<%BANNER%>

Monitoring Polymorphism and Inhibitor-Induced Conformational Ensemble Shifts in Hiv-1 Protease Via Pulsed Electron Param...

Permanent Link: http://ufdc.ufl.edu/UFE0041425/00001

Material Information

Title: Monitoring Polymorphism and Inhibitor-Induced Conformational Ensemble Shifts in Hiv-1 Protease Via Pulsed Electron Paramagnetic Resonance
Physical Description: 1 online resource (306 p.)
Language: english
Creator: Blackburn, Mandy
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: conformational, deer, distance, drug, ensemble, epr, flaps, hiv, inhibitor, peldor, polymorphism, protease, resistance, shift
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Human Immunodeficiency Virus Type 1 protease (HIV-1 PR) is an enzyme required for viral replication, and as such represents a major drug target in the treatment of AIDS. HIV-1 PR is responsible for the cleavage of the viral polyproteins gag and gag-pol, and inhibition of this enzyme results in the formation of immature non-infectious virus particles. The structure and function of HIV-1 PR has been studied for over 20 years. However, questions regarding the conformations, flexibility, dynamics and motion of the beta-hairpin turns, also known as flaps, which cover the active site cavity, remain. Pulsed EPR methods are emerging as a powerful method for monitoring protein conformations and flexibility. Here, double electron-electron resonance (DEER) EPR of spin labeled constructs of HIV-1 PR is used to characterize changes in flap conformations and domain rotations as a function of various substrates and inhibitors. The most probable distances and the breadth of the distance distribution profiles provide insights regarding conformational mobility and flexibility. We also show that drug pressure selected mutations alter the average conformation of the flaps and the degree of opening of the flaps. Distance profiles obtained from EPR are validated by molecular dynamic (MD) simulations performed by the Simmerling group, which provide structural models needed to fully interpret the EPR results. By combining experiment and theory to understand the role that altered flap dynamics/conformations play in the mechanism of drug resistance, key insights are gained that maybe useful in the rational development of new inhibitors of this important enzyme.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Mandy Blackburn.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Fanucci, Gail E.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041425:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041425/00001

Material Information

Title: Monitoring Polymorphism and Inhibitor-Induced Conformational Ensemble Shifts in Hiv-1 Protease Via Pulsed Electron Paramagnetic Resonance
Physical Description: 1 online resource (306 p.)
Language: english
Creator: Blackburn, Mandy
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: conformational, deer, distance, drug, ensemble, epr, flaps, hiv, inhibitor, peldor, polymorphism, protease, resistance, shift
Chemistry -- Dissertations, Academic -- UF
Genre: Chemistry thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Human Immunodeficiency Virus Type 1 protease (HIV-1 PR) is an enzyme required for viral replication, and as such represents a major drug target in the treatment of AIDS. HIV-1 PR is responsible for the cleavage of the viral polyproteins gag and gag-pol, and inhibition of this enzyme results in the formation of immature non-infectious virus particles. The structure and function of HIV-1 PR has been studied for over 20 years. However, questions regarding the conformations, flexibility, dynamics and motion of the beta-hairpin turns, also known as flaps, which cover the active site cavity, remain. Pulsed EPR methods are emerging as a powerful method for monitoring protein conformations and flexibility. Here, double electron-electron resonance (DEER) EPR of spin labeled constructs of HIV-1 PR is used to characterize changes in flap conformations and domain rotations as a function of various substrates and inhibitors. The most probable distances and the breadth of the distance distribution profiles provide insights regarding conformational mobility and flexibility. We also show that drug pressure selected mutations alter the average conformation of the flaps and the degree of opening of the flaps. Distance profiles obtained from EPR are validated by molecular dynamic (MD) simulations performed by the Simmerling group, which provide structural models needed to fully interpret the EPR results. By combining experiment and theory to understand the role that altered flap dynamics/conformations play in the mechanism of drug resistance, key insights are gained that maybe useful in the rational development of new inhibitors of this important enzyme.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Mandy Blackburn.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Fanucci, Gail E.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041425:00001


This item has the following downloads:


Full Text

PAGE 1

1 MON IT ORING POLYMORPHISM AND INHIB I TOR INDUCED CONFORMATIONAL ENSEMBLE SHIFTS IN HIV 1 PROTEASE VIA PULSED ELECTRON PARAMAGNETIC RESONANCE By MANDY ELIZABETH BLACKBURN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Mandy Elizabeth Blackburn

PAGE 3

3 To my loving parents and my amazing wife

PAGE 4

4 ACKNOWLEDGMENTS I want to thank my wife, Pamela Morris, and my parents, Robert and Sandra Blackburn, for their encouragement, patience, love and support I also want to thank my Paul and Susanne Morris and Joel and Autumn Morris for their support an d encouragement. I want to thank my advisor, Dr. Gail Fanucci, with my deepest gratitude Her faith in me inspired faith in myself and helped me to accomplish more than I thought I could. She also challenged me to constantly improve myself, both as a sci entist and a human. Her generosity also made it possible for me to attend an unusually large number of conferences where I was able to meet and make contacts with many important people in our field. I would also like to thank all the professors and teache rs along the way who helped inspire and encourage me to aim higher. Miss Reilly a high school teacher, for finding the shy kid at the back of the room and pushing me into honors classes. Drs. Deborah Evans, Mark Ondrais, Richard Watts, and Joe Ho, at th e University of New Mexico, for inspiring my love of science. Dr Philip Laipis at the University of Florida School of Medicine, for his advice and mentoring. Drs. Nicole Horen stein and Phil Brucat at the University of Florida Chemistry Department, for their encouragement and many engaging conversations. I would like to thank our collaborators in this project, specifically Dr. Carlos Simmerling and Ding Fangyu for Molecular Dynamics simulations on spin labeled Human Immunodeficiency Virus ( HIV 1 ) proteas e, and Dr. Ben Dunn and Dr. Roxana Coman for the DNA for preliminary studies on HIV 1 protease and their insight on kinetic characterization of aspartic proteases. I would also like to thank Mrs. Dawn Zbell oup at the University of Virginia, as well as Dr. Ralph Weber (Bruker Biospin) for their help with the pulsed electron paramagnetic resonance experiments. Most importantly, I would like to thank

PAGE 5

5 Dr. Angerhofer for graciously sharing his instrument with me and for his patience and willingness to help train me on the spectrometer. I would like to express my gratitude to all the members in my committee, Dr. Joanna Long, Dr. Ben Dunn, Dr. Maureen Goodenow, Dr. Alex Angerhofer and Dr. Adrian Roitberg for many v aluable discussions and their support. I also wish to express my gratitude to both the current and past members of the Fanucci research group, especially Thomas Frederick, Luis Galiano, Jordan Mathias, Natasha Hurst, Jamie Kear, Jeff Carter, Stacey Ann Be njamin for their friendship and patience.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ......... Error! Bookmark not defined. LIST OF FIGURES ................................ ................................ ................................ ....................... 10 LIST OF ABBREVIATIONS ................................ ................................ ................................ ........ 11 ABSTRACT ................................ ................................ ................................ ................................ ... 24 CHAPTER 1 INTROD UCTION ................................ ................................ ................................ .................. 26 Scope of this Work ................................ ................................ ................................ ................. 26 Introduction to Human Immunodeficiency Virus (HIV) and HIV Protease (HIV PR) .......... 27 Acquired Immunodeficiency Syndrome ................................ ................................ .......... 27 Human Immunodeficiency Virus ................................ ................................ .................... 28 Genome ................................ ................................ ................................ .................... 28 Virion contents ................................ ................................ ................................ ......... 30 Structure and organization of the virion ................................ ................................ ... 32 Role of the HIV 1 Proteas e in the Viral Life Cycle ................................ ........................ 33 Structure of HIV 1 Protease ................................ ................................ ............................ 35 HIV 1 Protease as a Drug Target ................................ ................................ .................... 41 Protease Inhibitors ................................ ................................ ................................ ........... 42 Characterization of Inhibitor Binding ................................ ................................ ............. 45 HIV 1 PR Subtype Polymorphisms and Rates of Muta tion ................................ ............ 47 Drug Pressure Selected Mutations ................................ ................................ .................. 49 Understanding the Flaps of HIV 1 PR ................................ ................................ ............ 50 Protein Structure and Flexibility ................................ ................................ ............................. 58 Protein Structure ................................ ................................ ................................ .............. 58 Protein Motion and Flexibility ................................ ................................ ........................ 61 Energy Landscapes, Conformational Ensembles, and Ensemble Shifts ......................... 63 Experimental Methods for Characterizing Ensembles and Ensemble Shifts .................. 67 Nuclear Magnetic Resonance ................................ ................................ ................... 68 Hydrogen/deuterium exchange ................................ ................................ ................ 70 Fluorescence ................................ ................................ ................................ ............. 71 X ray diffraction ................................ ................................ ................................ ....... 72 Cryogenic studies ................................ ................................ ................................ ..... 72 Computational Methods ................................ ................................ ........................... 73 Scope of Dissertation ................................ ................................ ................................ .............. 73 2 ELECTRON PARAMAGNETIC RESONANCE ................................ ................................ .. 75

PAGE 7

7 Introduction to Electron Paramagnetic Resonance (EPR) ................................ ...................... 75 Spin Hamiltonian ................................ ................................ ................................ ............. 75 Electron Zeeman ................................ ................................ ................................ ...... 76 Hyperfine inter action ................................ ................................ ................................ 78 Nuclear Zeeman ................................ ................................ ................................ ....... 80 Zero field splitting ................................ ................................ ................................ .... 80 Nuclear quadrupole ................................ ................................ ................................ .. 80 Nuclear spin spin interactions ................................ ................................ .................. 81 Electron electron dipole interactions ................................ ................................ ....... 81 Exchange coupling ................................ ................................ ................................ ... 81 Site D irected S pin L abeling ................................ ................................ ............................ 82 History of site directed spin labeling ................................ ................................ ....... 82 Spin labels ................................ ................................ ................................ ................ 84 Selection of labeling sites ................................ ................................ ......................... 85 Nitroxide line shapes ................................ ................................ ................................ 85 Spin label conformations ................................ ................................ .......................... 88 Line shape analysis ................................ ................................ ................................ ... 89 Introduction to Distance Measurements Via Pulsed EPR ................................ ...................... 90 Dipolar Interaction ................................ ................................ ................................ ........... 91 Electron Spin Echo Techniques ................................ ................................ ....................... 91 Double Quantum Coheren ce (DQC) ................................ ................................ ........ 91 Double electron electron resonance (DEER) ................................ ........................... 92 Experimental Considerations for Pulsed EPR ................................ ................................ ...... 100 Instrumental Requirements ................................ ................................ ............................ 100 Spin Relaxation ................................ ................................ ................................ ............. 101 T m ................................ ................................ ................................ ........................... 102 T 1 ................................ ................................ ................................ ............................ 105 Cryoprotectants and Glassing Agents ................................ ................................ ........... 107 Temperature Selection ................................ ................................ ................................ ... 107 Sample Concentration ................................ ................................ ................................ ... 108 Analysis of DEER Data ................................ ................................ ................................ ........ 109 Converting the Dipolar Evolution Curve into a Distance Profile ................................ .. 109 Curve Fitting Approaches ................................ ................................ .............................. 110 Tikhonov Regularization Method ................................ ................................ .................. 111 DeerAnalys is Software Package ................................ ................................ .................... 111 L curve ................................ ................................ ................................ ................... 111 Approximate Pake transformation ................................ ................................ ......... 113 M odel fitting ................................ ................................ ................................ ........... 113 Background correction options ................................ ................................ .............. 114 Validation module ................................ ................................ ................................ .. 115 Verification of Background Subtraction and Gaussian Reconstruction ........................ 116 Zero point selection ................................ ................................ ................................ 116 Self consistent analysis of the backgr ound subtraction level ................................ 117 Gaussian reconstruction ................................ ................................ ......................... 120 Experimental Considerations and the Corresponding Effect on the Results ................. 123 Signal to noise ratio ................................ ................................ ............................... 123

PAGE 8

8 Maximum dipolar evolution time, t max ................................ ................................ ... 125 Spin labe ling efficiency ................................ ................................ ......................... 127 3 DISTANCE MEASUREMENTS FOR HIV 1PR SUBTYPE B ................................ ......... 128 Introduction ................................ ................................ ................................ ........................... 128 Experimental Design ................................ ................................ ................................ ..... 128 Previous Studies ................................ ................................ ................................ ............ 129 Ensembles Shifts of HIV 1 PR ................................ ................................ ...................... 132 Materials and Methods ................................ ................................ ................................ ......... 133 Protein Expression and Purification ................................ ................................ .............. 133 DEER Samples ................................ ................................ ................................ .............. 135 Protonated matrix ................................ ................................ ................................ ... 135 Deuterated matrix ................................ ................................ ................................ ... 136 DEER Experiment ................................ ................................ ................................ ......... 136 Data Analysis ................................ ................................ ................................ ................. 137 Results ................................ ................................ ................................ ................................ ... 137 Apo and Substrate Mimic ................................ ................................ .............................. 137 Inhibitors ................................ ................................ ................................ ........................ 139 Estimation of Error in Distance Profiles and Population Analysis ............................... 145 Discussion ................................ ................................ ................................ ............................. 147 Conclusion ................................ ................................ ................................ ............................ 151 4 DEER RESULTS FOR HIV PROTEASE SUBTYPE C AND CLINICAL ISOLATE V6, A DRUG RESISTANT VARIANT ................................ ................................ .............. 153 Introduction ................................ ................................ ................................ ........................... 153 Subtype C ................................ ................................ ................................ ...................... 154 Clinical Isolate V6 ................................ ................................ ................................ ......... 159 Materials and Methods ................................ ................................ ................................ ......... 163 Protein Expression and Purification ................................ ................................ .............. 163 DEER Experiments ................................ ................................ ................................ ....... 164 Results ................................ ................................ ................................ ................................ ... 164 DEER Results for Subtype C HIV 1 PR ................................ ................................ ....... 164 DEER Results for the V6 Variant of HIV 1 PR ................................ ............................ 172 Discussion ................................ ................................ ................................ ............................. 180 Conclusion ................................ ................................ ................................ ............................ 181 5 SOLUTE EFFECTS ON SPIN LABEL MOBILITY AND PROTEIN CONFORMATIONS ................................ ................................ ................................ ............ 182 Introduction ................................ ................................ ................................ ........................... 182 Solute Effects on Solution Properties and Protein Structure and Function ................... 183 Important considerations for proteins ................................ ................................ .... 183 Important considerations for water ................................ ................................ ........ 184 Important considerations for solute s ................................ ................................ ...... 185 Solute exclusion ................................ ................................ ................................ ..... 189

PAGE 9

9 Macromolecular crowding and confinement ................................ .......................... 190 Os motic pressure and water activity ................................ ................................ ...... 190 Viscosity ................................ ................................ ................................ ................. 191 Solute Effects on Continuous Wave ( CW ) EPR and Pulsed EPR Data for HIV 1 Protease ................................ ................................ ................................ ...................... 193 CW EPR line shapes ................................ ................................ .............................. 193 Pulsed EPR distance measurements ................................ ................................ ....... 193 Materials and Methods ................................ ................................ ................................ ......... 194 Materials ................................ ................................ ................................ ........................ 194 Solute Solutions ................................ ................................ ................................ ............. 195 Fluorophore Labeling ................................ ................................ ................................ .... 195 Steady State Fluorescence Anisotropy ................................ ................................ .......... 195 Continuous Wave EPR ................................ ................................ ................................ .. 196 15 N Hetronuclear Single Quantum Coherence ( HSQC ) NMR ................................ ...... 196 Results ................................ ................................ ................................ ................................ ... 197 Effect of solutes on spin label correlation times ................................ ........................... 1 99 Protein solute interactions ................................ ................................ ............................. 204 Discussion ................................ ................................ ................................ ............................. 212 Conclusion ................................ ................................ ................................ ............................ 213 6 FUTURE WORK ................................ ................................ ................................ .................. 215 Improving the Data Analysis Process for DEER Experiments ................................ ............ 215 Investigation of Poi nt Mutations ................................ ................................ .......................... 215 Method Validation via Model Systems ................................ ................................ ................ 216 Isothermal Titration Calorimetry and Differential Scanning Calorimetry ........................... 217 NMR ................................ ................................ ................................ ................................ ..... 218 APPENDICES A HIV 1 PR DNA AND PROTEIN SEQUENCES ................................ ................................ 219 Protein Sequen ces ................................ ................................ ................................ ................. 219 Inhibitor Structures ................................ ................................ ................................ ............... 220 B SUPPLEMENTAL INFORMATION FOR DEER EXPERIMENTS AND DATA ANALYSIS ................................ ................................ ................................ ........................... 222 Subtype B ................................ ................................ ................................ .............................. 222 Subtype C ................................ ................................ ................................ .............................. 244 Variant V6 ................................ ................................ ................................ ............................ 268 LIST OF REFERENCES ................................ ................................ ................................ ............. 290 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 306

PAGE 10

10 LIST OF TABLES Table page 1 1 FDA approv ed protease inhibitors for treatment of HIV 1 ................................ ............... 45 3 1 Pulse sequence parameters used with the Xepr software package from Bruker. ............. 137 3 2 Parameters of Gaussian shaped populations used to reconstruct distance profiles. ........ 144 3 3 Inhibition constant and dissociation constants for subtype C HIV 1 PR. ........................ 147 3 4 Comparison of percentage closed populations for each FDA approved inhibitor with published values of K D K I G, H, T S (25 C), and the number of non water mediated hydrogen bonds in the crystal structures (excluding residue D25). ................. 149 4 1 Kinetic parameters for subtype C HIV 1 PR. ................................ ................................ .. 157 4 2 Inhibition constant and dissociation constan ts for subtype C HIV 1 PR. ........................ 159 4 3 Kinetic parameters for variant V6 HIV 1 PR. 60 ................................ ............................... 161 4 4 Inhibition constants for variant V6 HIV 1 PR. 60 ................................ ............................. 161 4 5 Parameters of Gaussian shaped populations used to reconstruct distance profiles for subtype C. ................................ ................................ ................................ ........................ 172 4 6 Paramet ers of Gaussian shaped populations used to reconstruct distance profiles for clinical isolate, V6. ................................ ................................ ................................ .......... 179 A 1 E. coli codon optimized HIV 1 Protease Variant Sequence Alignment Residues 1 40. 219 A 2 E. coli codon optimized HIV 1 Protease Variant Sequence Alignment Residues 41 70. ................................ ................................ ................................ ............................... 219 A 3 E. coli codon optimized HIV 1 Prot ease Variant Sequence Alignment Residues 71 99. ................................ ................................ ................................ ............................... 219

PAGE 11

11 LIST OF FIGURES Figure page 1 1 Schematic diagram of the complete HIV 1 genome ................................ .......................... 29 1 2 Pictorial representation of the structural assembly of immature and mature HIV Virions with various proteins labeled accordingly. ................................ ........................... 32 1 3 HIV viral life cycle. ................................ ................................ ................................ ........... 34 1 4 HIV 1 PR topology. ................................ ................................ ................................ ........... 37 1 5 1 PR active site. ................................ ........ 38 1 6 Structure of HIV 1PR PDB ID 2bpx. ................................ ................................ .............. 40 1 7 X ray crystal structure of the semi open and the MD structure of the wide open confor mation with a space filling model of an inhibitor placed in the active site pocket to illustrate the relative sizes of the inhibitor and the gap between the flap tips. ................................ ................................ ................................ ................................ ..... 41 1 8 Inhibitor structu res and scaffolds. ................................ ................................ .................... 43 1 9 Reaction scheme for competitive inhibition. ................................ ................................ ..... 46 1 10 Phylogenic tree of HIV 1 ................................ ................................ ................................ ... 48 1 11 Protein sequence variation among protease inhibitor nave patients and patients who have undergone protease inhibitor therapy. ................................ ................................ ..... 49 1 12 Struct ures of two x ray crystal structures of HIV of the flaps in the semi open and closed conformations. ................................ ................. 51 1 13 Four x ray crystal structures illustrating the major conformations of HIV 1 PR. ........... 52 1 14 Structures of three predominant HIV 1 protease conformations. A) Closed conformation (PDB ID 1HVR). B) Semi open conformation (PDB ID 1HHP). C) Wi de open conformation. ................................ ................................ ............................. 54 1 15 A) MD structure from a simulation of HIV 1 protease illustrating the curled or tucked conformation. B) Illustration of the interactions between Ile 50 and residu es 79 81 and 32. ................................ ................................ ................................ ..................... 55 1 16 MD structures from a simulation of HIV 1 protease illustrating the curled or tucked conformations. ................................ ................................ ................................ ................... 55 1 17 A) DEER dipolar evolution curves for apo and inhibitor bound HIV 1 PR. B) Corresponding distance profiles. 81 ................................ ................................ ..................... 56

PAGE 12

12 1 18 Inter spin label distances from DEER experiments and MD simulations. ...................... 57 1 19 Conformational parameters from Chou et al. illustrating the preference of certain amino acids for a particular secondary structural element. ................................ ................ 60 1 20 Schematic diagram of an energy landscape for a protein with two major conformations, A and B. ................................ ................................ ................................ .. 65 2 1 Energy level diagram for the Zeeman and hyperfine splittin gs of an S = 1/2 spin on a nucleus with I=1. ................................ ................................ ................................ ................ 79 2 2 Structure of spin labels before and after binding to Cys side chain. ................................ 84 2 3 Schematic diagram of the possible energy levels (A) for a spin on a nucleus with spin 1 and corresponding absorption (B) and derivative (C) spectra. Sample line shapes for nitroxide radicals undergoing rapid motion (D), moderate motion (E), and no motio n (F). ................................ ................................ ................................ .................... 86 2 4 Illustration of the three co rrelation times of a spin label ................................ ................... 87 2 5 Illustration of the 4 / 5 model. The S inter acts with the H on the C (indicated by dotted line), which restricts rotations to the 4 and 5 torsional angles. .......................... 89 2 6 Parameters for quantifying the breadth of a CW EPR line shape. A) Ce ntral line width, H pp and the ratio of the center field transition to the low field transition, I CF/LF B) Second moment, H 2 ................................ ................................ ....................... 90 2 7 puls e DEER, and C) Four pulse DEER. Pulse spacings labeled with remain constant and spacings labeled with t are incremented. ................................ ................................ ................................ ................. 93 2 8 Absorption spectra for a nitroxide spin label with the low f ield transition marked as the observe frequency and the center field transition marked as the pump frequency. ..... 95 2 9 Sample dipolar evolution curves before and after background subtraction. ...................... 97 2 10 Effect of the breadth of the distance profile and the most probable distance on the dipolar evolution curve. ................................ ................................ ................................ ..... 98 2 11 Mi nimum t max for a given interspin distance based on Equation 2 38. ............................ 100 2 12 Pulse sequence for an echo decay experiment and T m curves for HIV 1 protease in various buffer conditions. ................................ ................................ ................................ 104 2 13 Inversion recovery pulse sequence and corresponding sample data. ............................... 106 2 14 Plot of the maximum concentration of spins as a functio n of the inter spin distance. .. 109

PAGE 13

13 2 15 Example of an L curve and the corresponding distance profiles and dipolar evolution curves. ................................ ................................ ................................ .............................. 112 2 16 Illustration of the Self Consistent Analysis process developed to optimize the background subtraction in the DEER dipolar evolution curves. ................................ ...... 119 2 17 Illustration of the populatio n validation process used to interrogate the validity of population containing less than 15% of the total population. ................................ .......... 122 2 18 Example of an L curve and the corresponding distance profiles and dipolar evolution curves with a high SNR. ................................ ................................ ................................ .. 124 2 19 Example of an L curve and the corresponding distance profiles and dipolar evolution curves with a low SNR. ................................ ................................ ................................ ... 124 2 20 Illustration of impact that the length of t max has on the resulting distance profile. .......... 126 3 1 A) Dipolar evolution curves for apo (black) and RTV bound (grey) subtype B HIV 1 PR (curves offset vertically for clarity). B) Corresponding distance profiles. C) L curve for apo HIV 1 PR. D) L curve for RTV bound HIV 1 PR. ................................ .. 129 3 2 Dipolar evoluti on curves for apo (black) and RTV bound (grey) subtype B HIV 1 PR labeled with A) MTSL, B) MSL, C) IAP, and D) IASL. E F) Corresponding distance profiles generated. ................................ ................................ .............................. 130 3 3 A) Dipolar evoluti on curves for apo subtype B (black), V6 (grey), and MDR769 (light grey) (curves are vertically offset for clarity). B) Corresponding distance profiles generated by TKR. ................................ ................................ .............................. 131 3 4 Cartoon illust ration of possible energy landscapes for a HIV 1 PR with the three major conformations ................................ ................................ ................................ ........ 132 3 5 The four nomenclature in t he Xepr software package. ................................ ................................ ... 136 3 6 DEER data for HIV 1 PR subtype B. ................................ ................................ .............. 138 3 7 Dipolar Evolution curves for HIV 1 PR subtype B in t he presence of various FDA approved inhibitors. ................................ ................................ ................................ ....... 140 3 8 Distance profiles for HIV 1 PR subtype B in the presence of various FDA approved inhibitors ................................ ................................ ................................ .......................... 141 3 9 Gaussian shaped populations used to fit the distance profiles for HIV 1 PR subtype B. ................................ ................................ ................................ ................................ .... 142 3 10 Gaussian shaped populations used to reconstruct the distance prof iles for HIV 1 PR subtype B. ................................ ................................ ................................ ........................ 144

PAGE 14

14 3 11 DEER data for HIV 1 PR subtype B. ................................ ................................ ............ 145 3 12 DEER data for apo HIV 1 PR subtype B collecte d with various values. ................... 146 3 13 Average distance profile (grey) and average Gaussian shaped populations for the data shown in Figure 3 12. ................................ ................................ ............................. 147 3 14 Distance distributions for synthesized HIV 1 PR in the absence and presence of inhibitors which mimic various stages of the substrate in the catalytic process. ........... 151 4 1 Ribbon diagram of apo subtype C HIV 1 PR (PDB ID 2R8N) highlighting the locations of naturally occurring polymorphisms relative to subtype B. .......................... 155 4 2 Ribbon diagrams of apo subtype C HIV 1 PR (blue) (PDB ID 2R8N) overlaid with apo subtype B HIV 1 PR (gold) (PDB ID 1HHP). ................................ ........................ 156 4 3 Ribbon diagram illustrating the difference in packing between residues 36 and 15 in (A) subtype C HIV PR (PDB ID 1SGU) and (B) subtype B HIV PR (PDB ID 2BPX). ................................ ................................ ................................ ............................. 157 4 4 Ribbon diagram illustrating the locations of the drug pressure selected mutations in V6. ................................ ................................ ................................ ................................ .. 160 4 5 Overlay of x ray structures for V6 PR (I84V, I54V) (yellow) (PDB ID 1SGU) and B PR (PDB ID 1HSG) (blue) bound to IDV. ................................ ................................ .. 162 4 6 Overlay of x ray structure s for V6 PR (I84V, I54V) (PDB ID 1SH9) (yellow) and B PR (PDB ID 1HXW) (blue) bound to RTV. ................................ ................................ .. 163 4 7 A) DEER Dipolar evolution curves for apo subtype C and subtype B HIV 1 PR. B) Correspondi ng distance profiles generated by TKR. ................................ .................. 165 4 8 A) DEER Dipolar evolution curves for apo and CA p2 bound subtype C HIV 1 PR. B) Corresponding distance profiles generated by TKR. ................................ .................. 165 4 9 A) DEER Dipolar evolution curves for CA p2 bound subtype C subtype B HIV 1 PR. B) Corresponding distance profiles generated by TKR. ................................ .......... 166 4 10 DEER dipolar evolution curves for C PR in the presence of nine FDA approved inhibitors overlaid with the dipolar evolution curve for apo C PR. ................................ 167 4 11 DEER distance profiles for C PR in the presence of nine FDA approved inhibitors (solid line) ................................ ................................ ................................ ........................ 168 4 12 Distance profiles for C PR (solid) and B PR (dashed) in the presence of nine FDA approved inhibitors. ................................ ................................ ................................ ......... 169

PAGE 15

15 4 13 Gaussian shaped populations used to reconstruct the distance profiles of apo and CA p2 bound C PR. ................................ ................................ ................................ ........ 170 4 14 Gaussian shaped populations use d to reconstruct the distance profiles for C PR in the presence of nine FDA approved inhibitors ................................ ................................ .... 171 4 15 DEER dipolar evolution curves and distance profiles for apo V6 PR and B PR. ........... 172 4 16 DEER dipolar evolution curves and distance profiles for apo and CA p2 bound V6 PR. ................................ ................................ ................................ ............................. 173 4 17 DEER dipolar evolution curves and dis tance profiles for CA p2 bound V6 PR and B PR. ................................ ................................ ................................ ................................ 174 4 18 DEER dipolar evolution curves for V6 PR in the absence and presence of nine FDA approved inhibitors. ................................ ................................ ................................ ......... 175 4 19 DEER distance profiles for the V6 variant of HIV 1 PR in the presence of nine FDA approved inhibitors ................................ ................................ ................................ .......... 176 4 20 DEER distance profiles for the V6 variant o f HIV 1 PR in the presence of nine FDA approved inhibitors ................................ ................................ ................................ .......... 177 4 21 Gaussian shaped populations used to reconstruct the distance profiles of apo and CA p2 bound V6 PR. ................................ ................................ ................................ ...... 178 4 22 Gaussian shaped populations used to reconstruct the distance profiles for V6 PR in the presence of nine FDA approved inhibitors. ................................ .............................. 179 5 1 Struc tures of water and various solutes. ................................ ................................ .......... 183 5 2 Illustration concentration regimes for polymers. ................................ ........................... 187 5 3 Illustration of several c ommon polymers structures. ................................ ....................... 188 5 4 Illustration of potential size variations between solutes and proteins which lead to changes in the translation and rotational diffusion of proteins. ................................ ....... 192 5 5 Plots of (A) viscosity and (B) osmolality as a function of percent content for sucrose (dark grey, circle), glycerol (light grey, star), PEG 3000 (triangle, grey), and Ficoll400 (square, blac k). (C) Plot of osmolality versus viscosity. .............................. 198 5 6 CW EPR line shapes for MTSL labeled HIV 1 PR in the presence of various solutes with incremented concentrations ................................ ................................ ..................... 200 5 7 CW EPR line shapes (100 G scans) for MTSL labeled HIV 1 PR in the presence of various sizes of PEG and EG with incremented concentrations ................................ ...... 200

PAGE 16

16 5 8 CW EPR line shapes (100 G scans) for spin labeled HIV 1 PR in the presence (blue) and absence (red) of a tight binding inhibitor, ritonavir (RTV) ................................ ...... 201 5 9 Fluorescence anisotropy measurements for BODIPY labeled HIV 1 PR at sites (A) T74C and (B) K55C. ................................ ................................ ................................ ...... 202 5 10 Plots of the percent change in the fluorescence anisotropy BODIPY labeled HIV 1 PR variants T74C and K 55C in presence of four solutes ................................ ................ 203 5 11 Hydrophobic surfaces of HIV 1 PR in the (A) closed (PDB ID 2pbx), (B) semi open (PDB ID 3hvp), and (C) wide open (structure from Hornak et al. 50 ) conformations. ... 205 5 12 15 N HSQC NMR spectra of HIV 1 PR with the assignments determined by comparison to Ref. 72 ................................ ................................ ................................ ........ 206 5 13 A) 15 N HSQC NMR spectra of H IV 1 PR in the absence (red) and presence (blue) of 150 g/L of Ficoll400. B) 15 N HSQC NMR spectra of HIV 1 PR titrated with Ficoll400. ................................ ................................ ................................ ......................... 208 5 14 A) 15 N HSQC NMR spectra of HIV 1 PR in the absence (red) and presence (blue) of 100 g/L of PEG 8000. B) 15 N HSQC NMR spectra of HIV 1 PR titrated with PEG 8000. ................................ ................................ ................................ ................................ 209 5 15 A) 15 N HSQC NMR spectra of HIV 1 PR in the absence (red) and presence (b lue) of 250 g/L of sucrose. B) 15 N HSQC NMR spectra of HIV 1 PR titrated with sucrose. .... 210 5 16 Comparison of the A) hydrophobic surface of HIV 1 PR to resonances changes in HIV 1 PR (mapped on to PDB ID 2pbx) in the presence of B) PEG 8000, C) sucrose, and D) Ficoll400. ................................ ................................ ................................ ........... 211 A 1 Structures for the nine FDA approved inhibitors used in this work. ............................... 220 A 2 Stick and space filling models for the nine FDA approved inhibitors used in this work. ................................ ................................ ................................ ................................ 221 B 1 DEER dat a for apo HIV 1 PR subtype B. ................................ ................................ ...... 222 B 2 DEER data for CA p2 bound HIV 1 PR subtype B. ................................ ..................... 223 B 3 DEER data for IDV bound HIV 1 PR subtype B. ................................ ......................... 224 B 4 DEER data for NFV bound HIV 1 PR subtype B. ................................ ........................ 225 B 5 DEER data for ATV bound HIV 1 PR subtype B. ................................ ........................ 226 B 6 DEER data for SQV bound HIV 1 PR subtype B. ................................ .......................... 227 B 7 DEER data for RTV bound HIV 1 PR subtype B. ................................ ........................ 228

PAGE 17

17 B 8 DEER data for LPV bound HIV 1 PR subtype B. ................................ ........................ 229 B 9 DEER data for APV bound HIV 1 PR subtype B. ................................ ........................ 230 B 10 DEER data for DRV bound HIV 1 PR subtype B. ................................ ........................ 231 B 11 DEER data for TPV bound HIV 1 PR subtype B. ................................ ........................ 232 B 12 Error Analysis for populations < 15% in HI V 1PR Apo. ................................ ................ 233 B 13 Error Analysis for populations < 20% in HIV 1PR CA p2. ................................ ............ 234 B 14 Error Analysis for populations < 20% in HIV 1PR IDV. ................................ ............. 235 B 15 Error Analysis for populations < 20% in HIV 1PR NFV. ................................ ............... 236 B 16 Error Analysis for populations < 20% in HIV 1PR ATV. ................................ ............ 237 B 17 Error Analysis for populations < 20% in HIV 1PR SQV. ................................ ............. 238 B 18 Error Analysis for populations < 20% in HIV 1PR RTV. ................................ ............. 239 B 19 Error Analysis for populations < 20% in HIV 1PR LPV. ................................ ............... 240 B 20 Error Analysis for populations < 20% in HIV 1P R APV. ................................ ............... 241 B 21 Error Analysis for populations < 20% in HIV 1PR DRV. ................................ .............. 242 B 22 Error Analysis for populations < 20% in HIV 1PR TP V. ................................ ............. 243 B 23 DEER data for apo HIV 1 PR subtype C (collected 5/09). ................................ ........... 244 B 24 DEER data for apo HIV 1 PR subtype C (collected 6/09). ................................ ........... 245 B 25 DEER data for CA p2 bound HIV 1 PR subtype C. ................................ ..................... 246 B 26 DEER data for IDV bound HIV 1 PR subtype C. ................................ ......................... 247 B 27 DEER data for NFV bound HIV 1 PR subtype C. ................................ ........................ 248 B 28 DEER data for ATV bound HIV 1 PR subtype C. ................................ .......................... 249 B 29 DEER data for APV bound HIV 1 PR subtype C. ................................ .......................... 250 B 30 DEER data for LPV bound HIV 1 PR subtype C. ................................ .......................... 251 B 31 DEER data for RTV bound HIV 1 PR subtype C. ................................ .......................... 252 B 32 DEER data for SQV bound HIV 1 PR subtype C. ................................ ........................ 253

PAGE 18

18 B 33 DEER data for DRV bound HIV 1 PR subtype C. ................................ .......................... 254 B 34 DEER data for TPV bound HIV 1 PR subtype C. ................................ .......................... 255 B 35 Error Analysis for populations < 20% in apo subtype C HIV 1PR (collected 5/09) ....... 256 B 36 Error Analysis for populations < 20% in apo subtype C HIV 1PR (collected 6/09). ...... 257 B 37 Error Analysis for populations < 20% in CA p2 bound subtype C HIV 1PR. .............. 258 B 38 Error Analysis for populations < 20% in IDV bound subtype C HIV 1PR. .................. 259 B 39 Error Analysis for populations < 20% in NFV bound subtype C HIV 1PR. ................. 260 B 39 Error Analysis for populations < 20% in ATV bou nd subtype C HIV 1PR. .................. 261 B 40 Error Analysis for populations < 20% in APV bound subtype C HIV 1PR. ................. 262 B 41 Error Analysis fo r populations < 20% in LPV bound subtype C HIV 1PR. ................... 263 B 42 Error Analysis for populations < 20% in RTV bound subtype C HIV 1PR. ................. 264 B 43 Error Analysis for populations < 20% in SQV bound subtype C HIV 1PR. ................. 265 B 44 Error Analysis for populations < 20% in DRV bound subtype C HIV 1PR .................. 266 B 45 Error Analysis for populations < 20% in TPV bound subtype C HIV 1PR. ................... 267 B 46 DEER data for apo HIV 1 PR V6. ................................ ................................ ................... 268 B 47 DEER data for CA p2 bound HIV 1 PR V6. ................................ ................................ 269 B 48 DEER data for IDV bound HIV 1 PR V6. ................................ ................................ .... 270 B 49 DEER data for NFV bound HIV 1 PR V6. ................................ ................................ ...... 271 B 50 DEER data for ATV bound HIV 1 PR V6. ................................ ................................ ... 272 B 51 DEER data for APV bound HIV 1 P R V6. ................................ ................................ .... 273 B 52 DEER data for LPV bound HIV 1 PR V6. ................................ ................................ ...... 274 B 53 DEER data for RTV bound HIV 1 PR V6. A ................................ ................................ 275 B 54 DEER data for SQV bound HIV 1 PR V6. ................................ ................................ ...... 2 76 B 55 DEER data for DRV bound HIV 1 PR V6. ................................ ................................ ... 277 B 56 DEER data for TPV bound HIV 1 PR V6. ................................ ................................ ...... 278

PAGE 19

19 B 57 Error Analysis for populations < 20% in apo HIV 1PR V6. ................................ ......... 279 B 60 E rror Analysis for populations < 20% in NFV bound HIV 1PR V6. ............................ 282 B 61 Error Analysis for populations < 20% in ATV bound HIV 1PR V6. ............................ 283 B 62 Error Analysis for populations < 20% in APV bound HIV 1PR V6. ............................ 284 B 63 Error Analysis for populations < 20% in LPV bound HIV 1PR V6. .............................. 285 B 64 Error Analysis for populations < 20% in RTV bound HIV 1PR V6. .............................. 286 B 65 Error Analysis for populations < 20% in SQV bound HIV 1PR V6. ............................ 287 B 66 Error Analysis for populations < 20% in DRV bound HIV 1PR V6. ........................... 288 B 67 Error Analysis for populations < 20% in TPV bound HIV 1PR V6. ............................ 289

PAGE 20

20 LIST OF ABBREVIATIONS HIV H uman immunodeficiency virus PR P rotease EPR E lectron paramagnetic resonance RT Re verse transcri ptase IN I nte grase CW C ontinuous wave DEER D ouble electron electron resonance AIDS A cquired immune de ficiency syndrome PCP Pneumocystis jirovecii pneumonia ORF Open reading frame PIC Pre integration complex RNA Ribonucleic acid DNA Deoxyribonucleic acid Gag Group specific antigen gene Pol Polymerase gene Env Envelope gene Nef Negative factor protein Rev A nti repression transactivator protein Tat Transactivating regulatory protein Vif Virion infectivity factor Vpr Viral protein R Vpu Viral protein U MA Matrix protein CA Capsid protein

PAGE 21

21 NC Nucleocapsid protein RT Reverse transcriptase SU, gp120 Surface protei n, glycoprotein 120 TM, gp41 Transmembrane protein, glycoprotein 41 MHC Major histocompatibility complex CD4 Cluster of differentiation 4 PI Protease inhibitor II Integrase inhibitor FI Fusion inhibitor NRTI Nucleoside reverse transcriptase inhibitor NNRTI Non Nucleoside reverse transcriptase inhibitor PDB Protein DataBank MD Molecular dynamics CW EPR Continuous wave electron paramagnetic resonance CD Circular dichroism PMPR Pentamutated protease IPTG I sopropyl D thiogalactoside EDTA Ethylenediaminetetraa cetic acid BME Mercaptoethanol, 2 Mercaptoethanol diGly Diglycine, 2 [(2 aminoacetyl)amino]acetic acid MTSL (1 Oxyl 2,2,5,5 Tetramethyl Pyrroline 3 Methyl) Methanethiosulfonate IASL 4 (2 Iodoacetamido) TEMPO IAP 3 (2 Iodoacetamido) PROXYL MSL 4 Maleim ido TEMPO NHFML National High Field Magnetic Lab

PAGE 22

22 DQC Double quantum coherence SQC Single quantum coherence ZQC Zero quantum coherence T m Phase memory T 1 Spin lattice relax ation time T 2 Spin spin relaxation time APV Amprenavir TPV Tipranavir IDV Indivavir S QV Saquinavir LPV Lopinavr FPV Fosamprenavir RTV Ritonavir DRV Darunavir ATZ Atazanavir NFV Nelfinavir TKR Tikhonov Regularization MC Montecarlo CRF Circular Recombinant Form ITC Isothermal Titration Calorimetry DSC Differential Scanning Calorimetry NMR Nu clear Magnetic Resonance e Bohr magneton ( 9.274009 49 10 24 J T 1 ) N Nuclear magneton ( 5.050783 24 10 27 J T 1 ) e proton electric charge (1.60217653 10 C )

PAGE 23

23 1.054571628 10 3 4 J s) m e Electron rest mass ( 9.10938215 10 31 kg ) g e Spectroscopic (Land) splitting factor g N Nuclear spli tting factor A 0 Hyperfine splitting Kinematic viscosity ee Electron electron dipolar coupling R Spin label diffusion tensor Correlation time M s Total electronic spin; Scaled mobility F B Fraction of spins excited by the pump pulse in the DEER experiment Regularization parameter L Tikho nov Matrix K Tikhonov regularization kernel P CDC Center for Disease Control

PAGE 24

24 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial F ulfillment of the Requirements for the Degree of Doctor of Philosophy MONITORING POLYMORPHISM AND INHIBITOR INDUCED CONFORMATIONAL ENSEMBLE SHIFTS IN HIV 1 PROTEASE VIA PULSED ELECTRON PARAMAGNETIC RESONANCE By Mandy Elizabeth Blackburn August 2010 Chai r: Gail Elizabeth Fanucci Major: Chemistry Human Immunodeficiency Virus Type 1 protease (HIV 1 PR) is an enzyme required for viral replication, and as such represents a major drug target in the treatment of AIDS. HIV 1 PR is responsible for the cleavage o f the viral polyproteins gag and gag pol, and inhibition of this enzyme results in the formation of immature non infectious virus particles. The structure and function of HIV 1 PR has been studied for over 20 years. However, questions regarding the conform ations, flexibility, dynamics and motion of the beta hairpin turns, also known as flaps, which cover the active site cavity, remain. Pulsed EPR methods are emerging as a powerful method for monitoring protein conformations and flexibility. Here, double ele ctron electron resonance (DEER) EPR of spin labeled constructs of HIV 1 PR is used to characterize changes in flap conformations and domain rotations as a function of various substrates and inhibitors. The most probable distances and the breadth of the dis tance distribution profiles provide insights regarding conformational mobility and flexibility. We also show that drug pressure selected mutations alter the average conformation of the flaps and the degree of opening of the flaps. Distance profiles obtain ed from EPR are validated by molecular dynamic (MD) simulations performed by the Simmerling group, which provide structural models needed to fully interpret

PAGE 25

25 the EPR results. By combining experiment and theory to under stand the role that altered flap dynami cs/conformations play in the mechanism of drug resistance, key insights are gained that maybe useful in the rational development of new inhibitors of this important enzyme.

