Insights into the Mechanisms of HIV-1 Protease Drug Resistance from Pulsed Electron Paramagnetic Resonance and Nuclear M...

MISSING IMAGE

Material Information

Title:
Insights into the Mechanisms of HIV-1 Protease Drug Resistance from Pulsed Electron Paramagnetic Resonance and Nuclear Magnetic Resonance Spectroscopy
Physical Description:
1 online resource (240 p.)
Language:
english
Creator:
De Vera, Ian Mitchelle Sayo
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Chemistry
Committee Chair:
Fanucci, Gail E
Committee Members:
Horenstein, Nicole A
Dunn, Ben M
Angerhofer, Alexander
Maupin, Julie A

Subjects

Subjects / Keywords:
aids -- epr -- hiv -- nmr -- protease
Chemistry -- Dissertations, Academic -- UF
Genre:
Chemistry thesis, Ph.D.
Electronic Thesis or Dissertation
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )

Notes

Abstract:
The human immunodeficiency virus type 1 protease (HIV-1 PR) is an attractive target in AIDS therapies that utilize protease inhibitors (PIs), because this enzyme is required for viral maturation.  However, the rapid emergence of drug-pressure selected mutations in the HIV genome after exposure to PIs often results in drug resistance.  The understanding of the molecular mechanism by which the accumulation of mutations imparts resistance to PIs is important for future drug design.  We used double electron-electron resonance (DEER)spectroscopy to illuminate how amino acid substitutions in the apo protease and in the presence of inhibitors result in perturbations of the equilibrium fractional occupancies of HIV-1 PR conformational populations.  We found that drug resistance emerges from combinations of mutations that stabilize open-like conformations while destabilizing the closed state and retaining the semi-open population seen in native protease. Changes in protease flap conformations were also monitored in the presence of inhibitors in the multi-drug resistant HIV-1 PR variant MDR769.  Specifically, variations in inhibitor IC50 values compared to the native enzyme are related to the relative change in the inhibitor-induced shift to the closed state, |DC|. A linear correlation was found between |DC| and the fold-change in IC50 when inhibitor-binding is not too weak.    In 1H-15N heteronuclear single quantum coherence (HSQC) titration experiments for MDR769, asymmetric inhibitors that have longer residence time in the protease binding pocket showed peak splitting of several HSQC resonances because of slow exchange in the NMR timescale.  Peak splitting coincided with strong induction of flap closure based on DEER results. On the other hand, several resonances were missing in the HSQC spectra when an inhibitor undergoes intermediate exchange in the NMR timescale with the binding cleft.  Peak disappearances due to broadened line widths of several HSQC peaks were consistent with weak inhibitor-induced flap closure. Finally, we compared the residue-specific backbone flexibility of MDR769 with subtype B protease by measuring 15N spin relaxation parameters, namely T1, T2 and NOE.  Results suggest that the flap elbow and T80 loop of MDR769 are more rigid relative to subtype B.
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.
Thesis:
Thesis (Ph.D.)--University of Florida, 2012.
Local:
Adviser: Fanucci, Gail E.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-12-31
Statement of Responsibility:
by Ian Mitchelle Sayo de Vera.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0044917:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 INSIGHTS INTO THE MECHANISMS OF HIV 1 PROTEASE DRUG RESISTANCE FROM PULSED ELECTRON PARAMAGNETIC RESONANCE AND NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY By IAN MITCHELLE SAYO DE VERA 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 2012

PAGE 2

2 2012 Ian Mitchelle Sayo de Vera

PAGE 3

3 To my mother Cynthia de Vera and to my sister Mar ia Zahs Nu ez With a special dedication to my grandmother Godofreda Sayo

