E., and R. W. Carrell. 1995. W h a t d o dysfunctional s e r p i n s tell u s about m o l e c u l a r mobility a n d d i s e a s e ? N a t Struct Biol 2:96-113. Steinkuhler, C , L. Tomei, and R. De Francesco. 1996. In vitro activity of hepatitis C virus p r o t e a s e N S 3 purified from recombinant Baculovirus-infected S f 9 cells. J Biol C h e m 271:6367-73. Steinkuhler, C , A. Urbani, L. Tomei, G. Biasiol, M. Sardana, E. Bianchi, A. Pessi, and R. De Francesco. 1 9 9 6 . Activity of purified hepatitis C virus p r o t e a s e N S 3 o n peptide substrates. J Virol 70:6694-700. Steinkuhler, C , G. Biasiol, M. Brunetti, A. Urbani, U. Koch, R. Cortese, A. Pessi, and R. De Francesco. 1998. P r o d u c t inhibition of the hepatitis C virus N S 3 protease. Biochemistry 37:8899-905. Steinkuhler, C , U. Koch, F. Narjes, and V. G. Matassa. 2 0 0 1 . Hepatitis C virus protease inhibitors: current p r o g r e s s a n d future c h a l l e n g e s . C u r r M e d C h e m 8 : 9 1 9 932. Tanji, Y., M. Hijikata, S. Satoh, T. Kaneko, and K. Shimotohno. 1995. Hepatitis C v i r u s - e n c o d e d nonstructural protein N S 4 A h a s versatile functions in viral protein p r o c e s s i n g . J Virol 69:1575-81.  107  Teasdale, R. D., and M. R. Jackson. 1996. S i g n a l - m e d i a t e d sorting of m e m b r a n e proteins between the e n d o p l a s m i c reticulum a n d the golgi apparatus. A n n u R e v C e l l D e v Biol 12:27-54. Tomei, L , C. Failla, R. L. Vitale, E . Bianchi, and R. De Francesco. 1996. A central hydrophobic d o m a i n of the hepatitis C virus N S 4 A protein is n e c e s s a r y a n d sufficient for the activation of the N S 3 protease. J G e n Virol 77 ( Pt 5): 1065-70. Travis, J . , and G. S. Salvesen. 1983. H u m a n p l a s m a proteinase inhibitors. A n n u R e v B i o c h e m 52:655-709. Tsai, C. S., R. Ritch, D. H. Shin, J . Y. Wan, and T. Chi. 1992. A g e - r e l a t e d decline of d i s c rim a r e a in visually normal subjects. O p h t h a l m o l o g y 99:29-35. Tsuji, A., T. Ikoma, E. Hashimoto, and Y. Matsuda. 2 0 0 2 . D e v e l o p m e n t of selectivity of alphal-antitrypsin variant by m u t a g e n e s i s in its reactive site loop against proprotein c o n v e r t a s e . A crucial role of the P 4 arginine in P A C E 4 inhibition. Protein E n g 15:123-30. Urbani, A., E . Bianchi, F. Narjes, A. Tramontano, R. De Francesco, C. Steinkuhler, and A. Pessi. 1997. Substrate specificity of the hepatitis C virus serine protease N S 3 . J Biol C h e m 272:9204-9. Velazquez-Campoy, A., M. J . Todd, S. Vega, and E . Freire. 2 0 0 1 . Catalytic efficiency a n d vitality of HIV-1 p r o t e a s e s from African viral subtypes. P r o c Natl A c a d S c i U S A 98:6062-7. Vishnuvardhan, D., N. Kakiuchi, P. T. Urvil, K. Shimotohno, P. K. Kumar, and S. Nishikawa. 1997. E x p r e s s i o n of highly active recombinant N S 3 protease d o m a i n of hepatitis C virus in E. coli. F E B S Lett 402:209-12. Warshel, A., and J . Aqvist. 1 9 9 1 . Electrostatic energy function. A n n u R e v B i o p h y s B i o p h y s C h e m 20:267-98.  and  macromolecular  Whisstock, J . , R. Skinner, and A. M. Lesk. 1998. A n atlas of serpin conformations. T r e n d s B i o c h e m S c i 23:63-7. Wilczynska, M., M. Fa, P. I. Ohlsson, and T. Ny. 1995. T h e inhibition m e c h a n i s m of serpins. E v i d e n c e that the mobile reactive center loop is c l e a v e d in the native protease-inhibitor c o m p l e x . J Biol C h e m 270:29652-5. Wright, H. T., and J . N. Scarsdale. 1995. Structural b a s i s for serpin activity. Proteins 22:210-25.  108  inhibitor  Yang, S. H., C. G. Lee, M. K. Song, and Y. C. Sung. 2 0 0 0 . Internal c l e a v a g e of hepatitis C virus N S 3 protein is d e p e n d e n t o n the activity of N S 3 4 A protease. Virology 268:132-40. Yaron, A., A. Carmel, and E. Katchalski-Katzir. 