PAGE 26

26 CHAPTER 1 I NTRODUCTION Scope of this Work The goal of this research was to investi human immunodeficiency virus ( HIV ) protease (HIV 1 PR) using pulsed and continuous wave (CW) electron paramagnetic resonance (EPR) It has been hypothesized that the flexibility of the flaps, the hairpin turns that control access to the active site of the protease, is modulated by drug pressured selected mutations in the protein and that th is altered flexibility imparts resistance to the inhibitors. Additionally, the effect of the naturally occurr ing polymorphisms among the various subtypes on inhibitor binding and on flap flexibility has not been fully characterized. To characterize the flap conformations a pulsed techn ique called d ouble e lectro n e lectron r esonance (DEER) was utilized to measure the distance between spin labels incorporated i nto the flaps The aims of this work were to (1) optimize this technique for addressing these questions and (2) to determine if the flap flexibility is altered by drug pressure selected mutations and natura lly occurring polymorphisms This chapter will provide an introduction to HIV and the important factors of the virus that contribute to drug resistance. The function, structure, and characteristics of the protease will be discussed in detail. Finally, th e concept of conformational ensemble s in proteins will be introduced as it relates to the flexibility of the flaps. Chapter two will provide an introduction to both CW and pulsed EPR spectroscopy the primary techniques utilized in this work. In chapter t hree, the distance measurements for HIV 1 PR subtype B will be presented and discussed. In chapter four, the results of an investigation into the effects of solutes on the conformation and spin label mobility of HIV 1 PR will presented and discussed. Cha pter five will present and

PAGE 27

27 discuss the results of investigations using subtype C PR and a drug pressure selected variant, V6. Chapter six will discuss future directions that this work can take. Introduction to HIV and HIV Protease Acquired Immunodeficienc y Syndrome Acquired Immunodeficiency Syndrome (AIDS) is the final stage of an infection caused by t he Human Immunodeficiency Virus (HIV) 1; 2 AIDS is defined a s a clinical condition in which the patient develops v arious opportunistic infections and rare cancers resulting from a n HIV related HIV virions bind and enter cells that contain CD4 protei ns in their cellular membrane including T helper c ells, regulatory T cells, dendri tic cells, monocytes, and macrophages. The CD4+ T cells are essentia l for defense against infections in the host A s HIV eventually redu ces the number of CD4+ T cells from over 800 cells/mm 3 to less than 200 cells/mm 3 the immune system of the host is pr o gressively diminished 3 The immune system is considered compr om ised when the patient becomes susceptible to opportunistic infections, 4 such as Pneumocysti s jirovecii pneumonia (PCP) 5 a rare form of pneumonia 6 a rare form a cancer, both of which are com mon among people with AIDS It is believed that HIV first entered human populations from a primate source as early as 30 100 years ag o. 7 The earliest known HIV infection was isolated from a serum sample taken from a patient in Congo in 1959 8 The identification of AIDS in 1981 resulted from multiple patients PCP, in rapid succession This statistical anomaly was noticed by the Center for Disease Contro l (CDC) and prompted further investigation. Two years later, the HIV virus was identified as the potential cause of AIDS. Since the discovery of AIDS and HIV the number of HIV infections around the globe has increased at astonishing rates. In the last decade, with the advent of anti retroviral therapy

PAGE 28

28 and aggressive efforts for preventing the spread of HIV, the rate of increase for new infections has decreased. In 2007, there were 2.7 million new infections (down from 3 million in 2001) for a total of 33 million infections worldwide, with 2 million AIDS related deaths worldwide (up from 1.7 million in 2001) 9 Human Immunodeficiency Virus There are two human immunodeficiency vi ruses, HIV 1 (from chimpanzees) and HIV 2 ( from sooty mangabeys ) 7 Although the HIV 1 and HIV 2 genomes are very similar, their DNA seque nces differ by about 55% 3 T he antibodies used for detecting HIV 1 are typically non reactive towards HIV 2 and thus HIV 2 require s separate antibody based testing for confirming infect ion in patients 10 HIV 2 is less prevalent than HIV 1 because it has lower transmissibility and delayed pathogenesis relative to HIV 1. One significant result of this difference is that patients infected with HIV 2 tend to have a longer delay between infection a nd the onset of AIDS 11 Both HIV 1 and HIV 2 are classified as members of the genus Lentivirus a subgroup of the Retroviridae family. 7 such that several years can pass between initial infection and disease development. Retroviruses are characterized by their single stranded positive sense RNA (rib onucleic acid) genome that is reversed transcribed into DNA (deoxyribonucleic acid) for integration into the host cell genome. Genome Like most viruses, HIV 1 has a compact genome (Figure 1 1) where the percen tage of the DNA encoding genes versus the DNA not encoding genes is much higher than in other o rganisms whose ratios can be as low as 50% The viral genome contains several open reading frames (ORF) that are transcribed into the viral proteins. The three largest genes are transcribed into polyprotei n chains that are proteolytically processed into the individual proteins.

PAGE 29

29 The major structural proteins matrix (MA), capsid (CA), and nucleocapsid (NC) are encoded in the g roup s pecific a ntigen ( gag ) g ene The viral enzymes ( protease (PR), reverse trans criptase (RT), and integrase (IN) are encoded in the polymerase ( pol ) gene The envelope ( e nv ) gen e encodes the surface (SU) and transmembrane (TM) proteins involved in recognizing and binding the host cell. Additionally, the nef vpr vpu tat vif and rev genes encode accessory proteins which are responsible for suppression of the cell immune response formation and localization of the preintegration c omplex ( PIC ), degradation of CD4 and enhancement of virion release Figure 1 1. Schematic diagra m of the complete HIV 1 genome (modified from Levy 3 ) Like most retroviruses, the gag and pol genes of HIV 1 are expressed as polyprotein chains where transcription of the pol gene occurs only from a fra meshift during gag transcription. 3 It has been found that frameshifting in several organisms tend to occur at sites with multiple U bases or stable hairpin structures that result in ribosomal pauses duri ng transcription. The frequency of the frameshifting is determined by several factors which include, but are not limited

PAGE 30

30 to, the number of consecutive U bases, the stability of the hairpin, and the presence of proteins that bind and stabilize the hairpin structure. According the numbering convention used by Wain Hobson et al. 12 there are six uracil (U) bases beginning at bp 1631. In the gag reading frame these bases are part of the codons for Asn Phe Leu. In the pol reading frame, these bases are part of the codons for Phe Phe Arg. Additionally, there is a GC rich stretch of DNA downstream of the U bases that has been iden tified as a potential hairpin. Some viruses have been shown to have frameshifting events as frequently as 1:4 ; 13 however the rate for HIV 1 has been shown to be about 1:8 14 It has been suggested that the frameshifting mechanism contributes to regulation of the pol expression (since it lacks a promoter region and thus the opportunity to regulated in a traditional fashion) because the virus should not need as many copies of the enzymes as it does the structural proteins. Interestingly, the myeloblastosis associated virus (MAV) encode s the protease in the gag gene ; 15 however, the MAV PR has been shown to be significantly less active than the HIV 1 PR (as a result of Ser in the catalytic triad instead of the Thr). It has been proposed that the myeloblastosis associated viruses have e volved to produce more PR because it is less efficient 16 Virion contents Matrix (p17) Matrix, the N terminal protein in the gag poly protein (p55), coats the inner leaflet of the lipid bilayer of the virion and i s responsible for targeting p55 to the membrane via the myristoylation at the N terminus. Capsid (p24) Capsid multimerizes to form the capsid, or core of the virion. Nucleocapsid (p7) Nucleocapsid interacts with the viral RNA genome in addition to part icipating in the inter p55 interactions during viral assembly on the plasma membrane. Reverse transcriptase (p66, p 51) Reverse transcriptase generates the DNA intermediate for integration into the host cell genome. Integrase (p32) Integrase splices t he DNA intermediate into the host cell genome.

PAGE 31

31 Protease (p10) The protease cleaves the poly peptide chains into the individual proteins during the viral maturation step. Envelope su r face (gp120) The envelope surface protein binds CD4, a host cell surface protein, the first step in the binding and fusion process. Envelope transmembrane (gp41) The envelope transmembrane protein binds gp120 attaching it to the surface of the virion and participates in the binding to coreceptors on the host cell surface re quired for membrane fusion. Tat (p14) Tat (transactivating protein) upregulates HIV replication by b inding the Tat response element Rev (p19) Rev (regulator of viral protein expressison) binds the Rev res ponse element, a cis acting RNA loop in the envelope mRNA. Rev inhibits splicing of the viral mRNA and also shuttles incompletely spliced mRNA s from the nucleus to the cytoplasm 17 Nef (p27) Nef (negative factor) is a myristoylated 27 kDa protein that interacts with a v ariety of host cell proteins to increase the viral infectivity. Nef is also believed to pl ay a role in T cell activation and down regulat ion of CD4 expression. As one of the first viral proteins expressed following infection, Nef has also been suggested to play a role in setting the tempo or determining the magnitude of the infection 18 Vif (p23) Vif (viral infectivity factor) increases the viral infectivity by antagonizing the cellular protein APOBECC3G, a cytidine deaminase that is packaged into viral particles in the absence of Vif. APOBEC3G causes G to A hypermutation and thus prevents further spread of the infection as a result of genetic corruption 18 Vpr (p15) Vpr is a 96 residue protein that plays several important roles in the HIV 1 lifecycle. Vpr causes G 2 arrest in the host cell, transactivates the LTR in the integrated provirus, induces apoptosis, and is part of the preintegration complex 18 Vpu (p16) Vpu is an 81 residue homodimeric integral membrane protein that binds CD4 in the ER (endoplasmic reticulum) thus inducing CD4 turnover. Vpu also plays a role in virus release in certain cell types 18 and in down regulation of MHC Class I cell surface proteins p6 p6, a 6 kDa cleavage produ ct from the Gag polyprotein, contains the L domain which interacts with ESCRT complexes during assemb ly to facilitate budding p2 p2 is a 2 kDa cleavage product from the Gag polyprotein. During viral ass embly, the p2 domain within the Gag polyprotein interacts with other p2 domains leading to higher order multimerization (>1000 kDa) 19 p1 p1 is a 1 kDa cleavage product from the Gag polyprotein.

PAGE 32

32 Structure and organization of the virion The s tructure of mature HIV 1 vir i ons shown in Figure 1 2 A, is similar to that of many ot her viruses. The virions are essentially spherical with an overall diameter 100 120 nm 20 The two main components of the virion are the cone shaped core, which contains and protects the genomic content of the virus, and the envelope, which is comprised of a lip id bilayer and the viral surface proteins. The shell of the core is composed of many copies of the capsid protein (CA). Inside the core, t he genomic RNA is bound by multiple copies of the nucl eic acid binding protein, nucleo ca psid ( NC) The NC proteins help maintain the integrity of the nucleic acid by binding to the RNA and sterically preventing access to the nucleic acid The viral enzymes RT, IN, and PR are also located with in the core The surface of the virus is a lipid bilayer known as the viral envelope, which was transferred from the host cell during the budding event. Embedded in the outer leaflet of the lipid bi layer are up to 72 copies of gp41. 21 Each gp41 can bind three copies of gp120 forming an Env spike that oelectron microscopy tomography. 22 The inner leaflet of the bilayer is coated with a layer of matrix (MA) proteins. The MA proteins are myristo yl ated, 23 which Figure 1 2 Pictorial representation of the structural assembly of immature and mature HIV Virions with various proteins labeled accordingly. Image courtesy of National Institute of Allergy and Infectious Dis eases.

PAGE 33

33 means they have been modified with a 14 carbon (myristoyl) fatty acid. The hydrocarbons region of the myristoyl moiety is embedded in the inner leaflet of the bilayer, which acts to anchors MA to the bilayer. The remaining viral proteins and some contained within the virion. 24 I mmature v irions dif fer from the mature virions in the organization of the interior of the particle. The immature virions do not contain a core because the gag and gag pol polyproteins have not been proteolytically processed into the individual components The myr istoyl group on the N terminus of MA, which is the N terminal protein in the gag and gag pol polyprotein s is imbedded in the bilayer The gag and gag pol polyproteins extend towards the interior of the virion. The polyprotein self associate to coat the in ner leaflet of the bilayer. The RNA genome is bound to the NC region of the gag and gag pol polyproteins. The gp120 and gp41 are the same as in mature virions because they are processed by intracellular prot eases before the virion buds Role of the HIV 1 Protease in the Viral Life Cycle Figure 1 3 illustrates the steps in the HIV 1 viral lifecycle. The lifecycle begins with a binding even t between a mature (infectious) virus particle and a host cell. The gp120 protein on the virion surface binds to a CD 4 protein on the host cell surface ( step 1) Upon recruiting ( step 2) an additional host cell receptor protein (usually CCR5 or CXCR4), the v iral membrane begins to fuse ( step 3) with the host cell membrane eventually releasing the contents of the virion into the host cell. After cell entry, the viral capsid, which contains the viral RNA genome, is uncoated (step 4). The genomic RNA is reverse transcribed by RT (step 5) into double stranded DNA, which is then imported (step 6) into the nucleus. The doub le stranded DNA is then integrated (step 7) into the host cell genome. The integrated DNA remains dormant until the region is activated for transcription. The host cell transcribes (step 8) the mRNA, which is used

PAGE 34

34 Figure 1 3 HIV viral life cycle. Fi gure courtesy of National Institute of Allergy and Infectious Diseases. Numbered steps are described in detail in the text. in the production ( step 9) of the viral poly proteins. The viral proteins are trafficked ( step 10) to the host cell membrane where they begin to bud ( step 11) and form new viral particles, which are immature and thus non infectious. This immaturity refers to the state of the viral proteins, which are still polypeptide chains consisting of multiple proteins. The viral protease cleav es itself from the polypeptide chain by an unknown process and then cleaves the remaining polypeptide chain into the individual proteins. These proteins are then rearranged within the viral particle. The cleavage and reorganization is called the maturati on step ( step 12) and is the final step in the viral life cycle.

PAGE 35

35 Structure of HIV 1 Protease Although, we know now that HIV pr otease is a dimeric aspartic protease, the following provides a brief history of the identification and characterization of HIV 1 PR. I n 1984 viral genome was sequenced 12; 25 and many viral proteins were identified by homology to other retroviruses. It was determined that a protease was likely pol gene as in several other retroviruses 26 Similar to most retroviral proteases, HIV 1 PR was homologous to several cellular aspartic proteases such as pepsin 27 S equence analysis also revealed that HIV 1 PR was most likely a dimer because the sequence only included one of the two necessary catalytic triads (Asp Thr Gly) 27 In 1986 and 1987, several researchers expressed a portion of the gag pol polyprotein including th e putative HIV 1 PR in E. coli and demonstrated that the sequence contained the protein responsible for proteolytic processing of the gag pol polyprotein 28; 29 It was unclear from the sequence analysis alone, whi ch residues comprised the active protease because the protease cleaves itself from the polyprotein chain at two then unknown cleavage sites. In 1988, two groups synthesize d and characterize d a 99 amino acid sequence corresponding to the proposed minimum s equence necessary for HIV 1 PR. 30; 31 Both groups demonstrated that a pro tein containing just these 99 a mino a cids had enzymatic function for the proteolytic ally processing the gag pol polyprotein. In order to de termine the exact sequence of the HIV 1 PR, several researchers purified and characterize d HIV 1 PR from partial gag pol polyproteins expressed in E. coli cells 28; 32 They also discovered that a 10 kDa protein co rresponding to the HIV 1 PR sequence was sufficient for proteolytically processing the gag pol polyprotein. B y substituting the Asp25 for either Asn, Ala, or Thr residues and monitoring gag pol polyprotein processing 33; 34; 35 several other researchers established Asp25 of HIV 1 PR was necessary for activity. Furthermore, it was demonstrated in an in vitro assay of recombinantly

PAGE 36

36 expressed gag pol that pepstatin A a naturally occurring inhibitor to pepsin, could p revent gag pol processing 33 In 1989, the investigation into potential inhibitors for HIV 1 PR continued vigorously. Many potential transition state mimetics were identified that could inhibit the HIV 1 PR in vitro 36; 37; 38 furthermore, it was demonstrated that some inhibitors could prevent HIV replication in cell culture 39 while other s could not 40 In 1989, t he first crystal structures of HIV 1 PR were published (PDB ID 2hvp) 41; 42; 43 In 1996, the first structure from data of NMR experiments was determined 44 HIV 1 PR was studied in the presence of a symmetric cyclic urea based inhibitor. The results were compared to the structure obtained from crystallographic studies of HIV 1 PR bound to the same i nhibitor T he two structure s were similar but significant differences were observed in the loops and the two structure s Specifically the NMR data revealed more flexibility of the flaps than can be inferred from static x ray models. T he structure determination of HIV 1 PR confirmed the sequence homology of HIV 1 PR to other viral and cellular aspartic proteases. The cellular proteases are monomeric ; however, they contain two highly symmetric domains that are analogous to t he monomers of viral proteases. 45 Thus the structural features of each domain in the cellular proteases and each monomer from the viral proteases are similar and can be described by a general template. 45 Th is general template for aspartic proteases is illustrated in Figure 1 4 A and these features are mapped onto the HIV structure in Figure 1 4C The protease monomers have a pseudo twofold symmetry axis. The structural components are named alphabetically with the pseudo symmetry partners designated T he N terminal (residues 1 5, sheet A) of HIV 1 PR com prise the outer edges of the four sheet dimerization domain. Residues 9 15 form the strand B which,

PAGE 37

37 along with sheet C ( residues 18 24) form the fulcrum. The active site triad (Asp25, Thr26, and Gly27) sits on a loop between strand C and strand D ( residues 30 35). In the cellular proteases, the D strand would be followed by a helix H. However, in the smaller viral proteas es, like HIV 1 PR the D strand is followed by short disordered loop H ( residues 36 42) called the strand ( re sidues 43 strand ( residues 52 66) form the Figure 1 4 HIV 1 PR topology. A) Schematic diagram of secondary structural elements labeled according the aspartic protease template (modified from Wlodawer et al. 45 ) B) Top view of HIV 1 PR ribbon diagram color coded t o match the aspartic protease template (PDB ID 3HVP). C) Front view of HIV 1 PR ribbon diagram. strand (residues 69 78) form the cantilever. strand (residues 83 85) forms part of the active site wall and leads into helix ( residues 86 94). The C terminal is the Q strand (residues 95 99) which sits between the A strands to complete the dimer interface. The HIV 1 PR dimer is stabilize d by inter monomer interactions, with t he majority occu rring in the dimerization domain where the four terminal sheet s from each monomer intercalate, form ing 12 hydrogen bonds. A dditional inter monomer interactions in the active site further stabilize the dimer structure. Thr26 participate s in what is know (Figure

PAGE 38

38 1 5 ) where hydrogen bonds are formed between a side chain hydroxyl and both the backbone carbonyl of Leu24 and the backbone amide of the Thr26 from the other monomer. Investigation has confirmed the importance of Thr26 by substitution of other amino acids that are not capable forming the hydrogen bonding network resulting in a shift of the monomer dimer equilibrium in favor of the monomer 46 Figure 1 5 HIV 1 PR active site. Residues from The active site of HIV 1 PR has a geometry that is highly conserved among both viral and cellular aspartic proteases. The catalytic triad, Asp25, Thr26, Gly27, form the same interactions stabilized by additional hydrogen bonds. The backbone carbonyl of Asp25 hydrogen bond s with the backbone amide of Gly27. The Asp25 residues are typically depicted with one residue being protonated and the other being deprotonated with the proton being shared between the two closest oxygen atoms. A water molecule further stabilizes this i nteraction by hydrogen bonding with an oxygen from each Asp residue 47 The end result of this hydrogen bonding network is a

PAGE 39

39 stable, geometrically conserved active site, with the catalytic Asp in the necessary position for prot eolytic activity. HIV 1 PR contains two hairpin turns that sit over the active sit e In the inhibitor bound crystal structures, the flaps make extensive contact with the inhibitor. On average, the flaps contribute roughly half of the interactio ns between the inhibitor and HIV 1 PR. 48 When bound to an inhibitor or subs trate, the flaps are pulled in toward the active site by the interaction with the ligand. There are also interflap interactions such as the hydrogen bonds typically observed between the backbone atoms of residue 51. This conformation is typically referre d to as the closed conformation. In the absence of ligand, the flaps sit slightly further away from the active site in what is termed the semi open conformation For the first few years after the initial crystal structures were solved, researchers debat ed which flap conformation was the most stable in solution and which, if any, were stabilized by crystal contacts. The HIV 1 PR structures were compared to apo PR structures from other viruses, some of which were closed (SIV) or semi open (HIV 2) ; however some also contained disordered flaps that were unresolved (Rous Sarcoma Virus) Researchers eventually concluded that closed structures of apo proteases were stabilized by crystal contacts and that the semi open form is the most thermodynamically stable in solution 45 The other major questio n addressed by researchers was the issue of substrate access to the active site. The space filling model (Figure 1 6 B) and the top view (Figure 1 6 C) clearly show that with the flaps closed, there is insufficient sp ace for the substrate to enter from eith er the top or side of the protein. Researchers were quick to identify that in order for substrates or inhibitors to bind, it is necessary for the protein to undergo a large conformational change to move the flaps out of the way. Figure 1 7 shows the semi open and wide open conformations of HIV 1 PR with

PAGE 40

40 an inhibitor placed in the active. This figure illustrates that the interflap distance in the semi open conformation not sufficiently large to allow access to the active site and that a wide open conforma tion is necessary. Several experimental techniques as well as computational approaches were preformed to address this issue This topic will be discussed in greater detail later in this dissertation of HIV 1 PR Figure 1 6 Structure of HIV 1PR PDB ID 2bpx. A) Ribbon diagram highlighting the various structural regions of HIV 1PR. B) Space filling model of HIV 1PR illustrating the size of the active site cavity and the inability of a polypeptide substrate to s lide through the active site with flaps closed. C) Top view of the ribbon diagram illustrating the relative horizontal position of the flaps and how the flaps block access to the active site. Based on the results of several molecular dynamic simulations (MD) and by crystal structure analyses, the main structural components of HIV 1 PR are divided based on rigid domain movements during flap opening. The active site floor and the active site wall undergo a small motion during flap movement where symmetry partners move closer together upon flap closing and slightly farther away during flap opening. The elbows, cantilever, and fulcrum (Figure 1 6) are also involved in opening the flaps. As the flaps open, the elbows and cantilever shift down (away from the flaps). The fulcrum pivots slightly around its inner most point (point closest to

PAGE 41

41 the active site) to accommodate the shift in the cantilever. These motions allow significant movement in the outer most edges of the protease (flaps, cantilever, and fulcr um) while the core of the HIV 1 PR remains relatively immobile. 49; 50 Figure 1 7 X ray crystal structure of the semi open and the MD structure of the wide open conformation with a space filling model of an inhi bitor placed in the active site pocket to illustrate the relative sizes of the inhibitor and the gap between the flap tips Additional studies have revealed that the sequence of the flaps is highly conserved in patients that have never been exposed to pro tease inhibitors. However, there is an increase in the number of amino acid substitutions in the flap sequence found in patients that have been treated with protease inhibitor. T h i s correlation between protease inhibitor therapy and mutations in flaps in dicates that there is a strong possibility for a connection between mutations in the flaps and drug resistance. HIV 1 Protease as a Drug Target S everal viable strategies are currently utilized for drug therapy in H IV positive and AI DS patients; most target viral proteins essential in the viral life cycle. 51 For HIV 1, this includes the proteins involved in recognizing and binding to the host cell, the proteins involved in fusing the

PAGE 42

42 viral and host membranes, RT IN and PR all of which are targets for FDA approved antiretroviral drugs that are currently being used to treat HIV positive patients. Part of the success of targeting these proteins relies on the fact that the small molecules are sp ecific to viral proteins which d o not have homologous proteins in the host cell. Unfortunately, there are usually host cell proteins with some degree of homology, whic h leads to some inhibition of normal function within the host cell and can cause side effects in the patients 52 An e xample of side effects resulting from inhibition of cellular enzymes is the inhibition of DNA polymerase by nucleotide inhibitors fo r reverse transcriptase (RT). The nucleoside reverse transcriptase inhibitors (NRTI) are ef fective in inhibiting RT but typically have no noticeable effect on most of the cellular polymerases because of difference s in the selectivity of the polymerases for the nucleotides. However, DNA polymerase the only polymerase for mitochondrial DNA (mt DNA), is strong ly inhibited by NRTIs resulting in a depletion of mtDNA 53 Consequently, the expression of proteins encoded in the mtDNA, such as those involved in oxidative phosphorylation, is decreased, which results in clinical symptoms associated wi th genetic disorder of the mtDNA. Symptoms include loss of muscle coordination, muscle weakness, heart disease, liver disease, kidney disease, gastrointestinal disorders, among others. Protease Inhibitors In 1988, the search for protease inhibitors began when it was shown that preventing HIV 1 PR activity via a D25N mutation inhibited the viral lifecycle. 35 Within the next year, Deinhardt et al. 39 showed that viral replication could be halted by the addition of an aspartic acid specific inhibitor, pepstatin A (Figure 1 8 B) and Richards et al. 36 demonstrated that HIV 1 PR could be inhibited in vitro by several genera l aspartic protease inhibitors Wlodawer et al. 43 showed that a

PAGE 43

43 non hydrolysable substrate peptide could bind tightly to the HIV 1 PR active site and inhibit the HIV 1 PR act ivity. The first investigations into inhibitors designed specifically for HIV 1 PR began in 19 89 and w ere based upon the knowledge gained from studies o n other aspartic protease s like rennin. 40 These early inhibitors were based upon Pepstatin A a native inhibitor of pepsin, which contains a n uncommon amino acid statine (Figure 1 8 A). Statine (Sta) has the unique structure of a tetrahedral intermediate or transition state analogue. Thus many peptide mimetics containing a Figure 1 8 Inhibitor structures and scaffolds. A) Structure of the non standard amino acid statine. B) Structure of pepstatin containing two Sta amino acids. C) Structure of the hydroxyethylene scaffold. D) Structure of the hydroxyethylamine scaffold. E) Structure of the reduced amide scaffold. F) Structure of saquinavir (based on the hydroxyethylamine scaffold. G) Structure of Tipranavir, a non peptidomimetic inhibitor.

PAGE 44

44 variety of tetrahedral intermediates substituted for the scissile amide bond 36; 38; 40 were tested for HIV 1 PR inhibition. It was also found that compounds with a specific tetrahedral intermediate, a hydroxyethylene isostere ( [CH(OH)CH 2 ]) (Figure 1 8), inhibited HIV 1 PR more effectively than other s 54 Other transi tion state analogues such as the reduced amide (Figure 1 8 E), phosphinat es, and fluoroke tones were tested but few inhibited HIV 1 PR as efficiently as the hydroxyethylene scaffold 38 As with all drug design and discovery attempts, transitioning the small molecules which successfully inhibited HIV 1 PR in vitro to viable drugs was complicated by solubility, bioavailability, stabil ity, and side effect issues. The drug design was additionally stymied by the fact that in vitro inhibition did not necessarily result in inhibition of the viral replication in cell cult ures 40 The first FDA approved protease inhibitor for HIV in 1995 was saquinavir (SQV) (Figure 1 8 F) 55 followed shortly thereafter by ritona vir (RTV) and indin a vir (IDV) There are currently 10 FDA approved protease inhibitors ( structures are shown in A ppendix A ( Figure A 1 ) properties listed in Table 1 1). SQ V and most of the FDA approved protease inhibitors are peptidomimetic inhibitor s meaning that the structure s mimic th at of a peptide. The structure of the transition state analogue varies slightly between inhibi tors but they are al l based on either the hydroxyeth ylamine (HEA) (Figure 1 8 D) or the hydroxyethy lene (HEE) scaffold (Figure 1 8 C) Tiprana vir (Figure 1 8 G) the onl y non peptidomimetic inhibitor approved by the FDA was discovered from refinement of the lead compound phenprocoumon 56 and is th us considered to have a coumarin scaffold. Peptidomimetic inhibitors are not the only option for inhibiting HIV 1 PR. Technically, any molecule that can prevent the protease from functioning is an inhibitor. This means that

PAGE 45

45 An inhibitor which binds un der the elbow regions could stabilize the closed conformation and prevent flap molecules which prevent dimer formation could be used as inhibitors, since the monomeric protein is essentially inactive. 57 It has been hypothesized that small molecules designed to bindopen ing, thus preventing substrate binding to the active site. 58 Conversely, large inorganic compounds have been co crystallized with the protease and hold the flaps in a wide open conformation, thus preventing substr ate binding and substrate flap interactions. 59 Table 1 1. FDA approved protease inhibitors for treatment of HIV 1 Inhibitor Abbreviation Year Approved Scaffold Type K d (pM) a,b K I (nM) b,c # non water mediated hydr ogen bond d,e,f,g Amprenavir APV 1999 HEA 220(27) 0.17 c 5 d Tipranavir TPV 2005 Coumarin 19 b 0.019 b 6 g Indinavir IDV 1996 HEE 590(93) a 3.9 c 3 d Saquinavir SQV 1995 HEA 280(22) a 1.3 c 7 d Lopinavir LPV 2000 HEE 36(7) a 0.05 c 3 f Fosampre navir FPV 2003 HEE (modified) NA h NA h NA h Ritonavir RTV 1996 HEE 100(11) a 0.7 c 7 e Darunavir DRV 2006 HEA 10 b 0.010 b 6 d Atazanavir ATV 2003 HEA NA h 0.48 c 3 b Nelfinavir NFV 1997 HEA 670(110) a 1.2 c 2 d a ) Data from Clemente et al. 60 d) Data from Muzammil et al. 61 c ) Data from Yanchunas et al 62 d) Data from Prabu Jeyabalan et al. 63 e) Data from Prabu Jeyabalan et al. 64 f) Data from Reddy et al. 65 g) Data from Nalam et al. 66 h) Data not available Characterization of Inhibitor Binding The FDA approved p rotease inhibitors are competitive inhibitors, which means they bind HIV 1 PR stronger than the substrate, the substrate binding and inhibitor binding are mutual ly exclusive, and that the binding is reversible. Competitive inhibition follows the reaction scheme in Figure 1 9 where E, S, I, and P are the enzyme, substrate, inhibitor, and product respectively ES and EI are the enzyme substrate and enzyme inhibito r complexes respectively.

PAGE 46

46 Figure 1 9 Reaction scheme for competitive inhibition. The binding affinity of an inhibitor is specified in terms of the association constant, K a (Equation 1 1), which is in turn determined by the standard Gibbs free energ y, G (Equation 1 2) where R is the molar gas constant, T is the absolute temperature, H is the standard (1 1) (1 2) change in enthalpy, and S is the standard change in entropy. As competitive inhibitors, they must bind with very high affinity to preven t the substrate from binding. For binding to be thermodynamically favorable, G must be negative with larger absolute magnitudes corresponding to tighter binding. M ost of the first generation of inhibitors (IDV, SQV, and NFV) have small positive H val ues and large positive S values to achieve the large negative G values needed for effective competitive inhibition 67 R TV, APV, LPV, and ATV have small negative H values and moderate positive S values that combine to give large negative G values. The most recent inhibitors DRV and TPV, have large negative H values with moderate to small S values that combine to give

PAGE 47

47 large negative G values. At first glance, one might think these differences are trivial bec ause regardless of the size and sign of the ent r opic and enthalpic contributions, the free energies are all large and negative. However, these variations are significant because they indicate a considerable di versity in how the inhibitors bind and determi ne what conditions affect the binding. HIV 1 PR Subtype Polymorphisms and Rates of Mutation Like most viruses, HIV 1 has a high rate of mutation which gives rise to genetic variability The viral reverse transcriptase lacks the proof reading capability o f most eukaryotic polymerases resulting in a significantly higher rate of mutation relative to eukaryotic cells The mutation rate of human DNA polymerase is estimated to be roughly 5x10 11 mutations/bp/replication 68 whereas the mutation rate for HIV 1 is 3.4 x 10 5 mutations/bp/ replica tion 69 Since the viral genome contains about 10 kb, there is approximately one mutation for every three replication cycles. Th e genetic variability of HIV is manifested in the large n umber of differing viral genomic sequences that ap pear all over the world. The sequences have been divided into groups and subtypes as illustrated in Figure 1 10 based on their similarities and differences The sequences in e ach group (M Major, O, N) dif fer by at leas t 25% in the gag and env genes. Each group is subdivided into subtypes (A L). Each subtype refers to sequences from a common ancestor that differ by more than 20% in the env gene and 15% in the gag gene. 3 The variations in any given genomic sequence are typically characterized by comparing it to the consensus sequence for appropriate subtype. The most studied subtype is B, which is most prevalent in North America and Western Europe. The c onsensus sequence for subtype B is from the LAI isolate, 3 a sample taken from a

PAGE 48

48 Figure 1 10 Phylogenic tree of HIV 1 (reproduced from Peeters et al. 70 ) The scale bar represents 10% divergence in sequence from the most recent common ancestor. Group N and O are more closely related to SIV strains than to HIV M strains. patient designated LAI, and is thus referred to as the LAI sequence. It is listed as B.FR.83.HXB2 in the Los Alamos HIV database. Worldwide, subtype C is responsible for the highest number and is predominantly found in sub Saharan Africa. It is important to note that the natural sequence variation (not drug pressure selected) in a subgroup is not random and is found in specific regions of the HIV 1 PR. The functionally or structurally important regions of HIV 1 PR have lit tle or no significant variation A comparison of HIV 1 PR protein sequences (Stanford HIV Datab a se http://hiv db.stanford.edu ) from both untreated (protease inhibitor nave) and treated (protease inhibitor therapy) patients (Figure 1 1 1 )

PAGE 49

49 Figure 1 1 1 Protein sequence variation among protease inhibitor nave patients and patients who have undergone protease inh ibitor therapy. The data are averaged values of subtypes A, B, C, D, F, G, AE, and AG. The % variation is relative to the consensus sequence for each subtype. The residue number corresponds to the position in the amino acid sequence of HIV 1 PR. reveal s that there are multiple conserved regions for the protease in the untreated patients Several of these regions are no longer conserved in protease inhibitor treated patients. As illustrated in Figure 1 1 1 the conserved regions correspond to the dimer ization domain, the flaps, the active site, and the active s ite wall. These domains are functionally important for the protease. The dimerization domain is essential because the protease monomer is effectively inactive. Likewise, the flaps are nec essary in binding the substrate and holding the substrate during cat alysis in addition to regulating substrate access to the active site. The active site walls are also involved in binding and recognizing the substrate and are necessary for enzymatic activity and specificity Drug Pressure Selected Mutations The high mutation ra te of the HIV genome provides a mechanism of evolving for adapt ion to changing environments which results in immune system evasion and drug resistance. When the

PAGE 50

50 virus attempts to rep licate in the presence of an inhibitor, viruses containing random mutations that allow the virus to propagate will survive The viral fitness of th ese variant s will be honed over many generations to produce viable drug resistant virus es For HIV 1 PR the drug pressure selected mutations can be divided into two categories: active site and non active site mutations. The ability of the active site mutations to confer drug resistance is easily explained because these mutations can reduce inhibitor binding by changing the shape of the active site pocket. The ability of the non active site mutations to confer resistance is not as straightforward. The currently accepted explanation is that the non active site mutations affect the dy namics and/ or flexibility of HIV 1 PR either restricting HIV 1 PR from reaching the wide open conformation needed t o bind the inhibitor or preventing the flaps from closing before the inhibitor can optimize its geometry for high affinity binding. 71 Thus understanding the role of the flaps in inhibitor binding can potentially result in the design of better drugs that are less susceptible to resistance. Understanding the Flaps of HIV 1 PR As an important factor in both drug design and drug resistance, many methods (including NMR, x ray crystallography, molecular dynamics (MD) simulations, fluorescence, and others ) have been used to study the structure and flexibility of the HIV 1 PR flaps Together, these techn iques have elucidated two keys features of the flaps: the conformations and the flexibility of the flaps the ability of the flaps to transition between conformations. The semi open form has been shown to be the most stable conformation in apo HIV 1 PR by N MR, x ray crystallography, and MD studies. NMR relaxation measurements and NMR nuclear Overhauser effect (NOE) spectroscopy have show n that the flaps are relatively flexible in the absence of substrate or inhibitors. The hairpin turn structure (Figure 1 4) is maintain ed 72 but

PAGE 51

51 the flaps undergo large amplitude low frequency motion on the s to ms timescale indicative of conformational exchange Additionally, the lifetime of these states is approximately 100 s. The flap tips, residues 48 52, are even more flexible, undergoing rapid conformational exchange 73 As result of this flexibility, the flap conformations have been described as an ensemble of semi open conformations that can transiently sample closed and wide open conformations. The binding of substrate or inhibitor stabilizes the closed conformations because the flaps also interact with the substrate or inhibitor and move closer to the active site. These interactions also result in the flaps becoming rigid 72 and only experiencing slight motion in residues 50 and 51 73 the relative position of one flap in front of the other changes upon binding substrate or inhibitor as illustrated in Fi gure 1 1 2 Figure 1 1 2 Structures of two x ray crystal structures of HIV of the flaps in the semi open and closed conformations. A) Top view of the closed conformation (PDB ID 2BPX). B) Front view of the closed conf ormation. C) Top view of the semi open conformation (PDB ID1HHP). D) Front view of the semi open conformation. In addition to the closed and semi open conformations, MD simulations have also revealed two other possible flap conformations: wide open and curled (or tucked). Although the wide open form has been known to exist for years, the MD simulations were the first to provide a

PAGE 52

52 structure. 74 An x ray structure of a drug resistant clinical isolate, MDR769, 75 also shows an open conformation but MD simulations 76 have de monstrated that this conformation is an artifact of crystal contacts. However, it is possible that this structure contributes to a small fraction of the conformational ensemble of HIV 1 PR The x ray structure of MDR769 is shown in Figure 1 1 3 along with 3 other HIV 1 PR x ray structures illustrating the conformations seen by x ray crystallography. The closed form was Figure 1 1 3 Four x ray crystal structures illustrating the major conformations of HIV 1 PR. The closed conformation (PDB ID 2BPX) is shown in green and the semi open crystallized from a drug resistant variant. (It should be noted that the flap conformations are stabilized by a crystal crystal contact). The blue structure (1ZTZ) was crystallized in the presence of a bulky inorganic inhibitor that restricted the flaps from closing. crystallized in the presence of an inhibitor and is the conformation in which the flaps sit the closest to the active site. The semi open form was crystallized without substrate or inhibitor.

PAGE 53

53 The flaps in this state sit slightly farther away from active site but still interact with each other. con formations (PDB ID 1ZTZ) was crystallized in the presence of bulky inorganic molecules in addition to an inhibitor. These molecules stabilized a horizontal opening of the flaps. The other open form is MDR769 and was crystallized in the absence of substra te or inhibitor. This structure also displays more of a horizontal opening (as opposed to the more vertical opening events seen in MD simulations). The lengths of all atom MD simulations are typically restricted to less than 100 ns because of high computa tional demands. However, most large conformational changes occur on the s to ms timescale. To capture these events in a computationally feasible timescale, many computations utilize activated or restricted MD simulations 58; 74 which typically involve the use of harmonic force constants on specific atoms to initiate the conformational change. Although these simulations can provide accurate models, the forces used to activate or restrict the simulation can obscur e relevant information about the protein such as correlations between domains or the relative populations of each state. Thus, the MD simulations of Simmerling et al. 50; 77 are particularly important because their use of a low viscosity implicit solvent allowed them capture apo HIV 1 PR transitioning between the semi open closed, and wide open conformations and to extract valuable information about the correlations of various domains throughout these events. Figur e 1 1 4 shows the front and top views of the closed and semi open conformations used as starting points for the simulations, the wide open conformation from their simulation, and an overlay of these structures to highlight the changes in various domains. I n their simulation, the opening of the flaps correlated with a downward shift in the cantilever and fulcrum. At the peak

PAGE 54

54 of the opening event, the distance between the tips was greater than 30 An additional MD simulation 77 reveals that this open structure will close around an inhibitor placed in the active site to form the closed conformation. Two of the significant achievements of this work are that HIV 1 PR returned to the semi open form after reaching the wide open state and transiently sampled the closed conformation in addition to the wide open state. These results corroborate the work of Ishima et al 73 which determined that in the absence of substra te or inhibitors, the flap conformations are predominantly semi open but can also sample closed and wide open forms. Figure 1 1 4 Structures of three predominant HIV 1 protease conformations. A) Closed conformation (PDB ID 1HVR ). B) Semi open conforma tion (PDB ID 1HHP) C) Wide o pen conformation. Figure modified from Simmerling et al. 50 Several MD simulations have also shown that the flaps adopt a curled or tucked conformation These conformations include a broad range of flap conformations as illustrated in Figure s 1 1 5 and 1 1 6 In the more open curled states, 78 the flap tips curl back towards the active site and the monomers undergo a domain rotation that opens up the active site pocket. In these

PAGE 55

55 Figure 1 1 5 A) MD structure from a simulation of HIV 1 protease illust rating the curl ed or tucked conformation. B) Illustration of the interactions between Ile 50 and residues 79 81 and 32 Figure reprinted with permission from Scott et al. 79 structures, Ile50 typically interacts with residues 79 8 1 and 32 as shown in Figure 1 1 5 B. Other curled structures are more closed, 79 with the flaps curling into the active site pocket but without the domain rotation that opens the pocket as illustrated in Figure 1 1 6 Ishima et al. 80 recently demonstrated via NMR relaxation measurements and chemical shift analysis that the Figure 1 1 6 MD structures from a simulation of HIV 1 protease illustrating the curled or tucked conformations. A) X ray structure (PDB ID 1HHP), the starting structure of the simulation. B) Structure at 2500 ps. C) Structure at 5000 ps. Figure reprinted with permission from Toth et al. J. Mol. Graph. Model. 2006.