PAGE 4

4 ACKNOWLEDGMENTS First of all, I would like to thank God for the abundance of blessing s for more. Everything I do is for Your greater glory. I also would like t o thank my family for being my source of inspiration and strength to survive 4 years in grad uate school. Next, I would like to thank individuals who have been instrumental t o my success in grad uate school: My indefatigable mentor and committee chair Dr. Gail Fanucc i. My committee members, Dr. Ben Dunn, Dr. Nicole Horenstein, Dr. Alexander Angerhofer and Dr. Julie Maupin Furlow. d best friend. I would also like to thank her for assisting me with mass spectrometry experiments. Dr. Joanna Long for her wonderful insights regarding my NMR experiments. Dr. Alexander Ang erhofer for his help with setting up EPR experiments. Dr. Mandy Blackburn for training me on DEER spectroscopy. Dr. Luis Galiano for starting the HIV 1 protease project. Our dear collaborators, Dr. Ben Dunn and Dr. Julie Maupin Furlow. All of Fanucci grou p members, especially Jackie Esquiaqui, Francis Agama, Otonye Braide Xi Huang and Andrea Medina, for their wonderful friendship. Thanks to my closest friends, Nathaniel Hepowit, Jean Palmes, Wendy Anderson and Melissa Smith for keeping me sane. Spec ial thanks to my spiritual adviser, Pastor Stuart M cCutcheon

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 15 ABSTRACT ................................ ................................ ................................ ................... 22 CHAPTER 1 INTRODUCTION AND RESEARCH OVERVIEW ................................ ................... 24 Human Immunodeficiency Virus and the AIDS Pandemic ................................ ...... 24 HIV 1 vs. HIV 2 ................................ ................................ ................................ 24 Genetic Variability of HIV 1 ................................ ................................ .............. 24 The HIV 1 Genome ................................ ................................ .......................... 27 Structure of the HIV 1 Virion ................................ ................................ ............. 28 The HIV 1 Life Cycle ................................ ................................ ........................ 29 HIV 1 Therapeutic Regimens ................................ ................................ ........... 31 HIV 1 Prote ase ................................ ................................ ................................ ....... 31 Structure of HIV 1 Protease ................................ ................................ ............. 32 HIV 1 Protease Inhibitors ................................ ................................ ................. 38 HIV 1 Protease Subtype Polymorphisms ................................ ......................... 40 Drug pressure Selected Mutations in HIV 1 Protease ................................ ...... 41 Drug Resistance Mechanisms ................................ ................................ .......... 45 Multi drug Resistant Patient Isolate MDR769 ................................ ................... 46 Identical Transition States in Native and Drug resistant HIV 1 Protease ......... 47 Research Overview ................................ ................................ ................................ 47 Challenges to the Analysis of Conformational Sampling ................................ .. 47 Spe cific Aims ................................ ................................ ................................ .... 49 Scopes and Limitations ................................ ................................ .................... 49 2 PULSED EPR DISTANCE MEASUREMENTS IN SOLU BLE PROTEINS BY SITE DIRECTED SPIN LABELING ................................ ................................ ........ 52 Introduction ................................ ................................ ................................ ............. 52 Site directed Spin labeling ................................ ................................ ...................... 54 Background Informati on ................................ ................................ ................... 54 Spin labeling of muscle fibers ................................ ................................ .... 54 Synthetic and orthogonal labeling strategies for incorporating unnatural amino acid spin labels ................................ ............................. 55