1979. Intramolecularly q u e n c h e d fluorogenic substrates for hydrolytic e n z y m e s . A n a l B i o c h e m 95:228-35. Yasui, K., T. Wakita, K. Tsukiyama-Kohara, S. I. Funahashi, M. Ichikawa, T. Kajita, D. Moradpour, J . R. Wands, and M. Kohara. 1998. T h e native form a n d maturation p r o c e s s of hepatitis C virus c o r e protein. J Virol 72:6048-55. Young, S. D. 2 0 0 1 . Inhibition of HIV-1 integrase by s m a l l m o l e c u l e s : the potential for a n e w c l a s s of A I D S c h e m o t h e r a p e u t i c s . C u r r O p i n Drug D i s c o v D e v e l 4 : 4 0 2 - 1 0 . Zhang, R., J . Durkin, W. T. Windsor, C. McNemar, L. Ramanathan, and H. V. Le. 1997. P r o b i n g the substrate specificity of hepatitis C virus N S 3 serine protease by using synthetic peptides. J Virol 71:6208-13.  109  Chapter 9: Appendix  Residual activity of Elastase vs cci-AT inhibition  120 100  20  40 cci-AT  60  80  100  140  concentration (nM)  Appendix I. Tight binding titration of pancreatic elastase by ai-AT. Elastase (9 nM) was incubated with increasing amount of cci-AT for 3 hours at 31°C. MeoSuc-A-A-P-VA M C (50uM) was added to determine residual activity of elastase. Data analysis revealed a value of [E] = 9.0 nM, K, = 53 p M and SI = 2. The a i - A T concentration is in (nM), Eo is determined by the distance indicated in the graph converted to (nM) accordingly. D and d can be obtained from directly measuring the distance as indicated in the graph. A l l other Kj are estimated with the same method. 0  110  pET-21(+) Vector  TB063 12/98 pET-21(+) (Cat, No. 69770-3) is a transcription vector designed for expression from bacterial translation signals carried within a cloned insert, it therefore lacks the ribosorae binding site and ATC start codon present on the pET translation vectors. A C-terirunal His«Tag* sequence 'rs avnliable, Unique sites are shown on the circle map, Note that tlte sequence is numbered by the pBR322 convention, so the T7 expression region is reversed on the circular map. 'Hie clamng/expression region of the coding strand transcribed by T7 RNA polymerase is shown below. The fl origin is oriented so that Infection with lid per phage will produce virions containing single-stranded DNA that corresponds to the coding strand. Therefore, single-stranded sequencing should be performed using the T7 terminator primer (Gat. No. B9337-3).  Sly !(57) Bpu1102 l(S0i/  pET-Z1(+) sequence lantnartti T7 promoter 237-253 T7 transcription start 23G Multiple cloning sites (Pamlll'Xhol)158-203 His*Tag* coding sequence 140-157 T7 terminator 26-72 tacl coding sequence 610-1719 pBR322 origin 3153 lila codingsequence 3914-4771 fl origin 4903-5358  . A v a 1(158) >Xho 1(198) >Eag 1(166) ,Not 1(166) , H l n d 111(173)  , S a l 1(179)  , S a c 1(190) , E c o R 1(192) - B a m H l(19B) Dra lll(B1Z7).  7Bg\ ll(268) ;ph i(4Gs) EcoN 1(525) PflM 1(572) ApaB 1(6741  ,Mlll 1(900) - B e l 1(1004) ^BstE 11(1171) - B m g 1(110S) SApa 1(1201) Bsa 1(4046  pET-21(+)  BSSH 11(1401) IcdR V(1440) -Hpa 1(1496)  (5369bp)  Eam1105 1(3984)-  ••PshA 1(1339)  AlwN 1(3507) •>  faVPspS 11(2007) ^Bpu10 1(2107)  BspLU11 K3B91) I Sap 1(2875) ' Bst1107 1(2862)'  TO|111 l(2B36)r  IBspE 1(2617)  T7 promoter prim.r 009348-3 T7 promolrx  ^  lac opwator  TMT«a*rr«MTW«»»TOM«»T^ iMMTEmcrrraTME^erjB^^  BanH I gcoH I She I BpuWZ I  Srvl  T7 tomwifflor  T7torminatorprlm«r*ee337-3  pET-21 (+) cloning/expression region  Appendix II. Restriction map of pET-21 (Novagen). (A) pET-21 vector is ~5.4kbp with ampicillin resistance and multiple cloning sites. (B) pET-21 vector utilize the T 7 promoter and include the desired restriction sites, Xho 1 and B p u l 102 1 in the multiple cloning sites. Bpul 1021 and B L P I are Isoschizomer.  