PAGE 56

56 predominant conformation in the absence of substrate or inhibitors involves interactions between the flap tips. Thus, the curled or tucked conformations may constitute a small fraction of the confor mational ensemble but are not the dominant conformation. Recently our group utilized site directed spin labeling (SDSL) in conjuction with double electron electron resonance (DEER) to characterize the flap conformations. 81 These experiments measured the interflap distance between spin labels grafted at position 55 in the flaps. Figur e 1 1 7 shows the dipolar evolution curves and corresponding distance profiles for HIV 1 PR in the absence and presence of ritonavir (RTV). In the presence of RTV an inhibitor of HIV 1 PR the most probable distance was 32.6 corresponding to the distan ce predicted for the closed conformation. The breadth of the distribution was 3.0 which corresponds to the motion of the spin label about the flexible linker that attaches the label to the protein. In the absence of the inhibitor, the most probable dis tance was 35.5 which corresponds to the distance expected for the semi open conformation. The breadth of the profile was 10 and corresponds to the both the motion of the label and the motion of the flaps. Most significantly, this distance profile in cludes distances greater than 39 which likely correspond to a wide conformation. Figure 1 1 7 A) DEER dipolar evolution curves for apo and inhibitor bound HIV 1 PR. B) Corresponding distance profiles. 81

PAGE 57

57 These results were corroborated by additional simulation from Simmerling et al. 82 Figure 1 1 8 shows the overlay of the MD results with the DEER distance profiles in the absence and presence of RTV. Their results confirmed that the breadth of the profile for inhibited HIV 1 PR does correspond to the motion of the label about the flexible linker. Likewise, the breadth of the profile for apo HIV 1 PR results from a significant increase in the motion in the flaps. Their results also demonstrated that the larger distances (>39 ) can only be achieved in a wide open form. Figure 1 1 8 Inter spin label distances from DEER e xperiments and MD simulations. A) Inhibitor bound HIV 1 PR. B) Apo HIV 1 PR. MD simulation initiated from closed conformation. C) Apo HIV 1 PR. MD simulation initiated from semi open conformation. Figure reprinted with permission from Ref. 82 These results have collectively provided significant insight into the conformations and flexibility of HIV 1 PR However, they are mostly specific to subtype B. A few MD simulations 58; 74; 76; 83 have investigate d the effect of mutations on the flaps. The conclusion drawn from these simulations are that the various mutations alter the stability of each conformation in such a way that inhibitor binding becomes less favorable than with wild type HIV 1 PR In the c ase of the V82A I84V double mutant, the flaps preferred a more open conformation. Thus, inhibitor binding would include a larger entropic penalty relative to wild type. 58 In the case of M46I, 74 the flaps we re shown to prefer the closed conformation in the

PAGE 58

58 absence of an inhibitor. A large number of x ray structures exist for HIV 1 PR constructs including various single point mutations and combinations of mutations. These structures show that the largest eff ects are seen in the shape and size of the active site binding pocket and that the rest of the protein is relatively consistent. This trend implies that a significant portion of the drug resistance arises from changes in the flap flexibility and not in ju st the closed conformation. In order to gain a better understanding of the mechanisms of drug resistance, more investigations need to be performed to ascertain the effects of each mutation and various combinations of mutations. Protein Structure and Flexi bility Protein Structure Proteins cons ist of a polypeptide chain, many of which fold into a specific tertiary structure. The structure of proteins has been divided into four categories: primary, secondary, tertiary, and quaternary structure. The primary structure of the protein the sequence of amino acids in the polypeptide is determined by the DNA encoding for the protein. The secondary structur e of the protein is determined by the preferred rotomeric conformation s of the amino acid side chains about th e N C bond and the C C bond (as defined by the and angles respectively) The secondary structure is usually classified as being a heli x a sheet, a loop or disordered The helices can adopt several different turn radii which correspond to the helix, the 3 10 helix, and helix, although the helix is by far the most common. The tertiary structure is determined by the interactions between the various secondary structure elements in the protein the packing of helices or the stacking of sheets The quaternary structure if present is the interaction between multiple proteins in the formation of a macromolecular complex. References to a

PAGE 59

59 that the protein assumes in solution The secondary structural elements, namely the helix and the sheet, are stabilized by a network of hydrogen bonds between the atoms in the peptide backbone. Certain amino acids seem to have a preference for being structured and some unstructured Prolin e for example, is very rarely found in helices and is essentially limited to the ends of the helix which are more tolerant of non helical and angles. A data mining study of the structures in the PDB had revealed that disordered regions in proteins are typically enriched in charged residues ( l ys ine (Lys) a rg inine (Arg) a sp artate (Asp) g lu tamate (Glu) ) in addition to p ro line (Pro) and s er ine (Ser) The disordered regions are also typically lacking in hydrophobic residues such as tryptophan (Trp) phe nylalanine (Phe) iso le ucine (Ile), and Tyrosine 84 Additionally, certain amino acids seem to have a preference for the type of secondary structure in which they occur This preference results from balancing the optimization of the peptide backbone hydrogen bonding network and the packing of the bulky side chains together and is not as pronounced as the preference between order and disorder. This structural preference was initially investigated in 85; 86; 87 but has seen a resurgence in popularity as a result of efforts to engineer proteins that will fold into a desired structure 88 The known structures of 15 proteins were analyzed to determine the frequency of each amino acid occ urring in each of the major secondary structural elements. The total number of times an amino acid appeared in each element was divided by the total number of occurrences in all elements to give a normalized frequency ( f ). The frequency was divided by th e relative percent that each element occurred (36 % helix, 17.1 % sheet, and 46.9 % coil) to yield the conformational parameter (P). Figure 1 1 9 plots the conformational parameters from Chou et

PAGE 60

60 al. 85 for each amino acid. An amino acid such as Asp with roughly equivalent P values for each element would be said to have said to have no preference for a particular structural element. Whereas an amino acid such as Glu clearly has a preference for helical structures and Ile has a moderate preference for sheets. It should clarified that th ese numbers should not be extrapolated to large disordered regions in proteins as these numbers were determined from and thus apply to structur ally well defined regions and short loops Figure 1 1 9 Conformational parameters from Chou et al. illustrating the preference of certain amino acids for a particular secondary structural element. The tertiary structure of a protein is not as straigh tforward to determine as the secondary structure and has been the subject of numerous studie s. It has been determined that the tertiary structure is largely created from and stabilized by hydrophobic collapse. Essentially, the hydrophobic portions of t he protein self associate to reduce the number of ordered water molecules necessary to solvate these regions. This step results in a collapsed structure that lacks order called a molten globule 89 After the molte n globule is formed, additional interactions act to stabil ize the native state including electrostatic interactions and hydrogen bonds between polar groups on the protein s surface.

PAGE 61

61 It has long been understood that there is an important relationship betwe en the structure of a protein and its function. In 1893, Emil Fisher demonstrated that enzymes could selectively the ideas the structure of the enzyme and the substrate should fit together like a lock and key 90 In the late that proteins maintain a unique structure which is lost upon denaturation 91; 92 The first protein structures were determined in 1958 using x ray crystallography by Max Peru tz 93 and Sir John Cowdery Kendrew 94 who shared a Nobe l prize for their work. There are currently over 34,500 protein structure s deposited in the protein database (PDB) The formation of the protein structure function paradigm makes sense in the context of the available technology to the early protein scien tists. The first proteins studied were either the structural proteins which were typically sufficiently abundant to be easily purified, or enzymes, which could be easily identified and tracked through a purification process by their enzymatic function a ll with well defined structures which are critical to their functions. Protein Motion and Flexibility Although structures provide invaluable atomistic information about the protein and its function, it does not provide all of the information necessary to u nderstand the protein dynamics and how the protein converts from one structure to other. As Richard Feynman astutely recognized everything that living things can do can be 95 Stat ed another way by Dorothee Kern 96 Thus t he function of a protein although largely dependent upon its structure is also dependent upon are necessary for accomplishing their function O ther s rely upon changes in their vibrational

PAGE 62

62 freedom to bind ligands Without movement and flexibility, the se protein s would cease to function. Proteins undergo a large vari ety of motions that each occur within a certain timescale. The smallest motions a re the bond vibrations that occur on a ps f s time scale, f ollowed by bond rotations which occur on the ps s timescales ( depending on the size of the moieties at each end of the bond ) These motions result from the thermal energy in the system. There a re also small conformational changes with in the protein such as loop movement s or shifting the position of a helix which occur in the ns s timescale. The largest motions are the domain movements, where an e ntire domain in the protein shift s its location relative to the rest of the protein; these typically occur on the s m s timescale. Recent developments in computations and simulations have shown that motions associated with the larg er conformational changes (i.e. loop movements and larger ) frequently co rrespond to the coll ective low energy normal modes of the bond vibrations 97 Each motion in a protein carries with it a certain energy barrier that must be overcome. For the fast motions in the ns or faster timescales, the energy barriers are typically less the 1 kT (equivalent to 1 RT for a mole of the protein) where k is the Boltzman constant (1.380 650410 J K 1 ) and T is the absolute temperature (typically 298 K) Slow motions on the s ms timescale typically have energy barr iers on the order of several kT. Since t he thermal energy in the system is approximately 1 kT, the barriers less than 1 kT are easily and frequently crossed whereas the larger barriers are more difficult to cross and thus crossed less frequently. It is the magnitude of these energy barriers that determine s the r ate of inter conversion between the two states. The population of each state or conformation is determined by the relat ive free energy of each state according to Equation 1 3 where R is the gas constant, T is the absolute temperature,

PAGE 63

63 (1 3 ) [B] and [A] are the concentrations of the B and A states respectively, G b and G a are the corresponding molar free energies, and G b and G a are the molar free energie s of the standard state s At equilibrium, G=0, so E quation 1 3 can be rearranged to form Equation 1 4, which can also be rearranged to form Equation 1 5. According to Equation 1 5, if the relative (1 4 ) (1 5 ) populations of the states can be experimentally determined, then the free energy difference between the states can be estimated with reasonable certainty. These are the same fundamental equations used earlier to discuss inhibitor binding, but the focus here is on the population di fference between state A and state B as opposed the entropic and enthalpic contributions to each state. Energy Landscapes Conformational Ensembles, and Ensemble Shifts The concept of an energy landscape was initially applied to proteins in terms of protei n folding funnels The top of the folding funnel represented the protein in the unfolded state, which has a large number of different conformations available that have the same energy level. As the hydrophobic regions collapse, the energy of states decre ases and the number of accessible states decreases so the funnel deepens and narrows. As the specific stabilizing interactions are formed, the energy and number of states will continue to decrease so the funnel will continue to deepen and narrow. As the conformations sampled become more lik e the native state, the energy level s continue to decrease until finally the protein reaches the native state which has the lowest

PAGE 64

64 energy of all of the states. It has also been described that proteins in the process of folding can reach a meta stable state (a local minimum, but not the global minimum) and b ecome kinetically trapped where the protein c annot reach the global minimum which is the native state. The idea of energy landscapes applies to protein conformations as well Instead of starting at the top of the funnel, the focus is primarily on the bottom where the global and energetically accessible local minima are located. Each local minimum in the landscape corresponds to a stable conformation that the protein can form Although t he landscape s are typically portrayed as being smooth, e ach local minimum is actually composed of multiple local minima of approximately equal energy These local minima are divided by small energy barriers because the structures dif fer only by bond lengths and bond rotations ( Figure 1 20 ). The landscape is highly multidimensional because proteins have many atoms and many possible conformations. Each conformation or conformational state of the protein corresponds to a group of highly related structures. Collectively, all the conformational states form the protein conformational ensemble. Thus, the ensemble includes the predominant low energy conformations that dominate the average structure of the protein, as well as the higher ener gy conformations that the protein will rarely sample. The determination and characterization of protein conformational ensembles is an important aspect of understanding protein function. The energy landscape is specific for a certain set of conditions, such as solvent, temperature, or pH, in addition to the presence of a substrate, inhibitor, or other ligand. Changing any one of the above can, and most frequently does, remodel the energy landscape. This remodeling can be thought of in a variety of ways e.g. shifting the energy levels of the local minima to change the population distribution.

PAGE 65

65 Figure 1 20 Schematic diagram of an energy landscape for a protein with two major conformations, A and B. The black and grey trace s illustrate an ensemble s hift induced by some change in condition that stabilizes the B state and destabilizes the A s tate. Figure modified from Henzler Wild et al. 96 T he conformational change that a protein undergoes upon binding to a ligand can be discussed in terms of remodeling the energy landscape. One form of remodeling is the narrowing of breadth of the local minimum upon binding a ligand. This corresponds to an increase in the rigidity of the protein resulting from a decrease in the mobility of the atoms and residues although maintaining the same average structure. An example of this form of remodeling is the homodimeric cAMP binding transcriptional activator, which is highly dynamic upon binding the first cAMP but rigid afte r binding the second cAMP. 9 8 Another form of remodeling is the alteration of the relative e nergies of the local minima, thus shifting the relative populations of each state Thi s is a very common theme among most proteins that undergo a conformational change upon binding to a li gand The implication of this trend is tha t the bound form is energetic ally accessible to the apo protein and is stabilized by the interaction of the partner. An example of this is the human proto ocogene Vav, which has an

PAGE 66

66 binds to the guanidine nucleotide exchange factor (GEF) substrate binding domain resulting in autoinhibition. It was shown that Vav occasionally samples a conformation on the s ms timescale where the latch is displaced and autoinhibition is relieved. Phosp h orylation of Tyr174 also leads to the release of the latch and relieves the inhibition 99 Thus the uninhibited conformation is the higher energy state but it is still energetically accessible to the unphosphorylated protein. However, phosphorylation shifts the relative energy levels between the inhibited and uninhibited states making the un inhibit ed state the more populated form. Another mode of remodeling c an occur when the higher energy state is an intermediate in the pathway to the stable ligand bound state. In this case the ligand bound structure is not energetically accessible to the apoprotein and the minimum corresponding to the higher energy state is shifted to corresp ond to the new bound state 100 An example is maltose binding protein (MBP) which has been shown to have two unique predominant conformations (95 % open and 5% minor ) in the apo state and one predominant conformation (closed) in the maltose bound state. The dif ference between the open and closed states is a 35 domain rotation of the CTD (C terminal domain) relative to the NTD (N terminal domain). The minor state is very similar to the closed state but corresponds to only a 33 domain rotation. This differenc e results from an increased distance between the domain s because the interdomain interface is lined with positively charged residues on both sides. The binding of the maltose shields these charges and permits the domain s to move closer together to form th e closed state 101 In this case, the minor conformation is the higher energy state that is energetically accessible to the apoprotein and is an intermediate in the formation of the bound state upon ligand binding.

PAGE 67

67 Experiment al Methods for Characterizing Ensembles and Ensemble Shifts There is a variety of techniques suitable for characterizing protein conformations, ensembles, and ensemble shifts. Most of the techniques are only applicable for a set range of timescales and th us can only characterize certain regions of the energy landscape. Many techniques also have sensitivity limitations in that signals from sparsely populated states are sufficiently weak to not be detectable in the presence of signals from the highly popul ated states. Thus, the most complete characterization requires the combination of multiple techniques. The easiest motions to characterize a re the slow timescale motions, which interconvert slowly enough to either be studied directly Alternatively, the higher energy states can be trapped and studied via structure determination techniques such as x ray crystallography, NMR cryo electron microscopy, small angel x ray scattering, and hydrogen/deuterium exchange. The trapping of the higher energy states i s usually accomplished by careful selection of conditions (salt concentration, temperature, pH, etc) or by binding of a ligand (frequently an altered ligand that traps the protein in an intermediate state). Most of these techniques have limitations or di fficult to meet requirements. X ray crystallography requires that the state of interest be amenable to crystallization. For NMR the protein must weigh less than 100 kDa (although anything above 70 kDa is incredibly difficult). NMR also requires that th e protein can be isotopically labeled which typically involves recombinant expression and thus rules out proteins that must be isolated from natural sources. The motions occurring on the faster timescales are more difficult to characterize. They typica lly have be en discussed in terms of statistical distributions, because there are often many more states, which cannot us ually be trapped individually

PAGE 68

68 NMR Paramagnetic relaxation enhancement Paramagnetic relaxation enhancement (PRE) NMR utilizes a param agnetic center to increase the relaxation rates of nuclei within a certain radius. The paramagnetic moiety is typically introduced into the protein via site directed spin labeling (discussed in detail in Chapter 2), where an amino acid is replaced by a cy steine (Cys) and is subsequentia lly modified with a spin label a small molecule containing a stable nitroxide r adical or metal chelating group Alternatively, if the protein contains a diamagnetic ion cofactor a paramagnetic transition metal ion can some times be substituted for the native metal ion The spin relaxation rat es of the nuclei under investigation will be measured in the presence of the paramagnetic species and a corresponding diamagnetic species. The difference between the two is the enhance ment resulting from the paramagnetic species. The relaxation enhancement results from the dipolar interaction between the spins (discussed in Chapter 2) which is proportional to the 1/r 3 where r is the distance between the spins. The distance range vari es with the magnitude of the gyromagnetic ratio ( ) for the nuclei and can thus be selected by choosing the appropriate nuclei. Typically, the spin spin relaxation, T 2 (discussed in Chapter 2) is the preferred relaxati on measurement for this technique H owever, spin lattice relaxation, T 1 measurements are also suitable. Paramagnetic relaxation enhancement is especially useful because the range of distances over which it is sensitive can be changed by measuring the relaxation rates for a different nucl eus. However, the true power of PRE, in terms of characterizing an energy landscape for a protein, the sensitivity to sparsely populated states as long as the distance measured is shorter for the higher energy sta te r elative to the low energy state As a result, Clore et al. 101 w ere able to

PAGE 69

69 determine the structure of the minor conformation in MBP even though it represented only 5% of the population. Relaxation dispersion Relaxation disp e rsion (RD) 102 is another NMR technique that is gaining in popularity for the characteriz ation of conformational ensembles and sparsely po pulated states The major and minor conformations must have different chemical shift values and different relaxation rates, both resulting from the structural differences. The RD experiment measure s the apparent relaxation rate as a function of the spaci ng between refocusing pulses, which requires tha t the motions being studied occur on an intermediate timescal e relative to the experiment (exchange rate of the motion (1/s) chemical shift difference (rad/s) between the states). The apparent relaxation r ate is the combination of the relaxation rates of both states relative to the amou nt of time spent in each state. The data are typically plotted to give a relaxation profile ( R eff ( CMPG ) where R eff is related to change in signal intensity and CMPG is r epetition rate of the pulses in the Carr Purce ll Meiboom Gill pulse sequence ), from which the exchange rate between the two states and the ir chemical shift differences can be determined by fitting the data to the modified Bloch equations 103 Residual dipolar coupling Residual dipolar c oupling (RDC) NMR is a technique that uses residual alignment of the protein with the magnetic field to investigate the relative alignment of bonds within the protein. The protein alignment can be enhanced by adding large, non spherical moieties which align within the magnetic field to the sample such as lipid bicelles discoid shaped lipid bilayers resulting from the combination of short and long chain lipids which forms a planar bilayer with the long chain lipids and is ringed by the short chain lipids. The degree of ordering of the protein can be controlled by adjusting the concentration of the bicell e s 104

PAGE 70

70 Depending on the assumptions made during the data analysis, both structural and dynamical information can be elucidated 105 Order parameter determination NMR relaxation techniq ues can also be useful by fitting the data to determine the model free order parameter, S, and the correlation time e 106 The T 1 T 2 and nuclear Overhauser effect (NOE) values can be measured using inversion recovery, spin echo, and NOESY pulse sequences respectively. These data are fit to the equations described in Szabo et al. 106 to extract S and e which correspond to magnitude and rate of the motion respectively. Because this technique characterizes each residue in protein, different residues can compared to residues in regions of the protein that experience a larger change in motion upon ligand binding than other sites. It can also compare alterations in the motion at a particular site as a result of amino acid substitutio ns in the protein sequence, which can elucidate important residues for the motion and flexibility in the protein. Hydrogen/deuterium exchange Hydrogen/deuterium exchange utilizes the pH depend e nt exchange between the amide proton and water to replace amide protons with deuterium atoms. It has been shown 107 that the amide bonds involved in hydrogen bonds exchange at a slower rate relative to amide bonds solvated by water. It is also known that buried amides (the amides not exposed to t he solvent) are not accessible for exchange. The effective rate of deuterium exchange will be a convolution of the time spent by a particular proton in an exchangeable conformation vs. a non exchangeable conformation. For example, an amide proton in an a only be accessible when the lid is open. Thus measur ement of the exchange rate of such a proton would provide information on the dynamics of the lid domain. This technique requires a detection method sensiti ve to the small differences between hydrogen and deuterium and is thus

PAGE 71

71 predominantly performed using either NMR or mass spectrometry (MS). Fast timescales can be probed by NMR and either method can be used to probe the slow timescales The slower timesca les are typically monitored via quenching, i.e. the sample is allowed to exchange for a certain period of time at an acidic pH after which it is quenched by raising the pH above 7. Fluorescence Intrinsic tryptophan fluorescence A large variety of fluor escence based techniques can provide insight in to the conformations and ensembles in proteins. Intrinsic tryptophan (Trp) fluorescence is sensitive to the polarity of its environment and can distinguish between conformations with differing environments. Fluorescence lifetime measurements of the intrinsic Trp fluorescence can reveal the numbe r of different Trp environments If there is only one Trp and it has multiple lifetimes (i.e. the lifetime data are best fit by a multi exponent) then that Trp has multiple conformations. The lifetimes can be studied as a function of ligand concentration, pH, temperature, or other conditions to monitor the relative number of Trp in each conformation as well as changes to each conformation. For example, consider a p rotein has a Trp residue with two lifetimes, one short and one long. As ligand is added, the relative contribution from each state will change and be detected. It is also possible however, that the ligand can also change the lifetime of a state, making it either longer or shorter. This can also be identified and tracked as a function of ligand concentration. Single molecule fluorescence Another increasingly common technique is single molecule fluorescence. This method typically relies on modifying a residue on the protein with a very high quantum yield fluorescent probe (the quantum yield of Trp is too low for these experiments). By measuring the time dependence of the lifetime this method can reveal information about the conformation of the protein This technique can also be used for FRET (Forester s or Fluorescence Resonance Energy Transfer) in which case the time dependence of a

PAGE 72

72 single distance can be measured and directly related to a conformation. There are multiple benefits to observing sing le molecules. One is that it is possible to monitor both how long it takes for a conformational change to occur and how long the state persists. It has been observed that the conformational changes are typically faster than the observation time but that the benefit is that e vents can be observed in order Many techniques can show that a protein has multiple conformations, but few techniques can determine the exact order they occur in. For example s ingle molecule FRET reveal ed several additional rotational steps between the three major conformations and also reveal ed that the steps are reversed in ATP synthesis relative to ATP hydrolysis 108 X ray diffraction X ray crystallogra phy is powerful technique for discovering structural information about a protein on an atomistic scale. This technique is even powerful in conjunction with biochemical tricks for capturing or stabilizing higher energy or transient states. These tricks in clude the use of transition state mimetics, substrate or products mimetics, and non reactive substrates, as well as vari ant proteins that are enzymatic ally incompetent or impaired. Because the technique requires the protein to assume a conformation amenab le to crystallization, it is primarily limited to the study of slow timescales. However, faster motions can be studied by analyzing the B factor s as long as care is taken to avoid effects from crystal contacts and the lattice. Cryogenic studies Because t he transition between states relies on the thermal energy of the system, reducing the thermal energy available will reduce the rate at which the protein can switch states. This practice facilitates the study of transient states too short lived to study at room temperature. Enzymes are frequently studied by freezing quenching the sample immediately after mixing the enzyme with

PAGE 73

73 the substrate. For light triggered events, the sample is frozen in the dark and then exposed to light. Comput ational Methods The ability to calculate the position and trajectories of atoms as a function of time provides conjuction with experimental data form a particularly powerful c ombination because the computations can provide insight into the interpretation of the experimental results. Likewise, the experimental results can confirm the accuracy of the computation. It should be noted that the computations are only as good as the force fields used whereas the experimental results tend to be more reliable so long as they were preformed correctly and with the proper control experiments Scope of Dissertation This dissertation reports on the investigation into the conformations and fle xibility of the this technique for addressing these questions and (2) to determine if the flap flexibility is altered by drug pressure selected mutations and natur ally occurring polymorphisms. To optimize this technique, the data analysis process was thoroughly examined and expanded to include additional refinements in the final distance profiles. Additionally, the effects of solutes were investigated to determine the effect of solutes on the flap conformation and spin label mobility. This technique was applied to three HIV 1 PR variants subtype B, subtype C, and a clinical isolate with drug pressure selected mutations, V6. The distance between the spin label on the flaps was investigated for the apoenzyme and for the enzyme in the presence of nine FDA approved inhibitors and one substrate mimic. One of the major goals of this work was also to

PAGE 74

74 demonstrate that distance measurements via pulsed EPR can also be used to characterize the conformational ensemble of sampled proteins.

PAGE 75

75 CHAPTER 2 ELECTRON PARAMAGNETI C RESONANCE Introduction to Electron Paramagnetic Resonance Electron Paramagnetic Resonance (EPR) is a spectroscop ic technique for studying the interactions between electronic magnetic moments and their immediate environments including externally applied magnetic field s There are many excellent reviews for a deeper understanding of this material 109; 110; 111; 112; 11 3; 114 Presented here is a brief overview in sufficient detail for the reader to understand the topics presented in the later chapters. Spin Hamiltonian The Hamiltonian is an operator that corresponds to the energy of a system. In an externally applie d magnetic field which is independent of time, the energy of the spin system can be determined from the static spin Hamiltonian (Equations 2 1 and 2 2 ) as opposed to the oscillatory Hamiltonian that corresponds to the energy of the system in a time depende nt magnetic field Th e static Hamiltonian includes a variety of terms (d iscussed in subsequent sections) some of (2 1 ) (2 2 ) which contribute significantly only in p ulsed experiments and some which contribute significantly in both CW and pulsed EPR. In Equation 2 1 H EZ is the electron Zeeman term, H ZFS i s the zero field splitting term, H HF is the hyperfine interaction term, H NZ is the nuclear Zeeman term, H N Q is the n uclear quadrupole interactions H NN is the spin spin interaction term

PAGE 76

76 which includes all inter nuclear interactions H dd is the electron electron dipolar interaction, and H exch is the exchange coupling term. The oscill atory Hamiltonian results from the time dependence of the applied magnetic field expressed in Equation 2 3 Substitution of B MW (t) for B 0 in Equation 2 2 only affects the (2 3 ) electron and nuclear Zeeman terms, which can be rewritten as Equation 2 4 and Equation 2 5 respectively. (2 4) (2 5) Electron Zeeman All particles with spin angular momentum ( S ) also have a spin magnetic moment ( ) which is related to S by a proportionality cons tant, the gyromagnetic ratio ( = q /2 m ), where q and m are the charge and mass respectively of the particle and 6.626 10 J s) divided by 2 (Equation 2 6 ). For an electron, this relationship can be rewritten as Equation 2 7 where g is the spectroscopic g factor (or Land g factor) and e is the Bohr magneton (written as B in some literature) (Equation 2 8), where e is the charge on an electron ( e = 1.602 x 10 19 C), and m e is the rest mass of the electron (9.109 x 10 31 kg). The value of the g factor for a free electron is 2.0023193043737, but this value will deviate as a result of spin orbit coupling for bound electrons (2 6 )

PAGE 77

77 (2 7 ) (2 8 ) Additionally, the g factor is actually a tensor and can be anisotropic for some systems. These two factors make g a useful parameter for characterizing a spin system Equation 2 4 can be rewritten as Eq uation 2 9 to include the tensorial nature of g (2 9 ) In the presence of an applied magnetic field, the spin ma gnetic moment will interact with the field with the energy described by the Hamiltonian in Equation 2 10 where B 0 is the applied (2 10 ) f ield, and B 0 T is the transpose of B 0 This interaction causes the spins to align themselves parallel or antiparallel to the applied field. The spins aligned anti parallel to the field will be at a lower en ergy than the spins aligned parallel thus the dege neracy of the spin states is broken. Because the electron is negatively charged, the electron spin alignment is opposite of the nuclear spin alignment. Substitution with e yields Equation 2 11 where S is the spin vector for the (2 11 ) system. If B 0 is applied in z direction of the laboratory frame then the Hamiltonian can be simplified to Equation 2 12. For a single unpaired electron S = so t he corresponding (2 12 ) eigenstates are m s = The m s = + state is designated as and the m s = state is desig nated as The wave functions for these states are | and | respectively. The corresponding energies for the spin e nergy levels are calculated in Equations 2 13 and 2 14.

PAGE 78

78 (2 13) (2 14) Thus the difference in energy levels, E, is shown in Equation 2 15, which is also the reson ance condition for EPR. It should be noted the E and thus the resonant frequency depend on the magnitude of B 0 (2 1 5 ) Hyperfine interaction The hyperfine interaction is the result of the electron spin interacting with the additional magnetic field induced by the spin s of neighboring nuclei that have a net magnetic moment and is described by Equation 2 1 6 where I is the nuclear spin quantum nu mber A is the hyperfine (2 1 6 ) t ensor and the index k is for all interacting nuclei. This interaction results in a further s plitting of the energy levels based on the spin quantum number of the nucleus as shown in Figure 2 1 H HF is characterized by a tensor ( A ) and is anisotropic. The anisotropy makes the hyperfine interaction a useful parameter for characterizing a spin sy stem. The hyperfine interaction can also be expressed as Equation 2 17, which is composed of (2 17) the isotropic Fermi contact term ( H F ) and anisotropic electron nuclear dipole interaction ( H DD ). H F can be expressed as Equation 2 18, where a iso is the isotropic hyperfine coupling constant

PAGE 79

79 Figure 2 1 Energy level diagram for the Zeeman and hyperfine splittings of an S = 1/2 sp in on a nucleus with I=1. Selection rules require that only one quantum number changes during a transition so the m S number changes while the m I number remains constant resulting in the three possible transitions illustrated. (Equation 2 1 9 ) and | (0)| 2 is the electron spin density at the nucleus which typically includes (2 1 8 ) (2 1 9 ) only the contribution of the s orbital. More rigorous calculations should include contributions from configuration interactions and spin polarization mechanisms when the unpaire d electron is in a p, d, or f orbital 115 The electron nuclear dipole interaction H DD can be expressed as Equation 2 20 where r is (2 20 ) the vector connecting the electron spin and the nuclear spin. Integration of Equation 2 20 over the spatial distribution of the electron yields Equation 2 2 1 where T is the traceless and (2 2 1 ) symmetric di polar coupling tensor with the elements in Equation 2 2 2 (for spin only electron nuclear dipolar coupling).

PAGE 80

80 (2 2 2 ) S election rules dictate that only one quan tum number can change in a magnetic resonance transition so the only allowed transitions are between the m S states while the m I state remain s constant. For a nucleus with I = 1, t his leads to the three available transitions illustrated in Figure 2 1 Nu clear Zeeman The nuclear Zeeman term (Equation 2 2 3 ) results from the interaction of the nuclear spin (2 2 3 ) with the applied magnetic field, where g n ranges from 0.097 for 191 Ir to 5.58 for 1 H. This term is gener ally considered to be isotropic and is typically small and hard to detect in EPR experiments although it is responsible for spin flip transitions. Zero field splitting The zero field splitting term (Equation 2 2 4 ) where D is the zero fiel d interaction tensor, (2 2 4 ) results from the special cas e where the spin energy levels are not degenerate in the absence of an applied magnetic field. This only occurs with spins > and thus i s beyond the scope of this work. Nuclear quadr u pole The nuclear quadrupole term (Equation 2 2 5 ) where P is the nuclea r quadrupole tensor, (2 2 5 )

PAGE 81

81 results from nuclei with spin >1 that also have a non spherical charge distribution. The lack of symmetry with the charge distribution interacts with the electric field gradient resulting from nearby electrons and nuclei. The effects of this term are small and typically negligible for EPR experiments. Nuclear spin spin interactions The nuclear spin s pi n interaction term (Equation 2 2 6 ) where d is the nuclear dipole (2 2 6 ) coupling t ensor, results from the dipole dipole inte ractions between two nuclei. This term provides much of the structural information gained from NMR experiments but is negligible for most EPR applications Electron electron dipole interactions The electron electron dipole interaction term (Equation 2 2 7 ) results from the dipolar (2 2 7 ) interaction between electrons (the interaction of the electron with the magnetic field f rom a nearby electron ) This term is the basis for the distance measurements via pulsed EPR and will be discussed later in more detail. Exchange coupling The exchange coupling term (Equation 2 2 8) results from the Heisenberg exchange coupling (2 2 8 ) that occurs when there is orbital overlap between the orbitals of two electrons where J is the exchange coupling tensor Th e orbital overlap permits the electrons to swap orbitals thus confusing the knowledge of the respective spin states of each electron. This exchange occurs at

PAGE 82

82 distances much shorter than those used for DEER and thus does not contribute to the experiments performed for this wor k. Site D irected S pin L abeling History of site directed spin labeling The history of site directed spin labeling began i n 1965, when three papers were published which utilized a persistent nitroxide radical to enable EPR studies of bi ological molecules. Two of the papers combined the nitroxide radical with a reactive group to covalently attach the radical to a protein 116; 117 The third paper used a nitroxide radical with a 2,4 dinitrophenyl moiety to bind antibodies specific for the 2,4 dinitrophenyl moiety 118 ent dyes which would label covalently bind the biological molecule of interest. It was established that the nitroxide labels were inert stable in the presence of water or oxygen 119 sensitive to motion, and correspond to simple spectra 116 which ma k e the nitroxide labels ideal for probing the conformational changes in biomacromolecules. This technique incre ased in popularity over the next 20 years and was used to study a large variety of systems. However, there were t hree significant limitations th at hindered the technique from being almost universally appli cable selectivity, sensitivity, and versatility T he selectivity issue arose from the limited selection of suitable labeling sites in biomolecules. There was a small variety of labeling chemistries available that permitted some selection of the labeling sites. However, most proteins contain multiple lab eling sites reactive to each of the labeling chemistries. Consequently, it was very difficult to obtain protein selectively labeled at one site.

PAGE 83

83 The sensitivity issue arose from the difficulty of preparing samples in sufficient quantity. This issue was essentially solved in 1982 with the development of the loop gap resonator 120 This resonator permitted single scan spectra from picomole quantities of proteins. The versatility issue arose from the inability to selectively choose the site for labeling. Researchers were limited to using native sites w ithin the proteins for labeling. However, the domain s of interest do not always contain suitable labeling sites, which limited the applicability of the method The development of site directed mutagenesis a technique wherein the DNA bases can be manipulated at a specific site permitted researchers to remove all the native labeling sites and introduce a unique labeling site at the position desired. This combination of site directed mutagenesis and spin labeling was introduced in a 1989 manuscript by Altenbach et al. 121 Cys mutations were introduced into bacteriorhodopsin and subsequentially labeled with thiol selective spin label s. The ter ite directed spin labeling (SDSL) has been coined to describe this combination of site directed mutagenesis and spin labeling. In general, the DNA encoding the protein of interest is mutated such that the codon for a specific residue is mutated to en code a Cys residue. The mutant DNA is used to express the Cys variant in a recombinant expression system such as E. coli or yeast. The protein is purified and labeled with a thiol reactive spin label. Alternatively, small proteins and peptides can be sy nthesized with a Cys residue for labeling or with a n amino acid derivative that incorporates a spin label such as TOAC. 122 Site directed spin labeling can also refer to studies using spin labeled nucleic acids. A variety of labels have been developed for incorporation into nucleic acid strands that attac h via

PAGE 84

84 the base, the sugar, or phosphate group. Typically, the nucleic acid is synthesized and the label is incorporated via the addition of a derivatized nucleotide that contains the label. Spin labels By definition, a spin label is any molecule containin g an unpaired electron an EPR active probe and reactive moiety for binding to another molecule. The majority of spin labels used for investigating the structure and flexibility of biomolecules are nitroxide radicals because their simple line shape is high ly sensitive to motion making them well suited for investigating the structure and flexibility of a biomolecules. The nitroxide radical is protected by bulky methyl groups which prevent collisions. The geometry of the radical and surrounding methyl group s is typically preserved by inclusion in a five or six member ring The reactive moiety is typically attached to the ring structure via four five bonds as is necessitated by the synthesis process. A large variety of spin labels are available commerciall y including variations in the ring structure and in the reaction chemistry of the flexible linker. Figure 2 2 illustrates the structures of four common nitroxide spin labels with thiol based linkers as well as the structures of the modified Cys residues a fter reaction with the spin label. Figure 2 2 Structure of spin labels before and after binding to Cys side chain. A) and B) MTSL: (1 Oxyl 2,2,5,5 Tetramethyl Pyrroline 3 Methyl) Methane thiosulfonate; C ) and D) IAP: 3 (2 Iodoacetamido) PROXYL E ) an d F) MSL: 4 Maleimido TEMPO; and G ) and H) IASL: 4 (2 Iodoacetamido) TEMPO. Rectangular box represents protein backbone.

PAGE 85

85 Selection of labeling sites The c hoice of labeling sites within biological system s is not trivial. Obviously, the site should be chos en to report on the relevant aspect of the system under study. However, care must also be taken to preserve the structure and function of the system. For inter flap distance measurements in HIV 1 PR it is necessary to pick sites within the flaps that ha ve been sh own to be highly conserved and intolerant of amino acid substitution However, there is one site, position 55, that is able to tolerate a variety of amino acid substitutions and is a suitable choice for a labeling site. 123 Additionally, the extent of solvent exposure at the site should be considered. For distance measurements via pulsed EPR, solvent exposed sites that report only on the motion about the flexible linker, (e.g. the line shape do not chan ge upon ligand binding or other changes to the system) are ideal. However, for characterizing conformational changes via CW EPR, the label should be sensitive to changes in the conformation upon ligand binding (e.g. one conformation gives rise to motional ly restricted line shape and the other conformation to a highly mobile line shape). Nitroxide line shapes The unpaired electron in the nitroxide radical is strongly associated with the nitrogen atom (I=1). This lead s to hyperfine splitting with three tran sitions (m s =+1/2, m I =+1 m s = 1/2, m I =+1, m s =+1/2, m I =0 m s = 1/2, m I =0, m s =+1/2, m I = 1 m s = 1/2, m I = 1) as illustrated in Figure 2 1 and Figure 2 3A When the nitroxide is tumbling quickly in s olution, as is the case f or unattached nitroxides and n itroxi de s attached to small bio molecules (MW < 18 kDa) the transitions give rise to sharp features in the line shape as illustrated in Figure 2 3D. If the motion of the nitroxide is restricted, either as a result of slower tumbling from being attached to large

PAGE 86

86 mo lecule or from contacts with neighboring molecules (such as amino acid side chains), the transitions give rise to features that are broadened from incomplete averaging of anisotropic components in the spin Hamiltonian, namely the g and A tensors. Figure 2 3E shows a typical line shape for a nitroxide bound to a 20 kDa protein at a solvent exposed site. This site displays some broadening relative to the line shape in Figure 2 3D because of the slower rate of tumbling in solution. Figure 2 3F shows a line shape typical for completely immobile spin label. Figure 2 3 Schematic diagram of the possible energy levels (A) for a spin on a nucleus with spin 1 and corresponding absorption (B) and derivative (C) spectra. Sample line shapes for nitroxide radica ls undergoing rapid motion (D), moderate motion (E), and no motion (F). Three primary modes of motion affect the mo bility of the spin label and the resultant EPR spectrum: (1) the tumbling of the entire protein characterized by the correlation time, R ; (2) torsional oscillations about internal bonds within the nitroxide moiety and its attachment to the

PAGE 87

87 CYS residue side chain, defined at the internal correlation time, I ; and local macromolecular fluctuations of the protein at the labeling site, refer red to as local dynamics, B as illustrated in Figure 2 4 Figure 2 4 Illustration of the three correlation times of a spin label, where R is the overall tumbling of the protein, B is the movement of the protein backbone including local oscillation s and conformational changes, and I is the movement of the label about the flexible linker. Changes in R are useful for studying the binding of small molecules (peptide or RNA fragments) to larger biomolecules such as complex protein syst ems, membrane b ilayers, or RNA/protein complexes. However line shapes that are dominated by R do not contain site specific information about protein dynamics. I n order to gain site specific information, it is necessary that the line shape is not dominated by R Th i s condition is usually met for biomolecules greater than 18 kDa, and can be creatively circumvented for most biomolecules smaller than 18kDa. The major source of structural information in SDSL studies comes from line shape analysis from nitroxide labeled s ites where molecular motions are dominated by changes in I and B

PAGE 88

88 Spin label conformations It is important to understand the intrinsic flexibility and preferred conformations of the flexible linker that are used to attach the spin label because of the potential impact these factors can have on the mobility of the spin label. Spin labels can be attached to Cys residues in any secondary structural element. Each structural element has a different degree of backbone flexibility and tightness in the packing of t he side chains. These factors strongly a ffect the m obility of the spin label and thus the line shape. It has been found that on helical sites, the delta sulphur atom (S ) can form a stable interaction with the hydrogen atom (H ) on the alpha carbon (C ). This arrangement restricts rotations about the first three bonds in the linker thereby reducing the rotational freedom to the fourth and fifth bonds (corresponding to torsional oscillations about the 4 and 5 dihedral angles respectively). These constrained motions are the basis for the 4 / 5 model 124 illustrated in Figure 2 5 It has been demonstrated that this model is valid on heli ces in a variety of proteins. H owever, ever y protein can potentially have unique or rare factor s which cause the spin label motion to deviate from this model. Crystallographic studies of spin labeled proteins provide direct information regarding the pref erred conformation of the spin label. This is particularly useful for distinguishing between rotomeri c states of the spin label versu s different protein conformations when the CW line shape has multiple components. Additionally, computer simulations can be performed to predict the preferred spin label conformations. These computations can include full MD simulations of the entire protein or they can be simplified such that the atoms in the protein are motionally restricted still while the spin label is a llowed to move.