PAGE 6

6 Nitroxide Spin Labels ................................ ................................ ....................... 55 Selection of Labeling Sites ................................ ................................ ............... 56 S ite directed Spin labeling of Soluble Proteins ................................ ................. 58 Distance Measurements via DEER ................................ ................................ ......... 59 Preparing DEER Samples ................................ ................................ ................ 63 DEER Experimental Considerations ................................ ................................ 67 DEER Data Acquisition ................................ ................................ .................... 69 Analysis of DEER Data ................................ ................................ .................... 72 Critical DEER Parameters and Troubleshooting ................................ .............. 78 Anticipated Results ................................ ................................ ........................... 87 Time Considerations ................................ ................................ ........................ 89 3 UNRAVELING THE RELAT IONSHIP BETWEEN CONF ORMATIONAL SAMPLING AND DRUG RE SISTANCE IN HIV 1 PROTEASE .............................. 90 I ntroduction ................................ ................................ ................................ ............. 90 Materials and Methods ................................ ................................ ............................ 95 Nomenclature for HIV 1 Protease Variants ................................ ...................... 95 Cloning and Site directed Mutagenesis ................................ ............................ 96 Protein Expression, Purification and Spin labeling ................................ ........... 97 Confirmation of Ho mogeneous Spin labeling by ESI TOF MS ......................... 99 Secondary Structure Characterization by Circular Dichrorism (CD) Spectroscopy ................................ ................................ ................................ 99 Sample P reparation and DEER Data Acquisition ................................ ........... 100 DEER Data Processing ................................ ................................ .................. 100 Pearson Correlation ................................ ................................ ....................... 101 2 Error Analysis ................................ ................................ ............................. 101 Results and Discussion ................................ ................................ ......................... 102 Expression, Purification, Spi n labeling and Characterizati on of HIV 1 Protease ................................ ................................ ........................ 102 Conformational Sampling of Point Mutation Variants via DEER Spectroscopy ................................ ................................ .............................. 103 Gaussian Reconstruction Pro files and Hypothetical Folding Funnels ............ 106 Pearson Correlation of DEER Results with Enzymatic and Inhibition Parameters ................................ ................................ .................. 113 Inhib itor bound DEER Data for HIV 1 Protease with Point Mutations ............ 118 Conclusions ................................ ................................ ................................ .......... 123 4 CORRELATING CONFORMA TIONAL SHIFT INDUCTI ON WITH ALTERED INHBITOR POTENCY IN A MULTI DRUG RESISTANT HIV 1 PROTEASE VARIANT ................................ ................................ ................................ .............. 126 Introduction ................................ ................................ ................................ ........... 126 The MDR769 Patient Isolate ................................ ................................ .......... 126 Definition of the | C | Parameter ................................ ................................ ..... 127 Materials and Methods ................................ ................................ .......................... 128 Details on the MDR769 Variant ................................ ................................ ...... 128

PAGE 7

7 Cloning, Protein Expression, Spin labeling and Purification ........................... 129 Substrates and Inhibitors ................................ ................................ ................ 129 DEER Experiments and Data Processing ................................ ...................... 129 Results and Discussion ................................ ................................ ......................... 129 Ligand i nduced Flap Closure in MDR769 ................................ ....................... 129 Fractional Occupancies of Putative Conformations from Gaussian Reconstruction ................................ ................................ ............................ 131 Correlation of the | C | Parameter with IC 50 ................................ .................... 135 Conclusions ................................ ................................ ................................ .......... 137 5 ANALYSIS OF HIV 1 PROTEASE BACKBONE FLEXIBILITY AND PROT EIN LIGAND INTERACTION DYNAMICS BY NMR SPEC TROSCOPY ...................... 138 Introduction ................................ ................................ ................................ ........... 138 Protein Backbone Chemical Shift Assignment via Triple Resonance Experiments ................................ ................................ ................................ 138 15 N Spin Relaxation and Nuclear Overhauser Effect ................................ ...... 140 Previous NMR Relaxation Studies on HIV 1 Protease ................................ ... 141 Protein Ligand Exchange Dynamics ................................ .............................. 142 Materials and Methods ................................ ................................ .......................... 144 HIV 1 Protease Expression in M9 Media and NMR Sample Preparation ....... 144 NMR Data Acquisition and Processing ................................ ........................... 144 1 H 15 N HSQC Inhibitor Titration Experiments ................................ ................. 145 Backbone Chemical Shift Assignments ................................ .......................... 145 Relaxation Experiments ................................ ................................ .................. 145 Results and Discussion ................................ ................................ ......................... 146 1 H 15 N HSQC of Subtype B Single point Mutation Variants ............................ 146 Backbone Chemical Shift Assignment of MDR 769 ................................ ......... 148 1 H 15 N HSQC Ligand Titration Experiments for MDR769 ............................... 149 Comparison of NMR and Pulsed EPR Results ................................ ............... 153 Relaxation Measurements for MDR769 ................................ ......................... 157 Conclusions ................................ ................................ ................................ .......... 159 6 CHARACTERIZATION OF HIV 1 PROTE ASE BY MASS SPECTROM ETRY ..... 160 Introduction ................................ ................................ ................................ ........... 160 Materials and Methods ................................ ................................ .......................... 163 Determination of Intact Protein Average Molecular Weight by ESI TOF MS .. 163 In solution Trypsin Digestion and MALDI TOF MS Analysis .......................... 1 63 HPLC ESI Ion trap MS Analysis of Tryptic Digests ................................ ........ 164 Results and Discussion ................................ ................................ ......................... 164 Determination of MTSL Spin labeling Effi ciency by ESI TOF MS .................. 164 Average Molecular Weight by ESI TOF MS and Sequence Coverage by MALDI TOF MS ................................ ................................ ..................... 166 Shelf Life Studies of M TSL labeled HIV 1 Protease ................................ ....... 167 Tracking Amino Acid Substitutions by MALDI TOF MS ................................ 169 De Novo Sequencing of Modified Peptides by MA LDI TOF MS 2 ................... 171