Ill  K D E L constructs of a i - A T . M u t A & M u t D in pShuttle-CMV vector  Pat MCS 947-1000 BgUL  right arm homolofiy region left arm homology region  Pme\  ai-AT-KDEL  Mut A-KDEL  Mut D-KDEL  Appendix III. Restriction map of p - C M V shuttle vector (Stratagene). a i A T - K D E L was made and cloned into the p-CMV shuttle vector by our collaborator. Mut A and D - K D E L were made using quick change mutagenesis kit in the lab. Site-directed mutagenesis procedure is described in the Materials and Method section.  112  Appendix IV C L U S T A L W Result ( A l A T , Mutant A , B , C , D & E) GenomeNet CLUSTALW Server  (Kyoto Center) o n Wed O c t 23 04:01:28 J S T  CLUSTAL W (1.81) M u l t i p l e  Sequence  Sequence Sequence Sequence Sequence Sequence Sequence Sequence Sequence Start of Aligning, MutantB MutantE MutantD MutantC alAT MutantA  Alignments  type e x p l i c i t l y s e t t o P r o t e i n format i s Pearson MutantB 400 aa 400 aa MutantE 400 aa MutantD 400 aa MutantC alAT 400 aa MutantA 400 aa Pairwise alignments . His-tag  Flag-tag  MHHHHHHDPMDYKDDDDKEDPQGDAAQKTDTSHHDQDHPTFNKI TPNLAE FAFS LYRQLA M1HHHHH 3PM )YKDDDDKI!DPQGDAAQKTDTSHHDQDHPTFNKI T PNLAE FAFS LYRQLA M1HHHHH 3PM )YKDDDDKI1DPQGDAAQKTDTSHHDQDHPTFNKI TPNLAEFAFSLYRQLA M3HHHHH 3PM 3YKDDDDKI1DPQGDAAQKTDTSHHDQDHPTFNKI T PNLAE FAFS LYRQLA MfiHHHHH 3PM 3YKDDDDKi:DPQGDAAQKTDTSHHDQDHPTFNKI T PNLAE FAFS LYRQLA M H H H H H H n P M r l V K n n n n K H i n P n P i n A A Q K T n T H H H n Q n H P T T T M T ? T T PNLAE FAFS LYRQLA ************************************************************  M u t a n t B HQSNSTNIFFSPVSIATAFAMLSLGTKADTHDEILEGLNFNLTEIPEAQIHEGFQELLHT M u t a n t E HQSNSTNIFFSPVSIATAFAMLSLGTKADTHDEILEGLNFNLTEIPEAQIHEGFQELLHT M u t a n t D HQSNSTNIFFSPVSIATAFAMLSLGTKADTHDEILEGLNFNLTEIPEAQIHEGFQELLHT M u t a n t C HQSNSTNIFFSPVSIATAFAMLSLGTKADTHDEILEGLNFNLTEIPEAQIHEGFQELLHT alAT HQSNSTNIFFSPVSIATAFAMLSLGTKADTHDEILEGLNFNLTEIPEAQIHEGFQELLHT M u t a n t A HQSNSTNIFFSPVSIATAFAMLSLGTKADTHDEILEGLNFNLTEIPEAQIHEGFQELLHT ********************************************************** MutantB MutantE MutantD MutantC alAT MutantA  LNQPDSQLQLTTGNGLFLSEGLKLVDKFLEDVKKLYHSEAFTVNFGDTEEAKKQINDYVE LNQPDSQLQLTTGNGLFLSEGLKLVDKFLEDVKKLYHSEAFTVNFGDTEEAKKQINDYVE LNQPDSQLQLTTGNGLFLSEGLKLVDKFLEDVKKLYHSEAFTVNFGDTEEAKKQINDYVE LNQPDSQLQLTTGNGLFLSEGLKLVDKFLEDVKKLYHSEAFTVNFGDTEEAKKQINDYVE LNQPDSQLQLTTGNGLFLSEGLKLVDKFLEDVKKLYHSEAFTVNFGDTEEAKKQINDYVE LNQPDSQLQLTTGNGLFLSEGLKLVDKFLEDVKKLYHSEAFTVNFGDTEEAKKQINDYVE ************************************************************  MutantB MutantE MutantD MutantC alAT MutantA  KGTQGKIVDLVKELDRDTVFALVNYIFFKGKWERPFEVKDTEEEDFHVDQVTTVKVPMMK KGTQGKIVDLVKELDRDTVFALVNYIFFKGKWERPFEVKDTEEEDFHVDQVTTVKVPMMK KGTQGKIVDLVKELDRDTVFALVNYIFFKGKWERPFEVKDTEEEDFHVDQVTTVKVPMMK KGTQGKIVDLVKELDRDTVFALVNYIFFKGKWERPFEVKDTEEEDFHVDQVTTVKVPMMK KGTQGKIVDLVKELDRDTVFALVNYIFFKGKWERPFEVKDTEEEDFHVDQVTTVKVPMMK KGTQGKIVDLVKELDRDTVFALVNYIFFKGKWERPFEVKDTEEEDFHVDQVTTVKVPMMK ************************************************************  113  2002  MutantB MutantE MutantD MutantC alAT MutantA  RLGMFNIQHCKKLSSWVLLMKYLGNATAIFFLPDEGKLQHLENELTHDI ITKFLENEDRR RLGMFNIQHCKKLSSWVLLMKYLGNATAIFFLPDEGKLQHLENELTHDI ITKFLENEDRR RLGMFNIQHCKKLSSWVLLMKYLGNATAIFFLPDEGKLQHLENELTHDI ITKFLENEDRR RLGMFNIQHCKKLSSWVLLMKYLGNATAIFFLPDEGKLQHLENELTHDI ITKFLENEDRR RLGMFNIQHCKKLSSWVLLMKYLGNATAIFFLPDEGKLQHLENELTHDI ITKFLENEDRR RLGMFNIQHCKKLSSWVLLMKYLGNATAIFFLPDEGKLQHLENELTHDI ITKFLENEDRR * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *****************  MutantB MutantE MutantD MutantC alAT MutantA  SASLHLPKLS ITGTYDLKSILGQLGITKVFSNGADLSGVTEEAPLKLSKAVHKAVLTIDE S A S L H L P K L S ITGTYDLKSILGQLGITKVFSNGADLSGVTEEAPLKLSKAVHKAVLTIDE S A S L H L P K L S ITGTYDLKSILGQLGITKVFSNGADLSGVTEEAPLKLSKAVHKAVLTIDE S A S L H L P K L S ITGTYDLKSILGQLGITKVFSNGADLSGVTEEAPLKLSKAVHKAVLTIDE SASLHLPKLS ITGTYDLKSILGQLGITKVFSNGADLSGVTEEAPLKLSKAVHKAVLTIDE S A S L H L P K L S ITGTYDLKSILGQLGITKVFSNGADLSGVTEEAPLKLSKAVHKAVLTIDE  ************************************************************ R S L (P6-P'4)  MutantB MutantE MutantD MutantC alAT MutantA  KGTEAAGAMFbESIPCSIPPtVKFNKPFVFLMIEQNTKSP KGTEAAGAMF:)ESIPCPIPP::VKFNKPFVFLMIEQNTKSP KGTEAAGAMF:JESIPCSIPP!IVKFNKPFVFLMIEQNTKSP KGTEAAGAMFOEAIPCSIPP^VKFNKPFVFLMIEQNTKSP KGTEAAGAMF LEAIPMSIPP! SVKFNKPFVFLMIEQNTKSP KGTEAAGAMFLEAI PCS I P P E VKFNKPFVFLMT F.QNTKS P **********  *.