PAGE 89

89 Figure 2 5 Illustration of the 4 / 5 model. The S interacts with the H on the C (indicated by dotted line), which restricts rotations to the 4 and 5 torsional angles. Line shape analysis The line shape analysis typically involve s quantifying the breadth of the spectrum as a measure of the spin label mobility. If the spin label is tumbling rapidly and isotropically in solution, the features of the line shape will be very narrow and sharp resulting from the averaging of the anisot ropy in the A and g tensors. If the motion of the spin label is restricted, the averaging of the anisotropy in the A and g tensors will be incomplete and the spectrum will be broadened. The extent of the broadening correlates with the extent of restricti on in the motion. Figure 2 6 illustrates several methods commonly used for quantifying the breadth of the line shape including measuring the peak to peak distance ( H pp 125 or 124 ) the ratio between the intensity of the center field transition to the low field transition (I CF /I LF or H(0)/H(+1)), 126 and the second moment of the spectrum ( H 2 ). 124 The H pp can be normalized by conversion to the scaled mobility (M S ) as shown in Equation 2 29, where m is the most mobile H p p reported in the literature and im is the immobile. The H 2 is proportional to the sum of distance squared (2 29)

PAGE 90

9 0 between every point and the center of mass of the spectrum times the intensity at that point (Equation 2 30). (2 30) Figure 2 6 Parameters for quantifying the breadth of a CW EPR line shape. A) Central line width, H pp and the ratio of the center field transit ion to the low field transition, I CF/LF B) Second moment, H 2 Introduction to Distance Measurements Via Pulsed EPR All experimental techniques that measure distances rely on a distance dependant interaction between two moieties. For fluorescence reso na nce energy transfer (FRET), the distance depend e nt interaction between the fluoro phores is an energy transfer that stems from a long range dipole dipole interaction which i s distance depend e nt 127 For NMR, various experiments use the dipolar couplings, the J coupling, or spin polarization transfer between various nuclei all of which are distance depend e nt 128 For EPR, both p ulsed and continuous wave (CW), the distance depend e nt interaction is the dipolar coupling between two unpaired electrons. CW EPR is used for distances less than 15 20 129 and pulsed EPR is used for distances be tween 15 80 130 (although 60 is the upper limit for most non model systems) 131

PAGE 91

91 Dipolar Interaction The dipolar interaction between two spins is essentially the force of the magnetic f ield generated by one spin on the magnetic dipole of the other spin. The dipole interaction between two electrons can be described by the Hamiltonian in Equation 2 3 1 132 wher e r is the (2 3 1 ) inter electron vector. This can be rewritten as Equation 2 3 2 or Equation 2 3 3 by substituting (2 3 2 ) (2 3 3 ) the magnetic moment operators with the corresponding spin operators, which is usually represented as S T DS. Elec tron Spin Echo Techniques All electron spin echo (ESE) techniques have the same foundation which is simply that the dipolar interaction between two electrons takes the form of an additional modulation or decay/recovery. This manifestation is the result o f the instantaneous change in the effective magnetic field of spin when a coupled spin is flipped v ia a pulse 133 There are seve ral pulse sequences that utilize ESE to measure the dipolar interaction between two spins including double quantum coherence (DQC) and double electron electron resonance (DEER) (a l s o known as pulsed electron double resonance, PELDOR). Double Quantum Coherence (DQC) Multiple Quantum Coherence (MQC) EPR i s analogous to MQC NMR, which has been explained tho roughly elsewhere 111 The most significant difference has been the difficultly in

PAGE 92

92 adapting the technique to the instrument al limitations inherent in the EPR spectrometers 134 However, recent improvements in technology and improved pulse sequences have made DQC EPR a viable op tion for distance measurements in many systems, including biological systems. 135 Double e lectron e lectron r esonance (DEER) All DEER experiment s have the same general outline: microwave pulses are used to selectively excite two separate spin populations (generally referred to as s pins A and spins B). The B spins are flipped, which causes a perturbation on the coupled A spins that is manifested as an additional modulation on the A spins. 2+1 One significant instrumental limitation in the three and four pulse DEER experiments i s the need for two separate microwave frequencies and a sufficiently broad spectrum that two populations of spins exist that can be excited independently. When these conditions cannot be me sequence illustrated in Figure 2 7 A The 2+1 sequence needs only one microwave frequency but involves an additional complication resulting from the lack of separation between the A and B spins. These complications arise from the possibility of flippin g the A spins with the second pulse or flipping B spins with the third pulse and are manifested as phase changes for th e spins. In a DEER experiment with two frequencies the spins can only gain an additional phase but in a 2+1 experiment, the spins can either gain an additional phase or recover their original phase. Three pulse DEER. The f irst DEER experiment published 136 used the three pulse DEER se quence illustrated in Figure 2 7 B Th is sequence generates a Hahn echo using a two pulse Hahn echo sequence on 1 ) (the observe frequency) The intensity of the echo is measured as a

PAGE 93

93 function of the timing of a th ird pulse on 2 ) (the pump frequency). One of the major limitations of the three pulse DEER was the ~100 ns delay in the data acqu isition after the third pulse Figure 2 7 Pulse sequences fo pulse DEER, and C) Four pul se DEER. Pulse spacings labeled with remain constant and spacings labeled with t are incremented. because the highly sensitive detector cannot be switched on until the power from the microwave pulse has sufficiently dissipated from the resonator. Th is delay is commonly referred to as the

PAGE 94

94 experimental dead time. The largest consequence of the experimental dead time is that data from the early part of the dipolar evolution curve area is lost, which is the only region in the curve that contains inform ation on the shortest distances. This problem was overcome in the four pulse DEER sequence by refocusing the Hahn echo, using an additional observer pulse, at 2 2 after the Hahn echo appears (Figure 2 7C). 137 Because of the similarity to the four pul se DEER, the three pulse DEER will not be discussed in detail; most, if not all, of the discussion for four pulse will hold true for the three pulse DEER. Four pulse DEER. Figure 2 7C illustrates the pulse sequences used for four pulse DEER. Similar to t hree pulse DEER the sequence also begins with a two pulse Hahn echo sequence on 1 ) After the appearance of the Hahn echo, a pump pulse is applied on 2 ) with a varying time delay after the echo. At 2 the echo is refocu sed by an additional pulse on 1 ) Again, the echo intensity is recorded as a function of the time delay between the first echo and the pump pulse. The effect of the pump pulse is to flip the B spins at time t, which alters the effective magnetic field for the A spins that are c oupled to a B spin. This change in the magnetic field alters the precession frequency of the coupled A spins by ee (electron electron coupling,) which results in the magnetization being out of phase b y the angle ee = ee t. Thus ee can be determined by integrating the echo intensity as a function of t. Equation 2 3 4 defines ee (2 3 4 ) where r AB is the distance between the spins, AB is the angle between the static field B 0 and the vector between the spins, J is the exchange coupling, and dd is the dipolar coupling between the electrons. Equation 2 3 4 is valid as long as the position s of the electron spins are relatively well

PAGE 95

95 defined in relation to the distance between t hem (i.e. the point dipole approximation holds). This restriction is easily met for spins more than 15 apart, which is the lower limi t for a DEER experiment. The J coupling is significant only at smaller distances and is considered negligible for dista nces greater than 20 131 As with three pulse DEER, 1 ) is the observer frequency (co rresponding to A spins) and 2) is the pump frequency (B spins). Both frequencies are chosen such that there is no overlap (or minimal overlap) between the excitation windows of the pulses and that the most spins are exc ited. This condition is relatively simple to achieve for nitroxide spin labels because the low field and center field transitions are >26 G a part (which corresponds to ~72 MHz) as illustrated in Figure 2 8 Typically, the pump frequency is chosen to cor respond to the center field transition because it is the most populated region of the spectrum and the observe frequency is selected to correspond to the low field transition because it is the second most populated region Figure 2 8 Absorption spect ra for a nitroxide spin label with the low field transition marked as the o bserve frequency and the center field transition marked as the pump frequency. of the spectrum (that is at least 26 G from the center field transition). These positions can be re versed, to produce a stronger signal. However, pumping on the larger population leads to

PAGE 96

96 deeper oscillations in the dipolar evolution curve. 131 The presence of deeper oscillations improves the quality of the analysis (these topics w ill be discussed in more detail later). The theory behind DEER is typically explained by considering a single macromolecule, e.g. a protein, with two spins an A spin and a B spin. However, the reality is that there are a large number of macromolecules in the sample tube, most containing two spin labels but some containing only one label due to incomplete spin labeling. Thus not every protein will have both an A spin and B spin. Inspection of Figure 2 8 reveals that the majority of spins are neither A n or B spins (the portions of the spectrum not highlighted in grey) Thus most proteins will contain neither A n or B, some will contain only an A spin some will contain only a B spin and a small percent will contain both A and B spins This has two sign ificant consequences. First, only ~5 % of the sample will contain both A and B spins 131 The percentage can be determined by calculating the bandwidths of the pump and observe pulses determining the number of spins excited by each pulse and calculating the statistical probability of an A spin and B spin residing on the same macromolecule. These calc ulations assume that the sample was labeled with 100% efficiency. Incomplete labeling can drastically reduce the percent age of the sa mple that contributes to the dipolar evolution signal. Second, every A spin will be surrounded by other spins on other macromolecules, some of which will be B spins. These intermolecular interactions giv e rise to the background signal, which comes from a random distribution of large distances and thus takes the form of an exponential decay. The signal from the intramolecular interactions takes the form of a damped oscillation as illustrate d in Figure 2 9 The raw dipolar evolution curve is shown as the s olid black line. This

PAGE 97

97 signal is usually designated V(t). The background contribution is plotted as a dashed grey line and is represented by B(t). The background corrected signal, F(t), is plotted as a solid grey line. Figure 2 9 Sample dipolar ev olution curve s before and after background subtraction The relationship between the signal, the background, and the background corrected signal is given by Equation 2 3 5 The decay time for the oscillations in F(t), t decay and the maximum dipolar evol ution time, t max are also illustrated in Figure 2 9. The modulation depth, (also (2 3 5 ) represented by in some lite rature), is a correction factor that compensates for the incomplete excitation of all B spins by the pump pulse. As illustrated in Figure 2 1 0 the frequency and decay rate of the oscillations depend on the length of the most probable distance and the brea dth of the distance distribution respectively. By varying the breadth of a distance profile centered at 36 (Figure 2 10A) from 1 to 10 and generating the theoretical dipolar evolution curves (Figure 2 10B), it can be seen that the

PAGE 98

98 narrowest distributi ons have the most well defined oscillations, corresponding to the longest decay rates. The frequency of oscillations can likewise be illustrated by comparing the dipolar evolution curves (Figure 2 10D) corresponding to distance profiles (Figure 2 10C) tha t have the Figure 2 1 0 Effect of the breadth of the distance profile and the most probable distance on the dipolar evolution curve. same breadth (7 ) and vary in the most probable distance from 18 to 78 Th ese dipolar evolution curves have differ ent decay rates but because the frequency of the oscillations also changes, the curves have th e same number of oscillations before being completely damped. The inset in Figure 2 1 0 D highlights the dipolar evolution curves that decay within the first 2 s a t max frequently reported in the literature. These curves are plotted as solid lines and correspond to center distances of 36 or less. The dipolar evolution curves corresponding to center distances larger than 36 are plotted as dashed lines and do not decay within 2 s.

PAGE 99

99 Because the acquired signal V(t) is the combination of both the intramolecular signal F(t) and the background signal B(t), it is useful to be able to separate B(t) from V(t). If the V(t) is collected such that t decay is much shor ter than t max then the portion of the curve between t decay and t max will contain only contributions from B(t) and can thus be fit to the appropriate function to determine B(t). Alternatively, the ln ( V(t) ) can be used in an identical fashion to determine ln ( B(t) ) Strictly speaking B(t) should be an exponential function of the form in Equation 2 3 6 (2 3 6 ) where D is the dimen sionality of the background typically three dimensions for soluble proteins and two dimension s for membrane proteins. However, ln ( B(t) ) can be described as a simple low order polynomial such as the one in Equation 2 3 7 where q is the order of the polyn omial. (2 3 7 ) The distance limitations for DEER experiments are frequently cited as being from 15 to either 60 or 80 The lower limit arises from the requirement that the excitation bandwidth should exceed the electron electron coupling which can be met only for distances at or above 15 The upper limit is determined by the T m of the system. For nitroxide radicals that can be dissolved in a variety of organic solvents, the T m can be extended by using a deuterated solvent with no methyl protons, such as o terphenyl. Using this solvent, Jeschke et al. 131 were able to collect a dipolar evolution curve with t max = 24 s for a shape persistent biradical with an interspin distance of 75 For biological molecules, however, the solvent is typically restricted to aqueous solutions. Furthermore, the biological molecule typically contains many methyl protons which contribute to shorter T m values. These restrictions typically limit the T m to less

PAGE 100

100 than 5 s, although 3 4 s is much more common. These T m values correspond to distances of 60 or les s with 45 50 being a more practical limit. In gen eral, t max should be twice t decay so that the second half of the dipolar evolution curve can be fit to B(t). This minimum t max can be determined using Equation 2 3 8 131 and is plotted as (2 3 8 ) the solid line in Figure 2 11. The dashed line corresponds to the minimum t max at a given distance if detailed information is desired from the distance profile s. Figure 2 1 1 Minimum t max for a given interspin distance based on Equation 2 3 8 Experimental Consideration s for Pulsed EPR Instrumental Requirements A large variety of resonators have been employed in ESE experiments. The resonator (or cavity) is responsible for converting the microwave power into the B 1 field necessary for flipping the spins. There are three critical aspects of the resonator s performance for pulsed EPR experiments. First, it must efficiently convert the microwave power into th e largest B 1 field possible. Second, t he resonator must be able to handle the large bandwidth of microwave frequencies necessary to cover the difference in frequencies between the pump and observe

PAGE 101

101 pulses. Finally, the resonator mu st have a short ring dow n time i.e., it allows the microwave power to dissipate quickly so that the detector, which must be protected from the microwave pulses, can be turned on in time to capture the small signal. A bimodal resonator was used by Larsen et al. 138 so that each mode of the resonator needed to handle only the bandwidth of one pulse. However, this resonator suffered from a low filling factor a measure of the fraction of microwave energy tha t interacts with the sample and thus performed poorly. The commercially available dielectric and split ring resonators are currently the most popular. The dielectric resonators, E N 4118 X MD4 and E R 4118 X MD5 from Bruker Biospin, offer a large filling facto r and variable Q ratio of microwave power store d in the resonator to power lost via heat absorption which provides a high degree of sensitivity and adaptability for a var iety of experiments. The split ring resonators, ER 4118X MS5/3/2 from Bruker Biospin, however, generate the highest B 1 fields and have the largest bandwidths. Both the dielectric and spilt resonators are suitable for use with DEER although each offers distinct advantages. The sample volumes necessary for DEER vary depending on the resona tor. Generally, the signal to noise ratio is highest when the greatest number of spins is in the active area of the resonator. This criterion is met by using the largest sample tubes that will fit in the cavity and filling them with sufficient sample so that the active area is full. For a 4 mm (outer diameter) tube, this corresponds to ~100 L of sample. Spin Relaxation One of the major limiting factors in any pulsed magnetic resonance experiment is the time required by the magnetization to return to the rmal equilibrium. S pin relaxation is characterized by two time constants: T 1 or spin l attice relaxation time (or longitudinal relaxation) which is the

PAGE 102

102 time constant for relaxation between different Zeeman energy levels and equivalently the relaxation of the bulk magnetization along the z axis ; and T 2 or spin spin relaxation time (or transverse relaxation) which is the time constant for relaxation in the x y plane. There are also several other processes that affect the relaxation times. For example, T D or the spectral diffusion time is the time constant during which the spin moves between positions in an EPR spectrum. ( If an excited spin diffuses outside of the spectral window, then it appears to have relaxed. ) There is also the spin echo dephasing t ime, T m which includes all processes that lead to a loss of electron spin phase coherence (which includes T 2 ) In addition, there are the nuclear spin diffusion rate which is the rate of mutual nuclear spin flips and cross relaxation which is the mutu al spin flip of two unlike spins. The two relaxation times that will have the greatest impact on the ESE experiments are T m and T 1 T m At first glance, it might seem that T 2 should play a more prominent role in pulsed EPR experiments. However, the spin e cho dephasing time, T m which includes T 2 is the more relevant term because it encompasses all process es that a ffect the refocusing of spins into an echo, not just the time constant for spins dephasing in the x y plane. T m is highly temperature depend e nt and is the reason that ESE experiments are preformed at cryogenic temperatures where the T m is sufficiently long to accommodate a pulse sequence. The magnitude of the temperature dependence depends on the relaxation mechanism that is dominant in the ra nge of temperatures involved Two major contributing mechanisms to T m are instantaneous diffusion and nuclear spin diffusion. Instantaneous diffusion contributes only at high spin concentrations and results from incomplete excitation of the entire spectr um. Nuclear spin diffusion results from the flipping of the nuclear spins which are coupled to the electron spins. The simplest method for minimizing the effect of nuclear spin diffusion on the T m is to

PAGE 103

103 replace neighboring nuclei (usually in the solvent, but can also include isotopically labeling a protein) with nuclei that have smaller magnetic moments. Replacing protons with deuterons can extend the T m by almost 50 fold. T m is also dependent upon the solvent used. Solvents with methyl groups display T m curves with a different shape which is indicative of contributions from an additional relaxation process. One potential explanation for this could be that methyl group rotation can persist down to a few K 139 leading to an averaging of the couplings between the spin and the three protons (which is a dephasi ng process) Since any two pulses will create an echo, the spin dephasing can be measured by simply increasing the spacing between the two pulses and measuring the echo intensity as a function of that spacing (Figure 2 1 2 A) This simple method is typica lly referred to as an echo decay experiment because the echo intensity decays as the pulse spacing increases (Figure 2 1 2 B). This experiment is also used as a way to estimate the effective T 2 of a system with the understanding that it is not the true T 2 value. The data collected are typically fit to a stretched exponential of the form in Equation 2 39 where T m is the rate of decay and A, t 0 B, y 0 are the (2 39 ) constants to account for the initial intensity, time offset, the extent of stretching for the exponential function, and value of y as t approaches infinity respectively. The data in Figure 2 1 2 B were fit to Equation 2 39 with T m val ues of 3.8 s, 5.1 s, and 10.8 s for the same sample in H 2 O with 30 % glycerol, H 2 O with 30% deuterated glycerol, and D 2 O with 30% deuterated glycerol, respectively.. There is no standard rule that directly relates the measured T m to the longest accept able t max However, the t max needs to be less than T m or the data collection time becomes inordinately long.

PAGE 104

104 Figure 2 1 2 Pulse sequence for an echo decay experiment and T m curves for HIV 1 protease in various buffer conditions. A) Echo decay pulse sequence. B) Intensity normalized T m curves for HIV 1 PR in 2 mM NaOAc buffer pH 5.0 with H 2 O and 30% glycerol (black), 2 mM NaOAc buffer pH 5.0 with H 2 O and 30% deuterated glycerol (dark grey), 2 mM NaOAc buffer pH 5.0 with D 2 O and 30% deuterated glycer ol (light grey) with corresponding fits to Equation 2 39 This limitation on the length of the dipolar evolution curve also limits the amount of information collected from the experiment. As was discussed previously, if the dipolar evolution curve does n ot capture a sufficient number of oscillations, then the quality and resolution of the distance profile will be compromised. Thus efforts to extend the T m can improve the quality of the results garnered from a DEER experiment. The simplest way to extend the T m is through the selection of the matrix (solvents and cosolutes). The better solvents (and buffers) are the ones without methyl groups. Likewise, using deuterated water and buffers will also help dramatically as illustrated in Figure 2 1 2 B The s election of the cryoprotectant can also make a large difference. The use of glycerol extends the T m to a greater extent than other solutes including Ficoll400, sucrose, and various sizes of PEG. However, it should be noted that changes to the solvents an d solutes will only impact surface accessible sites. If the labeling site is buried within the protein or if the site is buried with in the hydrophobic portion of membrane, then the T m extension achieved by switching to D 2 O will be minimal.

PAGE 105

105 T 1 T 1 is the time constant for a spin to establish equilibrium between electron Zeeman energy levels. Because the m s value changes as transitions are made between levels, the energy of the system must also change by absorbing energy from or emitting energy to the latt ice According to the Bloch equations, T 1 is defined as the time constant for the z axis component of the bulk magnetization to return to equilibrium after the pulse These definitions are equivalent because the z axis component of the magnetization resu lts from population differences in the Zeeman energy levels. Theoretically it should be straightforward to measure T 1 experimentally H owever, there are other processes (primarily spectral diffusion and spin diffusion) that can take a spin off resonance with out changing the populations of the energy levels thus contributing to the apparent T 1 of the system. The parameter in the DEER experiment that depends on T 1 is the delay between pulse sequences usually referred to as the shot repetition time (SRT) Typically th e SRT is set to be five times the value of T 1 to allow ~99% of the magnetization to relax back to the z axis. The effective T 1 can be measured using an inversion recovery experiment although there are other options as well including satur ation recovery 134 or stimulated echo decay 132 As illustrated in Figure 2 13, i nversion recovery uses a three pulse sequence to measure the rate of recovery The first pulse inverts the magnetization which is then allowed to relax fo r a time t which is incremented A spin echo is then formed using a /2 pulse followed by a pulse. The intensity of the echo will change as a function of t This experiment is very straightforward to perform and provides a convenient measure of the e ffective T 1 H owever, it is very susceptible to spectral diffusion because of the short inversion pulse (the inversion pulse can be lengthened but the decreased spectral excitation window will decrease the signal

PAGE 106

106 strength). The data from an inversion rec overy experiment can be fit using either single or multiple exponential decays. Fitting to a single exponential (Equation 2 4 0 ) is usually sufficient for estimating the effective T 1 (2 4 0 ) for determining the appropriate shot repetition time for a pulsed experiment, although the double exponential frequently provides the superior fit. Figure 2 1 3 Inversion recovery pulse sequence and corresponding sample data. Similar to T m T 1 is strongly temperature depend e nt with different mechanisms dominating in different temperature ranges. These various mechanisms and their temperature dependence were reviewed thoroughly in Berliner et al 134 In general, the T 1 value will increase as the temperatu re decreases although the extent to which T 1 changes will depend on the dominating mechanism. T 1 is also solvent dependent Hydrogen bonded s olvents have longer T 1 values than solvent s with little to no hydrogen bonding. This results from the decreased molecular motion in a solvent system with more hydrogen bonding (for glassy systems < 200 K).

PAGE 107

107 Cryoprotectants and Glassing Agents The addition of a solute to a protein sample can have several benefits, including reducing the likelihood of the sample tub e cracking upon being frozen. One of the most commonly cited reasons for adding the solute is its function as a glassing agent. A glassing agent is an agent (is this case, solute) that reduces the glass transition temperature so that the sample remains in a glass state throughout the experiment. The key benefit to keeping the sample in a glass state is the prevention of protein aggregation during the process of the sample freezing (W ater tends to reject impurities as it freezes, thus concentrating the impurities which includes salts and proteins). An increase in the effective protein concentration is the decrease in the distance between neighboring molecules which can complicate the analysis for measuring the intramolecular distances. Glycerol is by far the most common glassing agent and cryoprotectant used, although many other solutes can be used, including sucrose, ethylene glycol, and PEG. The minimal requirement for a solute to be a glassing agent is that it reduce s t he glass transition temperatu re. H owev er, it is also important that the solute does not significantly alter the conformational ensemble of the protein. A simple way to check for this is to perform DEER experiments using a variety of solutes to see how the distance profile varies for each solute or to perform other biophysical characterizations in the presence of the solute. Temperature Selection The t max of the DEER experiment is limited strongly by the T m which in turn is strongly dependent upon temperature as previously discussed The ideal temperature is the one at which the T m is most strongly affected by the spin diffusion of nuclear spins as opposed to being dominated by the modulation of the hyperfine or g tensor resulting f rom molecular reorientation 130 For nitroxide radicals in aqueous buffers, this point is usually at or below 80 K.

PAGE 108

108 Most research groups typically choose temperatures between 55 K and 80K. Although the T m will continue to in crease at lower temperatures, the T 1 will also increase which lengthens the necessary experiment time Sample Concentration Typically, the simplest way to improve the signal to noise ratio (SNR) in an experiment is to increase the sample concentration. Unfortunately, increasing the spin co ncentration in ESE experiments does not always translate into greater sensitivity The most intuitive limitation is that a t high concentrations, the distance between neighboring molecules is reduced, leading to increased contributions to the signal from i ntermolecular dipolar couplings which complicate the mathematical separation of the two contributions. However, t here is an additional concentration limitation imposed by the phase memory loss associated with instantaneous diffusion at high concentration s 131 For most soluble proteins, these concentration limitations will not apply as the protein will most likely aggregate before these concentrations are reached However, t he limitations are relevant for membrane proteins becaus e the localization of the protein in the membrane drastically increases th e local effective concentration Figure 2 1 4 shows the plots of the concentration limits as a function of the inter spin distance for both the intermolecule distance restrictions (do tted line) and the instantaneous diffusion restrictions (solid lines). The intermolecular distance restriction assumes that the spin labels are on the surface of a protein with a 60 diameter. The optimal concentration based on the instantaneous diffusi on restriction (black line) is calculated from Equation 2 41 131 where t max can be expressed in terms of r AB using Equation 2 38 and f A, =0.25 (for a 32 ns observer pulse). For extracting information about the shape of the profile for narrower profiles, the t max should be three times as long, corresponding to a concentration one third as large (grey line).

PAGE 109

109 (2 41) Figure 2 1 4 Plot of the maximu m concentration of spins as a function of the inter spin distance. Solid black line corresponds to the restriction imposed by instantaneous diffusion. Solid grey line corresponds to the restriction imposed by instantaneous diffusion for minimum concentra tion when attempting to recover detailed information from the distance profile. Dashed line corresponds to the restriction imposed by intermolecular distances. Analysis of DEER Data For a complete understanding of how to properly collect and analyze dat a from a DEER experiment, the reader is directed to a number of excellent sources 130; 131; 140; 141; 142; 143 including (available online at www.epr.ethz.ch ) for those wishing to analyze their data with this software. This section will cover the material in sufficient detail to understand the process and ramifications of certain choices on the results. Converting the Dipolar Evolution Curve into a Distance Profile The dipolar evol ution curve is the manifestation of the additional modulation imposed upon the A spins by their coupling with the B spins. The frequency of this additional modulation can

PAGE 110

110 be determined in a variety of ways. Most simply, it can be Fourier transformed into a frequency domain spectra where the splitting between the singularities in the P ake pattern is proportional to 1/r 3 where r is the interspin distance. However, this only gives an estimate of the most probable distance and works better for narrower dista nce distributions. To extract a distance profile from the dipolar evolution, a more sophisticated technique is required. A handful of these techniques are currently employed but each has shortcomings in one aspect or another. The source of difficulty in converting the dipolar evolution curve into a distance profile is that there is no unique distance profile. Significantly distinct distance profiles can give rise to strikingly similar dipolar evolution curves. Additionally, the solution can be non trivi ally influenced by small amounts of noise in the data Several groups have come up with ways to either deal with the ill posedness or to circumvent it. These solutions typically involve solving the inverse problem, which is finding a distance profile tha t satisfies the experimental data. Curve Fitting Approaches A common approach to finding the best distance profile involves the use of curve fitting to optimize the solution There are many variations of this method, but all include a modeling of the dist ance profile based on current knowledge of the system and generating the corresponding theoreti cal dipolar evolution curve for comparison with the experimental data. The process involves changing the distance profile to optimize the fit between the theore tical and experime ntal dipolar evolution curves. One variation on this approach utilizes Monte C arlo (MC) methods to generate a distance profile with an assumed form, such as a Gaussian or Lorentizian shape by combining random functions. T he theoretical a nd experimental dipolar evolution curves are compared and the process repeated until the fit between the curves meets the pre defined level These approaches can yield high quality feasible results but have the limitations that they are either model

PAGE 111

111 depen dant or assume a form for the distance profile, which can impose an arti ficial symmetry on the solution. Tikhonov Regularization Method Tikhonov regularization (TKR) 144 is a mathematical method used for solving an ill posed problem by introducing a penalty for smoothness. TKR u ses the function in Equation 2 4 2 to balance the quality of fit to the experimental data (first term) with the smoothness of the solution (2 4 2 ) (second term) by vary ing the magnitude of the regularization parameter (also denoted in some literature), where P is the probability distribution of the interspin distance, K is an operator that maps the function P onto S, S is the experimental data vector, L is an operator, usually the identity of second derivative operators, and is the regularization parameter DeerAnalysis Software Package The DeerAnalysis software uses a combination of shell factorization ( to simulate the dipolar evolution cur ves ) and Tik h onov regularization (to optimize the solution) to convert dipolar evolution curves into distance profiles. It also includes several other options for data analysis including approximate P ake transformation (APT) and model fitting (to either G aussian functions or user defined functions). There are also a variety of options for background corrections. L c urve The L curve is the v isual result of the TKR process (Equation 2 4 3 ) where the log of ( ) (Equation 2 4 4 ) is plotted against the log of ( ) (Equation 2 45 ). As illustrated in Figure 2 1 5 A, e ach point on the curve corresponds to a particular value of the regularization (2 4 3 )

PAGE 112

112 (2 4 4 ) (2 45 ) par ameter. For small G(P) is dominated by ( ) the quality of the fit between the experimental dipolar evolution curve and the theoretical curve. As increases, the contribution by ( ) increases more than the contribution from ( ) so the L curve has a steep slope. At large values of ( ), G(P) is dominated by ( ) the smoothness of the solution. As increases, the contribution from ( ) is minimal because the solution is already smooth so the contribution from ( ) dominates and the slope is flat. There is a point between these two regime s where the slope changes. This point corresponds to the optimal solution because it is the point where the smoothness of the solution and the quality of fit are balanced. The value that corresponds to this poin t is the optimal value of regularization parameter and the corresponding distances profiles and dipolar evolution curve fit are the optimal distance profile and fit. Figure 2 1 5 Example of an L curve and the corresponding distance profiles and dipolar evolution curves.

PAGE 113

113 Figure 2 1 5 also illustrates the effect of non optimal values on the distance profile and the TKR fit. The distance profile in Figure 2 1 5 B is under smoothed and corresponds to a value that is too small. The corresponding dipolar ev olution curve in Figures 2 1 5 C is over fit, meaning that the theoretical dipolar evolution curve was fit to some of the noise. The distance profile in 2 1 5 D is qualitatively similar to the distance profile in Figure 2 1 5 B in that the most probable distanc e is the same and the breadth at the base of the first two peaks in Figure 2 1 5 B is only slightly smaller than the breadth of the peak in Figure 2 1 5 D. However, the distance profile in Figure 2 1 7 D is smoother and corresponds to the optimal value The dipolar evolution curve in Figures 2 1 5 D is optimal. The distance profile in Figure 2 1 5 F is over smoothed and thus overly broad. The corresponding dipolar evolution curve in Figures 2 1 5 G is under fit, meaning that some of the oscillations in the signal are neglected in the TKR fit. Approximate Pake transformation The approximate Pake transformation (APT) uses a discrete integral transformation analogous to the Fourier transformation that has been corrected for cross talk to convert the dipolar evolution curve to a frequency domain spectrum. The distance profile is extracted from the shape of the frequency domain spectrum. This method is very quick computationally also for a quick estimation of the distance profile while determining the optimal backgrou nd subtraction. However, APT frequently includes negative probabilities for some distances in the distribution, which lesse ns the overall accuracy of the profile. Model f itting In some cases, the use of a model based analysis can provide better results th an a model free analysis. For example, a semi rigid polymer with spin labels at the end will not give a Gaussian distribution of distances. However, the form of the distribution has been defined accurately in

PAGE 114

114 the worm like chain model. F itting a distrib ution to this model uses only two degrees of freedom the persistence length, L p and the end to end distance L instead of the large number of degrees of freedom associated with the model free analysis the number of r values for which P(r) is defined. This reduction in the degrees of freedom reduces artifacts from noise but it can also impose a shape on the distribution profile that may not be valid. Care must be used to ensure the model accurately represents the system. The DeerAnalysis software includes a module to perform such analyses. The included models include fitting a one or two Gaussian functions, the surface of a sphere, a random coil, a semi rigid worm like chain, and a semi rigid worm like chain that includes a Gaussian distribution of the lab el. Users can also define their own models by following the template in the Background c orrection o ptions The dipolar evolution curve is described by V(t) which is a function of t he dipolar evolution time t. V(t) is the p roduct of the two functions shown in Equation 2 51 F(t) is the form factor analogous to the form factor in scattering techniques and B(t) is the background factor. (2 51 ) Likewise, the spin spin interactions can be divided into intramolecular interactions between spins on the same macromolecule and intermolecular interactions between spins on different macromolecules. T he distance between the intram olecular spins should be shorter than the distances between the intermolecular spins. The only exceptions are for very large intramolecular spin spin distances or very concentrated samples where the intermolecular distance is short. Thus F(t) corresponds to the intramolecular spin spin interactions and B(t) corresponds to the intermolecular spin spin interactions.

PAGE 115

115 It follows that F(t) and B(t) will have different forms because F(t) corresponds to the distribution of distances between labels on a macromolecule while B(t) corresponds to the distribution of the macromolecules. As a result, many algorithms require the separation of B(t) to correct ly analyze F(t). The DeerAnalysis software comes with three modules for subtractin g the background. One option i s to experimentally determine the background contribution by perfor ming an identical experiment with singly labeled sample. However, this cannot be p er formed for all samples, in particular polymers and oligomeric proteins. The other two modules include fitting the background to a simple polynomial of various degrees and fitting the background to an exponential function corresponding to a homogeneous background with variable dimensions which is particularly useful for membra ne samples where the background signal is restricted to two dimensions Subtracting out the correct level of background contribution to the dipolar evolution curve is not trivial. Because the background contribution results from intermolecular spin spin interactions, the corresponding distances are large and the effect of the background contribution is distributed over the whole dipolar evolution curve. The intramolecular spin spin interactions which should be shorter distances than the intermolecular i nteractions contribute only high frequency oscillations to the dipolar evolution and thus strongly affect the dipolar evolution curve at small t values. As long as the dipolar evolution curve is collected such that t max > 2 t decay the background can be d etermine from the portion of the curve at t > t decay Validation m odule Because the error in TKR does not propagate in a manner that can be predicted analytically, the error must be determined by other means The DeerAnalysis2008 includes a Val idation m odule that estimates the error by determining distance profi les for all the combinations of the

PAGE 116

116 variables that contribute to the error background density, modulation depth, and background dimensionality. The user selects the upper and lower limits for eac h variable and the number of data points to test between limits and the Validation module determines the distance profile and corresponding dipolar evolution curve for each combination of variables. Because of the large number of calculations, this proces s is typically slow. The module contains two options for narrowing down the number of reasonable distance profiles. The user can trim out distance profiles with corresponding dipolar evolution curves that are a poor fit to the experimental data. The use r can also discard distance profiles with populations at large distances as these profiles tend to correlate with incorrect background. The remaining profiles can then be compared to give an estimate of the error in the distance profile. Figure 2 15 disp lays a sample image of a distance profile with error estimates. Verification of Background Subtraction and Gaussian Reconstruction In addition to the processes included in the DeerAnalysis software package, we developed additional steps in the analysis to further refine our results and to extend the analysis into conformational ensemble analysis. These steps include manual selection of the zero point in the dipolar evolution spectra, a self consistent analysis of the background subtraction level, and a Gau ssian function based fitting of the distance profiles called Gaussian reconstruction. Zero point selection In pulsed magnetic resonance experiments, data is frequently collected as a function of t which is the spacing between select pulses in the sequence In cases where t should vary between 0 t max t actually starts at a small negative time so that the data can be corrected for any discrepancy between the instrumental zero time and that actual zero time. The correction utilizes the symmetry in the dipo lar evolution curve the curve is symmetric about the zero point to mathematically find the true zero point.

PAGE 117

117 The DeerAnalysis software contains a subroutine that uses a first moment analysis for identifying the zero point and shifting the data accordingly 142 The first moment analysis utilizes the fact that the zero point should be a local maximum and that the function should be symmetric about the maximum. T hus the first moment of small range of distance center ed on the zero point should be zero. Accordingly, the data near t=0 are divided in small overlapping regions and the center of the region with the smallest first moment is selected as the true zero point. This subroutine has been tested and is reasonably r obust for accurately correcting the data except in datasets that are negatively phased, e.g data that has been phased such that the DEER echo intensity is negative. Since a large portion of the data in this work was collected in this fashion, it was nec essary to manually find the zero point and correct the data accordingly. To accomplish this correction, the raw dipolar evolution curve s are imported into Origin7.5 or Origin8.0. Both the first derivative method and fitting to a Gaussian function have bee n tested on several datasets. It was found that fitting all the data points < 400 ns to a Gaussian function and using the center of that Gaussian function for the zero point is more consistent than taking the first derivative of the data and using the poi nt where the first derivative was zero. This was largely due to the small amount of noise in the data near the zero point in some datasets that skewed the results. The zero point determined is manually entered into the DeerAnalysis software as the zero p oint. Self consistent analysis of the background subtraction level In a DEER experiment the background results from intermolecular interactions the dipolar interactions between spin labels on different protein molecules. The most accurate analysis of a di polar evolution curve depends on having the most accurate background subtraction possible. The DeerAnalysis software package utilizes the APT module to optimize the background subtraction as was discussed previously. Although the APT generates a good esti mate of the

PAGE 118

118 correct level of background to subtract, it a ctually tends to generate a small range of acceptable background It can be shown that the typically small variations in the background subtraction levels can lead to small differences in the distan ces profiles. These small differences typically do not affect the most probable distance in the profile, nor do they affect the full width at half maximum (FWHM). The small differences are typically manifested in the fine features of the distance profile s which correspond to lowly populated states in the protein. This variation in the fine features is problematic because it is these lowly populated states that are of interest. The best solution is to identify the most accurate level of background subtra ction which should correspond to the most accurate distance profile. We have dev eloped a self consistent method to analyze the level of background subtraction and identify the most accurate level. As illustrated in Figure 2 16, t he first step is to analyz e the data using the DeerAnalysis software with the recommended level of background subtraction. This will generate several output files including a distance profile and a background subtracted dipolar evolution curve. The distance profile is input into the DeerSim software package which has been incorporated into the DeerAnalysis software for ease of use. The DeerSim uses the distance profile to generate a corresponding theoretical dipolar evolution which is completely free of any background contributio ns. This theoretical curve is compared with the background subtracted curve (also called the TKR fit) from DeerAnalysis. If the two curves do not overlay exactly as illustrated in Figure 2 16 then the background subtraction was not correct and the proce ss needs to be repeated beginning with a slightly different background subtraction in DeerAnalysis. Th ese steps are repeated until the TKR fit and theoretical dipolar evolution curve overlay exactly.

PAGE 119

119 Figure 2 16 Illustration of the Self Consistent A nalysis process developed to optimize the background subtraction in the DEER dipolar evolution curves. Steps e xplained in detail in the text.

PAGE 120

120 If necessary, this process can be modified to probe the validity of different populations within a distance pro file. For example, if the level of background subtraction is too low, meaning that not enough background was subtracted, then populations can appear at distances too long to be relevant to the system. These populations can be suppressed in the distance p rofile that is used to generate the theoretical echo. In this case, the theoretical curve will not overlay with the TKR fit so the level of background subtraction selected in DeerAnalysis would be adjusted accordingly. If the resulting theoretical curve and TKR fit overlay, then the population at the large distance can be assigned to background and can be neglected. This process is extremely helpful in cases where t max is too short to capture more than one or two oscillations or in cases where t max is mo derately short but the oscillations are strongly damped from a broad distribution. Gaussian reconstruction The distance profiles of various samples can be compared in a variety of different ways. The earliest papers published discussed the distance profil es in terms of the most probable distance and the FWHM of the profile. This type of analysis is sufficient for demonstrating that a conformational change took place, resolving which direction the change occurred in, and providing insight into the magnitud e of the change. Although this analysis is simple, it was appropriate considering the available methods for converting the dipolar evolution curve into a distance profile. In other words, the distance profiles lacked a high degree of accuracy because of assumptions used in the analysis process, such as fitting to model or forcing an artificial symmetry onto the distance profile. It has been shown that the most probable distance and FWHM are very robust to a variety of assumptions that are frequently used and thus these analyses were sufficiently accurate.

PAGE 121

121 In light of the improvements made in the process of converting the dipolar evolution curves to distance profiles, the accuracy of the distance profiles have increased. Thus it is possible to discuss the distance profile s in more detail especially for distance profiles with fine feature s instead of a featureless peak One option includes fitting the distance profile to a series of Gaussian functions that will sum to regenerate the original profile. Thi s fitting can be accomplished in a variety of ways, including a brute force manual fitting. However, Origin8.0 service release 1 (sr1) and newer contains an easy to use process to fit the profile to any number of Gaussian functions with independent breadt hs and intensities using a variety of peak finding techniques. This is especially useful for finding hidden peaks when one peak has significant overlap with a larger peak and thus appears as one distorted peak. For the work presented in this dissertation the distance profiles were fit to Gaussians using the second derivative method in Origin8.0 sr5 unless stated otherwise. The second derivative method takes the second derivative of the distance profile and assumes th at each local minimum corresponds to a Gaussian function. In general, any Gaussian function corresponding to less than 1 % of the total population was discarded as unnecessary. One caveat to including the fine features of a distance profile in the interpretation of the data is that TKR dis tance profiles are known to include occasional artifacts, which are manifested as small populations. Thus, it is necessary to verify that any population contributing to less than 15% of the total population is actually representative of an actual conforma tion in the protein. This validation can be accomplished by suppressing the population in question and generating the corresponding dipolar evolution curve. If any portion of the theoretical curve lies outside of the noise in the experimental dipolar evo lution curve then the population can be attributed to a protein conformational state. Conversely, if the theoretical curve lies within the noise then the

PAGE 122

122 population can be attributed to part of the background signal and discarded. Occasionally, a theoret ical curve will lie at the edge of the noise in the experimental. In this case, it remains unclear whether the population is real. This process is illustrated in Figure 2 17. The Gaussian shaped populations used to reconstruct the distance profile is s hown in Figure 2 17A and the distance profile is shown in Figure 2 17B. The background subtracted experimental dipolar evolution curve and the TKR fit are shown in Figure 2 17 C The theoretical dipolar evolutions curves with the populations at 20 and 4 2 suppressed are shown in Figures 2 17 D E respectively and the theoretical curve for Figure 2 17 Illustration of the population validation process used to interrogate the validity of population containing less than 15% of the total population A) Gaussian shaped populations used to reconstruct the distance profile. B) Distance profile with (solid) and without (dashed) the 20 population. C) Background subtracted experimental dipolar evolution curve (black), fit from TKR (red), and theoretical cu rve generated by Gaussian reconstruction. D) Same as C with the 20 population suppressed in the Gaussian reconstruction. E) Same as C with the 42 population suppressed in the Gaussian reconstruction. F) Same as C with both the20 and 42 population s suppressed in the Gaussian reconstruction.

PAGE 123

123 both populations suppressed simultaneously is shown in Figure 2 17F.. It can clearly be seen that the theoretical curve for the 42 population is on the edge of the experimental noise and that the theoretical curve for the 20 population is inside of the noise. However, suppressing both populations results in a theoretical curve that lies outside of the experimental noise. Thus, the 20 population can be regarded as an artifact and the 42 can be deemed re al because it cannot be suppressed in conjuction with the 20 population. It should also be noted that the suppression of multiple peaks can have a smaller net effect on the theoretical curve because of cancellation. Essentially, the suppression of the s ingular peaks might be significant but because each peak shifts the theoretical dipolar evolution curve in opposite directions, the result of simultaneous suppression is that the theoretical curve remains in the noise. For this reason, it is also prudent to test all linear combinations of peaks that constitute less than 15% of the total population. Experimental Considerations and the Corresponding Effect on the Results Signal to noise ratio The signal to noise ratio (SNR) has a large impact on the quality and accuracy of the distance profile. In a dataset with low SNR, the oscillations in the noise mask oscillations in the data. This effect can clearly be seen by comparing the L curves for identical data sets with different levels of noise. Figure 2 1 8 s hows the L curve for a data set with a SNR of 28 Each of the three distances profiles correspond to values that are too low (under smoothed), optimal, and too high (over smoothed). The arrows indicate the location of each value on the L curve. Addi tionally, the background subtracted experimental dipolar evolution curve (thin grey line) is shown overlaid with the three TKR fits (thick black line) corresponding to the values. Figure 2 19 shows same data with a SNR of 14.

PAGE 124

124 Figure 2 18 Example o f an L curve and the corresponding distance profiles and dipolar evolution curves with a high SNR. Figure 2 19 Example of an L curve and the corresponding distance profiles and dipolar e volution curves with a low SNR A comparison of the L curves reve als that the lower SNR has a strong effect on the shape of the L curve. The corner is obscured making it more difficult to identify the optimal value. This distortion primarily occurs at small values where the curve is dominated b y changes in log ( ) ( the smoothing function) with only small changes in log ( ) (the fit function). The increased noise results in larger changes in log ( ) thus distorting the curve.

PAGE 125

125 A comparison of the TKR fits for both SNR reveals that the oscillations in the lower SN R data are more damped, which translates into broader distance profiles, for values. Furth er more, the TKR fits for the lower SNR contain additional features, such as the shoulder at 0.25 s that is present in all three TKR fits and the shifted minimum n ear 1.0 s for the lowest value. These features contribute to the additional populations seen in the distance profiles. It should be noted that these additional populations are larger than the populations arising from artifacts in the TKR process. Add itionally, the population validation method used to test the validity of these populations is not reliable if the SNR is too low because the suppression of populations will not shift the theoretical dipolar evolution curves outside of the experimental nois e. Maximum dipolar evolution time, t max The maximum dipolar evolution time, t max as was discussed previously, is strongly limited by the T m of the system. As t max increases, the length of the pulse sequence increases and the echo intensity decreases. Th is decrease in the signal requires that data collection times must be extended or the SNR will decrease. Figure 2 20 A shows two Gaussian shaped distance distributions centered at 48 with FWHM values of 1 and 7 The corresponding theoretical dipolar evolution curves are shown in Figure 2 20B. These theoretical curves were analyzed by TKR with t max values of 10, 4, and 2 s. White noise was convoluted onto the curves using the white noise generator in Origin8.0 to achieve a SNR of 40. These curves are shown in Figures 2 20C E with the corresponding distance profiles in Figures 2 20F H. The most probable distances for all curves were 48.00 0.01 The breadths for the 7 FWHM distance profiles were 7.0 0.1 However, for t max = 2 s, the breadth of the 1 FWHM profile was broadened to 6.8 The distances profiles for the 1 FWHM curves were more accurate for

PAGE 126

126 t max = 4 and 10 s with breadths of 1.4 and 1.1 respectively. These results clearly demonstrated that the accuracy of distance profile breadth depends strongly on collecting the dipolar evolution curve to a sufficiently large t max value. Figure 2 2 0 Illustration of impact that the length of t max has on the resulting distance profile. A) Two Gaussian shaped distance profiles centered at 48 with FWHM = 1 (solid), 7 (dashed) and B) the corresponding theoretical dipolar evolution curves. C E, and G ) The same echo curves as in (B) truncated to 10, 4, and 2 s respectively with random SNR of 40 convoluted onto the curve and D E, and F ) the distance profiles resulting from analysis of the echo curves in (C) (E), and (F) respectively with DeerAnalysis2008. Although it is unlikely to encounter distance profiles as narrow as 1 in a biological system, these results illustra te the extreme effect that short t max values can have on the breadth of a distance distribution. The dipolar evolution curves at t max = 2 s contain less than one half of an oscillation while the curves at t max = 4 s contain almost two full oscillations. Thus, it is advisable to collect experimental dipolar evolution curves with sufficiently long t max such that at least two oscillations are captured.