PAGE 8

8 Confirmation of Site specific Spin labeling in HIV 1 Protease via ESI Ion trap MS n ................................ ................................ ......................... 172 Conclusions ................................ ................................ ................................ .......... 173 7 FUTURE DIRECTIONS ................................ ................................ ........................ 175 Model free Analysis of MDR769 ................................ ................................ ........... 175 Relaxation Measurem ents for Single point Mutant Variants ................................ 175 DEER Analysis of Active HIV 1 Protease ................................ ............................. 175 Testing other Drug resistant HIV 1 Protease V ariants and Inhibitors in | C | Correlations with IC 50 ................................ ................................ ............. 176 APPENDIX A SUPPLEMENTAL INFORMA TION FOR DEER DATA A NALYSIS ....................... 177 B 2 ER ROR ANALYSIS FOR POP ULATION VALIDATION ................................ ... 203 C 1 H 15 N HSQC TITRATION FIG URES FOR MDR769 ................................ ............ 207 LIST OF REFERENCES ................................ ................................ ............................. 217 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 240

PAGE 9

9 LIST OF TABLES Table page 1 1 Mutations associated wit h protease inhibitor trea tments ................................ .... 43 3 1 Changes in flap distance distribution profile in subtype B variants ................... 108 3 2 Summary of DEER distance parameters for subtype B HIV 1 PR variants ...... 109 3 3 Average distance and range for HIV 1 PR conformations determined for seven subtype B variants with D30N, M36I, and A71V mutations ............... 109 3 4 Summary of Pearson product moment coefficients ( r ) for DEER relative percentages correlated to inhibition constants ( K i ) ................................ ........... 116 3 5 Ligand induced chan ges in flap distance distribution profile in subtype B variants determined by SDSL DEER ................................ ................................ 122 3 6 Summary of DEER distance parameters for D30N, M36I, A71V and D30N/M36I/A71V with inhibitor or sub strate mimic ................................ .......... 123 3 7 Average distance and range for HIV 1 PR conformations determined for D30N, M36I, A71V and D30N/M36I/A71V with inhibitor or substrate mimic ..... 123 4 1 Inhibitor induced changes in MDR769 flap distance distribution profile ............ 132 4 2 Summary of DEER distance parameters for free and inhibitor bound MDR769 samp les ................................ ................................ ............................. 133 4 3 Average distance and range for HIV 1 PR conformations determined for MDR769 apo and in the presence of inhibitors. ................................ ........... 133 5 1 Comparison of ligand exchange dynamics in MDR769 and subtype B PR. ..... 154 5 2 Summary of HSQC peak count and DEER percentage of closed population for HIV 1 PR variants with inhibitors ................................ ................................ 155 6 1 Molecular weights of intact protease via ESI TOF MS and percent amino acid coverage of trypsin digests via MALDI TOF MS ............................ 166 B 1 Values of 2 for Gaussian regenerated echo curves for subtype B variants with suppressed populations ................................ ................................ ............ 203 B 2 2 e rror analysis for subtype B variants w ith or without ligands ........................ 204 B 3 Values of 2 for Gaussian regenerated echo curves for MDR 769 with or without ligands ................................ ................................ ...................... 205