**  ***********************  114  Appendix V C L U S T A L W Result (oclAT, Mutant A , B , C , D & E) GenomeNet CLUSTALW Server  (Kyoto Center) o n Thu O c t 24 03:36:41 J S T  2002  CLUSTAL W (1.81) M u l t i p l e S e q u e n c e A l i g n m e n t s S e q u e n c e t y p e e x p l i c i t l y s e t t o DNA Sequence f o r m a t i s P e a r s o n S e q u e n c e 1: oclAT 12 01 b p S e q u e n c e 2: M u t a n t A 1201 bp S e q u e n c e 3: M u t a n t B 1201 bp S e q u e n c e 4: M u t a n t E 1201 bp S e q u e n c e 5: M u t a n t D 1201 bp S e q u e n c e 6: M u t a n t C 1201 bp Start o f Pairwise alignments Aligning... OclAT AT GCATCACCATCACCATCAC GAT C C CAT G GAC TACAAG GAC GAC GAT GACAAGGAAGAT M u t a n t A ATGCATCACCAT CAC CAT CAC GAT C C CAT G GAC TACAAG GAC GAC GAT GACAAG GAAGAT M u t a n t B ATGCATCACCAT C AC CAT C AC GAT C C C AT GGAC T AC AAGG AC G AC GAT GAC AAG GAAGAT M u t a n t E ATGCATCACCATCAC CAT CAC GAT C C CAT GGAC TACAAG GAC GAC GAT GACAAG GAAGAT M u t a n t D ATGCATCACCATCACCATCACGATCCCATGGACTACAAGGACGACGATGACAAGGAAGAT M u t a n t C ATGCATCACCATCACCATCACGATCCCATGGACTACAAGGACGACGATGACAAGGAAGAT *************************************************** alAT CCCCAGGGAGATGCTGCCCAGAAGACAGATACATCCCACCATGATCAGGATCACCCAACC M u t a n t A CCCCAGGGAGATGCTGCCCAGAAGACAGATACATCCCACCATGATCAGGATCACCCAACC M u t a n t B C C C CAGG GAGAT G C T G C C CAGAAGACAGATACAT CCCACCAT GAT CAG GAT CAC C CAAC C M u t a n t E C C C CAG G GAGAT GC TG C C C AGAAGAC AG AT AC AT CCCACCAT GAT C AG GAT C AC C C AAC C M u t a n t D CCCCAGGGAGATGCTGCCCAGAAGACAGATACATCCCACCATGATCAGGATCACCCAACC M u t a n t C CCCCAGGGAGATGCTGCCCAGAAGACAGATACATCCCACCATGATCAGGATCACCCAACC ************ *.* ********************************************** alAT MutantA MutantB MutantE MutantD MutantC  TTCAACAAGATCACCCCCAACCTGGCTGAGTTCGCCTTCAGCCTATACCGCCAGCTGGCA TTCAACAAGATCACCCCCAACCTGGCTGAGTTCGCCTTCAGCCTATACCGCCAGCTGGCA TTCAACAAGATCACCCCCAACCTGGCTGAGTTCGCCTTCAGCCTATACCGCCAGCTGGCA TTCAACAAGATCACCCCCAACCTGGCTGAGTTCGCCTTCAGCCTATACCGCCAGCTGGCA TTCAACAAGATCACCCCCAACCTGGCTGAGTTCGCCTTCAGCCTATACCGCCAGCTGGCA TTCAACAAGATCACCCCCAACCTGGCTGAGTTCGCCTTCAGCCTATACCGCCAGCTGGCA ************************************************************  alAT MutantA MutantB MutantE MutantD MutantC  CACCAGTCCAACAGCACCAATATCTTCTTCTCCCCAGTGAGCATCGCTACAGCCTTTGCA CACCAGTCCAACAGCACCAATATCTTCTTCTCCCCAGTGAGCATCGCTACAGCCTTTGCA CACCAGTCCAACAGCACCAATATCTTCTTCTCCCCAGTGAGCATCGCTACAGCCTTTGCA CACCAGTCCAACAGCACCAATATCTTCTTCTCCCCAGTGAGCATCGCTACAGCCTTTGCA CACCAGTCCAACAGCACCAATATCTTCTTCTCCCCAGTGAGCATCGCTACAGCCTTTGCA CACCAGTCCAACAGCACCAATATCTTCTTCTCCCCAGTGAGCATCGCTACAGCCTTTGCA  115  alAT MutantA MutantB MutantE MutantD MutantC  ATGCTCTCCCTGGGGACCAAGGCTGACACTCACGATGAAATCCTGGAGGGCCTGAATTTC ATGCTCTCCCTGGGGACCAAGGCTGACACTCACGATGAAATCCTGGAGGGCCTGAATTTC ATGCTCTCCCTGGGGACCAAGGCTGACACTCACGATGAAATCCTGGAGGGCCTGAATTTC ATGCTCTCCCTGGGGACCAAGGCTGACACTCACGATGAAATCCTGGAGGGCCTGAATTTC ATGCTCTCCCTGGGGACCAAGGCTGACACTCACGATGAAATCCTGGAGGGCCTGAATTTC ATGCTCTCCCTGGGGACCAAGGCTGACACTCACGATGAAATCCTGGAGGGCCTGAATTTC  ************************************************************  alAT MutantA MutantB MutantE MutantD MutantC  AACCTCACGGAGATTCCGGAGGCTCAGATCCATGAAGGCTTCCAGGAACTCCTCCATACC AACCTCACGGAGATTCCGGAGGCTCAGATCCATGAAGGCTTCCAGGAACTCCTCCATACC AACCTCACGGAGATTCCGGAGGCTCAGATCCATGAAGGCTTCCAGGAACTCCTCCATACC AACCTCACGGAGATTCCGGAGGCTCAGATCCATGAAGGCTTCCAGGAACTCCTCCATACC AACCTCACGGAGATTCCGGAGGCTCAGATCCATGAAGGCTTCCAGGAACTCCTCCATACC AACCTCACGGAGATTCCGGAGGCTCAGATCCATGAAGGCTTCCAGGAACTCCTCCATACC *******************************************  alAT MutantA MutantB MutantE MutantD MutantC  CTCAACCAGCCAGACAGCCAGCTCCAGCTGACCACCGGCAATGGCCTGTTCCTCAGCGAG