PAGE 127

127 Spin labeling efficiency Poor spin labeling efficiency can drastically increase the length of time needed to collect the data to a desired SNR. As illustrated in Figure 2 8, the majority of the spins are not affected by either the pump or observe pulses and can are designated C spins Thus, the majority of the proteins will contain non productive spin pair combinations such as A A, B B, A C, B C, and C C. Because the dipolar evolution signal only results from proteins that have both an A spin and B spin, only a small fraction of the sample contributes to the DEER signal. If the spin labeling efficiency is low then the amount of singly labeled protein increases. This in turn decreases the number of proteins with both A and B spins which in turn decreases the signal. It is not necessary to quantify exactly what the spin labeling efficiency is, but it is wor th the effort to optimize the labeling protocols to ensure that the protein is labeled with the greatest efficiency possible. There is no standard set of conditions that will result in maximal labeling for all sites on all proteins, but the pH and tempera ture can usually be optimized for a protein in general based on the stability of the protein. In general, the labeling reaction will proceed faster at higher temperatures however, not all proteins are sufficiently stable at room temperature and thus must be labeled at colder temperatures. Likewise, the labeling reactions tend in proceed faster at higher pHs however the reaction also tends to be less specific at higher pHs. Thus is necessary to confirm that only the desired sites were labeled. Solvent ex posed sites will naturally be easier to label than buried sites. It is possible to increase the labeling efficiency of buried sites by adding a small amount of denaturant to allow the protein to unfold a little thereby increasing solvent and thus spin lab el access to the interior of the protein. Alternatively, the protein can be labeled in the unfolded state and then refolded to the native state if the necessary folding conditions for the protein are known and if the presence of the spin label does not in terfere with folding significantly.

PAGE 128

128 CHAPTER 3 D ISTANCE MEASUREME N T S FOR HIV 1PR SUBTYPE B Introduction As discussed in Chapter 1 the flaps of HIV 1 protease ( HIV 1 PR) play an important role in the function of the enzyme It has been shown through seve ral experimental methods that the flaps are inherently flexible and that this flexibility is necessary for proper func tion. T he flaps are characterized by three major conformations closed, semi open, and wide open. Ishima et al 73 summarized the flap conforma tions in solution as being an ensemble of conformations, which are predominantly semi open but can also include small populations of open and closed forms. Results of r ecent molecular dynamics (MD) simulations 50 o f HIV 1 PR are consistent with the NMR results. Experimental Design This chapter describes the characterization of the conformational ensembles of HIV 1 PR the distances between two spin labels on the flaps by four pulse double electron electron resonance ( DEER ) experiments. Protein constructs containing the lysine (K) to cystei ne (C) mutations at position 55 ( K55C ) in the flaps were expressed purified, and label e d with methanethiosulfonate spin label ( MTSL ) It has been demonstrated that HIV 1 PR is to lerant to amino acid substitutions at the K55 position so the substitution to C ys and the incorporation of the spin label should not significantly alter the structure of the HIV 1 PR n or interfere with the activity of the HIV 1 PR A comparison of t he di stance between the terminal amine nitro gen atoms in the available x ray structures of HIV 1 PR at K55 varies from 25 to 36 which is well suited for DEER experiments.

PAGE 129

129 Previous Studies Galiano et al. recently showed that the DEER technique can be used to measure distances between spin labels incorporated into the flaps of HIV 1 PR and that these distances provide insight into the flap conformations. 81 The dipolar evolution curves, distance profiles, and L curves for HIV 1 PR in the absence and presence of ritonavir (RTV) are shown in Figure 3 1. Apo HIV 1 PR yielded a broad distan ce profile with a most probable distance of 35.4 corresponds to the distance expected for the semi open conformation. The breadth of th is profile, 12.7 is larger than the value predicted for just the semi open conformation and is indicative of the f laps sampling conformations ranging from the closed form to wide open conformations. Figure 3 1 A) Dipolar evolution curves for apo (black) and RTV bound (grey) subtype B HIV 1 PR (curves offset vertically for clarity). B) Corresponding distance profi les. C) L curve for apo HIV 1 PR D) L curve for RTV bound HIV 1 PR 81

PAGE 130

130 The dist ance profile for RTV bound HIV 1 PR shows a most probable distance of 32.6 which is consistent with predictions for the closed conformation based on x ray structures. The breadth of the profile is 3.6 which is narrower than that for the apoenzyme. B ecause the flaps undergo only very small and rapid oscillations when bound to an inhibitor, 73 the breadth of the RTV bound distance profile can be attributed to the motion of the spin label about the flexible linker. In this study, Galiano and co workers utiliz ed four different spin labels to compare the effect of varying the linker flexibility on the distance profiles (Figure 3 2) The most probable distances for spin labels closely approximated the distances for MTSL with the slight differences being attribut able to the different lengths of the linkers. The breadths of the profiles varied more noticeably but were consistent with predictions based on the known flexibility of the linkers. Figure 3 2 Dipolar evolution curves for apo (black) and RTV bound ( grey) subtype B HIV 1 PR labeled with A) MTSL, B) MSL, C) IAP, and D) IASL. E F) Corresponding distance profiles generated. 81

PAGE 131

131 The IAP linker is the same length as the MTSL and MSL linkers but is more flexible. The IASL linker contains an additional bond and is also more flexible than the MTSL and MSL linkers Galiano et al. 145 extended this study to include two drug resistant variants, V6 60 and MDR769, 146 which are clinical isolates from patients who have received protease inhibitor therapy. The distance profiles shown in Figure 3 3 demonstrated that the semi open conformation of V6 is more closed has a shorter inte rflap distance than subtype B whereas MDR769 is more open has a longer interflap distance. This trend is consistent with predic tions based on the x ray crystal structures. This important result showed that the flap s motion and flexibility is affected b y polymorphisms in the protein sequence. Figure 3 3 A) Dipolar evolution curves for apo subtype B (black), V6 (grey), and MDR769 (light grey) (curves are vertically offset for clarity). B) Corresponding distance profiles generated by TKR. 145 These results were also confirmed by molec ular dynamic (MD) simulations preformed by Simmerling et al. 82; 145 The distance profiles for apo and RTV bound protease were very similar to the experimental results and most importantly, the change in the bread ths of the profiles was also observed Likewise, the MD simulations for V6 and MDR769 also captured the differences in the semi open conformations.

PAGE 132

132 T he goal of the present work is to characterize the flap conformations of the inhibitor bound protease to d etermine if the effect of inhibitor binding on the flaps is universal for all inhibitors or if the effect varies among the inhibitors. T he nine inhibitors have similar binding affinities but they show varying enthalpic and entropic contributions to the fr ee energy of binding It is possible that these differences could result in different e ffects on the flap conformations monitored by DEER. Ensembles Shifts of HIV 1 PR Because the apo HIV 1 PR has been characterized as having some population in the closed and open conformations 73 it follows that the addition of an inhibitor c ould simply shift the ensemble to favor the closed conformation. This concept is illustrated in Figure 3 4 Two energy landscapes are shown, one corresponding the apoenzyme (Figure 3 4 A) and one Figure 3 4 Cartoon illustration of possible energy landscapes for a HIV 1 PR with the th ree major conformations (closed, right, semi open, middle, wide open, left ) in the presence and absence of an inhibitor. A) The semi open conformation i s the lowest in energy and thus the most populated but the closed and wide open conformations are close enough in energy and have low enough energy barriers that they will be sampled as well. B) The ensemble shifts in the presence of an inhibitor so that the energy of the closed conformation is the lowest. The closed conformation will now be the most populated with a small sampling of the semi open and wide open states.

PAGE 133

133 corresponding to an inhibitor bound protease (Figure3 4 B) Each local minimum corre sponds to a group of highly related structures which compose the major conformations of HIV 1 PR In the apoenzyme, the semi open conformation is the most populated and lowest in energy. The closed and wide open conformations are energetically accessible but have a slightly higher energy than the semi open form as illustrated in Figure 3 4 A. Upon binding an inhibitor, the closed conformation becomes the most stable. The presence of the inhibitor in the active site remodels the energy landscape so that t he energy of the closed conformation is lower than that of the semi open and wide open forms The closed conformation is the same regardless of whether the inhibitor is present or absent, so the conformational coordinates of the minimum do not change, ju st the relative energy level. Materials and Methods Protein Expression and Purification The codon and expression optimized DNA encoding HIV 1 PR was purchased from DNA 2.0 (Menlo Park, CA) and was cloned into the pET 23 a vector (Novagen, Gibbstown, NJ) us ing standard cloning techniques This DNA also included three additional amino acid substitutions (Q7K, L33I, L36I) to reduce auto proteolysis and two amino acid substitutions to remove the native cysteines (C67A, C95A). The active site of HIV 1 PR was r emoved via a D25N substitution and the labeling site was introduced (K55C) using site directed mutagenesis. The HIV 1PR was expressed into inclusion bodies using BL21*(DE3) pLysS E. coli cells (Invitrogen, Carlsbad, CA) Cells were grown at 37 C with sh aking (250 rpm) to OD 600 (optical density at 600 nm) = 1.0 at which time the protein expression was induce d with 1 mM IPTG (isopropyl D thiogalactoside). Expression was allowed to continue at 37 C for 4 to 6 hours until OD 600 = 1.6. Cells were pellet ed via centrifugation at 7500 x g for 20 mins at 4 C and the pellet resuspended in 40 mL Buffer A (20 mM Tris, 1 mM EDTA, 10 mM BME pH 7.5) Cells

PAGE 134

134 were lysed via sonication (2 mins total with 5 sec on and 5 sec off ) on ice, followed by three passes thr ough a 35 mL French pressure cell (Thermo Scientific). The inclusion bodies and cellular debris w ere pelleted by centrifugation at 18500 x g for 30 mins at 4 C. The inclusion bod y pellet was resuspend ed in 40 mL of B uffer 1 ( 25 mM Tris HCl 2.5 mM EDTA, 0.5 M NaCl, 1 mM diGly, 50 mM BME, pH 7.0 ) The inclusion bodies were then sonicated as before, homogenized usin g a 40 mL dounce homogenizer, and re pelleted via centrifugation. The inclusion bodies were washed twice more using two different buffers ( Buf fer 2: 25 mM Tris HCl 2.5 mM EDTA, 0.5 M NaCl, 1 mM diGly, 50 mM BME pH 7.0, 1 M Urea; Buffer 3: 25 mM Tris HCl 1 mM EDTA, 1 mM diGly, 50 mM BME, pH 7.0 ). The inclusion bodies were then solubilized in 9 M urea (also containing 25 mM TrisHCl, 5 mM NaCl, 1 mM EDTA, 1 mM diGly, 50 mM BME ) with pH adjusted to be about 0.5 pH units below the pI of the HIV 1 PR (pI = 9.35, pH = 8.85) to m inimize binding to the Q column. The solubilized inclusion bodie s were then passed through two 5 mL Q columns (Amersham Biosc iences) T he HIV 1 PR flowed through the column while the majority of the contaminating protein bound the column. The column flow through was acidified by the addition of formic acid ( to a final concentration of 25 mM) an d allowed to set overnight at 4 C to maximize precipitation of contaminating proteins. The precipitate was removed by centrifugation at 38500 x g for 30 mins at RT. The supernatant was added drop wise to 10 mM formic acid (pH ~ 3.0) solution on ice in a tenfold dilution for optimal refo lding of HIV 1 PR. The refold ed protein was then allow ed to equilibrate to roughly 25 C at which time the pH was adjusted to 5.0 by adding 2.5 M sodium acetate (NaOAc) buffer pH 5.5. Samples were spun again at 18500 x g for 25 mins at 4 to remove preci pitated contaminants.

PAGE 135

135 The HIV 1 PR was then concentrated to OD 280 = 0.5 using a n Amicon 8200 stirred cell with a polyethersu l fone (PES) membrane (Millipore) (MWCO 10 kDa) and was then buffer exchanged into 1 0 mM T ris buffer pH 6.9 using a 53 mL HiP rep desa lting column (Amersham Biosciences). The sample was then concentrated back to OD 280 = 0.5. A single crystal of spin label (approximately a 20 fold molar excess) was dissolved in 100% ethanol (EtOH) and was added to the sample. The reaction was allowed t o proceed in the dark for 8 16 hours at room temperature (20 to 24 C). The labeled protein was desalted into 10 mM NaOAc buffer pH 5.0 (to prevent non specific interactions between the protein and the column) diluted to reduce the buffer concentration to 2 mM NaOAc buffer pH 5.0 (to maximize protein stability and minimize aggregation) and concentrated to OD 280 > 1.0 using the stirred cell with a PES membrane The sample was subsequently stored at 20 C. Labeling effici ency was checked by collecting CW E PR spectra. DEER Samples Protonated m atrix The HIV 1 PR was concentrated to OD280 > 1.6 using an Amicon Ultra 4 Centrifugal Filter Unit with Ultracel 3 membrane To avoid multiple freeze/thaw cycles, 70 L of the protein was aliquoted into 0.2 mL PCR tu bes for storage at 20 C To prepare for a DEER experiment, one aliquot was allowed to thaw at room temperature. For substrate bound or inhibited samples, a 4:1 molar excess of substrate or inhibitor (typically 3 to 6 L) was added to the 70 L of prot ei n and allowed to s i t for 15 to 30 mins to ensure sufficient time for binding. The substrate mimic used in this work is a nonhydrolyzable synthetic peptide corresponding to the CA p2 cleavage site in the gag pol polyprotein. The g lycerol (either deuterate d or protonated) was then added to the sample to give a final concentration of 30% ( typically 30 to 32.6 L). The sample was

PAGE 136

136 mixed thoroughly with a pipet tip and transferred to a 4 mm quartz EPR tube using a syringe fitted with a ~30 cm long tube (outer diameter ~2 mm). Deuterated m atrix The HIV 1 PR was concentrated to OD280 > 1.6 using an Amicon Ultra 4 C entrifugal Filter Unit with Ultracel 3 membrane A 5 mL HiP rep desalting colum n was equilibrated with 11 mL of 50 mM deuterated NaOAc buffer pH 5.0 in D 2 O The sample was loaded onto the column and 1.5 mL of flow through was collected and discarded. The sample was collected in the next 2.0 mL. The column was then washed with three column volumes each of water, 1 M NaCl, water, 0.5 M NaOH, and water. The sample was concentrated back to OD280 > 1.6 and EPR signal was checked again. Th e sample was then treated as described above for protonated matri c es. DEER Experiment All DEER experiments were per formed on a Bruker E580 spectrometer at 65 K with a n ER 4118X MD5 resonator ing placed in to the resonator. The four pulse DEER sequence discussed in chapter two (Figure 3 5 ) was used in all experiments unless noted otherwise. The pulse pa rameters used with the Xepr Figure 3 5 The four pulse DEER sequence with the pulse spac nomenclature in the Xepr software package.

PAGE 137

137 Table 3 1. Pulse sequence parameters used with the Xepr software package from Bruker. Matrix type d 0 d 1 d 2 d 3 PG dx Protonated 100 200 3000 100 220 12 Deuterated 100 400 3000 1 00 220 12 All times listed in units of ns. software package from Bruker Biospin are listed in Table 3 1. The length of the d1 spacing varies depending on whether or not the matrix is protonated or deuterated. The 200 ns spacing is used for protonated ma trices and the 400 ns spacing is used for deuterated matrices. Data Analysis All DEER dat a were analyzed as discussed in C hapter two. Briefly, the dipolar evolution curves were analyzed via Tikhonov regularization (TKR) to generate distance profiles and background subtracted dipolar evolution curves. The level of background subtraction was optimized via the self consistent analysis (SCA). The distance profile from TKR was used to generate a theoretical dipolar evolution curve that was compared to the b ackground subtracted dipolar evolution curve. If the two dipolar evolution curves did not match, then the TKR analysis was repeated with an adjusted level of background. This process was repeated until the theoretical and background dipolar evolution cur ves matched. The final distance profile was fit to series of Gaussian functions. Results Apo and Substrate Mimic The background subtracted dipolar evolution curve s for the apo and substrate mimic ( CA p2 ) bound HIV 1 PR subtype B with amino acid substitu tions D25N, Q7K, L33I, L36I, and K55C and labeled with MTSL are shown in Figure 3 6 A. It is clear that the dipolar evolution curves for the apo and substrate bound forms of HIV 1 PR are very different. The first minimum in the apo curve is shifted to rig ht relative to the CA p2 bound curve indicating that the most probable

PAGE 138

138 Figure 3 6 DEER data for HIV 1 PR subtype B. A) Background subtracted d ipolar evolution curves for apo and CA p2 bound HIV 1 PR overlaid with the fit from TKR B) The correspon ding distance profile generated via analysis with TKR for apo and CA p2 distance is larger. Also the oscillations in the CA p2 bound curve have a shorter periodicity indicative of a narrower distribution. These trends are also seen in the distance prof iles generated by TKR. The most probable distance for the apoenzyme is 36.3 compared to 33.0 for the CA p2 bound HIV 1 PR. Likewise, the FWHM (full width at half maximum) of the apoenzyme is 5.2 relative to 3.2 for the CA p2 bound HIV 1 PR. This change in the d istance profile results from an alteration in the flap conformations of HIV 1 PR upo n binding to a substrate mimic as seen in the x ray structures The ~3 shift to a smaller distance is indicative of the flap conformation changing from predominantly semi open to predominantly closed in the presence of the substrate thus bringing the spin labels closer together. Likewise, the decrease in the FWHM is indicative of the decrease in the flexibility of the flaps as they close around the subs trate. The 3 breadth of the CA p2 distance is consistent with the smal l cone of motion of the spin label around the flexible linker even when the flaps are essentially immobilized.

PAGE 139

139 Inhibitors A comparison of the dipolar evolution curves for HIV 1 PR in the presence of the nine FDA approved inhibitors (Figure 3 7 ) reveals that the inhibitors do not all have the same effect on the interfl ap distances. Some inhibitors ( RTV APV, LPV, DRV, TPV, and SQV ) shift the first minimum to the left and increase the f requency of the oscillations (Figure 3 7 D I). These changes are indicative of a narrower distance distribution centered at a short er distances. Other inhibitors ( IDV NFV, and ATV) have little to no effect on the dipolar evolution curves (Figure 3 7 A C). The distance profiles resulting from the TKR analysis of the dipolar echo curves for HIV 1 PR subtype B in the presence of the nine FDA approved inhibitors are not as identical as might be expected considering their similar binding affinities (nM level) Figure 3 8 shows the distance profiles for the PR in the presence of the nine inhibitors, as well as the apoprotease, grouped according to the e ffect on the interflap distance measurements. It can clearly be seen that the inhibitors in Figure 3 8 D I, have a strong e ffect on the flaps, shifting the most probable distance from 36 to 33 and narrowing the breadth from 5 to 3 in the same fashion as the substrate mimic. Likewise, it can clearly be seen that the inhibitors in Figure3 5A C ha ve little to no effect on the flaps, with the largest shift being from 36 to 35 and the breadth actually increasing from 5.2 to 5.5 Interestingly, the distance profiles for all three of these inhibitors contain a pronounced shoulder at 33 (indi cated by the dashed vertical lines). This shoulder is indicative of a population in the closed conformation. The distance profiles can be fit to a series of Gaussian functions as discussed in Chapter 2. The Gaussian functions used to reconstruct these TK R distance profiles for both the apo and CA p2 samples are centered at distances that correspond to the predominant conformations of

PAGE 140

140 Figure 3 7 Dipolar Evolution curves for HIV 1 PR subtype B in the presence of various FDA approved inhibitors. Apo pr otease is shown in black and the inhibitor bound protease is shown in grey A) Nelfinavir (NFV), B) Indinavir (IDV ), C) Atazanavir (ATV), D) Amprenavir (APV), E) Lopinavir (LPV ), F) Daurunavir (DRV), G) Saquinavir ( SQV), H) Ritonavir (RTV ), I) Tip ranaivr (TPV ). HIV 1 PR (Figure 3 9). The 33 distances correspond to the closed conformation, the 35 36 distances to the semi open conformation, and the 41 40 distances to the wide open conformation. The profile for the apoprotease was fit to four Gaussian functions. The Gaussian function centered at 36 with a full width at half max (FWHM) value of 4.9 corresponds to the expected distance and breadth expected for the semi open conformation. This population

PAGE 141

141 Figure 3 8 Distance profiles fo r HIV 1 PR subtype B in the presence of various FDA approved inhibitors (dashed vertical at 33 highlight s the shift of the most probable distance) Apo distance profile is represented as a dashed line. A) Ne lfinavir (NFV ), B) Indinavir (ID V),) C) Ata zanavir (ATV ), D) Amprenavir (APV), E) Lopinavir (L PV), F) Daurunavir (DRV ), G) Saquinavir (SQV), H) Ritonavir (RTV), I) Tipranaivr (TPV) accounts for 86% of the total population. The Gaussian function centered at 33 with a FWHM of 3 correspon ds to the distance and breadth expected for the closed conformation with a 3% population. The Gaussian function centered at 41 with a FWHM of 3.4 corresponds with the wide n at 25 with a FWHM of 2.3 does not correspond with a well known conformation of the

PAGE 142

142 Figure 3 9 Gaussian shaped populations used to fit the distance profiles for HIV 1 PR subtype B. A) Apo HIV 1 PR. B) CA p2 bound HIV 1 PR. protease. However, s everal MD simulations have obtained results that indicate the flaps can tuck or curl into the active site cavity. The distance for the 25 population is consistent with the at th is although more data is needed for a confirmed assignment. The profile for the CA p2 bound protease was also fit to four Gaussian functions. In this sample the most populated conform ation was the closed form constituting 80 % of the total. The corresponding Gaussian function was centered at 33 with a FWHM of 2.7 The two Gaussian functions near 36 were combined to form the population for the semi open conformation and had a mo st probable distance of 36.3 and a FWHM of 3.3 The fourth Gaussian function was cent ered at 41.7 with a FWHM of 2.1 and corresponds to the wide open conformation. These results can be interpreted to mean that 86 % of the apoprotease is in

PAGE 143

143 the sem i open conformation but upon binding an inhibitor, the closed conformation becomes more stable and thus the closed conformation becomes populated with 80 % of the protein. Figure 3 10 shows the Gaussian reconstruction for each of the inhibitor bound HIV 1 PR samples overlaid with the distance profile generated by TKR. Table 3 2 lists the most probable distance, FWHM, and percent population for each of the Gaussian functions used in the reconstruction. Interestingly, populations with similar characteristic s appear in all the samples and vary primarily in their relative populations. For example, in the distance profile for CA p2 bound HIV 1 PR, the most probable distance is 33 which corresponds to the expected distance for the closed conformation of the flaps. The apo HIV 1 PR distance profile has a most probable distance of 36 which corresponds to the semi open conformation of the flaps. However, the CA p2 distance profile also contains a small population at 36 and the apo profile contains a small p opulation at 33 The Gaussians for all the samples can be divided into four categories that correspond to the four major conformations of the flaps: populations between 25 30 which correspond to tucked or curled conformations, populations around 33 which correspond to the closed conformation, populations around 36 that correspond to the semi open form, and populations between 39 45 that correspond to the wide open conformation. Additionally, the breadths of the populations correspond to the expe cted flexibility of the spin labels in each conformation. It is known that the flaps are more rigid in the closed conformation than in the semi open conformation 73 and this trend is reflected in the distance profiles. The populations at 33 are narrower in b readth than are the populations for the open and semi open populations.

PAGE 144

144 Figure 3 10 Gaussian shaped populations used to reconstruct the distance profiles for HIV 1 PR subtype B. Populations were assigned to the closed (black), semi open (grey), and w ide open (white) conformations. A) Nelfinavir (NFV), B) Indinavir (IDV), C) Atazanavir (ATV), D) Amprenavir (APV), E) Lopinavir (LPV), F) Daurunavir (DRV), G) Saquinavir (SQV), H) Ritonavir (RTV), I) Tipranaivr (TPV). Ta ble 3 2. Parameters of Gau ssian shaped populations used to reconstruct distance profiles. Sample Curled Closed Semi open Wide open Center FWHM % Center FWHM % Center FWHM % Center FWHM % Apo 24.7 2.3 4 33.0 3.0 3 36.4 4.9 86 41.3 3.4 7 CA p2 33.0 2.7 80 36.3 3.3 16 41.7 2.1 4 NFV 32.8 2.6 14 35.9 4.9 78 41.8 3.2 8 IDV 32.8 2.6 14 35.9 4.8 79 41.2 3.3 7 ATV 32.8 3.2 41 35.9 3.9 49 41.3 2.5 6 APV 33.2 2.8 76 37.0 2.4 18 41.7 2.0 6 LPV 32.9 2.9 84 36.2 3.9 11 42.8 2.8 5 DRV 33.2 2.9 87 36.6 2.5 13 TPV 32.9 2.3 91 36.3 1.7 9 RTV 33.0 2.6 90 37.4 2.1 10 SQV 32.9 2.8 93 37.4 2.1 7 Error is estimated to be 5% in th e percent population, 0.5 for the center distances, and 0.5 for the FWHM values.

PAGE 145

145 Estimation of Error in Distance Profiles and Population Analysis The Figure 3 11 shows the dipolar evolution curves for apo subtype B HIV 1 PR in H 2 O (discussed above) a nd D 2 O (both contain 30% D 8 glycerol and 2 mM NaOAc buffer pH 5.0). The use of deuterated water shifts the first minimum in the dipolar evolution curve to 46 ns lower, leading to a slightly shorter most probable distance. The most probable distance for H IV 1 PR in H 2 O is 36.3 compared to a most probable distance of 35.2 in the presence of D 2 O. Although this difference is small, it is most likely real given the high SNR ratio of both dipolar evolution curves. The most likely cause for the difference in the distance distributions is Figure 3 11 DEER data for HIV 1 PR subtype B. A) Background subtracted dipolar evolution curves for apo and CA p2 bound HIV 1 PR overlaid with the fit from TKR. B) The corresponding distance profile generated via ana lysis with TKR for apo and CA p2. the slightly higher viscosity of the D 2 O solution relative to the H 2 O solution. This data also suggests that given sufficiently high SNR, subtle changes in the conformational ensemble can be detected via DEER. Figure 3 1 2 show the dipolar evolution curves for apo subtype B HIV 1 PR in D 2 O collected as function of (the length of the dipolar evolution curve) and the corresponding distance profiles. These data were collected on the sample without thawing the sample to ens ure the

PAGE 146

146 highest degree of similarity possible between the curves. The curves were also analyzed with the same level of background subtraction (global analysis). Collecting and analyzing the data in this fashion provides an estimate of error in the distan ce profiles in addition to confirming that the dipolar evolution curves are independent of Figure 3 12 DEER data for apo HIV 1 PR subtype B collected with various values. A) Background subtracted dipolar evolution curves for apo HIV 1 PR vertica lly offset for clarity. B) The corresponding distance profile generated via analysis with TKR for apo HIV 1 PR. Figure 3 13 shows the average of the distance profiles in Figure 3 12B in addition to the average Gaussian shaped populations. The individua l distance profiles were fit to individual series of Gaussian functions to generate the Gaussian shaped populations, which were subsequentially averaged. The error bars i ndicate the local error propagated through the calculation of the average. It can be seen that the largest relative errors correlate with the least populated conformations. This figure provides an estimate of the error in the distance profiles that arises from the TKR process but not sample to sample variation.

PAGE 147

147 Figure 3 13 Average di stance profile (grey) and average Gaussian shaped populations for the data shown in Figure 3 12 Error bars indicate the local error propagated through calculating the average. Table 3 3. Inhibition constant and dissociation constants for subtype C HIV 1 PR. Most Probable Distance () Std Dev 95% CI FWHM () Std Dev 95% CI % Population Std Dev 95% CI Unassigned 1 5 5 2 5 2 0 4 1 1 7 1 4 1.9 1.7 1.4 Curled 2 4 7 1 9 1 6 4 1 1 8 1 5 3.2 1.3 1.1 Tucked 2 8 5 0 5 0 4 3 5 0 4 0 3 1.2 2.0 1.7 Closed 3 3 1 0 20 0 17 4 2 0 7 0 6 26 8 6 Semi open 3 9 8 0 24 0 20 5 0 0 4 0 4 64 7 6 Wide open 3 9 6 0 7 0 6 7 .0 1 1 9 4.8 3.5 2.9 Discussion These DEER results provide a measure of the percent of the protease in the closed conformation for comparison to other me Table 3 3 lists the inhibitors in order of increasing percent closed and contains literature values for the inhibition constant (K I ), the dissociation constant (K D ), the number of non water mediated

PAGE 148

148 hydrogen bonds in the crystal structure (excluding hydrogen bonds to the D25 residues) and the thermodynamic parameters of the inhibitor binding ( G, H, and S). There is a strong correlation between the percent closed population and the number of non water mediated hydrogen bonds (excluding the hydrogen bonds to the D25 residues) in the x ray structures. The inhibitors with the largest percent closed populations also have the most hydrogen bonds. There is a weak correlation between the percent close d population and the K D in that all inhibitors with a K D value > 300 pM also ha ve a percent closed value < 50 %. There is also a weak correlation between the thermodynamic parameters ( G, H, and T S ) and the percent closed conformation with the excepti on of TPV, RTV and SQV. The inhibitors with the most unfavorable G, H, T S values (most positive or least negative) also have the smallest percent closed populations. The exception of TPV can potentially be explained by the non peptidomimetic structur e of TPV. This structural difference also give s rise to a unique feature of TPV, which is that the H for TPV binding drug resistant protease variants is actually more favorable than the H than the binding to the wild type protease 61 However, there is no obvious explanation for the exception o f RTV and SQV since the thermodynamic parameters are almost identical to those for IDV and NFV but the two groups have drastically different percent closed populations. Additionally, the percent closed populations do not track with the K I values. The c aveat with comparing these DEER results to parameters from the literature is three fold. First, and probably most important, is that the DEER data were collected using protease constructs with the D25N mutation that renders the protease catalytically inco mpetent. All of the inhibitors contain a transition mimic that specifically hydrogen bonds with both of the D25 residues. Replacing the D25 residues with Asn (N) disrupts the formation of these hydrogen

PAGE 149

149 Table 3 4 Comparison of percentage closed populat ions for each FDA approved inhibitor with published values of K D K I G, H, T S (25 C) and the number of non water mediated hydrogen bonds in the crystal structures (excluding residue D25) Inhibitor Abbrevia tion % K d K I # of G H T S Closed (pM) a ,b (nM) b ,c Hydrogen Bonds d,e,f,g ( Kcal/mol ) d, i,j, Nelfinavir NFV 14 670 a 1.2 c 2 d 9.7 i 2.8 i 14.2 i Indinavir IDV 14 590 a 3.9 c 3 d 11.8 i 13.1 d 3.9 i 1.3 d 15.7 i 14.4 d Atazanavir ATV 41 NA h 0.48 c 3 b 14.3 d 4.2 d 10.1 d Lopinavir LPV 84 36 a 0.05 c 3 f 14.3 d 2.4 d 11.9 d Amprenavir APV 76 220 0.17 c 5 d 13.2 j 13.3 d 6.9 j 6.7 d 6.9 j 6.6 d Darunavir DRV 87 10 b 0.010 b 6 d 15.0 d 12.7 d 2.3 d Tipranavir TPV 91 19 b 0.019 b 6 g 14.6 d 0.7 d 13.9 d Ritonavir RTV 90 100 a 0.7 c 7 e 13.5 i 2.3 i 11.2 i Saquinavir SQV 93 280 a 1.3 c 7 d 11.8 i 2.2 i 14.0 i a) Data from Clemente et al. 60 d) Data from Muzammil et al. 61 c) Data from Yanchunas et al 62 d) Data from Prabu Jeyabalan et al. 63 e) Data from Prabu Jeyabalan et al. 64 f) Data from Reddy et al. 65 g) Data from Nalam et al. 66 h) Data not available. i) Data from Todd et al. 147 j) Data from Ohtaka et al. 67 bond s and reduces the K d from the nanomolar range to the micromolar range. It is possible but unlikely that the D25N mutation could also lead to other differences in the interaction of the inhibitors with the HIV 1 PR, as crystal structures of both the D25 and D25N protease have been co crystallized with DRV and the structures contain no significant differences 148 Second, the DEER data were collected at 65K which requires that the samples be frozen. The freezing occurs over 2 to 3 seconds, which is significantly slower than the timescale of the conformational changes. Thus it is possible that the pop ulations in the lowest energy conformational are larger in the frozen sample than at ambient temperatures. Third, the DEER data is collected in the presence of 30% w/v glycerol. The presence of the glycerol, which is preferentially excluded from the prot ein surface, can potentially shift the ensemble of conformations. Exploring the effects of the solutes on the flap conformations and spin label mobility is the focus of Chapter 5

PAGE 150

150 However, the correlation between percent closed populations with the numb er of non water mediated hydrogen bonds in the crystal structures, as well as the weak correlation with the K d and thermodynamic parameters, suggest that, although the percent populations determined by DEER may not absolutely correlate with the ensemble in solution, the trends in ensemble shifts reflect the changes at ambient temperatures. Recently, Kent et al. 149 utilized DEER distance measurements to investigate the effects of three inhibitors that mimic different phases of the substrate throughout the cata lytic process. Their study used both active (D25) and inactive (D25N) chemically synthesized HIV 1 PR labeled at site K55C with M TSL. Similar to our results, their apo HIV 1 PR had a broadest distance profile with three distinct populations corresponding to the closed, semi open, and wide open conformations. Figure 3 14 shows the distance profiles for both active (D25) and inactive (D25N) HIV 1 PR in the absence and presence of three inhibitors. In the presence of the MVT 101 inhibitor, which mimics an shifted to the closed conformation with a distribution breadth comparable to the breadth seen in our RTV bound HIV 1 PR In the presence of the KVS 1 inhibitor, which mimics the full tetrahedra l intermediate, the flaps are also strongly shifted to the closed conformation but the breadth of the distribution is slightly broader than was seen for MVT 101. In the presence of the JG is shifted approximately 1 to left and has the same breath as the profile for apo HIV 1 PR These results also corroborate that apo HIV 1 PR is predominantly in the semi open form and transiently samples the wide open and closed conformations. Addition ally, their works demonstrates that even inhibitors which do not strongly close the flaps in D25N HIV 1 PR have a similar effect in active HIV 1 PR

PAGE 151

151 Figure 3 14 Distance distributions for synthesized HIV 1 PR in the absence and presence of inhibitors which mimic various stages of the substrate in the catalytic process. Figure reproduced with permission from Torbeev et al. 149 Conclusion Our results are signi ficant because they demonstrate that careful analysis of DEER data with sufficiently high SNR ca n provide information on the conformational ensembles of proteins, which is significantly more information than just the most probable distance and breadth of the distance distribution. Additionally, these results demonstrate that DEER is sensitive to l es s populated conformational states as long as the SNR is sufficiently high. The ability to detect these l ess populated states means that pulsed EPR can be added to a very short list of techniques that are capable of detecting the signal from a small minor ity of the protein population. In light

PAGE 152

152 of the growing interest in the dynamics and conformation al ensembles of protein s and macromolecular complexes, this technique will prove to be incredibly useful. Additionally, many proteins which are not amenable t o study by other powerful techniques such as x ray crystallography and NMR can be successfully st udied via pulsed EPR.

PAGE 153

153 CHAPTER 4 DEER RESULTS FOR HIV PROTEASE SUBTYPE C A ND CLINICAL ISOLATE V6, A DRUG RESISTANT VARIANT Introduction As discussed in Chap ter 1 HIV has a high rate of mutation which gives rise to the large variety of protease sequences observed that include both naturally occurring polymorphism s found in the various subtypes and d rug pressure selected mutations found in clinical isolates from patients receiving protease inhibitor therapy. The effect s of both naturally occurring polymorphisms and drug pressure selected mutations on the structure and function of the protease have been studied by a variety of techniques although the drug pr essure selected mutations have been studied more thoroughly than the naturally occurring polymorphisms The polymorphisms are typically grouped according to their location within the HIV 1 PR structure and are divided into active site (AS) and non active site (NAS) mutations 150 T he effect of the AS mutations o n the inhibitor binding is intuitive ly obvious; they alter the shape and size of the active s ite binding pocket. Because inhibitors are rigid and designed to match the shape and size of the active site pocket, the ir binding affinities are more strongly af fected by changes in the active site than the more flexible substrates. 151 T he effect s of the NAS mutations however, are not so obvious. M any x ray crystallography studies and several MD simulations have shown t hat some, but not all, of the NAS mutations alter the packing of the hydrophobic core in such a way that the shape and size of the active site is altered. It has also been shown that some NAS mutations alter the flexibility of the flaps or shift the confo rmational ensembles of HIV 1 PR in favor of either more closed or more open forms. 150 T he effects of the NAS mutations are especially important because the majority of naturally occurring polymorphisms occur outside of the active site. In order to understand and predict

PAGE 154

154 how the various subtypes will interact with the inhibitors, the mechanisms by which the NAS mutations alter the effectiveness of the inhibitors must be understood. This work seeks to characterize the flap conformations of sub type C and clinical isolate, V6 which is a drug resistant variant in th e absence and presence of FDA approved inhibitors. These results can be combined with thermodynamic, kinetic, x ray crystallogra phy, and MD studies to further understand the effects of mutations in these HIV 1 PR variants and to develop more effective the rapies Subtype C World wide, subtype C is responsible for more HIV infections than any other subtype and is predominantly found in sub Saharan Africa. However, most of the HIV related research has focused on subtype B, which is most prevalent in North A merica and Western Europe. Additionally, all of the FDA approved protease inhibitors (PIs) have been developed against subtype B. Because the protease sequences of the various subtypes can vary by up to 30% relative to subtype B, it cannot be assumed tha t currently available PIs will be effective against these other subtypes. Thus, the susceptibility to FDA approved inhibitors and the relative vitality of subtype C HIV 1 PR are of great interest. Figure 4 1 shows the x ray crystal structure of apo subtyp e C HIV 1 PR ( C PR ). The naturally occurring polymorphisms typically found in C PR relative to subtype B, are highlighted by spheres and include T12S, I15V, L19V, M36I, S37A, H69K, L89M, and I93L. As with most naturally occurring polymorphisms, these oc cur outside of the active site. Of particular interest, is occurrence M36I, which has been associated with drug resistance in subtype B. Additionally, two polymorphisms occur in the elbow (M36I and S37A) three in the fulcrum (T12S, I15V, and L19V) and one in the cantilever (H69K) These regions are invol ved in the opening of the flaps; thus, we hypothesize that the flexibility of the flaps in C PR differ s

PAGE 155

155 Figure 4 1 Ribbon diagram of apo subtype C HIV 1 PR (PDB ID 2R8N) highlighting th e locations of naturally occurring polymorphisms relative to subtype B from the flexibility of the flaps in B PR. In particular, the side chain of L89 packs into the hydrophobic core and mutations at this position could alter the packing of the core subsequently af fect ing the shape of the active site. The L93 side chain faces towards the cantilever. Mutations at this site could alter the flap flexibility by altering the interactions of the cantilever with the core of the protease. Figure 4 2 shows an overlay of th e C PR structure with a structure of apo subtype B HIV 1 PR (B PR). Both structures were crystallized in the P4 1 2 1 2 space group. Overall, these structures are highly similar with an average root mean squared deviation (RMSD) value of 1.09 152 ; however, there are several regions with potentially significant differences including polymorphisms in these regions, it is not surprisi ng that they differ from those in B PR. This also reinforces the prediction that the flap flexibility in C PR may differ from that observed in B PR. Additionally, the flap conformations are dramatically different as can be seen in top view of the overlay (A). The closest atomic distance between the flap tips in B PR is 4.4 but is

PAGE 156

156 increased to 12.2 for C PR. Interestingly, apo subtype B has also been crystallized in a structure (PDB ID 2PC0) 153 almost identical to that of apo subtype C with an average RMSD of 0.28 152 Figure 4 2 Ribbon diagrams of apo subtype C HIV 1 PR (blue ) (PDB ID 2R8N) overlaid with apo subtype B HIV 1 PR (gold ) (PDB ID 1HHP). A) Top view. B) Front view. These figures are generated using Chimera 154 and overlays were generated by aligning the alpha carbons of one monomer. Th ese difference s could result from the altered side chain packing between the M36I and I15V polymorphisms M36I is located near the elbow and the side chain faces towa rds the fulcrum. I15V is in the fulcrum and the side chain faces towards the elbow. Both n aturally occurring polymorphisms involve replacing larger side chains with smaller residues thus decreasing the steric hindrance between the flap elbows and the ful crum. M36I in C PR h as two fold less van d er Waals contacts than M36 in B PR 152 This variation is illustrated in Figure 4 3, which shows the space filling models for the residues at position 36 and 15 for C PR ( A) and B PR (B) It can clearly be seen that there is more crowding for these residue in B PR relative to C PR One potential consequence of this disparity is that the flaps of C PR might be more likely to favor an open conformation.

PAGE 157

157 Figure 4 3 Ribb on diagram illustrating the difference in packing between residues 36 and 15 in (A) subtype C HIV PR (PDB ID 1SGU) and (B) subtype B HIV PR (PDB ID 2BPX). The enzymatic properties of C PR are similar to those of B PR and are summarized in Table 4 1. The Michaelis Menten constants (K m ), an indicator of substrate binding affinity, are almost identical for C PR and B PR, with values of 17 M and 19.5 M, respectively. The catalytic constants (k cat ) show a slightly larger difference. The k cat is 5.6 s 1 for C PR but is 10 s 1 for B PR which is an almost two fold decrease. The catalytic efficiencies (k cat /K m ) were 0.32 M 1 s 1 and 0.55 M 1 s 1 for C PR and B PR respectively. 155 These values are comparable to thos e determined for two other subtype C variants, C and C SA by Velzquez Campoy et al. 156 They found K m values of 11.6 M 5.4 M, and 14 M, k cat values of 5.8 s 1 7.7 s 1 and Table 4 1. Kinetic parameters for subtype C HIV 1 PR. C oman et al. 155 Velzquez Campoy et al. 156 K m ( M) k cat (s 1 ) k cat /K m ( M 1 s 1 ) K m ( M) k cat (s 1 ) k cat /K m ( M 1 s 1 ) B PR 19.5 10 0.55 14 8.9 0.64 C PR 17 5.6 0.32 C SA 11.6 5.8 0.5 C 5.4 7.7 1.43

PAGE 158

158 8.9 s 1 and k cat /K m values of 0.32 M 1 s 1 1.4 3 M 1 s 1 and 0.64 M 1 s 1 for C SA, C, and B PR, respectively. The sequence of their C SA variant which is more prolific in South Africa, i s closer to the C PR variant used by Coman et al. than the variant designated C. The results of these two stud ies show that the naturally occurring polymorphisms in these variants of subtype C do not provide a catalytic advantage over subtype B, although other variants have been shown to have such an advantage. 157 Coman et al. 155 also determined the inhibition con stants (K I ) for eight o f the FDA approved PIs for C PR; this work is summarized in Table 4 2. The K I values for most of the inhibitors were higher for C PR than for B PR, which is indicative of weaker inhibitor binding. This result is not surprising sinc e the inhibitors were designed to inhibit B PR The largest difference was seen for RTV, with K I increased 3.8 fold relative to B PR However, the K I values for all inhibitors remained in the nM range suggesting they would still be effective at inhibiti ng C PR Interestingly, the K I values for APV and TPV were lower for C PR than for B PR indicating tighter inhibitor binding. Velzquez Campoy et al. 156 measured the dissociation constants (K d ) for IDV, SQV, NFV, RTV, APV, and LPV for C PR which are also summarized in Table 4 2 They found that for all inhibitors tested, the K d values were larger for C PR than for B PR indic ating the inhibitors bind more weakly to C PR The smallest increase seen was 3.1 fold increase for APV ( 0 .0465 nM for C PR compared to 0.015 n M for B PR ). Similar to the trend seen for the K I constants all the K d values remained in the nM range indicating the inhibito rs still bind sufficiently to inhibit the enzyme.