PAGE 10

10 LIST OF FIGURES Figure page 1 1 Phylogenetic tree of HIV 1 ................................ ................................ .................. 25 1 2 Geographical distribution of HIV 1 protease variants ................................ ......... 26 1 3 Schematic d iagram of the H IV 1 genome ................................ ........................... 27 1 4 Structural assembly of viral proteins in an HIV 1 virion ................................ ...... 29 1 5 The HIV viral life cycle ................................ ................................ ........................ 30 1 6 HIV 1 PR topology ................................ ................................ .............................. 33 1 7 T 1 PR active site. ................................ ................. 35 1 8 Structure of HIV 1 PR showing the restricted access to the binding pocket in the semi open conformation ................................ ................................ ............... 36 1 9 Predominant flap conformations of HIV 1 protease ................................ ............ 37 1 10 Apoenzyme X ray crystal structure of MDR769 overlaid with wild type (LAI) ..... 38 1 11 Protease inhibitor scaffolds and structures ................................ ......................... 39 1 12 Sites of amino acid substitutions in HIV 1 protease ................................ ............ 42 1 13 Subtype B HIV 1 PR mutation prevalence in PI na ve and treated patient isola tes ................................ ................................ ................................ ............... 44 1 14 Mutation sites i n patient isolate MDR769 relative to the LAI consensus sequence ................................ ................................ ................................ ............ 46 2 1 Chemical structures of four com mon nitroxide spin labels before and after reacting to a cysteine side chain ................................ ................................ ......... 56 2 2 Four pulse DEER and the corresponding p ump and observe frequencies for nitroxide labels at X band. ................................ ................................ .................. 59 2 3 Sample dipolar evolution curve before and after applying the background subtraction function ................................ ................................ ............................ 63 2 4 Plots of the maximum spin concent ration as a function of the inter spin distance ................................ ................................ ................................ .............. 64 2 5 P reliminary experiments prior DEER experiment set up ................................ .... 71

PAGE 11

11 2 6 The four pulse DEER sequence with the pulse spacings labeled according to Bruker Xepr software package nomenclature ................................ ..................... 72 2 7 Schematic diagram of three methods utilized for obtaining distance information from t he background subt racted dipolar evolution curve ................. 74 2 8 Example of an L curve and the corresponding distance profiles and dipolar modulation curves for low optimal and high regularization para meters ( ) ........ 75 2 9 Self consistent procedure for determinin g the appropriate level of background subtraction ................................ ................................ ...................... 77 2 10 Effect of the breadth and most probable distance on the dipolar modulation curves ................................ ................................ ................................ ................. 79 2 11 The influence of max length to the corresponding distance profile ...................... 81 2 12 DEER data for subtype B HIV 1 protease acquired at variable max ................... 81 2 13 Intensity normalized T m curves and corresponding exponential decay fits for MTSL labeled HIV 1 PR ................................ ................................ ................ 85 2 14 DEER da ta processing for HIV 1 PR ................................ ................................ .. 88 3 1 Ribbon diagram of HIV 1 PR showing the m utation sites D30, M36 and A71 .... 94 3 2 Protein sequence for subtype B HIV 1 PR ................................ ......................... 96 3 3 SDS PAGE of HIV 1 PR ................................ ................................ ................... 102 3 4 Circular dichroism spectra for MTSL labeled HIV 1 PR variants ...................... 103 3 5 Background subtracted dipolar modulation curves in the time domain for subtype B variants and corresponding TKR dista nce profiles .......................... 104 3 6 Gaussian reconstruction profile for M36I/A71V and Gaus sian reconstruction analysis of TKR distance profiles for subtype B variants ................................ .. 107 3 7 Hypothetical protein folding funnels for WT subtype B and the D30N/M36I/A71V variants ................................ ................................ ................ 111 3 8 Relative percentage change in the population of each conformation fo r the variants studied ................................ ................................ ................................ 112 3 9 Correlation plots for trends of closed an d semi open conformation percen tages with kinetic parameters ................................ ................................ 115