CTCAACCAGCCAGACAGCCAGCTCCAGCTGACCACCGGCAATGGCCTGTTCCTCAGCGAG CTCAACCAGCCAGACAGCCAGCTCCAGCTGACCACCGGCAATGGCCTGTTCCTCAGCGAG CTCAACCAGCCAGACAGCCAGCTCCAGCTGACCACCGGCAATGGCCTGTTCCTCAGCGAG CTCAACCAGCCAGACAGCCAGCTCCAGCTGACCACCGGCAATGGCCTGTTCCTCAGCGAG CTCAACCAGCCAGACAGCCAGCTCCAGCTGACCACCGGCAATGGCCTGTTCCTCAGCGAG  alAT MutantA MutantB MutantE MutantD MutantC  ********************************  GGCCTGAAGCTAGTGGATAAGTTTTTGGAGGATGTTAAAAAGTTGTACCACTCAGAAGCC GGCCTGAAGCTAGTGGATAAGTTTTTGGAGGATGTTAAAAAGTTGTACCACTCAGAAGCC GGCCTGAAGCTAGTGGATAAGTTTTTGGAGGATGTTAAAAAGTTGTACCACTCAGAAGCC GGCCTGAAGCTAGTGGATAAGTTTTTGGAGGATGTTAAAAAGTTGTACCACTCAGAAGCC GGCCTGAAGCTAGTGGATAAGTTTTTGGAGGATGTTAAAAAGTTGTACCACTCAGAAGCC GGCCTGAAGCTAGTGGATAAGTTTTTGGAGGATGTTAAAAAGTTGTACCACTCAGAAGCC ************************************************************  alAT TTCACTGTCAACTTCGGGGACACCGAAGAGGCCAAGAAACAGATCAACGATTACGTGGAG M u t a n t A TTCACTGTCAACTTCGGGGACACCGAAGAGGCCAAGAAACAGATCAACGATTACGTGGAG M u t a n t B TTCACTGT CAAC T T C G GG GACAC C GAAGAGG C CAAGAAACAGAT CAAC GAT TACGTGGAG M u t a n t E TTCACTGTCAACTTCGGGGACACCGAAGAGGCCAAGAAACAGATCAACGATTACGTGGAG M u t a n t D TTCACTGT CAAC T T C G GG GACAC C GAAGAG G C CAAGAAACAGAT CAAC GAT TAC GT G GAG M u t a n t C TTCACTGTCAACTTCGGGGACACCGAAGAGGCCAAGAAACAGATCAACGATTACGTGGAG ************************************************************ alAT MutantA MutantB MutantE MutantD MutantC  AAGGGTACTCAAGGGAAAATTGTGGATTTGGTCAAGGAGCTTGACAGAGACACAGTTTTT AAGGGTACTCAAGGGAAAATTGTGGATTTGGTCAAGGAGCTTGACAGAGACACAGTTTTT AAGGGTACTCAAGGGAAAATTGTGGATTTGGTCAAGGAGCTTGACAGAGACACAGTTTTT AAGGGTACTCAAGGGAAAATTGTGGATTTGGTCAAGGAGCTTGACAGAGACACAGTTTTT AAGGGTACTCAAGGGAAAATTGTGGATTTGGTCAAGGAGCTTGACAGAGACACAGTTTTT AAGGGTACTCAAGGGAAAATTGTGGATTTGGTCAAGGAGCTTGACAGAGACACAGTTTTT ************************************************************  alAT MutantA MutantB MutantE MutantD MutantC  GCTCTGGTGAATTACATCTTCTTTAAAGGCAAATGGGAGAGACCCTTTGAAGTCAAGGAC GCTCTGGTGAATTACATCTTCTTTAAAGGCAAATGGGAGAGACCCTTTGAAGTCAAGGAC GCTCTGGTGAATTACATCTTCTTTAAAGGCAAATGGGAGAGACCCTTTGAAGTCAAGGAC GCTCTGGTGAATTACATCTTCTTTAAAGGCAAATGGGAGAGACCCTTTGAAGTCAAGGAC GCTCTGGTGAATTACATCTTCTTTAAAGGCAAATGGGAGAGACCCTTTGAAGTCAAGGAC GCTCTGGTGAATTACATCTTCTTTAAAGGCAAATGGGAGAGACCCTTTGAAGTCAAGGAC ************************************************************  116  alAT ACCGAGGAAGAGGACTTCCACGTGGACCAGGTGACCACCGTGAAGGTGCCTATGATGAAG M u t a n t A ACCGAGGAAGAGGACTTCCACGTGGACCAGGTGACCACCGTGAAGGTGCCTATGATGAAG M u t a n t B ACCGAGGAAGAGGACTTCCACGTGGACCAGGTGACCACCGTGAAGGTGCCTATGATGAAG M u t a n t E ACCGAGGAAGAGGACTTCCACGTGGACCAGGTGACCACCGTGAAGGTGCCTATGATGAAG M u t a n t D ACCGAGGAAGAGGACTTCCACGTGGACCAGGTGACCACCGTGAAGGTGCCTATGATGAAG M u t a n t C AC C GAG GAAGAG GAC TTCCACGTG GAC CAGGT GAC CAC C G T GAAG GT G C C TAT GAT GAAG ********************************************* alAT MutantA MutantB MutantE MutantD MutantC  CGTTTAGGCATGTTTAACATCCAGCACTGTAAGAAGCTGTCCAGCTGGGTGCTGCTGATG CGTTTAGGCATGTTTAACATCCAGCACTGTAAGAAGCTGTCCAGCTGGGTGCTGCTGATG CGTTTAGGCATGTTTAACATCCAGCACTGTAAGAAGCTGTCCAGCTGGGTGCTGCTGATG CGTTTAGGCATGTTTAACATCCAGCACTGTAAGAAGCTGTCCAGCTGGGTGCTGCTGATG CGTTTAGGCATGTTTAACATCCAGCACTGTAAGAAGCTGTCCAGCTGGGTGCTGCTGATG CGTTTAGGCATGTTTAACATCCAGCACTGTAAGAAGCTGTCCAGCTGGGTGCTGCTGATG ************************************************************  alAT MutantA MutantB MutantE MutantD MutantC  AAATACCTGGGCAATGCCACCGCCATCTTCTTCCTGCCTGATGAGGGGAAACTACAGCAC AAATACCTGGGCAATGCCACCGCCATCTTCTTCCTGCCTGATGAGGGGAAACTACAGCAC AAATACCTGGGCAATGCCACCGCCATCTTCTTCCTGCCTGATGAGGGGAAACTACAGCAC AAATACCTGGGCAATGCCACCGCCATCTTCTTCCTGCCTGATGAGGGGAAACTACAGCAC AAATACCTGGGCAATGCCACCGCCATCTTCTTCCTGCCTGATGAGGGGAAACTACAGCAC AAATACCTGGGCAATGCCACCGCCATCTTCTTCCTGCCTGATGAGGGGAAACTACAGCAC ************************************************************  alAT MutantA MutantB MutantE MutantD MutantC  C T G GAAAAT GAAC TCACCCACGATATCAT CAC CAAG T T C C T GGAAAAT GAAGACAGAAG G C T G