PAGE 159

159 Table 4 2 Inhibition constant and dissociation constants for subtype C HIV 1 PR. B PR C PR K I 155 (nM) K D 156 (nM) K I 155 (nM) K D 156 (nM) Indinaivir ( IDV ) 1.8 0.48 3.3 2.784 Nelfinavir ( NFV ) 1.7 0.26 2.7 1.014 Atazanaivr ( ATV ) 0.07 0.13 Amprenavir ( APV ) 0.04 0.015 0.29 0.0465 Lopenavir ( LPV ) 0.11 0.008 0.19 0.0624 Ritonavir ( RTV ) 0.07 0.03 0.27 0.726 Saquinavir ( SQV ) 2.2 0.40 2.6 2.32 Tipranavir ( TPV ) 0.4 0.11 Clinical Isolate V6 As discussed in Chapter one, the high mutation rate of the HIV genome provides a n e volutionary mechanism for adapti ng to the presence of protease inhibitors The currently accepted explanation for why the non active site mutations affect the efficacy of protease inhibitors is that they change the dynamics and/or flexibility of HIV 1 PR by either preventing HIV 1 PR from reaching the wide open conformation needed to bind the inhibitor or by preventing the flaps from closing before the inhibitor can optimize its geometry for high affinity binding. 71 Thus, understanding the role of the flaps in inhibitor binding can potentially result in the design of better drugs that are less susceptible to drug selected resistance. The drug resistant variant V6 was isolate d from a pediatric patient undergoing RTV therapy. It contains two AS mutations, V32I and V82A, and six NAS mutations, K20R, L33F, M36I, L63P, A71V, and L90M. In addition to RTV resistance, these mutations are also associated with IDV and NFV resistance. The V82 mutation is frequently seen in resistance to all clinically used inhibitors. The position s of these mutations are highlight ed in Figure 4 4 The combination of V32I and V82A alter s the shape of the active site, which typically results in significant dec reases in inhibitor binding. The side chain of L90M faces into the hydrophobic core and can alter the shape of the active site by changing the packing of the hydrophobic core. R esidues L33, M36, and K20 pack together in the region between the fulcrum and the flap elbows. Mutating these

PAGE 160

160 residues most likely impacts the mechanics of flap opening. Likewise, residues A71 and L63 are located in the cantilever, which is also involved in the opening of the flaps. Thus mutations in these positions can also aff ect the ability of the flaps to open. Figure 4 4 Ribbon diagram illustrating the locations of the drug pressure selected mutations in V6. Active site mutations are shown in black and non active site mutations are shown in grey. Mutations on mapped on a structure of subtype B HIV 1 PR (PDB ID 2BPX). A previous study by Molla et al. 158 found that drug pressure selected mutations appeared in a particular order for patients on RTV therapy V82A is the first mutation to appear and is associated with the initial decrease in the effectiveness of RTV. It is most frequently followed by the mutations I54V, A71V, and, M36I, with the mutations I84V, K2 0R, L33F, and L90M appearing later Similar to many drug resistant HIV 1 PR variants, the V6 PR variant does not performs as well enzymatically as the wild type HIV 1 PR (subtype B). 60 The K m value for V6 PR shows a three fold increase relative to B PR (47 M for V6 PR compared to 18.2 M for B P R). The k cat value for V6 PR is 27 s 1 which is 1.3 fold higher than the k cat value for B PR, which is 21 s 1

PAGE 161

161 The catalytic efficiency of V6 PR is roughly half that of B PR ( 0 58 M s 1 for V6 PR compared to 1.2 M s 1 for B PR), which largely results from the three fold increase in K m Table 4 3 Kinetic parameters for variant V6 HIV 1 PR. 60 K m ( M) k cat (s 1 ) k cat /K m ( M 1 s 1 ) B PR 18 21 1.2 V6 PR 47 27 0.58 Error is estimated to be 5% in the percent population, 0.5 for the center distances, and 0.5 for the FWHM values. Clemen te et al. 60 determined the K I values for R TV, IDV, and NFV for V6 PR which are summarized in Table 4 4 As expected, V6 PR shows a large decrease in susceptibility to RTV because these mutations were selected for by RTV therapy. The K I value for RTV was 42 fold higher for V6 PR relative to B PR (30 nM for V6 PR versus 0.7 nM for B PR). However, V6 PR also shows some level of cross resistance to IDV and NFV with 22 fold and 14 fold increase, respectively. This result is also not unexpected given that a number of the mutations observed in V6 PR also occur in IDV selected and NFV selected variants. Table 4 4. Inhibition constants for variant V6 HIV 1 PR. 60 B PR V6 PR K I (nM) K I (nM) Indinaivir (IDV) 3.1 69 Nelfinavir (NFV) 1.2 17 Ritonavir (RTV) 0.7 30 Although there are no x ray crystal structures for V6 PR there are two x ra y structures for a V6 PR variant containing the additional mutations I84V and I54V 60 These structures can be compared to those of B PR to provide insight into the structural effects induced by the drug pressure selected mutations. However, the V6 and subtype B structures are from crystals with

PAGE 162

162 different symmetry point groups (V6 = P6 1 subtype B (1HSG, 1HXW) = P2 1 2 1 2) so some differences might be attributable to crystal contact artifacts. Figure 4 5 shows the IDV bound V6 PR structure overlaid with a comparable IDV bound B PR structure. Compar ison of the active sites reveals that the mutations lead to small changes in the shape of the active site pocket. Because the inhibitors are much more rigid than the peptide substrates, even subtle changes in the active site can have a large impact on inh ibitor binding with only small affects on the substrate binding. However, the structures are very similar overall including the position of the bound IDV Comparison of the C backbone trace (A) shows that the regions in V6 PR with the largest deviation s relative to B PR are the elbow and fulcrum and are the regions with the most NAS mutations. Figure 4 5 Overlay of x ray structures for V6 PR (I84V, I54V) (yellow) (PDB ID 1SGU) and B PR (PDB ID 1HSG) (blue) bound to IDV A) Backbone trace to high light regions of variation. B) Ribbon diagram to highlight extent of the similarities. Figure 4 6 shows the RTV bound x ray structure of C PR overlaid with RTV bound B PR. Although these structures have a high degree of similari ty, they are more differ ent than the structures also have significant differences in the fulcrum and cantilever domains. Additionally,

PAGE 163

163 the position of the bound RTV is not same for C PR a s for B PR. Because the overall structures for both RTV and IDV bound C PR are similar to B PR and the largest deviations are seen in regions that are involved in the flap opening, it has been hypothesized that the contributions of the NAS mutations to d rug resistance are via alterations in the mechanics of flap opening. Figure 4 6 Overlay of x ray structures for V6 PR (I84V, I54V) (PDB ID 1SH9) (yellow) and B PR (PDB ID 1HXW) (blue) bound to RTV. A) Backbone trace to highlight regions of variation. B) Ribbon diagram to highlight extent of the similarities Thus, characterizing the flaps conformations of variant V6 PR can elucidate the effect of the NAS mutations on the flap dynamics To this end, double electron electron resonance (DEER) was use d to measure the distance between spin labels covalently attached to K55C in the flaps. Distance distri butions were determined for apo and substrate ( CA p2 ) bound V6 PR in addition to V6 PR in the presence of nine FDA approved inhibitors. These distance profiles correlate with changes in the population profiles for the various flap conformations described below Materials and Methods Protein Expression and Purification Subtype C HIV 1 PR and variant V6 PR were expressed and purified as discussed in Cha pter three with the following the exception. The pH of buffer 4 (9 M urea, 25 mM TrisHCl, 5 mM

PAGE 164

164 NaCl, 1 mM EDTA, 1 mM diGly, 50 mM BME ) was adjusted to be about 0.5 pH units below the pI of the HIV 1 PR (for subtype C pH = 9 0 5 (pI = 9.55) was used and for V6 pH = 8. 42 (pI = 8.92) was used) to minimize binding to the Q column. DEER Experiments All samples were prepared as described in Chapter three. DEER data was collected as described in Chapter three. The dipolar evolution curves were converted to dista nce profiles via Tikhonov regularization (TKR) as implemented in the DeerAnalysis2008 software package. 142 Distance profiles were reconstructed using a series of Gaussian shaped populations. Any populations contr ibuting to less than 15% of the total population were validated using the method described in Chapter two. Results DEER Results for Subtype C HIV 1 PR Figure 4 7 shows the dipolar evolution curves and corresponding distance profiles for apo subtype C and a po subtype B HIV 1 PR. T he dipolar evolution curves are similar but not identical. The first minimum in the curve for C PR is shifted 33 n s higher than that of B PR which is indicative of a longer most probable distance. The oscillations in the curve f or C PR are similar to those for B PR but have a slightly smaller intensity leading to broader corresponding distance profile for C PR relative to B PR The most probable distance for apo B PR is 35.4 compared to 37.6 for apo C PR. The breadth of di stance distribution for C PR is 6.9 which is broader than the breadth seen for B PR (5.1 ). Figure 4 8 shows the dipolar evolution curves for apo and CA p2 bound C PR and the corresponding distance profiles. A comparison of the dipolar evolution curve s shows that the first minimum in the curve for CA p2 bound C PR is shifted 153 ns lower, leading to a shorter most probable distance. Additionally, the oscillations in the curve for CA p2 bound C PR have a

PAGE 165

165 Figure 4 7 A) DEER Dipolar evolution curves for apo subtype C and subtype B HIV 1 PR. B) Corresponding distance profiles generated by TKR. higher intensity than those of apo C PR which indicates the breadth of the distance distribution for CA p2 bound C PR is narrower than that of apo C PR The m ost probable distance for CA p2 bound C PR is 33.5 and the breadth of the distribution reduced to 3.9 The change in the distance profile for C PR upon binding CA p2 is similar to the change observed for B PR binding CA p2 ; for B PR the most prob able distance shifts from 35.4 to 33.0 and the breadth narrow s from 5.1 to 3.2 T hese changes represent a shift in the flap conformations from being predominantly semi open to predominantly closed. Figure 4 8 A) DEER Dipolar evolution curves for ap o and CA p2 bound subty pe C HIV 1 PR B) Corresponding distance profiles generated by TKR.

PAGE 166

166 Figure 4 9 shows the dipolar evolution curves and corresponding distance profiles for CA p2 bound C PR and B PR. It can clearly be seen that both the dipolar evol ution curves and distance profiles are nearly identical. This similarity suggests that substrate bound structure of C PR highly resembles the substrate bound structure of B PR. Figure 4 9 A) DEER Dipolar evolution curves for CA p2 bound subtype C su btype B HIV 1 PR. B) Corresponding distance profiles generated by TKR. Figure 4 10 shows the dipolar evolution curves for C PR in the presence of nine FDA approved protease inhibitors. It can clearly be seen that the inhibitors affect the dipolar evolut ion curve s differently. NFV, IDV, and ATV affect only minor changes, whereas the other inhibitors induce more significant changes Interestingly, this trend is almost identical to the trend observed for inhibitor binding in B PR as shown in Chapter three For all the inhibitors excep t IDV, the first minimum is shifted lower indicating smaller most probable distances. The magnitude of th is shift range s from 17 ns for IDV to 182 ns for TPV. Additionally, the intensity of the oscillations is higher for AP V, LPV, DRV, TPV, SQV, and RTV than those for apo C PR. As mentioned previously, the higher intensity oscillations correspond to narrower distance distributions. Figure 4 11 shows the distance profiles for C PR in the presence of nine FDA approved inhibi tors. It can clearly be seen that IDV, NVF, and ATV have minimal effects on the flap

PAGE 167

167 conformations. The distance profile for IDV is almost identical to that of apo C PR. NFV and ATV shift the most probable distance to a slightly shorter distance (~36 ) but do not change the breadth of the distribution. APV shifts the most probable distance to below 34 similar to the remaining inhibitors but does not narrow the distribution to the same extent. Likewise, LPV and Figure 4 10 DEER dipolar evolutio n curves for C PR in the presence of nine FDA approved inhibitors overlaid with the dipolar evolution curve for apo C PR. RTV affect less drastic reductions in the distribution brea dths than do SQV and DRV. TPV induces the largest shift in the most proba ble distance and narrowest distance distribution. These changes in the distance profiles for C PR can be compared to the inhibitor induced changes seen in the distance profiles from B PR. Figure 4 12 show the overlays of the distance

PAGE 168

168 profiles for C PR and B PR with each of the nine FDA approved inhibitors The differences seen between the distance profiles of C PR and B PR for IDV, NFV, and ATV are similar those observed for apo C PR and apo B PR. The differences between the profiles for C PR and B PR w ith APV, LPV, Figure 4 11 DEER distance profiles for C PR in the presence of nine FDA approved inhibit o rs (solid line) (vertical dotted line at 33 highlights the shift in the most probable distance) overlaid with the distance profile for apo C PR (da shed line). SQV, DRV, and TPV are strikingly minor. In each case, the most probable distances are essentially the same and the breadths of the distance distributions for C PR are slightly broader than th ose of the B PR profile s Interestingly, this is n ot the case for RTV. The most probable

PAGE 169

169 distance for C PR with RTV is 0.7 longer and the breadth of the profile significantly broader than for B PR. However, a comparison of the inhibition constants (Table 4 2) for C PR reveal s that largest change in th e inhibition constant between C PR and B PR is with RTV. Figure 4 12 Distance profiles for C PR (solid) and B PR (dashed) in the presence of nine FDA approved inhibitors. The distance profiles for apo and CA p2 bound C PR can be reconstructed using a series of Gaussian shaped populations, which are shown in Figure 4 13. The apo C PR distance profile was fit to four Gaussian shaped populations that are centered at 29, 33, 37, and 41 The 41

PAGE 170

170 population corresponds to the wide open conformation and constitutes 21 % of the total protein population. The population centered at 37 corresponds to the semi open form of C PR and contributes 58 % to the total population. The closed form corresponds to the population centered at 33 which is 9 % of the po pulation. The population centered at 29 (12%) does not correspond to any of the well known conformations of HIV 1 PR and is distinct from the populations that appear around 25 and are designated as curled as discussed in Chapter three However, it is likely that this population also represents a curled or tucked conformation that brings the flaps closer together. Thus, for now, populations centered near 30 are putatively assigned to the tucked conformation for the purposes of discussion The dista nce profile for CA p2 bound C PR was fit to three Gaussian shaped populations, which are centered at 31.8, 33.7, 38.1 These populations correspond to the tucked, closed and semi open conformations, respectively and represent 13%, 79%, 7% of the total p opulation, respectively. Figure 4 13 Gaussian shaped populations used to reconstruct the distance profiles of apo and CA p2 bound C PR Likewise, the distance profiles for C PR in the presence of the inhibitors can be reconstructed by a series of Gau ssian shaped populations. These reconstructions are shown in Figure 4 14 and are

PAGE 171

171 summarized in Table 4 5. For IDV, NFV, and ATV, the most populated conformation is the semi open form, with increasing ly large r populations of the closed conformation, respe ctively. Interestingly, the percentage of wide open conformation does not change appreciably between these inhibitors. The distance profiles for APV, RTV, and LPV are all predominantly closed but still have appreciable populations in the semi open form. The conformations for SQV, DRV, and TPV are almost entirely all closed. This trend is very similar that observed for B PR except that most of the inhibitors lead to a slightly smaller percent closed population for C PR relative to B PR. Figure 4 14 Gaussian shaped populations used to reconstruct the distance profiles for C PR in the presence of nine FDA approved inhibitors

PAGE 172

172 Table 4 5 Parameters of Gaussian shaped populations used to reconstruct distance profiles for subtype C. Sample Curled /Tucked Closed Semi open Wide open Center FWHM % Center FWHM % Center FWHM % Center FWHM % Apo 29.6 3.1 12 33.1 4.4 9 36.9 4.6 5 8 40.3 3.4 21 CA p2 31.8 1.9 13 33.7 3.4 80 38.1 3.4 7 NFV 32.7 3.9 1 2 36.5 4.7 7 8 39.6 2.9 10 IDV 30.5 2.6 4 33.0 3.2 6 37.3 5.4 8 1 41.0 3.1 9 ATV 33.0 3.7 17 36.0 5.2 68 39.6 4.3 15 APV 33.3 3.2 54 36.6 5.1 42 41.4 2.5 4 LPV 33.0 3.3 7 6 37.3 4.4 24 DRV 33.2 2.8 84 37. 0 2.2 1 2 41.3 1.6 4 TPV 32.8 2.8 9 7 36.3 1.2 3 RTV 31.7 1.7 8 33.8 3.4 7 7 38.0 3.4 15 SQV 32.0 1.8 21 34.0 3.0 7 2 39.0 2.0 7 Error is estimated to be 5% in the percent population, 0.5 for the cen ter distances, and 0.5 for the FWHM values. DEER Results for the V6 Variant of HIV 1 PR Figure 4 1 5 shows the dipolar evolution curves and corresponding distance profiles for apo V6 PR overlaid with those of apo B PR. A comparison of the dipolar evolu tion curves reveals that apo V6 PR and B PR should have similar distance profiles although the most probable distance for V6 PR should be slightly shorter than for B PR. The most probable distance for V6 PR is 35.7 which is just shorter than the 36.3 seen for B PR. The breadth of the profile for V6 PR is 7.1 which is broader than the breadth of 5.1 for B PR. Figure 4 15 DEER dipolar evolution curves and distance profiles for apo V6 PR and B PR.

PAGE 173

173 Figure 4 16 shows the dipolar evolution curves and distance profiles for apo and CA p2 bound V6 PR. Similar to C PR and B PR, the binding of CA p2 to V6 PR shifts the first min imum in the dipolar evolution 114 ns lower and increases the intensity of the oscillations, indicative of a shorter most prob able distance with a narrower distribution. The most probable distance is 32.9 for CA p2 bound V6 PR with a breadth of 3.5 Figure 4 16 DEER dipolar evolution curves and distance profiles for apo and CA p2 bound V6 PR Figure 4 17 shows the dipo lar evolution curves and distance profiles for CA p2 bound V6 PR and B PR. The first minimum in the dipolar evolution curve for V6 PR is shifted slightly lower by ~58 ns, indicative of a slightly shorter most probable distance. However, this difference not what is seen in the distance profiles. The most probable distances are the same but the distribution of distances around the most probable distances differ. In the case of V6 PR, there is more population for distances below the most probable distance Whereas, f or B PR, there is more population at distances greater than the most probable distance. However, the SNR for the CA p2 bound V6 PR is substantially higher than the SNR for B PR and it is likely that the differences seen in the dipolar evoluti on curves and distance profiles are a result of the noise.

PAGE 174

174 Figure 4 17 DEER dipolar evolution curves and distance profiles for CA p2 bound V6 PR and B PR. Figure 4 18 shows the dipolar evolutions curves for V6 PR in the presence of nine FDA approved inhibitors. Similar to C PR and B PR, IDV, NFV, and ATV have minimal effects in the dipolar evolution curves, shifting the first minimum by ~30 ns. However, RTV, which had a more drastic effect on C PR and B PR, has a much smaller effect on V6 PR. The f irst minimum is shifted by only 67 ns. This result is not unexpected because the drug pressured selected mutations in V6 were selected for under RTV therapy. The trend in the remaining inhibitors is similar to that seen with C PR and B PR. TPV, SQV, and DRV have a strong effect on the dipolar evolution curve, shifting the first minimum by ~115 ns. LPV and APV have slightly less effect in the dipolar evolution curves and shift the first minimum by only 86 ns. This trend is also seen in the distance profi les for V6 PR in the presence of the inhibitors, which are shown in Figure 4 19. NFV and IDV shift the most probable distance to ~34 and increase the breadth of the profile slightly. ATV and RTV shift the most probable distance to ~ 34 but do not cha nge the breadth of the profiles. APV shifts the most probable distance to 33.7 and narrows the breadth to 3.8 SQV and DRV shift the most probable distances

PAGE 175

175 Figure 4 18 DEER dipolar evolution curves for V6 PR in the absence and presence of nine FDA approved inhibitors. to~33.3 and narrow the distance profile significantly to ~3.4 LPV also shifts the most probable distance to 33.5 and narrows the breadth to 2.2 but retains a large population of distances around 30 TPV, as with C PR and B PR, has both the largest shift in the most probable distance, to 32.9 and the largest decrease in profile breadth, to 2.6 With the exception of RTV, this trend is similar to that seen for B PR. Figure 4 20 show the distance profiles for V6 PR and B PR in the presence of nine FDA approved inhibitors. In

PAGE 176

176 Figure 4 19 DEER distance profiles for the V6 variant of HIV 1 PR in the presence of nine FDA approved inhibitors (vertical dashed line at 33 highlights the shift in the most probable d istance) overlaid with the distance profile for apo V6 PR general, the shifts of the most probable distance are the same for V6 PR and B PR but the breadths for V6 PR are slightly broader than those for B PR. The exceptions are LPV, which is narrower f or V6 PR than for B PR, TPV, which is the same for V6 PR and B PR, and RTV, which is drastically different in both the shift and breadth of the profile as a result of the RTV resistance of V6 PR.

PAGE 177

177 Figure 4 20 DEER distance profiles for the V6 variant o f HIV 1 PR in the presence of nine FDA approved inhibitors (vertical dashed line at 33 highlights the shift in the most probable distance) overlaid with the distance profile for the same inhibitor bound to B PR Figure 4 2 1 shows the Gaussian reconstru ctions for the apo and CA p2 bound V6 PR distance profiles. The distance profile for apo V6 PR was fit to three Gaussian shaped populations corresponding to the tucked (29 ), closed (33 ), semi open (36 ) and wide open (39 ) conformations, representi ng 10%, 21 %, 61 % and 8 % of the total population, respectively. Interestingly, the CA p2 bound distance profile was fit to only one population, which corresponds to the closed conformation and is centered at 33

PAGE 178

178 Figure 4 21 Gaussian shaped populati ons used to reconstruct the distance profiles of apo and CA p2 bound V6 PR. Figure 4 22 shows the Gaussian reconstruction for the inhibitor bound distance profiles for V6 PR. The largest semi open populations are seen for NFV, IDV, RTV, and ATV and corre spond to 3 9 %, 54 %, 51 %, 6 3 %, respectively. The second largest populations for these inhibitors corresponds to the closed conformation and represent 30 %, 30 %, 34 %, and 30 %, respectively. The third largest population is the tucked conformation, which const itutes 14 %, 10 %, 15 %, and 7 %, respectively. NFV and IDV also have wide open populations that contribute 17 % and 6 % of the total population. The largest population for APV, SQV, DRV, LPV, and TPV is in the closed conformation, which represent s 72 %, 87%, 8 9 %, 83% and 92 %, respectively. APV has three additional populations corresponding to tucked ( 7 %), semi open ( 16 %), and wide open ( 4 %). SQV has one additional population in the semi open conformation (13%). DRV has four additional populations correspond ing to the curled (3%), tucked (3%), semi open (3%), and wide open (2%).. LPV has an additional tucked populations of 17% and TPV has three additional populations, curled (5%), tucked (2%), and wide open (2%).

PAGE 179

179 Figure 4 2 2 Gaussian shaped populations used to reconstruct the distance profiles for V6 PR in the presence of nine FDA approved inhibitors. Table 4 6. Parameters of Gaussian shaped populations used to reconstruct distance profiles for clinical isolate, V6. Sample Curled /Tucked Closed Semi op en Wide open Center FWH M % Center FWHM % Center FWHM % Center FWHM % Apo 29.6 3.9 1 0 33.2 3.8 21 36.4 4.8 8 61 39.9 3.1 8 CA p2 33.1 4.0 100 NFV 27.3 6.1 14 32.9 3.8 30 35.4 4.3 39 39.2 42.4 3.4 3.3 17 ID V 27.5 6.1 10 32.8 5.4 30 36.8 7.2 54 46.8 4.5 6 ATV 29.6 4.0 7 32.8 3.7 30 36.0 5.5 63 APV 29.8 2.8 8 33.3 3.1 72 37.7 3.5 16 45.0 2. 8 4 LPV 30.0 3.0 17 33.4 2.2 83 DRV 25.9 28.5 1.1 1.1 3 3 33.2 3.2 89 38.5 1.5 3 43.7 2.4 2 TPV 25.0 29.2 2.1 1.0 5 2 32.9 3.4 92 40.9 1.9 2 RTV 31.1 4.5 15 33.7 4.7 34 36.0 6.2 51 SQV 33.2 3.1 87 38.2 1.8 13 Error is estimated to be 5% i n the percent population, 0.5 for the center distances, and 0.5 for the FWHM values.

PAGE 180

180 Discussion In this work, we characterized the flap conformations of subtype C HIV 1 PR and a drug resistant variant of HIV 1 PR, V6, in the absence and presence of a substrate mimic, CA p2 and in the presence of nine FDA approved inhibitors. This was accomplished by measuring the distance between spin labels incorporated at site K55C in the flaps using DEER. Our results for apo C PR correlate with the differences s een between the x ray crystal structures for C PR and B PR. The most probable distance in the semi open conformation for C PR is longer than that of B PR by ~ 1 This difference is smaller than that seen in the x ray structures but the trend is consist ent. Additionally the apo C PR has a larger wide open population than B PR, which is consistent with the predictions that the M36I and I15V polymorphisms would increase the flexibility of the flaps in C PR. Our results also show that CA p2 bound C PR is highly similar to that of B PR. This indicates that the substrate bound conformations of C PR and B PR are most likely very similar as well. The changes in the distance profiles for C PR in the presence of the nine FDA approved inhibitors are similar to t hose seen for B PR. The noticeable difference is that the distance profiles are slightly broader for C PR, which likely results from the decreased binding affinity of the inhibitors for C PR. Also, the distance profile for RTV with C PR showed the larges t difference relative to B PR, which is consistent with RTV having a significantly larger increase in K I for C PR. Our results for apo V6 PR, which showed the V6 PR is less open then B PR, are also consistent with the predictions based on the locations and known effects of several of the polymorphisms in V6 PR. The side chains of L33, M36, and K20 pack together in the region between the fulcrum and elbows. Thus, the combined effect of mutating all three of these residues is most likely to alter the flexib ility of the flaps and to stabilize a slightly more closed

PAGE 181

181 conformation. This trend is also seen in the CA p2 bound V6 PR, which in more closed than that of B PR. Our results for V6 PR in the presence of the nine FDA approved inhibitors is also consistent with the trend seen in the inhibition constants for V6 PR. The distance profile for RTV shows the most largest difference relative to B PR, which correlates with the large increase seen in the K I value. This result is expected because the polymorphisms in V6 were selected for under RTV therapy. The K I values for other inhibitors also increased for V6 PR compared to B PR. This trend is also seen in our results. The distance profiles for all the inhibitors except DRV and TPV, showed an increased breadth relative to B PR. Although the K I s have not been determined for all the inhibitors, this indicates that these inhibitors might also have increased K I values. Furthermore, the lack of change between the distance profiles of DRV and TPV for V6 PR and B PR suggests that these inhibitors might be the most effective at inhibiting V6 PR. Conclusion These results demonstrate that DEER is sufficiently sensitive to detect subtle changes in the conformational ensembles resulting from polymorphisms in the protein s equence. Furthermore, the correlations seen between the changes in the K I values for C PR and V6 PR variants relative to B PR, suggests that DEER can potentially predict changes in the effectiveness of inhibitors for variants of HIV 1 PR. Additionally, t he details of the conformational ensembles of HIV 1 PR can confirm the predicted effects of polymorphisms on the flap flexibility and conformation and provide a me asure of the magnitude of the effects.

PAGE 182

182 CHAPTER 5 SOLUTE EFFECTS ON SP IN LABEL MOBILITY AND PROTEIN CONFORMATION S Introduction The addition of solutes to protein samples is often required for a variety of experimental techniques. In x ray crystallography, the solutes aid in the crystallization process and protect the crystals during the freezi ng process. In SDSL EPR studies of proteins and biological molecules solutes are frequently added to increase the correlation time for isotropic tumbling ( R ) of molecules under 20 kDa so that the line shape is not dominated by R (e.g., fully averaged) and can thus reveal local information about the spin a flexible linker ( I ) and about the flexibility of the protein backbone ( B ). For pulsed EPR studies, which are preformed at cryogenic temperatures, solutes are added a s glassing agents to prevent phase separation and protein aggregation However, it has been shown that solutes can alter the structure and function of proteins. Solutes can increase or decrease the rate of an enzymatic reaction, stabilize or destabilize the nativ change. These effects result from the differences between solute water, solute protein, protein water, and water water interactions. Consequently the effects of the solute on the protein will depend on the size, concentration, and chemical composition of both the solute and protein. To simplify the discussion, it is useful to deconstruct the various types of solutes and the various impacts the solutes have on the solution. Furthermore, it useful to deconstruct the various aspects of the protein in order to delineate how the various effects of solutes on the solution properties collectively affect the protein.

PAGE 183

183 Figure 5 1 Structures of water and various solutes A) Water, B) urea, C) ammon ia D) glycerol, E) sucrose, and F) poly(eythlene glycol). Solute Effects on Solution Properties and Protein Structure and Function Important considerations for proteins In general, proteins can be considered to have two prim ary components the peptide backbone and the amino acid side chains. The peptide backbone is hydrophobic and generally prefers to self associate in the presence of aqueous solutions. The amino acid side chains are typically grouped into four categories po lar, charged, aromatic, and hydrophobic (non polar). The surface of a folded protein is a combination of each of these groups. Typically, the more hydrophobic side chains and peptide backbone are buried in the core of the protein or buried i n the hydroph obic region of a bilayer with the charged and polar groups exposed to the solvent. However the solvent exposed sur face of the protein is heterogeneous and comprised of both polar and non polar regions that vary for every protein. The water and solutes i nteract with each region differently and the overall effect of the solute on the protein will depend on the

PAGE 184

184 contribution of each region to the whole. Thus, each protein can potentially have a unique response to various solutes. Important considerations fo r water Water is unique in its interactions and despite the long history of investigations into the structure and nature of water, a full understanding has yet to be accomplished. Few molecules if any can hydrogen bond as efficiently as water. Water h as the unique ability to either donate or accept hydrogen bonds at all four positions on the oxygen creating a tetrahedral structure. Additionally, water can donate the same number of hydrogen bonds as it can accept which maximizes the ability of water to form hydrogen bonding networks. Other small polar compounds such as alcohols and acetone contain carbon residues and thus are only polar on one side of the molecule. Other polar compounds like ammon ia which can also donate and accept up to four hydroge n bonds lacks the symmetry of water because it can donate three hydrogen bonds but only accept one. Consequently, the addition of any compound to an aqueous solution raises the chemical potential ( ) of the solution by disrupting the hydrogen bonding netw ork. This disruption is typically manifested as an increase in the structure of the water near the molecule. For non polar molecules, the water molecules are structured around the molecule to form a cage. For ions, the water molecules strongly hydrogen bond with ion creating a layer of rigid water. For polar compounds, there is typically no significant ordering of water molecules, just a small disruption in the hydrogen bonding network. The chemical potential for each species in a solution can be descri bed by Equation 5 1, ( 5 1) where j is the index for each component, j sln is the chemical potential in solution, j is the c hemical potential in the pure state, a j is the activity of component j, R is the molar gas constant,

PAGE 185

185 and T is the absolute temperature. For an ideal solution, the activity of component j is equal to j the mole fraction of component j. Thus, the additio n of a solute to an aqueous solution reduces the mole fraction of water and the activity of water which decreases the chemical potential of water. Important considerations for solutes Solutes can vary greatly in their size, structure, and chemical composit ions and each facet impacts the effects of the solute on the protein and the solution differently. Additionally, certain combinations of features can also have colle ctive effects that differ from effects of either feature alone. In general, the size, den sity, polarity, hydrophobicity, concentration, and number of hydrogen bond donors and acceptors are considered to be the most important aspects of the solutes. Size Solutes can be divided into two important categories based on their size small molecules whose size is on the same order as water and larger molecules whose size varies from from the surface of the protein resulting from steric restraints. Howeve r, the polarity/hydrophobicity of the solute can also contribute to the effect of the solute on the protein. The effect of small solutes is determined solely by their chemical composition and concentration. Polarity / hydrophobicity The polarity/hydropho bicity of the solute will determine whether the solute interacts more favorably with the water or with the protein. A more favorable solute water interaction results in the solute being excluded from the surface of the protein (generally referred to as pr eferential hydration meaning the protein is preferentially hydrated by water, which will be discussed in more detailed below). Solutes like urea interact more favorably with the protein which tends to destabilize the protein structure. Solutes like glyce rol

PAGE 186

186 and sucrose tend to interact more favorably with the water which tends to stabilize the protein structure. Hydrogen bond donors and acceptors Each solute has a set number of hydrogen bond donors and acceptors. The number and ratio of these sites det ermines to a large extent how the solute will behave in each of the concentration regimes. Glycerol, for example, has three hydroxyl groups which can each donate and accept one hydrogen bond. At low concentrations, each hydroxyl group will hydrogen bond to two water molecules and the rest of the glycerol molecule will be solvated by a cage of water molecules. All the water molecules involved in solvating the glycerol are removed from the bulk solvent. As the concentration increases, the bulk will be dep leted as the glycerol molecules are solvated. At low concentrations of bulk water, the water molecules from the s o hydroxyl groups and will be replaced by other glycerol molecules. Thus solutes with more hydrogen bonding sites deplete the bulk water faster. Concentration The effect of each solute depends on the concentration of the solute, which can be divided into three regimes based on the solvation of the solutes as illustrated in Figure 5 2 At lo w concentrations (dilute regime Figure 5 2 A ), the solute will effectively only interact with bulk water. The net amount of bulk water will decrease as some of the water molecules will be required to solvate the solute. Polymers in the dilute regime occu py a space determined by their radius of gyration (Rg). As the concentration increases (semi dilute regime Figure 5 2 B ), the amount of bulk water decreases and the solute becomes more likely to interact with other solutes. Polymer molecule s will begin t o intertwine with other polymer molecules as the solution becomes more crowded and as the available bulk water for solvation decrease. The

PAGE 187

187 concentration at which the polymers changes from predominantly interacting with only itself to intertwining with oth er polymer Figure 5 2 Illustration concentration regimes for polymers. Molecules and can be a protein ligand pair or two proteins that bind. A) Dilute regime where polymers interact predominantly with themselves. The dotted line around the polym ers indicates the radius of gyration (Rg). B) Crossover concentration, defined as the concentration where the polymers switch from interacting with themselves to interacting with other polymer molecules. C) Semi dilute regime where the polymers interact with other polymer molecules and begin to intertwine. D) Concentrated regime where polymer molecules are tightly intertwined. Figure modified from Kozer et al. 159 molecules is called the crossover concentration (Figure 5 2C) and depends primarily on the Rg of the polymer. At high concentrations (concentrated regime Figure 5 2D ), there is insufficient bulk water to solvate all of the solutes so the solutes must in teract with one another. Polymers will be highly intertwined. It is intuitively obvious that the effect of solutes on the solution and the protein in each of these concentration regimes will differ. In the dilute regime, diffusion limited processes lik e protein protein and protein ligand interactions will be hindered because of the decrease in translation diffusion. Changes in the protein conformation may be affected as a result of

PAGE 188

188 preferential hydration. At the crossover concentration and in the semi dilute regime, protein interactions will tend to be increased as the solutes form effective cages around the protein. The cage tends to increase the lifetime of the collision complex which increases the amount of time available for the protein and its bi nding partner to find the correct binding geometry. In the concentrated regime, protein interactions will tend to be decreased as the tightly intertwined polymer network can occupy the space between the protein and its binding partner. Density (polymer st ructure) Polymers can by synthesized with a variety of structures density of each structure that matters. Several polymer structures are illustrated in Fig ure 5 3. The linear polymers are the least dense and will be more likely to intertwine at lower concentrations. The dendritic polymers are the most dense and, depending on the number of generations present (number of layers extending from the central mon omer) may be unable to intertwine because the density at the edge of the protein is already maximized. Other structures, such as block or brush polymers, can have a range of densities that can be as low as the linear polymers. The density of the polymer determines the concentration at which the polymers begin to interact with polymer molecules (crossover concentration). Polymers which lower crossover concentrations will tend to be more excluded strongly and excluded at lower concentrations. Figure 5 3 Illustration of several common polymers structures. A) Dendritic polymer with three generations. B) Brush polymer. C) Linear polymer in the random coil conformation. D) Branched polymer.

PAGE 189

189 Solute exclusion There are several factors that can contribute to a solute being excluded from the surface of a protein, including preferential hydration, steric exclusion, and surface tension. The exclusion of the solute from the surface requires energy which is proportional to the surface area. Consequently, the energy is minimized when the surface area is minimized which means that solute exclusion favors the most compact form of the protein. Any conformational change in the protein that exposes more surface area will be hindered whereas any conformational chan ge that reduces the surface area will be favored. Preferential hydration means that the protein is predominantly solvated by water molecules with a corresponding exclusion of solute molecules. In a protein solute solution, there is a balance among the wat er water interactions and the protein water, solute water, and protein solute interactions. If the protein water or solute water interactions are more favorable than the protein solute interactions, then the solute will be excluded from the surface of the protein. Steric exclusion results from the solute being excluded from the protein surface because the solute is significantly larger than the water and is sterically hindered from approaching the protein. When steric hindrance is a factor, the protein c onformation with the smallest amount of excluded volume is stabilized. Surface tension was believed to be an important factor in the stabilizing effect that many solutes have on protein structure. The water molecules at the air water interface are unable to maintain the favorable tetrahedral hydrogen bonding network of bulk water and thus experience a higher degree of order. This increase in the ordering of the water leads to the exclusion of ions and solutes from the surface similar to the solute exclusi on from the surface of proteins. According to this theory, any solute that increases the surface tension of water should have a

PAGE 190

190 stabilizing effect on the protein and any solute that decreases the surface tension of water should have a destabilizing effect on the protein structure. It has been shown that sucrose increases the surface tension and has a stabilizing effect. However, glycerol reduces the surface tension but also has a stabilizing effect on protein structure. Thus surface tension may play a r ole in the effect of solutes on the protein, but is not a reliable predictor of the effect. Macromolecular crowding and confinement Macromolecular crowding is a term used to describe the intracellular environment and refers to the high concentration of mac romolecules present. The effects of macromolecular crowding on proteins are generally mediated by reduced volume effects. Essentially the crowders occupy a fraction of the solution volume and therefore increase the effective concentrations of other speci es in the solution. Osmotic pressure and water activity A decrease in the chemical potential of water also results in a decrease in the water activity and an increase in the osmotic pressure as demonstrated in Equation 5 2 160 where p W and p W are the vapor pressure of water in the solution and in the pure state ( 5 2) re spectively, m sln is the solute concentration in molal units, sln is the osmotic coefficient of the solution, is the partial molar volume of water, and is the osmotic pressure. Thus osmotic pressure can be used to measure the activity of water. The activity of water is essentially a measure of the effective concentration of water, or in other words, the am ount of bulk water. The availability of bulk water in a protein solute solution will affect the balance between the protein water and solute interactions can lead to preferential hydration.

PAGE 191

191 Viscosity The viscosity of a solution is defined as the resistanc e of a fluid to deformation under stress. Solutions with low viscosity, like water, deform easily whereas solutions with high viscosity, like molasses, do not deform easily Because the viscosity of water is so low, the addition of any solute tends to in crease the viscosity of the solution. The magnitude of the increase is dependent upon the size and shape of the solute added. Large polymers increase the viscosity more than smaller polymers or monomers such as glycerol. However, the shape of the polyme r is also important to consider because a linear polymer will increase the viscosity to a greater extent than a branched polymer of equal mass. This trend results from the effective density of the solute in solution, the branched polymers are more dense t han a linear polymer and thus do not contribute as effectively to resisting shear forces. The change in viscosity upon solute addition is also concentration dependant. The higher the concentration of solute, the greater the solution viscosity will be. As suming the solute is inert, meaning it has no significant interactions with the protein, the primary effect of solution viscosity on the protein is to slow down the rate of diffusion. Proteins are considered to have two kinds of diffusion, translational diffusion and rotational diffusion. The translational diffusion rate of molecules is described by the translational diffusion coefficient (D t ) in Equation 5 3 and can be predicted from the Stokes Einstein (SE) relation where k is the Boltzman constant, T is the temperature, is the viscosity of the solution, and R is the radius of the particle (assumed to be spherical) The rotational diffusion coefficient (D r ) in Equation 5 4 is also predicted from the SE relation. It has been demonstrated that D t for ( 5 3)

PAGE 192

192 ( 5 4) poly(ethylene glycol) (PEG) polymers of various sizes as well as monomers such as glycerol and sucrose follow this SE relation. However, D r only follows the SE relation for very small polymers and monomer s, 161 and deviates from the SE relation for larger polymers. This deviation is typically attributed to different microscopic environments create d by large and small solutes in solution as illustrated in Figure 5 4 S mall solutes like glycerol are much closer to size to that of the solvent when compared to the size of the protein as illustrated in Figure 5 4 A. However, p as the protein (Figure 5 4 B). The large polymers r educe translational diffusion but because the protein occupies a small pocket that is free of solute, the rotational diffusion is only slightly damped. The viscosity of the solution in these pockets differs from the viscosity measured for the bulk solutio n and is referred to as the microviscosity. The viscosity of the bulk solution is referred to as both the bulk viscosity and macroviscosity. Because the term bulk viscosity has other connotations as well, the term macroviscosity will be used in this work Figure 5 4 Illustration of potential size variations between solutes and proteins which lead to changes in the translation and rotational diffusion of proteins.