PAGE 12

12 3 10 C orrelation plots for K i values of the inhibitors nelfinavir, ritonavir, and indinavir against the ratio of the percentage closed to the percentage of open like conformations ................................ ................................ ............... 117 3 11 Background subtracted dipolar evolution curves and TKR fits for MTSL labeled subtype B variants D30N, M36I, A71V and D30N/M36I/A71V bound to CA p2 or RTV, overlaid with apo ................................ ....................... 120 3 12 Area n ormalized TKR distance distribution profiles for subtype B variants D30N, M36I, A71V and D30N/M36I/A71V for apo, CA p2 bound a nd RTV bound protease ................................ ................................ ........................ 121 4 1 Ribbon diagram of MDR769 showing pri mary and compensatory mutations relative to wild type HIV 1 PR ................................ ................................ ........... 127 4 2 Amino acid sequence alignment of subtype B and MDR769 HIV 1 PR ............ 128 4 3 Background subtracted DEER dipolar evolution curves with fits from Tikhonov regularization analysis for MDR769 ................................ .................. 130 4 4 Comparison of s ubtype B and MDR769 distance distribution profi les in the free and ligand bound state ................................ ................................ .............. 134 4 5 D egree of flap closure measured as percentage occupancy of the closed state (% closed) in subtype B PR and MDR769 ................................ ............... 135 4 6 Logarithmic plot of half maximal inhibitory concentration (IC 50 ) a gainst DEER % closed for MDR769 and p lot of the log IC 50 fold change versus the magnitude of % difference in DEER closed population (| C |) between MD R769 and subtype B. ................................ ................................ .................. 136 5 1 Triple resonance experiments ................................ ................................ .......... 139 5 2 The effect of ligand exchange rate (k ex ) on resonance intensity a nd chemical shift ................................ ................................ ................................ .... 143 5 3 1 H 15 N HSQC spectra for D30N, M36I, and A71V overlaid with wild type subtype B and the corresponding chemical shift perturbation plots .................. 147 5 4 1 H 15 N HSQC spectrum of HIV 1 PR MDR769 at 600 MHz and 293 K with peak assignments from triple resonance experiments ................................ ...... 150 5 5 Overlay of 1 H 15 N HSQC spect ra of subtype B and MDR769 ........................... 151 5 6 Representative 1 H 15 N HSQC spectra for 1:1 inhibitor to protease samples illuminating two types of exchange dynamics in HIV 1 PR ............................... 154 5 7 Plots of HSQC peak count against DEER% closed ................................ ........ 156