GAAAAT GAAC T CAC C CAC GATAT CAT CAC CAAG T T C C T G GAAAAT GAAGACAGAAG G C T GGAAAAT GAAC T CAC C CAC GATAT CAT CAC CAAG T T C C T GGAAAAT GAAGACAGAAG G C T GGAAAAT GAAC T CAC C CAC GATAT CAT CAC CAAG T T C C T G GAAAAT GAAGACAGAAGG C T GGAAAAT GAAC T CAC C CAC GATAT CAT CAC CAAG T T C C T G GAAAAT GAAGACAGAAGG CTGGAAAATGAACTCACCCACGATATCATCACCAAGTTCCTGGAAAATGAAGACAGAAGG ************************************************************  alAT MutantA MutantB MutantE MutantD MutantC  TCTGCCAGCTTACATTTACCCAAACTGTCCATTACTGGAACCTATGATCTGAAGAGCATC TCTGCCAGCTTACATTTACCCAAACTGTCCATTACTGGAACCTATGATCTGAAGAGCATC TCTGCCAGCTTACATTTACCCAAACTGTCCATTACTGGAACCTATGATCTGAAGAGCATC TCTGCCAGCTTACATTTACCCAAACTGTCCATTACTGGAACCTATGATCTGAAGAGCATC TCTGCCAGCTTACATTTACCCAAACTGTCCATTACTGGAACCTATGATCTGAAGAGCATC TCTGCCAGCTTACATTTACCCAAACTGTCCATTACTGGAACCTATGATCTGAAGAGCATC ************************************************************  alAT MutantA MutantB MutantE MutantD MutantC  CTGGGTCAACTGGGCATCACTAAGGTCTTCAGCAATGGGGCTGACCTCTCCGGGGTCACA CTGGGTCAACTGGGCATCACTAAGGTCTTCAGCAATGGGGCTGACCTCTCCGGGGTCACA CTGGGTCAACTGGGCATCACTAAGGTCTTCAGCAATGGGGCTGACCTCTCCGGGGTCACA CTGGGTCAACTGGGCATCACTAAGGTCTTCAGCAATGGGGCTGACCTCTCCGGGGTCACA CTGGGTCAACTGGGCATCACTAAGGTCTTCAGCAATGGGGCTGACCTCTCCGGGGTCACA CTGGGTCAACTGGGCATCACTAAGGTCTTCAGCAATGGGGCTGACCTCTCCGGGGTCACA ************************************************************  alAT MutantA MutantB MutantE MutantD MutantC  GAGGAGGCACCCCTGAAGCTCTCCAAGGCCGTGCATAAGGCTGTGCTGACCATCGATGAG GAGGAGGCACCCCTGAAGCTCTCCAAGGCCGTGCATAAGGCTGTGCTGACCATCGATGAG GAGGAGGCACCCCTGAAGCTCTCCAAGGCCGTGCATAAGGCTGTGCTGACCATCGATGAG GAGGAGGCACCCCTGAAGCTCTCCAAGGCCGTGCATAAGGCTGTGCTGACCATCGATGAG GAGGAGGCACCCCTGAAGCTCTCCAAGGCCGTGCATAAGGCTGTGCTGACCATCGATGAG GAGGAGGCACeCCTGAAGCTCTCCAAGGCCGTGCATAAGGCTGTGCTGACCATCGATGAG  ******************************  117  alAT MutantA MutantB MutantE MutantD MutantC  AAAGGGACTGAAGCTGCAGGCGCCATGTTTTTAGAGGCCATACCCATGTCTATCCCCCCC AAAGGGACTGAAGCTGCAGGCGCCATGTTTTTAGAGGCCATACCCTGCTCTATCCCCCCT AAAGGGACTGAAGCTGCAGGCGCCATGTTTGACGAGTCCATACCCTGCTCTATCCCCCCT AAAGGGACTGAAGCTGCAGGCGCCATGTTTGACGAGTCCATACCCTGCCCTATCCCCCCT AAAGGGACTGAAGCTGCAGGCGCCATGTTTTTAGAGTCCATACCCTGCTCTATCCCCCCT AAAGGGACTGAAGCTGCAGGCGCCATGTTTTTAGACGCCATACCCTGCTCTATCCCCCCT ****************************** ** ******** **********  alAT MutantA MutantB MutantE MutantD MutantC  GAGGTCAAGTTCAACAAACCCTTTGTCTTCTTAATGATTGAACAAAATACCAAGTCTCCC GAGGTCAAGTTCAACAAACCCTTTGTCTTCTTAATGATTGAACAAAATACCAAGTCTCCC GAGGTCAAGTTCAACAAACCCTTTGTCTTCTTAATGATTGAACAAAATACCAAGTCTCCC GAGGTCAAGTTCAACAAACCCTTTGTCTTCTTAATGATTGAACAAAATACCAAGTCTCCC GAGGTCAAGTTCAACAAACCCTTTGTCTTCTTAATGATTGAACAAAATACCAAGTCTCCC GAGGTCAAGTTCAACAAACCCTTTGTCTTCTTAATGATTGAACAAAATACCAAGTCTCCC ************************************************************  alAT MutantA MutantB MutantE MutantD MutantC  C C C C C C  118  Appendix V I C L U S T A L W Result ( K D E L - a l A T , Mutant A & D) GenomeNet CLUSTALW S e r v e r  (Kyoto Center) o n Wed O c t 23 0 6 : 1 3 : 3 6 J S T  2002  CLUSTAL W (1.81) M u l t i p l e S e q u e n c e A l i g n m e n t s Sequence t y p e e x p l i c i t l y s e t t o P r o t e i n S e q u e n c e 1: M u t a n t A - K D E L 429 a a S e q u e n c e 2: M u t a n t D - K D E L 429 a a S e q u e n c e 3: a l A T - K D E L 429 a a Start of Pairwise alignments Aligning... Flag-tag MutantA MPSSVSWGILLLAGLCCLVPVSLAEDPQGDAAQIYKDDDDKKTDTSHHDQDHPTFNKITP M u t a n t D MPSSVSWGILLLAGLCCLVPVSLAEDPQGDAAQI YKDDDDKKTDTSHHDQDHPTFNKITP alAT MPSSVSWGILLLAGLCCLVPVSLAEDP0GDAAQ1I) YKDDDDKKTDTSHHDQDHPTFNKITP ************************************************************ MutantA MutantD alAT  NLAEFAFSLYRQLAHQSNSTNIFFSPVSIATAFAMLSLGTKADTHDEILEGLNFNLTEIP NLAEFAFSLYRQLAHQSNSTNIFFSPVSIATAFAMLSLGTKADTHDEILEGLNFNLTEIP NLAEFAFSLYRQLAHQSNSTNIFFSPVSIATAFAMLSLGTKADTHDEILEGLNFNLTEIP **************************************************  MutantA MutantD alAT  