PAGE 193

193 Solute Effects on CW EPR and Pulsed EPR Data for HIV 1 Protease CW EPR line shapes As d iscussed in Chapter 2, the CW EPR line shapes are sensitive to the motion of the spin label, which is the combination of the protein tumbling in solution, the motion of the protein backbone at the labeling site, and the motion of the label about the flexib le linker. Each of these motions will be impacted differently by the various solute effects. The protein tumbling in solution will be most sensitive to the microviscosity. The backbone motion, which combines both the conformational changes of the protei n and the small oscillations of the backbone between energetically accessible bond angles, will be most sensitive to the changes in the preferential hydration and solute exclusion. In the case of HIV 1 protease (HIV 1 PR) it is difficult to separate the contributions from conformational changes in the flaps and the small oscillations of the backbone for labeling sites on the flaps because both motions change upon binding inhibitors. 73 The motion of the spin label about the flexible linker will also be sensiti ve to changes in preferential hydration and solute exclusion. As the motion of the spin label decreases, the features in the line shape will broaden. The increase in spectral breadth can be qualitatively compared by visual inspection of the high field t ransition because it is the most sensitive to motion. The spectral breadth can be quantitatively compared by measuring the change in the intensity of the central transition ( H pp ) or the ratio of the intensities of the low field and center field transitio ns (I LF /I CF ) or by calculating the second moment ( H 2 ) or scaled mobility (M S ) as discussed in Chapter 2. Pulsed EPR distance measurements The distance profiles for HIV 1 PR from DEER experiments are sensitive to the flexibility of the spin label about th e flexible linker and to changes in the flaps conformations as demonstrated in Chapter 3. The breadth of the distance distribution in an inhibitor bound sample represents

PAGE 194

194 the distribution of the spin label around the flexible linker because the flaps have been shown to be essentially rigid when bound to an inhibitor. If the motion of the spin label about the linker is decreased in the presence of solute, this change will most likely be manifested as a change in the breadth of the distribution. If a highe r energy conformation of a protein is more compact than the lowest energy conformation, then the presence of the excluded solutes will favor the higher energy conformation. This shift in the conformational ensemble will be manifested as a shift in the dis tance profile. For the apo form of HIV 1 PR the semi open conformation has both the smallest volume and the least surface area. In the presence of solutes which are excluded from the surface, the conformational ensemble will be shifted in favor of the se mi open conformation. However, the closed conformation of the inhibitor bound HIV 1 PR has the smallest volume and least surface area if the volume and surface of the inhibitor are in included in the calculations. Thus, in the presence of excluded solute s, the ensemble will be shifted in favor of the closed conformation. Materials and Methods Materials (1 Oxyl 2,2,5,5 Tetramethyl Pyrroline 3 Methyl) Methanethiosulfonate (MTSL) was purchased from Toronto Research Chemicals (North York, ON, Canada). 4 Ma leimido TEMPO (MSL), 3 (2 Iodoacetamido) PROXYL (IAP) and 4 (2 Iodoacetamido) TEMPO (IASL) were purchased from Acros Organics (Belgium) through Sigma Ald rich. Glycerol, sucrose, Ficoll 400 sodium acetate (NaOAc), Tris HCl, and dimethyl sulfoxide (DMSO) wer e purc hased from Fisher Bioreagents. All molecular weights of p oly(ethylene glycol) (PEG) polymers were purchased from Fluka through Sigma

PAGE 195

195 Solute Solutions Stock solutions of the solutes were prepared by dissolving 6 g of sucrose or PEG (all molecular weig hts) in acetate buffer (2 mM NaOAc pH 5.0) to a volume of 10 mL using volumetric flasks to yield 60% weight/volume (w/v) stock solutions. The 30% w/v Ficoll400 stock solution was prepared by dissolving 3 g in acetate buffer to a volume of 10 mL using volu metric flasks and incubating in a 42 C water bath until dissolved. The 60% w/v solution of glycerol was prepared by adding 6 g of glycerol to acetate buffer to a final volume of 10 mL using volumetric flasks. Fluorophore Labeling The Cys variants of HIV 1 PR were desalted into 10 mM Tris pH 7.0 for labeling with the fluorophores. Dipyrromethane boron difluoride (BODIPY) was dissolved in DMSO to make a 0.5 mg/mL solution. The 5 ((((2 iodoacetyl)amino)ethyl)amino)naphthalene 1 sulfonic acid (IAEDANS) (FW 434.25 g/mol) was dissolved in DMSO to make a 0.5 mg/mL solution. The fluorophore solutions were added to the protein sample to give an approximately 100 fold excess of label to protein. The solutions were incubated overnight at 4 C and excess label was removed via desalting column (Hiprep 26/10, Amersham Biosciences) equilibrated with 2 mM NaOAc pH 5.0. The labeled protein was concentrated to ~10 M and stored at 20 C. The purity of the labeled protein was estimated to be greater than 95% pure by SDS PAGE. Steady State Fluorescence Anisotropy Steady state fluorescence anisotropy measurements were made on a Fluoromax 3 (Horiba Jobin Yvon, Edison, NJ) equipped with Glan Thompson polarizers (#FL 1044) in an L format configuration. Excitation and emissio n wavelengths were set to 505 nm and 535 nm respectively for BODIPY and 336 nm and 490 nm for IEADANS. The silt widths were set to 5 nm and the

PAGE 196

196 integration time to 0.3 s. Measurements were made using a 4 mm x 4 mm quartz cuvettes and a theromstated cell holder set to 20C. Initial protein concentrations ranged from 0.15 to 0.3 mg/ml (7.5 to 12 M) as determined by absorption at 280 nm, using an extinction coefficient of 1.15 mg cm 1 ml 1 Stock solute solutions (made as described below ) were titrated into the protein sample using Hamilton Gastight syringes (Hamilton, Reno, NV) and mixed well by repeatedly transferring the solution between a large Hamilton syringe and the cuvette. Anisotropy values ( r ) were calculated from the E quation 5 5, where I vv I vh I hv I hh are the intensities of the light ( 5 5) when the excitation and emission polarizers, respectively, are in the vertical position or horizontal position as designated by the subscripts and G = I hv / I hh an d is a correction factor that compensates for the instrumental bias in the detection of polarized light. Continuous Wave EPR CW X band EPR spectra were obtained on a modified Bruker ER200 spectrometer with an ER023M signal channel, an ER 032M field control unit, and equipped with a loop gap resonator (Medical Advances, Milwaukee, W I). All EPR spectra were collected at 24 C. The temperature was controlled by utilizing a glass dewar (Wilmad Labglass, Buena,NJ) to surround the loop gap resonator and protein. N itrogen gas was flow ed through a copper coil submerged in a waterbath (Thermo Scientific Neslab RTE 7 digital one ( 25C to 150C 0.01C)) of 40% ethylene glycol in water and over the sample. 1 5 N HSQC NMR NMR samples were prepared by dissolving vario us amount of PEG 8000 in a volume of protein solution containing 125 M HIV 1 PR solution in 2 mM NaOAc buffer pH 5.0. The

PAGE 197

197 final PEG 8000 concentrations ranged from 0 to 300 g/L in 50 g/L increments. Sample pH was verified after the addition of PEG using litmus paper and was corrected if necessary. The 15 N HSQC NMR spectra were collected at 20 C on a Bruker Avance II 600 MHz NMR s pectrometer with a 14 T magnet and a 5 mm cyroprobe. Spectral widths were set to 13 ppm ( 1 H) and 38 ppm ( 15 N). The 1H chemic al shifts were referenced to 4,4 dimethyl 4 silapentane 1 sulfonic acid (DSS). Results Because there is a complex relationship between the various attributes of the solute and the ir effects on the solution properties, it is important to characterize severa l aspects of the solute solutions such as osmolality and viscosity so that connections can be made between solute properties and corresponding effects on the protein. The solution viscosities for various concentrations of each solute were measured using a Cannon Fenske viscometer. The osmotic pressure of the same solutions were measured using vapor pressure osmometry. Figure 5 5 A plots the measured viscosity for a series of solutions containing systematically varied amounts (percent content (w/v)) of Fic oll400, sucrose, PEG 3000, and glycerol in 2 mM NaOAc buffer pH 5.0. The data for each solute can be fit to an exponential expression such as Equation 5 6, where the 1/t 1 values were smallest for Ficoll400 (6.8 0.2), followed by PEG 3000 (13.0 0.9) and glycerol (36 1) with largest being sucrose (43.8 0.6). Thus the (5 6) dependence of the viscosity on the solute follows th e trend sucrose < glycerol < PEG 3000< Ficoll400. It is also noteworthy that the concentration dependence (percent content in terms of weight/volume) of the solution viscosity for glycerol and sucrose are very similar which corresponds with the similarity of their sizes relative to PEG 3000 and Ficoll400. The high

PAGE 198

198 structure. Figure 5 5 Plots of (A) viscosity and (B) osmolality as a function of percent content f or sucrose (dark grey, circle), glycerol (light grey, star), PEG 3000 (triangle, grey), and Ficoll400 (square, black). (C) Plot of osmolality versus viscosity. Figure modified from Galiano et al. 162 Viscosity measurements performed by LG and osmolality measurements performed by GEF. Figure 5 5 B plots the measured osmolality for the same solutions in Figure 5 5 A. The 30% and 40% glycerol solutions were not measureable via vapor pressure osmometry and are thus extrapolated from data in the literature measured via vapor pressure deficit. 163 The osmotic pressure follow s the trend glycero l > sucrose> PEG 3000 > Ficoll400, which is different than the trend for viscosity. It is also important to note that the measurements for Ficoll400 were on the order of the osmolality for the NaOAc buffer. The osmolality for glycerol however, spanned tw o orders of magnitude. At low concentrations (<10% w/v), glucose and sucrose have a strong effect on the osmolality of the solution and PEG 3000 and Ficoll400 have almost no effect. At

PAGE 199

199 intermediate concentration ranges (10% to 20% w/v), the osmolalities for glycerol and sucrose diverge with the glycerol values being almost double the values of the sucrose. The osmolality for PEG 3000 also begins to increase but is still roughly half of the sucrose values. Figure 5 5 C plots the measured osmolality versu s the measured viscosity. There is a clear distinction between glycerol and sucrose small viscogens and the polymers. The polymers exhibit large changes in the viscosity with only small changes in the osmolality and the small viscogens exhibit moderate c hanges in the viscosity with large changes in the osmolality. Effect of solutes on spin label correlation times A comparison of the CW EPR line shapes (Figures 5 6 and 5 7 ) for HIV 1 PR labeled with MTSL at site K55C in the presence of several solutes at v arious concentrations reveals that the motion of the spin label (the convolution of R I and B ) depends on the nature and concentration of the solute. The high field transition (marked by the arrow in Figure 5 6 B) is the most sensitive to the solute a nd can be used as an indicator for the reduction in the spin label motion. Of all the solutes tested, glycerol had the most pronounced effect in changing the line shapes, followed by PEG, ethylene glycol (EG), sucrose, and then Ficoll400. This trend does not correlate with the size, viscosity, or os molality of the solute. Instead the trend correlate s with the hydrophobic character of the solute. Of the solutes investigated, Ficoll400 has the most hydrophilic character followed by sucrose and both these solutes have minimal effects on the line shapes. The hydrophobicity of the PEG polymers scales with the size of the polymer, which is consistent with the line shape changes seen in Figure 5 7 EG has more hydrophobic character than PEG 300 but less hydro phobic character than glycerol.

PAGE 200

200 Figure 5 6 CW EPR line shapes for MTSL labeled HIV 1 PR in the presence of various solutes with incremented concentrations in 2 mM NaOAc buffer pH 5.0. Spectra were collected at 24 C. Data collected by LG. 162 Figure 5 7 CW EPR line shapes (100 G scans) for MTSL labeled HIV 1 PR in the presence of various sizes of PEG and EG with incremented concentrations in 2 mM NaOAc buffer pH 5.0 Spectra were collected at 24 C.

PAGE 201

201 T he CW EPR line sh apes are sensitive to changes in I R or B but EPR line shape analysis alone cannot distinguish which correlation times have changed However, it is well known the motion of the flaps ( B ) decreases significantly upon binding an inhibitor. Thus, comparing the line shapes of HIV 1 PR in the absence and presence of an inhibitor will determine if the line shape is reporting on changes i n B Shown in Figure 5 8 are the line shapes for apo and RTV bound HIV 1 PR labeled with four common spin labels. It can be seen that the line shapes do not change in the presence of inhibitor regardless of the spin label used. This result indicates th at the spin labels at site K55C are not sensitive to changes in B Figure 5 8 CW EPR line shapes (100 G scans) for spin labeled HIV 1 PR in the presence (blue) and absence (red) of a tight binding inhibitor, ritonavir (RTV) (2 mM NaOAc buffer, pH 5.0) Spectra were collected at 24 C. Data collected by LG. 162 In general, proteins larger than 18 kDa tumble sufficiently slow in solution that the X band EPR line shape s for nitroxide radicals are not sensitive to any decrease in R The HIV 1 PR dimer is roughly 22 kDa which should be sufficiently large that the line shapes will not report on

PAGE 202

202 R This assump tion can be tested by comparing the trends in the line shapes to the trends in the fluorescence anisotropy measurements of HIV 1 PR in the presence of the same solutes. The steady state fluorescence anisotropy measurements are sensitive to changes in the orientation that occur within the fluorescent lifetime of the probe fluorophore (~5 ns for BODIPY). Thus, the anisotropy is sensitive to changes in all three correlation times but is dominated by changes in R because the protein tumbling wi ll result in much larger changes in the orientation of the fluorophore Figure 5 9 plots the change in anisotropy of BODIPY labeled HIV 1 PR as function of the solution viscosity. HIV 1 PR was labeled at two sites, K55C and T74C a relatively rigid site on a strand near the core of the protease so that a comparison can be made between the more Figure 5 9 Fluorescence anisotropy measurements for BODIPY labeled HIV 1 PR at sites (A) T74C and (B) K55C. The error bars for these data are smaller than the symbols and are not shown. mobile K55C site on the flaps and a rigid site. The anisotropy for both sites increased the most for sucrose and glycerol, which is indicative of the largest decrease in motion. Both Ficoll400 and PEG 3000 resulted in a small er increase in the anisotropy. This trend is not the same as the trend observed for the CW EPR line shapes where glycerol and PEG 3000 induced the largest

PAGE 203

203 change in the line shape broadening From these results, we can conclude that the change seen in th e X band EPR line shapes do not result solely from changes in R Figure 5 10 plots the same anisotropy data as shown in Figure 5 7 but now grouped according to the solute inste ad of labeling site. A clear a difference can be seen between the T74C site an d the K55C site in the presence of glycerol (A) and PEG 3000 (C) that is absent in sucrose (B) and Ficoll400 (D) In the presence of high glycerol and PEG 3000 concentrations, K55C experiences a larger decrease in motion than T74C This difference could potentially result from changes in the fluorophore motion Figure 5 10 Plots of the percent change in the fluorescence anisotropy BODIPY labeled HIV 1 PR variants T74C and K55C in presence of four solutes, (A) glycerol, (B) sucrose (C) PEG 3000, and (D ) Ficoll400.

PAGE 204

204 about its flexible linker. The fluorophore is large and hydrophobic and has a longer linker region than the spin labels used in this work. Additionally, the fluorphore is more exposed to the solvent at the K55C site. Thus the fluorphore sh ould have a large degree of motion about the flexible linker at the K55C site. In the presence of excluded solutes, the preferential hydration could decrease these motions by constrain ing the fluorophore to lie against the protein. However, the diff erenc e in anisotropy between K55C and T74C can potentially result from preferential hydration induced changes in the flap motion. Measuring the anisotropy of the K55C for inhibitor bound HIV 1 PR would easily distinguish between these scenarios as the flaps in the inhibitor bound HVI 1 PR are essentially immobile Therefore, any change in the anisotropy would result from changes in the fluorophore motion about the linker. Protein solute interactions In several reports, PEG polymers were found to bind to hydrop hobic regions of the protein surface 164; 165; 166; 167 In t hese studies a variety of solution NMR techniques were utilized to detect the residue specific changes induced by the presence of PEG. A straightforwar d method to monitor residue level interactions is by 1 H 15 N HSQC NMR When applied to 15 N labeled cytochrome C as a function of PEG concentration it was determine d that PEG molecules specifically interact with a hydrophobic patch on the surface that cons titutes the ligand binding domain. 164 The surface of HIV 1 PR contains several hydrophobic patches including the flap tips as illustrated in Figure 5 11 In order to characterize sites of specifi c interactions between the polymers and the surface of HIV 1 PR, I preformed an analogous experiment on 15 N labeled HIV 1 PR. Figure 5 12 shows the HSQC spectrum for 15 N labeled HIV 1 PR (containing the D25N mutation to prevent autoproteolysis) with the re sonance assignments labeled. This spectrum is

PAGE 205

205 Figure 5 11 Hydrophobic surfaces of HIV 1 PR in the (A) closed (PDB ID 2pbx), (B) semi open (PDB ID 3hvp), and (C) wide open (structure from Hornak et al. 50 ) confo rmations. The colors range from blue to orange with the darker blue being the most polar and the darker orange being the most hydrophobic. Surfaces generated using Chimera. 154 very similar to the published HSQC spectrum for active 15 N labeled HIV 1 PR containing the same three stabilizing mutations utilized in our HIV 1 PR constructs (Q7K, L33I, and L6 3I). Our spectrum contains 93 resonances which correspond to the 93 non proline residues in HIV 1 PR The additional peaks correspond to nitrogen containing side chains. Based on the results of our CW EPR and fluorescence anisotropy experiments and the h ydrophilic nature of Ficoll400, it is reasonable to predict that the presence of Ficoll400 will not interact with the surface of HIV 1 PR. Figure 5 1 3 A shows an overlay of the HSQC spectra for 15 N labeled HIV 1 PR in the absence and presence of 150 g/L Fi coll400. It can be seen, that in the presence of the Ficoll400, some of the resonances are broadened and thus less intense. However, none of the resonances are shifted significantly. The overlay of the HSQC spectra for each of the six Ficoll400 concentr ations (0 300 g/L) is shown in Figure 4 1 3 B, with the 1 H axis shifted by 0.05 ppm and the 15 N axis shifted by 0.25 ppm per spectrum to highlight the concentration dependence of the broadening. Although the concentration dependence varies slightly among t he resonances, the broadening effects in the presence of Ficoll400 can be attributed to an increase in the rotational correlation time, R

PAGE 206

206 Figure 5 1 2 15 N HSQC NMR spectra of HIV 1 PR with the assignments determined by comparison to Ref. 72 The HSQC spectra are shown in Figure 5 1 4 A for 15N labeled HIV 1 PR in the absence and presence of 100 g/L of PEG 8000. In the presence of PEG, many of the resonances are broadened and some of the resonances are shifted. F igure 4 1 4 B shows an overlay of the spectra collected as a function of PEG concentration (0 300 g/L). Each spectrum is offset by 0.05 ppm in the 1 H axis and 0.25 ppm in the 15 N axis. The broadening of many of the resonances in the

PAGE 207

207 presence of PEG is also attributed to an increase in R However, the shifted resonances correspond to sites that specifically bind to the polymer. Spectroscopically, shifted resonances result from changes in the chemical environment of the nuclei that alter the magnetic shie lding. In a titration experiment, the gradual shifting of resonances between the chemical shifts of the start point and end point of the titration indicates that the nucleus in question in undergoing exchange on a timescale faster than the NMR experiment. If the exchange was slower than the NMR timescale, the intensity of the resonance would be split between the start and end points with the ratio changing as a function of concentration. Likewise, if the exchange is on similar timescale to the NMR experi ment, the resonance will be broadened as the intensity is blurred be tween the start and end points. Phenomenologically, the shifts indicate that the corresponding residues are interacting with the PEG molecules. This interaction alters the chemical envir onment of these residues resulting in a change in their chemical shifts. I also investigated the residue specific effects of sucrose on 15 N labeled HIV 1 PR. Figure 4 1 5 A shows the HSQC for 15 N labeled HIV 1 PR in the absence and presence of 250 g/L sucro se. Similar to the PEG 8000, the presence of the sucrose broadened many resonances and shifted many others. Likewise, the broadening of the resonances in the presence of sucrose can be attributed to an increase in R Interestingly, many of the resonanc es that shifted in the presence of PEG also shifted in the presence of sucrose but in the opposite direction. This result is likely due to the different chemical natures of PEG and sucrose. Although some of the resonances shifted in the presence of both PEG and sucrose, many resonances shifted in only one solute or the other. This result shows that the residue specific interactions of PEG and HIV 1 PR are not the same as the interacti on between sucrose and HIV 1 PR.

PAGE 208

208 Figure 5 1 3 A) 15 N HSQC NMR spectr a of HIV 1 PR in the absence (red) and presence (blue) of 150 g/L of Ficoll400. B) 15 N HSQC NMR spectra of HIV 1 PR titrated with Ficoll400. 0 g/L (black), 50 g/L (red), 100 g/L (blue), 150 g/L (green), 200 g/L (magenta), 250 g/L (purple), 300 g/L (cyan). 15 N axis offset by 0.25 ppm and 1 H axis offset by 0.05 ppm per spectra.

PAGE 209

209 Figure 5 1 4 A) 15 N HSQC NMR spectra of HIV 1 PR in the absence (red) and presence (blue) of 100 g/L of PEG 8000. B) 15 N HSQC NMR spectra of HIV 1 PR titrated with PEG 8000. 0 g/ L (black), 50 g/L (red), 100 g/L (blue), 150 g/L (green), 200 g/L (magenta), 250 g/L (purple), 300 g/L (cyan). 15 N axis offset by 0.25 ppm and 1 H axis offset by 0.05 ppm per spectra.

PAGE 210

210 Figure 5 1 5 A) 15 N HSQC NMR spectra of HIV 1 PR in the absence (red) and presence (blue) of 2 50 g/L of sucrose B) 15 N HSQC NMR spectra of HIV 1 PR titrated with sucrose 0 g/L (black), 50 g/L (red), 100 g/L (blue), 150 g/L (green), 200 g/L (magenta), 250 g/L (purple) 15 N axis offset by 0.25 ppm and 1 H axis offset by 0.0 5 ppm per spectra.

PAGE 211

211 The changes in the resonances can mapped onto the surface of the HIV 1 PR as illustrated in Figure 5 1 6 A comparison of these surfaces to the hydrophobic surface area of the HIV 1 PR reveals that the resonances that shifted in the pr esence of PEG 8000 correspond to the hydrophobic patches on the protease surface. The resonances that shifted in the presence of sucrose have some degree of correlation with the hydrophobic patches, but this correlation is not as clear as with PEG 8000. T hese results show that although PEG 8000 does not bind to a specific site on the protease, as was the case with cyt c 164 PEG 8000 does inter act more significantly with HIV 1 PR than do either suc rose or Ficoll400. Figure 5 1 6 Comparison of the A) hydrophobic surface of HIV 1 PR to resonances changes in HIV 1 PR (mapped on to PDB ID 2pbx) in the presence of B) PEG 8000, C) sucrose, and D) Ficoll400. A) The most hydrophobic regions shown in da rk orange and the most hydrophilic regions shown in dark blue. B D) Shifted resonances indicated in red and broaden resonance in blue with unaffected resonances in white.

PAGE 212

212 Discussion The goal of this work was to characterize the effects of solutes on two important aspects of our DEER experiments: the spin label mobility and flap conformations. We tested the effects of glycerol, sucrose, PEG, and Ficoll400 using CW EPR, steady state fluorescence anisotropy, and 1 H 15 N HSQC NMR spectroscopy. Our results sh ow that the X band EPR line shapes are dominated by change in I Furthermore, o ur results show that I is sensitive to the presence of excluded solutes. Thus, we have shown that the spin label mobility is sensitive to the effects of preferential hydrati on. In the presence of solutes that are preferentially excluded from the protein, the mobility of the spin label is decreased because the label is forced to lie against the protein surface. The magnitude of this reduction in mobility depends on the conce ntration of the solute and the extent to which the solute is excluded from the protein surface. Solutes like glycerol and PEG, which are more strongly excluded than sucrose and Ficoll400, result in larger reductions of the spin label mobility. We have al so shown that the effects of the solutes are the result of preferential hydration and not the result of osmotic pressure effects which have been seen in other proteins. 168 However, a decrease in the spin label mobility may not have a corresponding impact on the DEER distance distributions. If the local envi ronment of the spin label is such that the label is randomly oriented along the protein surface when preferentially hydrated then the distance profiles should not be noticeably different than in the presence of non excluded solutes. Conversely, if the loc al environment is such that the label has a preferred conformation, then it is possible that the DEER distance profiles will be narrower in presence of excluded volumes. Position K55C, the labeling site used in our DEER experiments is solvent exposed and has

PAGE 213

213 minimal steric hindrance resulting from neighboring side chains. Thus, it is likely that spin labels at this site will be randomly oriented when preferentially hydrated and should not result in differences in the DEER distance distributions in presenc e of excluded or non excluded solutes. Our fluorescence anisotropy results show that R is affected by all the solutes tested to varying degrees. However, we also demonstrated that changes in R are not reported in the X band CW EPR line shape. Likewise changes in R will not affect the DEER distance distributions because the samples are frozen and thus the proteins do not tumble during the DEER experiments. Furthermore, our fluorescence anisotropy results indicate that B could be affected by the pres ence of excluded solutes such as glycerol and PEG. Additionally, the NMR results demonstrate that PEG interacts specifically with the residues in the flap tips. Because of the correlation between the effects of PEG and glycerol in the CW EPR and fluoresc ence experiments, it is likely that glycerol also interacts with the flap tips. Consequently, it is possible that the presence of excluded solutes could alter either the conformational ensemble on the fla ps or conformation of the flaps. However, further studies need to be performed in order to determine the effect of the solutes on B and to elucidate how these changes could affect the DEER distance profile. Conclusion In this work, we have shown that solutes affect both I and R but that the spin label only reports on changes in I for X Band EPR experiments. Furthermore, we demonstrated that the solute effects seen for HIV 1 PR result from preferential hydration and not osmotic pressure. We also demonstrated that PEG interacts specifically with hydro phobic patches on the surface of

PAGE 214

214 HIV 1 PR which includes the flaps tips. This finding can potentially affect the flap conformation determined from DEER distance measurements.

PAGE 215

215 CHAPTER 6 FUTURE WORK Improving the Data Analysis Process for DEER Experiments This work utilized a large number of improvements in the data analysis process for DEER data, however, there are still many improvements needed. To date, no statistical analysis has been preformed to demonstrate the repeatability of our distance profiles L ikewise, the inherent variation in the distance profiles for the apo protease has not been fully investigated. We have demonstrated that a simple change such as replacing the H 2 O in the buffer with D 2 O can alter the most probable distance by approxima tely 1 This difference suggests that other environmental changes can alter the distance profile as well but has not been investigated further. Similarly, the distance profiles for HIV 1 PR in the presence of IDV or NFV, which strongly resemble the pro files of the apo protease could also potentially be sensitive to environmental change. Investigation of Point Mutations The high rate of mutation in HIV has lead to a large number of studies on the variations in the HIV DNA and protein sequences resultin g in several important findings. Two of t hese findings include the discovery that mutations occur sequentially in drug therapy patients and that some mutations tend to occur in combination with other mutations. As discussed in Chapter 2, the mutations in HIV 1 PR can be divided in active site and non active site mutations and that the non active site mutations have been hypothesized to affect the enzymatic activity by modulating the flexibility of the flaps. Although m any of these mutations have been cha racterized by kinetic and x ray crystallography studies which provides insight into how the mutations affect both the function and structure of the protease the effects of the mutations on the flap movement and flexibility remains unclear. This relation ship, between the mutations and

PAGE 216

216 the flap flexibility, can be investigated by using the inter flap DEER distance measurements discussed in this work. The M36I mutation occurs in several HIV subtypes as well as the two drug resistant variants studied by Gali ano et al. 145 In x ray structures the native M36 residue makes several van der Waals contacts with the neighboring residues. However, the shorter Ile residue cannot occupy the same volume, so the protein backbone shifts downward to partially compensate. 169 This has been hypothesized to make the flaps more rigid. The active site mutation, V82A is commonly seen in patients receiving most protease inhibitors. However, V82A frequently occurs in combination with the I54V mutation, which in turn freq uently occurs in combination with L90M, and M46I. I54 sits in a hydrophobic pocket lined by P79, I47, and V56. Substituting I54 for a Val residue is believed to affect the ability of the flaps to open and close. 60 These mutations can be investigated as single point mutations and combinations of mutations to explore the singular and combined effects on the flap conformational ensembles. Method Validation via Model Systems The use of distance distributions from DEER experiments to characterize the conformational ensembles of a protein in this work is a novel application of DEER. However, these experiments are preformed at cryogenic temperatures and the in the presence of solutes, both of which can affect the conformational distribution. The effect of these conditions on the conformational ensembl es of proteins can be investigated by using model systems with well characterized ensembles such as calmodulin (CaM) 170 or maltose binding protein (MBP). 101 CaM is well suited for validating the population analysis method because the conformations of CaM are well known and easily controlled. CaM contains four Ca 2+ binding EF hand motifs, two in the N terminal domain and two in the C terminal domain. These domains are connected

PAGE 217

217 by an helical linker region. Upon binding Ca 2+ each domain undergoes a conformational change which exposes hydrophobic surfaces. In this state, CaM has a high affinity for its many substrates. Upon binding a substrate, the linker helix bends and CaM wraps around the substrate bringing the N and C terminal domains closer together. This conformational change can be monitored by DEER by labeling a site i n each terminus. The populations of CaM in each state can be controlled by adjusting the Ca 2+ concentrations. A variety of distances ranging from 20 to 45 can be studied by using different combinations of labeling sites. MBP also contains well studie d conformations which are easily controlled. Upon binding maltose, the C terminal domain of MBP undergoes a rigid body rotation that moves it closer to the N terminal domain. In the apo state, roughly 5 % of MBP occupies a minor conformation similar to t he closed state. The conformations of MBP can be monitored by labeling a site in each terminus and the relative population of each concentration can be controlled by varying the concentration of maltose. Isothermal Titration Calorimetry and Differential S canning Calorimetry The DEER experiments in this work have provided novel insight into the role of the flaps i n binding various inhibitors. As this work expands into other protease subtypes and variants, it will be important to have other experimental dat a for comparison and for contributing to a better understanding of the distance distributions. Isothermal titration calorimetry (ITC) is a powerful technique for simultaneously determining the enthalpic and entropic contributions to free energy of inhibit or binding and can be used to determine the thermodynamic parameters of the protease constructs binding to the inhibitors. Additionally, differential scanning calormietry (DSC) can be used to monitor the effect of binding an inhibitor on the stability of the protein.

PAGE 218

218 NMR Nuclear magnetic resonance is an especially powerful tool for investigating the dynamics of proteins The inhibitor binding to a variety of HIV 1 PR constructs can be investigated by 1 H 15 N HSQC NMR. These experiments will reveal the spe cific residues involved in binding the inhibitors and can also provide information about how point mutations affect the inhibitor binding. Similarly, relaxation measurements can reveal information about the inhibitor binding affects the dynamics of variou s residues in the HIV 1 PR and how these changes vary among different HIV 1 PR constructs.

PAGE 219

219 APPENDIX A HIV 1 PR DNA AND PROTEIN SEQUENCES Protein Sequences Table A 1 E. coli codon optimized HIV 1 Protease Variant Sequence Alignment Residues 1 4 0. Subty pe B: PQITLW K RPL VTIKIGGQLK EALL N TGADD TV I EEMSLPG Subtype C: PQITLW K RPL V S IKVGGQ I K EALL N TGADD TV I EE I A LPG V6: PQITLWQRPL VTIKIGGQL R EALL N TGADD T I F EE I SLPG Table A 2 E. coli codon optimized HIV 1 Protease Variant Sequence Alig nment Residues 4 1 7 0. Subtype B: RWKPKMIGGI GGFI C VRQYD QI I IEI A GHK Subtype C: RWKPKMIGGI GGFI C VRQYD QI I IEI A G K K V6: RWKPKMIGGI GGFI C VRQYD QI P IEI A GHK Table A 3 E. coli codon optimized HIV 1 Protease Variant Sequence Alignme nt Residues 7 1 99. Subtype B: AIGTVLVGPT PVNIIGRNL L TQI G A TLNF Subtype C: AIGTVLVGPT PVNIIGRN M L TQ L G A TLNF V6: V IGTVLVGPT P A NIIGRNL M TQI G A TLNF

PAGE 220

220 Inhibitor Structures Figure A 1 S tructures for the ni ne FDA approved inhibitors used in this work

PAGE 221

221 Figure A 2 Stick and space filling models for the nine FDA approved inhibitors used in this work

PAGE 222

222 APPENDIX B SUPPLEMENTAL INFORMA TION FOR DEER EXPERI MENTS AND DATA ANALY SIS Subtype B Figure B 1 DEE R data for apo HIV 1 PR subtype B. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR ( grey ) and the theoretical curve generated from the Gaussian reconstruction ( light grey ). B) Raw dipolar evolution curve and back ground subtraction. C) The corresponding distance profile generated via analysis with TKR ( black ) and the sum of Gaussian functions used in the reconstruction ( grey dotted). D ) The individual Gaussian functions used in the reconstruction labeled accor ding to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 223

223 Figure B 2 DEER data for CA p2 bound HIV 1 PR subtype B. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions u sed in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 224

224 Figure B 3 DEER data for IDV bound H IV 1 PR subtype B. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtracti on. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corr esponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 225

225 Figure B 4 DEER data for NFV bound HIV 1 PR subtype B. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical c urve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconst ruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 226

226 Figure B 5 DEER data for ATV bound HIV 1 PR subtype B. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corre sponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conforma tion of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 227

227 Figure B 6 DEER data for SQV bound HIV 1 PR subtype B. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated fro m the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dot ted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 228

228 Figure B 7 DEER data for R TV bound HIV 1 PR subtype B. A) Background s ubtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 229

229 Figure B 8 DEER data for LP V bound HIV 1 PR subtype B. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian rec onstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The ind ividual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 230

230 Figure B 9 DEER data for A P V bound HIV 1 PR subtype B. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency doma in spectrum. F) L curve.

PAGE 231

231 Figure B 10 DEER data for DR V bound HIV 1 PR subtype B. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (ligh t grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 232

232 Figure B 11 DEER data for TPV bound HIV 1 PR subtype B. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis wit h TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 233

233 Figure B 12 Error Analysis for populations < 15 % in HIV 1PR Apo. A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis (red) overlain with the summed Gau being validated. C) Individual populations for the Gaussian reconstruction. D J) Background subtra cted dipolar echo curve (black) overlaid with the TKR fit (red) and the modified echo curve for the distance profile with one or more populations suppressed (blue). Note, the presence of 3 4% of closed and curled conformational populations is questionable However, given the biological significance of the closed conformation and computational and crystallographic evidence of the flaps curled in, we are leaving these populations in the APO state in the analysis but do admit their significance is questionabl e at this level of SNR.

PAGE 234

234 Figure B 13 Error Analysis for populations < 20% in HIV 1PR CA p2 A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis (red) overlain wi th the summed Gaussian population profile (blue dashed). C) Individual populations for the Gaussian reconstruction. D F) Background subtracted dipolar echo curve (black) overlaid with the TKR fit (red) and the modified echo curve for the distance profile with one or more populations suppressed (blue). (Inset highlights the difference between the modified dipolar echo curve and the background subtracted echo curve.) Note, as can be seen here, in certain cases, more than one Gaussian curve was needed for a given biological population state to regenerate the TKR profiles. In these cases, the multiple Gaussians were combined together to describe the conformational state. This seems reasonable as there is no reason why a given conformational state must have a d istance profile that is Gaussian in shape.

PAGE 235

235 Figure B 14 Error Analysis for populations < 20% in HIV 1PR IDV. A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis (red) overlain with the summed discarded. C) Individual populations for the Gaussian reconstruction. D J) Background subtracted dipolar echo curve (black) overlaid with the TKR fi t (red) and the modified echo curve for the distance profile with one or more populations suppressed (blue).

PAGE 236

236 Figure B 15 Error Analysis for populations < 20% in HIV 1PR NFV. A) Table of values summarizing the populations from TKR analysis and Gaussi an reconstruction procedure. B) Distance profile from TKR analysis (red) overlain with the summed discarded. C) Individual populations for the Gaussian reconstruction. D H) Backg round subtracted dipolar echo curve (black) overlaid with the TKR fit (red) and the modified echo curve for the distance profile with one or more populations suppressed (blue).

PAGE 237

237 Figure B 16 Error Analysis for populations < 20% in HIV 1PR ATV. A) Tabl e of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis (red) overlain with the summed discarded. C) Individual populations for the Gaussian reconstruction. D) Background subtracted dipolar echo curve (black) overlaid with the TKR fit (red) and the modified echo curve for the distance profile with open population suppressed (blue).

PAGE 238

238 Figure B 17 Error Analysis for populations < 20% in HIV 1PR SQV. A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis (red) overlain with the summed Gaussian population profile (blue discarded. C) Individual populations for the Gaussian reconstruction. D F) Background subtracted dipolar echo curve (black) overlaid with the TKR fit (red) and the modified echo curve for the distance profile wit h one or more populations suppressed (blue).

PAGE 239

239 Figure B 18 Error Analysis for populations < 20% in HIV 1PR RTV. A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analy sis (red) overlain with the summed Gaussian population profile (blue dashed). discarded. C) Individual populations for the Gaussian reconstruction. D F) Background subtracted dipolar echo curve (black) overlaid with the TK R fit (red) and the modified echo curve for the distance profile with one or more populations suppressed (blue).

PAGE 240

240 Figure B 19 Error Analysis for populations < 20% in HIV 1PR LPV. A) Table of values summarizing the populations from TKR analysis and Ga ussian reconstruction procedure. B) Distance profile from TKR analysis (red) overlain with the summed Gaussian population profile (blue dashed). nt dipolar echo curve for the distance profile containing this population is consistent with the background subtracted dipolar echo curve. C) Individual populations for the Gaussian reconstruction. D I) Background subtracted dipolar echo curve (black) overlaid with the TKR fit (red) and the modified echo curve for the distance profile with one or more populations suppressed (blue).

PAGE 241

241 Figure B 20 Error Analysis for popula tions < 20% in HIV 1PR APV. A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis (red) overlain with the summed Gaussian population profile (blue dashed). C) Individ ual populations for the Gaussian reconstruction. D F) Background subtracted dipolar echo curve (black) overlaid with the TKR fit (red) and the modified echo curve for the distance profile with one or more populations suppressed (blue).

PAGE 242

242 Figure B 21 Err or Analysis for populations < 20% in HIV 1PR DRV. A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis (red) overlain with the summed Gaussian population profile (blu e dashed). P discarded. C) Individual populations for the Gaussian reconstruction. D K) Background subtracted dipolar echo curve (black) overlaid with the TKR fit (red) and the modified echo curve for the distance profile wi th one or more populations suppressed (blue).

PAGE 243

243 Figure B 22 Error Analysis for populations < 20% in HIV 1PR TPV. A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analys is (red) overlain with the summed Gaussian population profile (blue dashed). discarded. C) Individual populations for the Gaussian reconstruction. D J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit (red) and the modified echo curve for the distance profile with one or more populations suppressed (blue).

PAGE 244

244 Subtype C Figure B 23 DEER data for apo HIV 1 PR subtype C (collected 5/09) A) Background subtracted dipolar evolution curve (black) ov erlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (bla ck) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 245

245 Figure B 24 DEER data for apo HIV 1 PR subtype C (collected 6/09). A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) R aw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 246

246 Figure B 2 5 DEER data for CA p2 bound HIV 1 PR subtype C. A) Background subtracted dipolar evolution curve (black) overl aid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 247

247 Fi gure B 2 6 DEER data for IDV bound HIV 1 PR subtype C. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evo lution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconst ruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 248

248 Figure B 2 7 DEER data for NFV bound HIV 1 PR subtype C. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 249

249 Figure B 2 8 DEER data for ATV bound HIV 1 PR subtype C. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 250

250 Figure B 2 9 DEER data for APV bound HIV 1 PR subtype C. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functio ns used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 251

251 Figure B 30 DEER data for LPV bo und HIV 1 PR subtype C. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subt raction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 252

252 Figure B 3 1 DEER data for RTV bound HIV 1 PR subtype C. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoret ical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the r econstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 253

253 Figure B 3 2 DEER data for SQV bound HIV 1 PR sub type C. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) Th e corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding c onformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 254

254 Figure B 3 3 DEER data for DRV bound HIV 1 PR subtype C. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve gener ated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (g rey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 255

255 Figure B 3 4 DEER data for TPV bound HIV 1 PR subtype C. A) Bac kground subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of H IV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 256

256 Figure B 3 5 Error Analysis for populations < 20% in apo subtype C HIV 1PR (collected 5/09) A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). Populations marked with J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 257

257 Figure B 36 Error Analysis for populations < 20% in apo subtype C HIV 1PR (collected 6/09) A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). Populations marked with ) Individual populations for the Gaussian reconstruction. D J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 258

258 Fig ure B 3 7 Error Analysis for populations < 20% in CA p2 bound subtype C HIV 1PR A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summ ed G aussian population profile (grey dashed). Populations marked with J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified e cho curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 259

259 Figure B 3 8 Error Analysis for populations < 20% in IDV bound subtype C HIV 1PR A) Table of values summarizing the populations from TKR analysis and Gaussian reconst ruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). Populations marked with J) Background su btracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 260

2 60 Figure B 3 9 Error Analysis for populations < 20% in NFV bound subtype C HIV 1PR A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). Populations marked with iscarded. C) Individual populations for the Gaussian reconstruction. D J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 261

261 Figure B 3 9 Error Analysis for populations < 20% in ATV bound subtype C HIV 1PR A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain wit h the summed G aussian population profile (grey dashed). Populations marked with J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 262

262 Figure B 40 Error Analysis for populations < 20% in APV bound subtype C HIV 1PR A) Table of values summarizing the populations from TKR analysis and Gaussi an reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). Populations marked with J) Bac kground subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 263

263 Figure B 41 Error Analysis for populations < 20% in LPV bound subtype C HIV 1PR A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). Populations marked with J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppre ssed ( grey ).

PAGE 264

264 Figure B 42 Error Analysis for populations < 20% in RTV bound subtype C HIV 1PR A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) ov erlain with the summed G aussian population profile (grey dashed). Populations marked with J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 265

265 Figure B 43 Error Analysis for populations < 20% in SQV bound subtype C HIV 1PR A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). Populations marked with D J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 266

266 Figure B 44 Error Analysis for populations < 20% in DRV b ound subtype C HIV 1PR A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). Populations marked with J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more popula tions suppressed ( grey ).

PAGE 267

267 Figure B 4 5 Error Analysis for populations < 20% in TPV bound subtype C HIV 1PR A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). Populations marked with J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 268

268 Variant V6 Figure B 46 DEER data for apo HIV 1 PR V6. A) Background subtracted dipolar evolution curve (black) overlaid with the fit fr om TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gau ssian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 269

269 Figure B 47 DEER da ta for CA p2 bound HIV 1 PR V6. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and backgro und subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled accordin g to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 270

270 Figure B 48 DEER data for IDV bound HIV 1 PR V6. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theore tical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 271

271 Figure B 49 DEER data for NFV bound HIV 1 PR V6 A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corr esponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conform ation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 272

272 Figure B 50 DEER data for ATV bound HIV 1 PR V6. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 273

273 Figure B 51 DEER data for APV bound HIV 1 PR V6. A) Background subtracted d ipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile gen erated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequen cy domain spectrum. F) L curve.

PAGE 274

274 Figure B 52 DEER data for LPV bound HIV 1 PR V6. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (ligh t grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 275

275 Figure B 53 DEER data for RTV bound HIV 1 PR V6. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR ( black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve

PAGE 276

276 Figure B 54 DEER data for SQV bound HIV 1 PR V6. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evol ution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstr uction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 277

277 Figure B 55 DEER data for DRV bound HIV 1 PR V6. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TK R (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background subtraction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to the corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 278

278 Figure B 56 DEER data fo r TPV bound HIV 1 PR V6. A) Background subtracted dipolar evolution curve (black) overlaid with the fit from TKR (grey) and the theoretical curve generated from the Gaussian reconstruction (light grey). B) Raw dipolar evolution curve and background sub traction. C) The corresponding distance profile generated via analysis with TKR (black) and the sum of Gaussian functions used in the reconstruction (grey, dotted). D) The individual Gaussian functions used in the reconstruction labeled according to th e corresponding conformation of HIV 1 PR. E) Frequency domain spectrum. F) L curve.

PAGE 279

279 Figure B 57 Error Analysis for populations < 20% in apo HIV 1PR V6 A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction proce dure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). discarded. C) Individual populations for the Gaussian reconstruction. D J) Background subtracted dipo lar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 280

280 Figure B 58 Error Analysis for populations < 20% in CA p2 bound HIV 1PR V6 A) Table of val ues summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). discarded. C) Indiv idual populations for the Gaussian reconstruction. D J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 281

281 Figure B 59 Error Analysis for populations < 20% in IDV bound HIV 1PR V6 A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian popul ation profile (grey dashed). discarded. C) Individual populations for the Gaussian reconstruction. D J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 282

282 Figure B 60 Error Analysis for populations < 20% in NFV bound HIV 1PR V6 A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) D istance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). discarded. C) Individual populations for the Gaussian reconstruction. D J) Background subtracted dipolar echo cu rve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 283

283 Figure B 61 Error Analysis for populations < 20% in ATV bound HIV 1PR V6 A) Table of values summarizi ng the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). discarded. C) Individual populat ions for the Gaussian reconstruction. D J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 284

284 Figure B 62 Error Anal ysis for populations < 20% in APV bound HIV 1PR V6 A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). discarded. C) Individual populations for the Gaussian reconstruction. D J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance pro file with one or more populations suppressed ( grey ).

PAGE 285

285 Figure B 63 Error Analysis for populations < 20% in LPV bound HIV 1PR V6 A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profi le from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). discarded. C) Individual populations for the Gaussian reconstruction. D J) Background subtracted dipolar echo curve (black) o verlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 286

286 Figure B 64 Error Analysis for populations < 20% in RTV bound HIV 1PR V6 A) Table of values summarizing the popula tions from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). discarded. C) Individual populations for the Gaussian reconstruction. D J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 287

287 Figure B 65 Error Analysis for popu lations < 20% in SQV bound HIV 1PR V6 A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed ). discarded. C) Individual populations for the Gaussian reconstruction. D J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 288

288 Figure B 66 Error Analysis for populations < 20% in DRV bound HIV 1PR V6 A) Table of values summarizing the populations from TKR analysis and Gaussian reconstruction procedure. B) Distance profile from TKR a nalysis ( black ) overlain with the summed G aussian population profile (grey dashed). discarded. C) Individual populations for the Gaussian reconstruction. D J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 289

289 Figure B 67 Error Analysis for populations < 20% in TPV bound HIV 1PR V6 A) Table of values summarizing the populations from TK R analysis and Gaussian reconstruction procedure. B) Distance profile from TKR analysis ( black ) overlain with the summed G aussian population profile (grey dashed). discarded. C) Individual populations for the Gaussian reco nstruction. D J) Background subtracted dipolar echo curve (black) overlaid with the TKR fit ( black ) and the modified echo curve for the distance profile with one or more populations suppressed ( grey ).