PAGE 13

13 5 8 15 N spin relaxation parameters R 1 R 2 and NOE for MDR769 .......................... 158 6 1 ESI TOF MS reveals optimized conditions to achieve homogeneous spin labeling of HIV 1 protease monitored with ESI TOF MS ........................... 165 6 2 Sequence coverage of subtype B HIV 1 protease variants via MALDI (+) and MALDI ( ) TOF/TOF MS ................................ ................................ ........... 167 6 3 Deconvoluted ESI TOF MS spectrum o f the triple mutant, after 8 months of storage at 20 o C ................................ ................................ ........................... 168 6 4 MALDI TOF MS of active subtype B (D25) variant after storage at 20 C for a week (control) and after heating at 30 C for up to 3 h .............................. 169 6 5 The MALDI TOF MS spectra of WT subtype B, M36I and D30N/M36I tryptic digests ................................ ................................ ................................ .... 170 6 6 MALDI TOF MS 2 spectra of m/z 2213.0 [M+H] + ion from M36I and m/z 2211.9 from D30N/M36I ................................ ................................ ............. 171 6 7 The ESI TOF MS spectrum showing the [M+H] + ion of the nitroxide labeled MIGGIGGFIXVR peptide at m/z 1406.6 ................................ ........................... 172 6 8 The ESI TOF MS spectrum shows the d oubly charged ion of the nitroxide labeled MIGGIGGFIXVR peptide and the tandem mass spectrum that confirmed the presence of amino acid X (Cys+R1). ................................ ......... 173 A 1 DEER data processing for the D30N v ar iant ................................ .................... 177 A 2 DEER data processing for D30N with CA p2 ................................ ................... 178 A 3 DEER data processing for D30N with ritonavir (RTV) ................................ ...... 179 A 4 DEER data p rocessing for the M36I variant ................................ ..................... 180 A 5 DEER data processing for M36I with CA p2 ................................ ..................... 181 A 6 DEER data processing for M36I with ritonavir (RTV) ................................ ........ 182 A 7 DEER data p rocessing for the A71V variant ................................ ..................... 183 A 8 DEER data processing for A71V with CA p2 ................................ .................... 184 A 9 DEER data processing for A71V with ritonavir (RTV) ................................ ....... 185 A 10 DEER data proces sin g for the D30N/M36I variant ................................ ........... 186 A 11 DEER data proces sing for the D30N/A71V variant ................................ ........... 187

PAGE 14

14 A 12 DEER data proces sing for the M36I/A71 V variant ................................ ............ 188 A 13 DEER data processing for the D30N/M36I/A71V varian t ................................ .. 189 A 14 DEER data processing for D30N/M36I/A71V with CA p2 ................................ 190 A 15 DEER data processing for D30N/M36I/A71V with ritonavir (RTV) .................... 191 A 16 DEER data processing for MDR769 apo ................................ .......................... 192 A 17 DEER data processing for MDR769 with indinavir (IDV) ................................ .. 193 A 18 DEER data processing for MDR769 with nelfinavir (NFV) ................................ 194 A 19 DEER data processing for MDR769 with atazanavir (ATV) .............................. 195 A 20 DEER data processing for MDR769 with saquinavir (SQV) ............................. 196 A 21 DEER data processing for MDR769 with amprenavir (APV) ............................ 197 A 22 DEER data processing for MDR769 with ritonavir (RTV) ................................ 198 A 23 DEER data processing for MDR769 with darunavir (DRV) ............................... 199 A 24 DEER data processing for MDR769 with lopinavir (LPV) ................................ 200 A 25 DEER data processing for MDR769 bound to tipranavir (TPV) ........................ 201 A 26 DEER data processing for MDR769 with CA p2 ................................ ............... 202 C 1 1 H 15 N HSQC spectra MDR769 titrated with amprenavir (APV). ...................... 207 C 2 1 H 15 N HSQC spectra MDR769 titrated with atazanavir (ATV) ......................... 208 C 3 1 H 15 N HSQC spectra MDR769 titrated with darunavir (DRV) .......................... 209 C 4 1 H 15 N HSQC spectra MDR769 titrated with saquinavir (SQV) ......................... 210 C 5 1 H 15 N HSQC spectra MDR769 titrated with lopinavir (LPV) ............................ 211 C 6 1 H 15 N HSQC spectra MDR769 titrated with nelfinavir (NFV) ........................... 212 C 7 1 H 15 N HSQC spectra MDR769 titrated with indinavir (IDV) ............................. 213 C 8 1 H 15 N HSQC spectra MDR769 titrated with ritonavir (RTV) ............................ 214 C 9 1 H 15 N HSQC spectra MDR769 titrated with tipranavir (TPV) ........................... 215 C 10 1 H 15 N HSQC spectra MDR769 titrated with CA p2 ................................ .......... 216