EAQIHEGFQELLRTLNQPDSQLQLTTGNGLFLSEGLKLVDKFLEDVKKLYHSEAFTVNFG EAQIHEGFQELLRTLNQPDSQLQLTTGNGLFLSEGLKLVDKFLEDVKKLYHSEAFTVNFG EAQIHEGFQELLRTLNQPDSQLQLTTGNGLFLSEGLKLVDKFLEDVKKLYHSEAFTVNFG ************************************************************  MutantA MutantD  DTEEAKKQINDYVEKGTQGKIVDLVKELDRDTVFALVNYIFFKGKWERPFEVKDTEEEDF DTEEAKKQINDYVEKGTQGKIVDLVKELDRDTVFALVNYIFFKGKWERPFEVKDTEEEDF  alAT  DTEEAKKQINDYVEKGTQGKIVDLVKELDRDTVFALVNYIFFKGKWERPFEVKDTEEEDF ************************************************************  MutantA MutantD alAT  HVDQVTTVKVPMMKRLGMFNIQHCKKLSSWVLLMKYLGNATAIFFLPDEGKLQHLENELT HVDQVTTVKVPMMKRLGMFNIQHCKKLSSWVLLMKYLGNATAIFFLPDEGKLQHLENELT HVDQVTTVKVPMMKRLGMFNIQHCKKLSSWVLLMKYLGNATAIFFLPDEGKLQHLENELT ************************************************************  MutantA MutantD alAT  HDIITKFLENEDRRSASLHLPKLSITGTYDLKSVLGQLGITKVFSNGADLSGVTEEAPLK HDIITKFLENEDRRSASLHLPKLSITGTYDLKSVLGQLGITKVFSNGADLSGVTEEAPLK  MutantA MutantD alAT  LSKAVHKAVLTIDEKGTEAAGAMFIiEAIPCSIPPIIVKFNKPFVFLMIEQNTKSPLFMGKV  HDIITKFLENEDRRSASLHLPKLSITGTYDLKSVLGQLGITKVFSNGADLSGVTEEAPLK ************************************************************ R S L (P6-F4)  LSKAVHKAVLTIDEKGTEAAGAMFIiESIPCSIPPIlVKFNKPFVFLMIEQNTKSPLFMGKV LSKAVHKAVLTIDEKGTEAAGAMFIEAIPMSIPPEVKFNKPFVFLMIEONTKSPLFMGKV ***************************** ****************************** KDEL-motif  MutantA MutantD alAT  VNPTQKDEL VNPTQKDEL VNPTOJKDEL. ***********  119  Appendix VII CLUSTALW Result (KDEL alAT, Mutant A & D) GenomeNet CLUSTALW Server  (Kyoto Center) on. Thu O c t 24 03:48:21 J S T  2002  CLUSTAL W (1.81) M u l t i p l e S e q u e n c e A l i g n m e n t s S e q u e n c e t y p e e x p l i c i t l y s e t t o DNA Sequence format i s P e a r s o n S e q u e n c e 1: a l A T - K D E L 1298 b p S e q u e n c e 2: MutantD-KDEL 1298 bp S e q u e n c e 3: M u t a n t A - K D E L 1298 bp Start of Pairwise alignments Aligning... alAT ATGCCGTCTTCTGTCTCGTGGGGCATCCTCCTGCTGGCAGGCCTGTGCTGCCTGGTCCCT M u t a n t D ATGCCGTCTTCTGTCTCGTGGGGCATCCTCCTGCTGGCAGGCCTGTGCTGCCTGGTCCCT M u t a n t A ATGCCGTCTTCTGTCTCGTGGGGCATCCTCCTGCTGGCAGGCCTGTGCTGCCTGGTCCCT ************************************************ alAT MutantD MutantA  GTCTCCCTGGCTGAGGATCCCCAGGGAGATGCTGCCCAGGACTACAAAGACGACGACGAC GTCTCCCTGGCTGAGGATCCCCAGGGAGATGCTGCCCAGGACTACAAAGACGACGACGAC GTCTCCCTGGCTGAGGATCCCCAGGGAGATGCTGCCCAGGACTACAAAGACGACGACGAC ************************************************************  alAT AAAAAGACAGATACAT CCCACCAT GAT CAG GAT CAC C CAAC C T T CAACAAGAT CAC C C C C M u t a n t D AAAAAGACAGATACAT C C CAC CAT GAT CAGGAT CAC C CAAC C T T CAACAAGAT CAC C C C C M u t a n t A AAAAAGACAGATACAT CCCACCAT GAT CAG GAT CAC C CAAC C T T CAACAAGAT CAC C C C C ************************************************************ alAT AACCTGGCTGAGTTCGCCTTCAGCCTATACCGCCAGCTGGCACACCAGTCCAACAGCACC M u t a n t D AACCTGGCTGAGTTCGCCTTCAGCCTATACCGCCAGCTGGCACACCAGTCCAACAGCACC M u t a n t A AACCTGGCTGAGTTCGCCTTCAGCCTATACCGCCAGCTGGCACACCAGTCCAACAGCACC  ***********************************  alAT AATATCTTCTTCTCCCCAGTGAGCATCGCTACAGCCTTTGCAATGCTCTCCCTGGGGACC M u t a n t D AATATCTTCTTCTCCCCAGTGAGCATCGCTACAGCCTTTGCAATGCTCTCCCTGGGGACC M u t a n t A AATATCTTCTTCTCCCCAGTGAGCATCGCTACAGCCTTTGCAATGCTCTCCCTGGGGACC ************************************************************ alAT AAGGCTGACACTCACGATGAAATCCTGGAGGGCCTGAATTTCAACCTCACGGAGATTCCG M u t a n t D AAGGCTGACACTCACGATGAAATCCTGGAGGGCCTGAATTTCAACCTCACGGAGATTCCG M u t a n t A AAGGCTGACACTCACGATGAAATCCTGGAGGGCCTGAATTTCAACCTCACGGAGATTCCG  ************************************************************  alAT GAGGCTCAGATCCATGAAGGCTTCCAGGAACTCCTCCGTACCCTCAACCAGCCAGACAGC M u t a n t D GAGGCTCAGATCCATGAAGGCTTCCAGGAACTCCTCCGTACCCTCAACCAGCCAGACAGC M u t a n t A GAGGCTCAGATCCATGAAGGCTTCCAGGAACTCCTCCGTACCCTCAACCAGCCAGACAGC ************************************************************  120  CClAT MutantD MutantA  CAGCTCCAGCTGACCACCGGCAATGGCCTGTTCCTCAGCGAGGGCCTGAAGCTAGTGGAT