PAGE 290

290 LIST OF REFEENCES 1. Sepkowitz, K. A. (2001). AIDS -The First 20 Years. N Engl J Med 344 1764 1772. 2. Weiss, R. A. (1993). How Does Hiv Cause Aids. Science 260 1273 1279. 3. Levy, J. A. (2007). HIV and the Pathogensis of AIDS 3rd edit, ASM Press, Washington DC. 4. Cahn, P., Coo per, D. A., Tindall, B., Reibnegger, G., Steiner, B., Wachter, H., Zangerle, R., Colebunders, R., Cortes, E., Cameron, D. W., Garber, G., Howleyfoss, N., Montaner, J. S. G., Schechter, M., Mathiesen, L., Pindborg, J. J., Charmot, G., Coulaud, J. P., Detruc his, P., Fegueux, S., Lafaix, C., Malkin, J. E., Maslo, C., Matheron, S., Prazuck, T., Rozenbaum, W., Brodt, H., Alvarez, M., Cantaluppi, P., Esposito, R., Lizioli, A., Goto, H., Ikegami, N., Oka, S., Shimada, K., Sugiyama, K., Yoshioka, K., Attas, H., Man daliya, D., Waweru, H., Mwansambo, P., Wirima, J., Deleon, S. P., Lange, J. M. A., Gonzalezlahoz, J., Najeramorrndo, R., Lallinger, G., Mbaga, I., Mwakyusa, D., Perenboom, R., Richter, C., Katabira, E., Mubiru, F., Mugerwa, R., Weber, J., Castro, K., Mills J., Moss, R. A., Redfield, R., Royce, R., Schreger, L., Wagner, K., Pokrovsky, V., Yurin, O., Bianco, N., Deibis, L., Perezecheverria, G., Almaviva, M., Ceravolo, C., Kalambay, K., Kapita, B., Piccioni, I., Welo, K., Burton, A., Chin, J., Farley, T., Hen nessey, K., Kristoffersson, S. P., Maj, M., Mann, J., Osmanov, S., Vonsonnenburg, F. & Tamashiro, H. (1993). Proposed World Health Organization Staging System for Hiv Infection and Disease Preliminary Testing by an International Collaborative Cross Secti onal Study. Aids 7 711 718. 5. Hirsch, H. H., Kaufmann, G., Sendi, P. & Battegay, M. (2004). Immune reconstitution in HIV infected patients. Clinical Infectious Diseases 38 1159 1166. 6. Cadogan, M. & Dalgleish, A. G. (2008). HIV Induced AIDS and Relat ed Cancers: Chronic Immune Activation and Future Therapeutic Strategies. Advances in Cancer Research, Vol 101 101 349 395. 7. Sharp, P. M., Bailes, E., Chaudhuri, R. R., Rodenburg, C. M., Santiago, M. O. & Hahn, B. H. (2001). The origins of acquired immu ne deficiency syndrome viruses: where and when? Philosophical Transactions of the Royal Society of London Series B Biological Sciences 356 867 876. 8. Nahmias, A. J., Weiss, J., Yao, X., Lee, F., Kodsi, R., Schanfield, M., Matthews, T., Bolognesi, D., Du rack, D., Motulsky, A., Kanki, P. & Essex, M. (1986). EVIDENCE FOR HUMAN INFECTION WITH AN HTLV III/LAV LIKE VIRUS IN CENTRAL AFRICA, 1959. The Lancet 327 1279 1280. 9. UNAIDS. (2008). 2008 Report on the Global AIDS Epidemic. UNAIDS.

PAGE 291

291 10. Obrien, T. R., George, J. R. & Holmberg, S. D. (1992). Human Immunodeficiency Virus Type 2 Infection in the United States Epidemiology, Diagnosis, and Public Health Implications. Jama Journal of the American Medical Association 267 2775 2779. 11. Whittle, H., Morris, J., Todd, J., Corrah, T., Sabally, S., Bangali, J., Ngom, P. T., Rolfe, M. & Wilkins, A. (1994). Hiv 2 Infected Patients Survive Longer Than Hiv 1 Infected Patients. Aids 8 1617 1620. 12. Ratner, L., Haseltine, W., Patarca, R., Livak, K. J., Starcich, B ., Josephs, S. F., Doran, E. R., Rafalski, J. A., Whitehorn, E. A., Baumeister, K., Ivanoff, L., Petteway, S. R., Pearson, M. L., Lautenberger, J. A., Papas, T. S., Ghrayeb, J., Chang, N. T., Gallo, R. C. & Wong Staal, F. (1985). Complete nucleotide sequen ce of the AIDS virus, HTLV III. Nature 313 277 284. 13. Jacks, T., Townsley, K., Varmus, H. E. & Majors, J. (1987). 2 Efficient Ribosomal Frameshifting Events Are Required for Synthesis of Mouse Mammary Tumor Virus Gag Related Polyproteins. Proceedings o f the National Academy of Sciences of the United States of America 84 4298 4302. 14. Jacks, T., Madhani, H. D., Masiarz, F. R. & Varmus, H. E. (1988). Signals for Ribosomal Frameshifting in the Rous Sarcoma Virus Gag Pol Region. Cell 55 447 458. 15. Cr aven, R. C., Bennett, R. P. & Wills, J. W. (1991). Role of the Avian Retroviral Protease in the Activation of Reverse Transcriptase during Virion Assembly. Journal of Virology 65 6205 6217. 16. Ingr, M., Uhlikova, T., Strisovsky, K., Majerova, E. & Konva linka, J. (2003). Kinetics of the dimerization of retroviral proteases: The "fireman's grip" and dimerization. Protein Science 12 2173 2182. 17. Greene, W. C. & Peterlin, B. M. (2002). Charting HIV's remarkable voyage through the cell: Basic science as a passport to future therapy. Nature Medicine 8 673 680. 18. Malim, M. H. & Emerman, M. (2008). HIV 1 Accessory Proteins -Ensuring Viral Survival in a Hostile Environment. Cell Host & Microbe 3 388 398. 19. Zhang, Y. Q. & Barklis, E. (1997). Effects of nucleocapsid mutations on human immunodeficiency virus assembly and RNA encapsidation. Journal of Virology 71 6765 6776. 20. Kuznetsov, Y. G., Victoria, J. G., Robinson, W. E. & McPherson, A. (2003). Atomic force microscopy investigation of human immunod eficiency virus (HIV) and HIV Infected lymphocytes. Journal of Virology 77 11896 11909.

PAGE 292

292 21. Gelderblom, H. R., Ozel, M., Hausmann, E. H. S., Winkel, T., Pauli, G. & Koch, M. A. (1988). FINE STRUCTURE OF HUMAN IMMUNODEFICIENCY VIRUS (HIV), IMMUNOLOCALIZA TION OF STRUCTURAL PROTEINS AND VIRUS CELL RELATION. Micron and Microscopica Acta 19 41 60. 22. Zhu, P., Liu, J., Bess, J., Chertova, E., Lifson, J. D., Gris, H., Ofek, G. A., Taylor, K. A. & Roux, K. H. (2006). Distribution and three dimensional struc ture of AIDS virus envelope spikes. Nature 441 847 852. 23. Resh, M. D. (2004). A myristoyl switch regulates membrane binding of HIV 1 Gag. Proceedings of the National Academy of Sciences of the United States of America 101 417 418. 24. Ott, D. E., Cor en, L. V., Kane, B. P., Busch, L. K., Johnson, D. G., Sowder, R. C., Chertova, E. N., Arthur, L. O. & Henderson, L. E. (1996). Cytoskeletal proteins inside human immunodeficiency virus type 1 virions. Journal of Virology 70 7734 7743. 25. Wainhobson, S., Sonigo, P., Danos, O., Cole, S. & Alizon, M. (1985). Nucleotide Sequence of the Aids Virus, Lav. Cell 40 9 17. 26. Sonigo, P., Alizon, M., Staskus, K., Klatzmann, D., Cole, S., Danos, O., Retzel, E., Tiollais, P., Haase, A. & Wainhobson, S. (1985). Nucl eotide Sequence of the Visna Lentivirus Relationship to the Aids Virus. Cell 42 369 382. 27. Toh, H., Ono, M., Saigo, K. & Miyata, T. (1985). Retroviral Protease Like Sequence in the Yeast Transposon Ty1. Nature 315 691 691. 28. Kramer, R. A., Schabe r, M. D., Skalka, A. M., Ganguly, K., Wongstaal, F. & Reddy, E. P. (1986). Htlv Iii Gag Protein Is Processed in Yeast Cells by the Virus Pol Protease. Science 231 1580 1584. 29. Debouck, C., Gorniak, J. G., Strickler, J. E., Meek, T. D., Metcalf, B. W. & Rosenberg, M. (1987). Human Immunodeficiency Virus Protease Expressed in Escherichia Coli Exhibits Autoprocessing and Specific Maturation of the Gag Precursor. Proceedings of the National Academy of Sciences of the United States of America 84 8903 8906. 30. Schneider, J. & Kent, S. B. H. (1988). Enzymatic activity of a synthetic 99 residue protein corresponding to the putative HIV 1 protease. Cell 54 363 368. 31. Nutt, R. F., Brady, S. F., Darke, P. L., Ciccarone, T. M., Colton, C. D., Nutt, E. M., Rod key, J. A., Bennett, C. D., Waxman, L. H., Sigal, I. S., Anderson, P. S. & Veber, D. F. (1988). Chemical Synthesis and Enzymatic Activity of a 99 Residue Peptide with a Sequence Proposed for the Human Immunodeficiency Virus Protease. Proceedings of the Nat ional Academy of Sciences of the United States of America 85 7129 7133.

PAGE 293

293 32. Lillehoj, E. P., Salazar, F. H. R., Mervis, R. J., Raum, M. G., Chan, H. W., Ahmad, N. & Venkatesan, S. (1988). Purification and Structural Characterization of the Putative Gag P ol Protease of Human Immunodeficiency Virus. Journal of Virology 62 3053 3058. 33. Seelmeier, S., Schmidt, H., Turk, V. & Vonderhelm, K. (1988). Human Immunodeficiency Virus Has an Aspartic Type Protease That Can Be Inhibited by Pepstatin A. Proceedings of the National Academy of Sciences of the United States of America 85 6612 6616. 34. Legrice, S. F. J., Mills, J. & Mous, J. (1988). Active Site Mutagenesis of the Aids Virus Protease and Its Alleviation by Trans Complementation. Embo Journal 7 2547 25 53. 35. Kohl, N. E., Emini, E. A., Schleif, W. A., Davis, L. J., Heimbach, J. C., Dixon, R. A. F., Scolnick, E. M. & Sigal, I. S. (1988). Active Human Immunodeficiency Virus Protease Is Required for Viral Infectivity. Proceedings of the National Academy o f Sciences of the United States of America 85 4686 4690. 36. Richards, A. D., Roberts, R., Dunn, B. M., Graves, M. C. & Kay, J. (1989). Effective Blocking of Hiv 1 Proteinase Activity by Characteristic Inhibitors of Aspartic Proteinases. Febs Letters 247 113 117. 37. Moore, M. L., Bryan, W. M., Fakhoury, S. A., Magaard, V. W., Huffman, W. F., Dayton, B. D., Meek, T. D., Hyland, L., Dreyer, G. B., Metcalf, B. W., Strickler, J. E., Gorniak, J. G. & Debouck, C. (1989). Peptide Substrates and Inhibitors of the Hiv 1 Protease. Biochemical and Biophysical Research Communications 159 420 425. 38. Dreyer, G. B., Metcalf, B. W., Tomaszek, T. A., Carr, T. J., Chandler, A. C., Hyland, L., Fakhoury, S. A., Magaard, V. W., Moore, M. L., Strickler, J. E., Debouck, C & Meek, T. D. (1989). Inhibition of Human Immunodeficiency Virus 1 Protease Invitro Rational Design of Substrate Analog Inhibitors. Proceedings of the National Academy of Sciences of the United States of America 86 9752 9756. 39. Vonderhelm, K., Gurt ler, L., Eberle, J. & Deinhardt, F. (1989). Inhibition of Hiv Replication in Cell Culture by the Specific Aspartic Protease Inhibitor Pepstatin A. Febs Letters 247 349 352. 40. Mcquade, T. J., Tomasselli, A. G., Liu, L., Karacostas, V., Moss, B., Sawyer, T. K., Heinrikson, R. L. & Tarpley, W. G. (1990). A Synthetic Hiv 1 Protease Inhibitor with Antiviral Activity Arrests Hiv Like Particle Maturation. Science 247 454 456. 41. Miller, M., Schneider, J., Sathyanarayana, B. K., Toth, M. V., Marshall, G. R., Clawson, L., Selk, L., Kent, S. B. H. & Wlodawer, A. (1989). Structure of Complex of Synthetic Hiv 1 Protease with a Substrate Based Inhibitor at 2.3 a Resolution. Science 246 1149 1152.

PAGE 294

294 42. Navia, M. A., Fitzgerald, P. M. D., Mckeever, B. M., Leu, C. T ., Heimbach, J. C., Herber, W. K., Sigal, I. S., Darke, P. L. & Springer, J. P. (1989). 3 Dimensional Structure of Aspartyl Protease from Human Immunodeficiency Virus Hiv 1. Nature 337 615 620. 43. Wlodawer, A., Miller, M., Jaskolski, M., Sathyanarayana, B. K., Baldwin, E., Weber, I. T., Selk, L. M., Clawson, L., Schneider, J. & Kent, S. B. H. (1989). Conserved Folding in Retroviral Proteases Crystal Structure of a Synthetic Hiv 1 Protease. Science 245 616 621. 44. Yamazaki, T., Hinck, A. P., Wang, Y. X., Nicholson, L. K., Torchia, D. A., Wingfield, P., Stahl, S. J., Kaufman, J. D., Chang, C. H., Domaille, P. J. & Lam, P. Y. S. (1996). Three dimensional solution structure of the HIV 1 protease complexed with DMP323, a novel cyclic urea type inhibitor, determined by nuclear magnetic resonance spectroscopy. Protein Science 5 495 506. 45. Wlodawer, A. & Gustchina, A. (2000). Structural and biochemical studies of retroviral proteases. Biochimica Et Biophysica Acta Protein Structure and Molecular Enzymolog y 1477 16 34. 46. Strisovsky, K., Tessmer, U., Langner, J., Konvalinka, J. & Krausslich, H. G. (2000). Systematic mutational analysis of the active site threonine of HIV 1 proteinase: Rethinking the "fireman's grip" hypothesis. Protein Science 9 1631 16 41. 47. Pearl, L. & Blundell, T. (1984). The Active Site of Aspartic Proteinases. Febs Letters 174 96 101. 48. Gustchina, A. & Weber, I. T. (1990). Comparison of Inhibitor Binding in Hiv 1 Protease and in Nonviral Aspartic Proteases the Role of the Fl ap. Febs Letters 269 269 272. 49. Harte, W. E., Swaminathan, S., Mansuri, M. M., Martin, J. C., Rosenberg, I. E. & Beveridge, D. L. (1990). DOMAIN COMMUNICATION IN THE DYNAMIC STRUCTURE OF HUMAN IMMUNODEFICIENCY VIRUS 1 PROTEASE. Proceedings of the Natio nal Academy of Sciences of the United States of America 87 8864 8868. 50. Hornak, V., Okur, A., Rizzo, R. C. & Simmerling, C. (2006). HIV 1 protease flaps spontaneously open and reclose in molecular dynamics simulations. Proceedings of the National Acade my of Sciences of the United States of America 103 915 920. 51. De Clercq, E. (2009). Anti HIV drugs: 25 compounds approved within 25 years after the discovery of HIV. International Journal of Antimicrobial Agents 33 307 320. 52. Carr, A. & Cooper, D. A. (2000). Adverse effects of antiretroviral therapy. Lancet 356 1423 1430.

PAGE 295

295 53. Rowe, T. C., Weissig, V. & Lawrence, J. W. (2001). Mitochondrial DNA metabolism targeting drugs. Advanced Drug Delivery Reviews 49 175 187. 54. Tomasselli, A. G., Olsen, M. K., Hui, J. O., Staples, D. J., Sawyer, T. K., Heinrikson, R. L. & Tomich, C. S. C. (1990). Substrate Analog Inhibition and Active Site Titration of Purified Recombinant Hiv 1 Protease. Biochemistry 29 264 269. 55. la Porte, C. J. L. (2009). Saquinavir, the pioneer antiretroviral protease inhibitor. Expert Opinion on Drug Metabolism & Toxicology 5 1313 1322. 56. Turner, S. R., Strohbach, J. W., Tommasi, R. A., Aristoff, P. A., Johnson, P. D., Skulnick, H. I., Dolak, L. A., Seest, E. P., Tomich, P. K., Bohanan, M. J., Horng, M. M., Lynn, J. C., Chong, K. T., Hinshaw, R. R., Watenpaugh, K. D., Janakiraman, M. N. & Thaisrivongs, S. (1998). Tipranavir (PNU 140690): A potent, orally bioavailable nonpeptidic HIV protease inhibitor of the 5,6 dihydro 4 hydroxy 2 pyrone sulfonamide class. Journal of Medicinal Chemistry 41 3467 3476. 57. Boggetto, N. & Reboud Ravaux, M. (2002). Dimerization inhibitors of HIV 1 protease. Biological Chemistry 383 1321 1324. 58. Perryman, A. L., Lin, J. H. & McCammon, J. A. (200 6). Restrained molecular dynamics simulations of HIV 1 protease: The first step in validating a new target for drug design. Biopolymers 82 272 284. 59. Rezacova, P., Pokorna, J., Brynda, J., Kozisek, M., Cigler, P., Lepsik, M., Fanfrlik, J., Rezac, J., S askova, K. G., Sieglova, I., Plesek, J., Sicha, V., Gruner, B., Oberwinkler, H., Sedlacek, J., Krausslich, H. G., Hobza, P., Kral, V. & Konvalinka, J. (2009). Design of HIV Protease Inhibitors Based on Inorganic Polyhedral Metallacarboranes. Journal of Med icinal Chemistry 52 7132 7141. 60. Clemente, J. C., Moose, R. E., Hemrajani, R., Whitford, L. R. S., Govindasamy, L., Reutzel, R., McKenna, R., Agbandje McKenna, M., Goodenow, M. M. & Dunn, B. M. (2004). Comparing the accumulation of active and nonactiv e site mutations in the HIV 1 protease. Biochemistry 43 12141 12151. 61. Muzammil, S., Armstrong, A. A., Kang, L. W., Jakalian, A., Bonneau, P. R., Schmelmer, V., Amzel, L. M. & Freire, E. (2007). Unique thermodynamic response of tipranavir to human immu nodeficiency virus type 1 protease drug resistance mutations. Journal of Virology 81 5144 5154. 62. Yanchunas, J., Langley, D. R., Tao, L., Rose, R. E., Friborg, J., Colonno, R. J. & Doyle, M. L. (2005). Molecular basis for increased susceptibility of is olates with atazanavir resistance conferring substitution I50L to other protease inhibitors. Antimicrobial Agents and Chemotherapy 49 3825 3832.

PAGE 296

296 63. Prabu Jeyabalan, M., King, N. M., Nalivaika, E. A., Heilek Snyder, G., Cammack, N. & Schiffer, C. A. (200 6). Substrate envelope and drug resistance: Crystal structure of RO1 in complex with wild type human immunodeficiency virus type 1 protease. Antimicrobial Agents and Chemotherapy 50 1518 1521. 64. Prabu Jeyabalan, M., Nalivaika, E. A., King, N. M. & Schi ffer, C. A. (2003). Viability of a drug resistant human immunodeficiency virus type 1 protease variant: Structural insights for better antiviral therapy. Journal of Virology 77 1306 1315. 65. Reddy, G. S. K. K., Ali, A., Nalam, M. N. L., Anjum, S. G., Ca o, H., Nathans, R. S., Schiffer, C. A. & Rana, T. M. (2007). Design and synthesis of HIV 1 protease inhibitors incorporating oxazolidinones as P2/P2' ligands in pseudosymmetric dipeptide isosteres. Journal of Medicinal Chemistry 50 4316 4328. 66. Nalam, M. N. L., Peeters, A., Jonckers, T. H. M., Dierynck, I. & Schiffer, C. A. (2007). Crystal structure of lysine sulfonamide inhibitor reveals the displacement of the conserved flap water molecule in human immunodeficiency virus type 1 protease. Journal of Vi rology 81 9512 9518. 67. Ohtaka, H. & Freire, E. (2005). Adaptive inhibitors of the HIV 1 protease. Progress in Biophysics & Molecular Biology 88 193 208. 68. Drake, J. W., Charlesworth, B., Charlesworth, D. & Crow, J. F. (1998). Rates of spontaneous m utation. Genetics 148 1667 1686. 69. Mansky, L. M. & Temin, H. M. (1995). Lower in Vivo Mutation Rate of Human Immunodeficiency Virus Type 1 Than That Predicted from the Fidelity of Purified Reverse Transcriptase. Journal of Virology 69 5087 5094. 70. Peeters, M., Toure Kane, C. & Nkengasong, J. N. (2003). Genetic diversity of HIV in Africa: impact on diagnosis, treatment, vaccine development and trials. Aids 17 2547 2560. 71. Rose, R. B., Craik, C. S. & Stroud, R. M. (1998). Domain flexibility in ret roviral proteases: Structural implications for drug resistant mutations. Biochemistry 37 2607 2621. 72. Freedberg, D. I., Ishima, R., Jacob, J., Wang, Y. X., Kustanovich, I., Louis, J. M. & Torchia, D. A. (2002). Rapid structural fluctuations of the free HIV protease flaps in solution: Relationship to crystal structures and comparison with predictions of dynamics calculations. Protein Science 11 221 232. 73. Ishima, R., Freedberg, D. I., Wang, Y. X., Louis, J. M. & Torchia, D. A. (1999). Flap opening an d dimer interface flexibility in the free and inhibitor bound HIV protease, and their implications for function. Structure 7 1047 1055.

PAGE 297

297 74. Collins, J. R., Burt, S. K. & Erickson, J. W. (1995). Flap Opening in Hiv 1 Protease Simulated by Activated Molecu lar Dynamics. Nature Structural Biology 2 334 338. 75. Martin, P., Vickrey, J. F., Proteasa, G., Jimenez, Y. L., Wawrzak, Z., Winters, M. A., Merigan, T. C. & Kovari, L. C. (2005). "Wide Open" 1.3 Structure of a Multidrug Resistant HIV 1 Protease as a Drug Target. Structure 13 1887 1895. 76. Layten, M., Hornak, V. & Simmerling, C. (2006). The open structure of a multi drug resistant HIV 1 protease is stabilized by crystal packing contacts. Journal of the American Chemical Society 128 13360 13361. 77 Hornak, V., Okur, A., Rizzo, R. C. & Simmerling, C. (2006). HIV 1 protease flaps spontaneously close to the correct structure in simulations following manual placement of an inhibitor into the open state. Journal of the American Chemical Society 128 281 2 2813. 78. Scott, W. R. P. & Schiffer, C. A. (2000). Curling of flap tips in HIV 1 protease as a mechanism for substrate entry and tolerance of drug resistance. Structure 8 1259 1265. 79. Toth, G. & Borics, A. (2006). Flap opening mechanism of HIV 1 pr otease. Journal of Molecular Graphics & Modelling 24 465 474. 80. Ishima, R., Torchia, D. A. & Louis, J. M. (2007). Mutational and structural studies aimed at characterizing the monomer of HIV 1 protease and its precursor. Journal of Biological Chemistry 282 17190 17199. 81. Galiano, L., Bonora, M. & Fanucci, G. E. (2007). Interflap distances in HIV 1 protease determined by pulsed EPR measurements. Journal of the American Chemical Society 129 11004 +. 82. Ding, F., Layten, M. & Simmerling, C. (2008). Solution structure of HIV 1 protease flaps probed by comparison of molecular dynamics simulation ensembles and EPR experiments. Journal of the American Chemical Society 130 7184 +. 83. Ode, H., Matsuyama, S., Hata, M., Neya, S., Kakizawa, J., Sugiura, W. & Hoshino, T. (2007). Computational characterization of structural role of the non active site mutation M36I of human immunodeficiency virus type 1 protease. Journal of Molecular Biology 370 598 607. 84. Dunker, A. K., Brown, C. J. & Obradovic, Z. (2002 ). Identification and functions of usefully disordered proteins. In Unfolded Proteins Vol. 62, pp. 25 49. Academic Press Inc, San Diego. 85. Chou, P. Y. & Fasman, G. D. (1974). Conformational Parameters for Amino Acids in Helical, Beta Sheet, and Random Coil Regions Calculated from Proteins. Biochemistry 13 211 222.

PAGE 298

298 86. Chou, P. Y. & Fasman, G. D. (1974). Prediction of Protein Conformation. Biochemistry 13 222 245. 87. Chou, P. Y. & Fasman, G. D. (1973). Structural and Functional Role of Leucine Resid ues in Proteins. Journal of Molecular Biology 74 263 281. 88. Koehl, P. & Levitt, M. (1999). Structure based conformational preferences of amino acids. Proceedings of the National Academy of Sciences of the United States of America 96 12524 12529. 89. Dill, K. A., Bromberg, S., Yue, K. Z., Fiebig, K. M., Yee, D. P., Thomas, P. D. & Chan, H. S. (1995). PRINCIPLES OF PROTEIN FOLDING A PERSPECTIVE FROM SIMPLE EXACT MODELS. Protein Science 4 561 602. 90. Fisher, E. (1894). Einfluss der configuration auf die wirkgin den enzyme. Berichte der Deutschen Chemischen Gesellschaft (Ber Dtsch Chem Ges) 27 2985 2993. 91. Anson, M. L. & Mirsky, A. E. (1925). On some general properties of proteins. Journal of General Physiology 9 169 179. 92. Mirsky, A. E. & Pau ling, L. (1936). On the structure of native, denatured, and coagulated proteins. Proceedings of the National Academy of Sciences of the United States of America 22 439 447. 93. Perutz, M. F., Steinrauf, L. K., Stockell, A. & Bangham, A. D. (1959). Chemic al and Crystallographic Study of the 2 Fractions of Adult Horse Haemoglobin. Journal of Molecular Biology 1 402 &. 94. Kendrew, J. C., Bodo, G., Dintzis, H. M., Parrish, R. G., Wyckoff, H. & Phillips, D. C. (1958). 3 Dimensional Model of the Myoglobin Mo lecule Obtained by X Ray Analysis. Nature 181 662 666. 95. Feyman, R. P. (1996). Six Easy Pieces: Essentials Of Physics By Its Most Brilliant Teacher Basic Books, New York, NY. 96. Henzler Wildman, K. & Kern, D. (2007). Dynamic personalities of protein s. Nature 450 964 972. 97. Tama, F. & Sanejouand, Y. H. (2001). Conformational change of proteins arising from normal mode calculations. Protein Engineering 14 1 6. 98. Popovych, N., Sun, S. J., Ebright, R. H. & Kalodimos, C. G. (2006). Dynamically dri ven protein allostery. Nature Structural & Molecular Biology 13 831 838.

PAGE 299

299 99. Li, P. L., Martins, I. R. S., Amarasinghe, G. K. & Rosen, M. K. (2008). Internal dynamics control activation and activity of the autoinhibited Vav DH domain. Nature Structural & Molecular Biology 15 613 618. 100. Smock, R. G. & Gierasch, L. M. (2009). Sending Signals Dynamically. Science 324 198 203. 101. Tang, C., Schwieters, C. D. & Clore, G. M. (2007). Open to closed transition in apo maltose binding protein observed by pa ramagnetic NMR. Nature 449 1078 U12. 102. Neudecker, P., Lundstrom, P. & Kay, L. E. (2009). Relaxation Dispersion NMR Spectroscopy as a Tool for Detailed Studies of Protein Folding. Biophysical Journal 96 2045 2054. 103. McConnell, H. M. (1958). REACTI ON RATES BY NUCLEAR MAGNETIC RESONANCE. Journal of Chemical Physics 28 430 431. 104. Bax, A. & Tjandra, N. (1997). High resolution heteronuclear NMR of human ubiquitin in an aqueous liquid crystalline medium. Journal of Biomolecular Nmr 10 289 292. 105 Tolman, J. R., Al Hashimi, H. M., Kay, L. E. & Prestegard, J. H. (2001). Structural and dynamic analysis of residual dipolar coupling data for proteins. Journal of the American Chemical Society 123 1416 1424. 106. Lipari, G. & Szabo, A. (1982). Model F ree Approach to the Interpretation of Nuclear Magnetic Resonance Relaxation in Macromolecules .1. Theory and Range of Validity. Journal of the American Chemical Society 104 4546 4559. 107. Hvidt, A. & Linderstromlang, K. (1954). Exchange of Hydrogen Atom s in Insulin with Deuterium Atoms in Aqueous Solutions. Biochimica Et Biophysica Acta 14 574 575. 108. Diez, M., Zimmermann, B., Borsch, M., Konig, M., Schweinberger, E., Steigmiller, S., Reuter, R., Felekyan, S., Kudryavtsev, V., Seidel, C. A. M. & Grab er, P. (2004). Proton powered subunit rotation in single membrane bound F0F1 ATP synthase. Nature Structural & Molecular Biology 11 135 141. 109. Poole, C. P. (1967). Electron Spin Resonance, a comprehensice Treatise on Experimental Techniques John Wile y and Sons, New York. 110. Berliner, L. J., Ed. (1976). Spin Labeling, Theory and Applications. Molecular Biology, an International Series of Monographs and Textbooks. Edited by Hornecker, B., Kaplan, N. O., Marmur, J., Scheraga, H. A. New York: Academic Press. 111. Slichter, C. P. (1963). Principles of Magnetic Resonance Springer Series in Solid State Sciences (Cardona, M., Fulde, P., von Klitzing, K., Queisser, H. Ed.), 1, Springer Verlag, New York.

PAGE 300

300 112. Ehrenstein, G., Lecar, H., Ed. (1982). Biophy sics. Vol. 20. Methods of Experimental Biophysics. Edited by Marton, C. New York: Academic Press. 113. Carrington, A., McLaughlin, A. D. (1967). Introduction to Magnetic Resonance Harper's Chemistry Series (Rice, S. A., Ed.), Harper & Row, New York. 114 Weil, J. A., Bolton, J. R., Wertz, J. E. (1994). Electron Paramagnetic Resonance, Elementary Theory and Practical Applications John Wiley and Sons, New York. 115. Mcgarvey, B. R. (1967). Isotropic Hyperfine Interaction. Journal of Physical Chemistry 71 51 &. 116. Stone, T. J., Buckman, T., Nordio, P. L. & Mcconnel.Hm. (1965). Spin Labeled Biomolecules. Proceedings of the National Academy of Sciences of the United States of America 54 1010 &. 117. Griffith, O. H. & McConnel.Hm. (1966). A NITROXIDE MA LEIMIDE SPIN LABEL. Proceedings of the National Academy of Sciences of the United States of America 55 8 &. 118. Stryer, L. & Griffith, O. H. (1965). A SPIN LABELED HAPTEN. Proceedings of the National Academy of Sciences of the United States of America 5 4 1785 &. 119. Hoffmann, A. K. & Henderson, A. T. (1961). NEW STABLE FREE RADICAL DI T BUTYLNITROXIDE. Journal of the American Chemical Society 83 4671 &. 120. Froncisz, W. & Hyde, J. S. (1982). The Loop Gap Resonator a New Microwave Lumped Circuit Electron Spin Resonance Sample Structure. Journal of Magnetic Resonance 47 515 521. 121. Altenbach, C., Flitsch, S. L., Khorana, H. G. & Hubbell, W. L. (1989). Structural Studies on Transmembrane Proteins .2. Spin Lableing of Bacteriorhodopsin Mutants a t Unique Cysteines. Biochemistry 28 7806 7812. 122. Hanson, P., Millhauser, G., Formaggio, F., Crisma, M. & Toniolo, C. (1996). ESR characterization of hexameric, helical peptides using double TOAC spin labeling. Journal of the American Chemical Society 118 7618 7625. 123. Shao, W., Everitt, L., Manchester, M., Loeb, D. D., Hutchison, C. A. & Swanstrom, R. (1997). Sequence requirements of the HIV 1 protease flap region determined by saturation mutagenesis and kinetic analysis of flap mutants. Proceeding s of the National Academy of Sciences of the United States of America 94 2243 2248. 124. Columbus, L. & Hubbell, W. L. (2002). A new spin on protein dynamics. Trends in Biochemical Sciences 27 288 295.

PAGE 301

301 125. McHaourab, H. S., Lietzow, M. A., Hideg, K. & Hubbell, W. L. (1996). Motion of Spin Labeled Side Chains in T4 Lysozyme. Correlation with Protein Structure and Biochemistry 35 7692 7704. 126. Belle, V., Rouger, S., Costanzo, S., Liquiere, E., Strancar, J., Guigliarelli, B., Fournel, A. & Longhi, S. (2008). Mapping alpha helical induced folding within the intrinsically disordered C terminal domain of the measles virus nucleoprotein by site directed spin labeling EPR spectroscopy. Proteins Structure Function and Bioinformatics 73 973 988. 127. Stryer, L. & Haugland, R. P. (1967). Energy Transfer a Spectroscopic Ruler. Proceedings of the National Academy of Sciences of the United States of America 58 719 &. 128. Keeler, J. (2007). Understanding NMR Spectroscopy John Wiley & Sons Inc, West Sussex, England. 129. Banham, J. E., Baker, C. M., Ceola, S., Day, I. J., Grant, G. H., Groenen, E. J. J., Rodgers, C. T., Jeschke, G. & Timmel, C. R. (2008). Distance measurements in the borderline region of applicability of CW EPR and DEER: A model study on a homologous series of spin labelled peptides. Journal of Magnetic Resonance 191 202 218. 130. Jeschke, G. (2002). Distance measurements in the nanometer range by pulse EPR. Chemphyschem 3 927 932. 131. Jeschke, G. & Polyhach, Y. (2007). Dist ance measurements on spin labelled biomacromolecules by pulsed electron paramagnetic resonance. Physical Chemistry Chemical Physics 9 1895 1910. 132. Schwieger, A., Jeschke, G. (2001). Principles of Pulsed Electron Paramagnetic Resonance Oxford Universi ty Press, New York. 133. Klauder, J. R. & Anderson, P. W. (1962). Spectral Diffusion Decay in Spin Resonance Experiments. Physical Review 125 912. 134. Berliner, L. J., Eaton, S. S., Eaton, G. E., Ed. (2000). Distance Measurements in Biological Systems by EPR. Vol. 19. Biological Magnetic Resonance. New York: Kluwer Academic/Plenum Publishers. 135. Borbat, P. P., Crepeau, R. H. & Freed, J. H. (1997). Multifrequency two dimensional Fourier transform ESR: An X/Ku band spectrometer. Journal of Magnetic Res onance 127 155 167.

PAGE 302

302 136. Milov, A. D., Salikohov, K. M. & Shirov, M. D. (1981). Application of Endor in Electron Spin Echo for Paramagnetic Center Space Distribution in Solids. Fizika Tverdogo Tela 23 975 982. 137. Pannier, M., Veit, S., Godt, A., Jesc hke, G. & Spiess, H. W. (2000). Dead time free measurement of dipole dipole interactions between electron spins. Journal of Magnetic Resonance 142 331 340. 138. Larsen, R. G. & Singel, D. J. (1993). Double Electron Electron Resonance Spin Echo Modulation Spectroscopic Measurement of Electron Spin Pair Separations in Orientationally Disordered Solids. Journal of Chemical Physics 98 5134 5146. 139. Mcbriert.Vj. (1974). N M R of Solid Polymers Review. Polymer 15 503 520. 140. Jeschke, G., Bender, A., Paulsen, H., Zimmermann, H. & Godt, A. (2004). Sensitivity enhancement in pulse EPR distance measurements. Journal of Magnetic Resonance 169 1 12. 141. Jeschke, G., Koch, A., Jonas, U. & Godt, A. (2002). Direct conversion of EPR dipolar time evolution d ata to distance distributions. Journal of Magnetic Resonance 155 72 82. 142. Jeschke, G., Chechik, V., Ionita, P., Godt, A., Zimmermann, H., Banham, J., Timmel, C. R., Hilger, D. & Jung, H. (2006). DeerAnalysis2006 a comprehensive software package for analyzing pulsed ELDOR data. Applied Magnetic Resonance 30 473 498. 143. Chiang, Y. W., Borbat, P. P. & Freed, J. H. (2005). The determination of pair distance distributions by pulsed ESR using Tikhonov regularization. Journal of Magnetic Resonance 172 279 295. 144. Hansen, P. C. (1992). Analysis of Discrete Ill Posed Problems by Means of the L Curve. Siam Review 34 561 580. 145. Galiano, L., Ding, F., Veloro, A. M., Blackburn, M. E., Simmerling, C. & Fanucci, G. E. (2009). Drug Pressure Selected Muta tions in HIV 1 Protease Alter Flap Conformations. Journal of the American Chemical Society 131 430 +. 146. Ohtaka, H., Schon, A. & Freire, E. (2003). Multidrug resistance to HIV 1 protease inhibition requires cooperative coupling between distal mutations Biochemistry 42 13659 13666. 147. Todd, M. J., Luque, I., Velazquez Campoy, A. & Freire, E. (2000). Thermodynamic basis of resistance to HIV 1 protease inhibition: Calorimetric analysis of the V82F/I84V active site resistant mutant. Biochemistry 39 11 876 11883.

PAGE 303

303 148. Sayer, J. M., Liu, F. L., Ishima, R., Weber, I. T. & Louis, J. M. (2008). Effect of the active site D25N mutation on the structure, stability, and ligand binding of the mature HIV 1 protease. Journal of Biological Chemistry 283 13459 1347 0. 149. Torbeev, V. Y., Raghuraman, H., Mandal, K., Senapati, S., Perozo, E. & Kent, S. B. H. (2009). Dynamics of "Flap" Structures in Three HIV 1 Protease/Inhibitor Complexes Probed by Total Chemical Synthesis and Pulse EPR Spectroscopy. Journal of the A merican Chemical Society 131 884 +. 150. Clemente, J. C., Hemrajani, R., Blum, L. E., Goodenow, M. A. & Dunn, B. A. (2003). Secondary mutations M36I and A71V in the human immunodeficiency virus type 1 protease can provide an advantage for the emergence o f the primary mutation D30N. Biochemistry 42 15029 15035. 151. Lafont, V., Armstrong, A. A., Ohtaka, H., Kiso, Y., Amzel, L. M. & Freire, E. (2007). Compensating enthalpic and entropic changes hinder binding affinity optimization. Chemical Biology & Drug Design 69 413 422. 152. Coman, R. M., Robbins, A. H., Goodenow, M. M., Dunn, B. M. & McKenna, R. (2008). High resolution structure of unbound human immunodeficiency virus 1 subtype C protease: implications of flap dynamics and drug resistance. Acta Crys tallographica Section D Biological Crystallography 64 754 763. 153. Heaslet, H., Rosenfeld, R., Giffin, M., Lin, Y. C., Tam, K., Torbett, B. E., Elder, J. H., McRee, D. E. & Stout, C. D. (2007). Conformational flexibility in the flap domains of ligand fr ee HIV protease. Acta Crystallographica Section D Biological Crystallography 63 866 875. 154. Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt, D. M., Meng, E. C. & Ferrin, T. E. (2004). UCSF chimera A visualization system for e xploratory research and analysis. Journal of Computational Chemistry 25 1605 1612. 155. Coman, R. M., Robbins, A. H., Fernandez, M. A., Gilliland, C. T., Sochet, A. A., Goodenow, M. M., McKenna, R. & Dunn, B. M. (2008). The contribution of naturally occu rring polymorphisms in altering the biochemical and structural characteristics of HIV 1 subtype C protease. Biochemistry 47 731 743. 156. Velazquez Campoy, A., Vega, S., Fleming, E., Bacha, U., Sayed, Y., Dirr, H.W., Freire, E. (2003). Protease Inhibitio n in African Subtypes. Aids Reviews 5 165 171. 157. Velazquez Campoy, A., Todd, M. J., Vega, S. & Freire, E. (2001). Catalytic efficiency and vitality of HIV 1 proteases from African viral subtypes. Proceedings of the National Academy of Sciences of the United States of America 98 6062 6067.

PAGE 304

304 158. Molla, A., Korneyeva, M., Gao, Q., Vasavanonda, S., Schipper, P. J., Mo, H. M., Markowitz, M., Chernyavskiy, T., Niu, P., Lyons, N., Hsu, A., Granneman, G. R., Ho, D. D., Boucher, C. A. B., Leonard, J. M., Nor beck, D. W. & Kempf, D. J. (1996). Ordered accumulation of mutations in HIV protease confers resistance to ritonavir. Nature Medicine 2 760 766. 159. Kozer, N., Kuttner, Y. Y., Haran, G. & Schreiber, G. (2007). Protein protein association in polymer solu tions: From dilute to semidilute to concentrated. Biophysical Journal 92 2139 2149. 160. Timasheff, S. N. (2002). Protein hydration, thermodynamic binding, and preferential hydration. Biochemistry 41 13473 13482. 161. Kuttner, Y. Y., Kozer, N., Segal, E., Schreiber, G. & Haran, G. (2005). Separating the contribution of translational and rotational diffusion to protein association. Journal of the American Chemical Society 127 15138 15144. 162. Galiano, L., Blackburn, M. E., Veloro, A. M., Bonora, M. & Fanucci, G. E. (2009). Solute Effects on Spin Labels at an Aqueous Exposed Site in the Flap Region of HIV 1 Protease. Journal of Physical Chemistry B 113 1673 1680. 163. Davis, D. J., Burlak, C. & Money, N. P. (2000). Osmotic pressure of fungal compatibl e osmolytes. Mycological Research 104 800 804. 164. Crowley, P. B., Brett, K. & Muldoon, J. (2008). NMR spectroscopy reveals cytochrome c poly(ethylene glycol) interactions. Chembiochem 9 685 688. 165. Szleifer, I. (1997). Polymers and proteins: Intera ctions at interfaces. Current Opinion in Solid State & Materials Science 2 337 344. 166. Sheth, S. R. & Leckband, D. (1997). Measurements of attractive forces between proteins and end grafted poly(ethylene glycol) chains. Proceedings of the National Acad emy of Sciences of the United States of America 94 8399 8404. 167. Tubio, G., Nerli, B. & Pico, G. (2004). Relationship between the protein surface hydrophobicity and its partitioning behaviour in aqueous two phase systems of polyethyleneglycol dextran. Journal of Chromatography B Analytical Technologies in the Biomedical and Life Sciences 799 293 301. 168. Fanucci, G. E., Lee, J. Y. & Cafiso, D. S. (2003). Spectroscopic evidence that osmolytes used in crystallization buffers inhibit a conformation chan ge in a membrane protein. Biochemistry 42 13106 13112.

PAGE 305

305 169. Sanches, M., Krauchenco, S., Martins, N. H., Gustchina, A., Wlodawer, A. & Polikarpov, I. (2007). Structural characterization of B and non B subtypes of HIV protease: Insights into the natural susceptibility to drug resistance development. Journal of Molecular Biology 369 1029 1040. 170. Gsponer, J., Christodoulou, J., Cavalli, A., Bui, J. M., Richter, B., Dobson, C. M. & Vendruscolo, M. (2008). A coupled equilibrium shift mechanism in calmod ulin mediated signal transduction. Structure 16 736 746.

PAGE 306

BIOGRAPHICAL SKETCH Mandy Elizabeth Blackburn was born in 1978 in Denver, Colorado. She obtained her Bachelor of Science degree in c hemistry at the University of New Mexico in Albuquerque in De cember 2000. After graduation, she served for a year in the United States Navy Reserve in Pensacola, Florida. Following a four year hiatus in industry of which she spent a year in the Quality Control division of Eppendorf 5 Prime in Boulder, Colorado, s he moved to Gainesville, FL. She in the Biochemistry Department of the m edical school at the University of Florida for a semester before applying to graduate school In January of 2005, she began graduate school in the Department of C hemistry at the University of Florida. In April of 2005, grou p.