CAGCTCCAGCTGACCACCGGCAATGGCCTGTTCCTCAGCGAGGGCCTGAAGCTAGTGGAT CAGCTCCAGCTGACCACCGGCAATGGCCTGTTCCTCAGCGAGGGCCTGAAGCTAGTGGAT ********************************************  alAT AAGTTTTTGGAGGATGTTAAAAAGTTGTACCACTCAGAAGCCTTCACTGTCAACTTCGGG MutantD AAGTTTTTGGAGGATGTTAAAAAGTTGTACCACTCAGAAGCCTTCACTGTCAACTTCGGG MutantA AAGTTTTTGGAGGATGTTAAAAAGTTGTACCACTCAGAAGCCTTCACTGTCAACTTCGGG ************************************************************ alAT MutantD MutantA  GACACCGAAGAGGCCAAGAAACAGATCAACGATTACGTGGAGAAGGGTACTCAAGGGAAA GACACCGAAGAGGCCAAGAAACAGATCAACGATTACGTGGAGAAGGGTACTCAAGGGAAA GACACCGAAGAGGCCAAGAAACAGATCAACGATTACGTGGAGAAGGGTACTCAAGGGAAA ************************************************************  alAT ATTGTGGATTTGGTCAAGGAGCTTGACAGAGACACAGTTTTTGCTCTGGTGAATTACATC MutantD ATTGTGGATTTGGTCAAGGAGCTTGACAGAGACACAGTTTTTGCTCTGGTGAATTACATC MutantA ATTGTGGATTTGGTCAAGGAGCTTGACAGAGACACAGTTTTTGCTCTGGTGAATTACATC ************************************************************ alAT MutantD MutantA  TTCTTTAAAGGCAAATGGGAGAGACCCTTTGAAGTCAAGGACACCGAGGAAGAGGACTTC TTCTTTAAAGGCAAATGGGAGAGACCCTTTGAAGTCAAGGACACCGAGGAAGAGGACTTC TTCTTTAAAGGCAAATGGGAGAGACCCTTTGAAGTCAAGGACACCGAGGAAGAGGACTTC ************************************************************  alAT MutantD MutantA  CACGTGGACCAGGTGACCACCGTGAAGGTGCCTATGATGAAGCGTTTAGGCATGTTTAAC CACGTGGACCAGGTGACCACCGTGAAGGTGCCTATGATGAAGCGTTTAGGCATGTTTAAC CACGTGGACCAGGTGACCACCGTGAAGGTGCCTATGATGAAGCGTTTAGGCATGTTTAAC ************************************************************  alAT MutantD MutantA  ATCCAGCACTGTAAGAAGCTGTCCAGCTGGGTGCTGCTGATGAAATACCTGGGCAATGCC ATCCAGCACTGTAAGAAGCTGTCCAGCTGGGTGCTGCTGATGAAATACCTGGGCAATGCC ATCCAGCACTGTAAGAAGCTGTCCAGCTGGGTGCTGCTGATGAAATACCTGGGCAATGCC ************************************************************  alAT MutantD MutantA  ACCGCCATCTTCTTCCTGCCTGATGAGGGGAAACTACAGCACCTGGAAAATGAACTCACC ACCGCCATCTTCTTCCTGCCTGATGAGGGGAAACTACAGCACCTGGAAAATGAACTCACC ACCGCCATCTTCTTCCTGCCTGATGAGGGGAAACTACAGCACCTGGAAAATGAACTCACC ************************************************************  alAT ' CAC GATAT CAT CAC CAAG T T C C T G GAAAAT GAAGACAGAAGG T CTGCCAGCT TACAT T TA MutantD CACGATATCATCACCAAGTTCCTGGAAAATGAAGACAGAAGGTCTGCCAGCTTACATTTA MutantA CACGATATCATCACCAAGTTCCTGGAAAATGAAGACAGAAGGTCTGCCAGCTTACATTTA ************************************************************ alAT MutantD MutantA  CCCAAACTGTCCATTACTGGAACCTATGATCTGAAGAGCGTCCTGGGTCAACTGGGCATC CCCAAACTGTCCATTACTGGAACCTATGATCTGAAGAGCGTCCTGGGTCAACTGGGCATC CCCAAACTGTCCATTACTGGAACCTATGATCTGAAGAGCGTCCTGGGTCAACTGGGCATC ************************************************************  alAT MutantD MutantA  ACTAAGGTCTTCAGCAATGGGGCTGACCTCTCCGGGGTCACAGAGGAGGCACCCCTGAAG ACTAAGGTCTTCAGCAATGGGGCTGACCTCTCCGGGGTCACAGAGGAGGCACCCCTGAAG ACTAAGGTCTTCAGCAATGGGGCTGACCTCTCCGGGGTCACAGAGGAGGCACCCCTGAAG ************************************************************  121  alAT MutantD MutantA  CTCTCCAAGGCCGTGCATAAGGCTGTGCTGACCATCGACGAGAAAGGGACTGAAGCTGCT CTCTCCAAGGCCGTGCATAAGGCTGTGCTGACCATCGACGAGAAAGGGACTGAAGCTGCT CTCTCCAAGGCCGTGCATAAGGCTGTGCTGACCATCGACGAGAAAGGGACTGAAGCTGCT **************************************************  alAT GGGGCCATGTTTTTAGAGGCCATACCCATGTCTATCCCCCCCGAGGTCAAGTTCAACAAA MutantD GGGGCCATGTTTTTAGAGTCCATACCCTGCTCTATCCCCCCTGAGGTCAAGTTCAACAAA MutantA GGGGCCATGTTTTTAGAGGCCATACCCTGCTCTATCCCCCCTGAGGTCAAGTTCAACAAA ****************** ******** *********** ****************** alAT CCCTTTGTCTTCTTAATGATTGAACAAAATACCAAGTCTCCCCTCTTCATGGGAAAAGTG MutantD CCCTTTGTCTTCTTAATGATTGAACAAAATACCAAGTCTCCCCTCTTCATGGGAAAAGTG MutantA CCCTTTGTCTTCTTAATGATTGAACAAAATACCAAGTCTCCCCTCTTCATGGGAAAAGTG ************************************************************ alAT GTGAATCCCACCCAAAAAGACGAGCTCTGAAGCTTCTA MutantD GTGAATCCCACCCAAAAAGACGAGCTCTGAAGCTTCTA MutantA GTGAATCCCACCCAAAAAGACGAGCTCTGAAGCTTCTA **************************************  122  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0090883/manifest

Comment

Related Items