PAGE 15

15 LIST OF ABB REVIATIONS Regularization parameter Angstrom (10 10 meter) Modulation depth | C | Change in the fractional occupancy of the closed state G Change in Gibbs free energy decay Decay time max Maximum dipolar evolution time f Final frequency i Initial frequency A Adenine AIDS Acquired immune deficiency syndrome Ala (A) Alanine APT Approximate Pake transformation APV Amprenavir Arg (R) Arginine Asn (N) Asparagine Asp (D) Aspartic acid ATV Atazanavir AZT D (+) azido deoxythymidine BME mercaptoethanol BMRB Biological Magnetic Resonance Bank B si Stabilized and inactive subtype B HIV 1 protease C Cytosine CA Capsid

PAGE 16

16 CCR5 C C motif chemokine receptor type 5 CD Circular dichroism CD4 Cluster o f differentiation 4 c opt Optimal concentration CPMG Carr Purcell Meiboom Gill CRF Circulating recombinant form CW EPR Continuous wave electron paramagnetic resonance CXCR4 C X C motif chemokine receptor type 4 Da Dalton DEER Double electron electron re sonance diGly (Gly Gly) glycylglycine DNA Deoxyribonucleic acid dNTP Deoxynucleotide triphosphate DQC Double quantum coherence DQFR Double quantum filtered refocused DRV Darunavir DSC Differential scanning calorimetry dsDNA d ouble stranded DNA DTT D ithiothreitol E. coli Escherichia coli EDTA Ethylenediaminetetraacetic acid Env Envelope EPR Electron paramagnetic resonance ESE EM Electron spin echo envelope modulation ESI Electrospray ionization

PAGE 17

17 FDA Food and Drug Administration FDNA 3 (5 fluoro 2, 4 dinitro anilino)proxyl FPV Fosamprenavir FWHM Full width at half maximum G Guanine Gag Group specific antigen Gln (Q) Glutamine Glu (E) Glutamic acid gp Glycoprotein HAART Highly active antiretroviral therapy His (H) Histidine HIV 1 Human immuno deficiency virus type 1 HIV 2 Human immunodeficiency virus type 2 HSQC Heteronuclear single quantum coherence IAP 3 (2 iodoacetamido) PROXYL IASL 4 (2 i odoacetamido) TEMPO IB Inclusion body IC 50 Half maximal inhibitory concentration IDV Indinavir Ile ( I) I soleucine IN Integrase IPTG Isopropyl D thiogalactoside ITC Isothermal titration calorimetry k cat Turnover number K d Dissociation constant

PAGE 18

18 kDa kiloDalton k ex Exchange rate K i Inhibition constant K m Michaelis constant LB Luria Bertani Leu (L) Leucine LPV Lopinavir Lys (K) Lys ine MA Matrix MALDI Matrix assisted laser desorption ionization MC Monte Carlo MD Molecular dynamics MDR Multi drug resistant Met (M) Methionine MHC Major histocompatibility complex mRNA messenger RNA MS Mass spectrometry MSL 4 m aleimido TEMPO MTSL ( 1 o xyl 2,2,5,5 t etramethyl p yrroline 3 m ethyl) m ethanethiosulfonate MWt Molecular weight N A 23 mol 1 ) NC Nucleocapsid Nef Negative factor protein NFV Nelfinavir NMR Nuclear magnetic resonance

PAGE 19

19 NNRTI Non nucleoside reverse transcriptase inh ibitor NOE Nuclear Overhauser effect NRTI Nucleos(t)ide reverse transcriptase inhibitor ns nanosecond OD Optical density ORF Open reading frame P Probability distribution PAGE Polyacrylamide gel electrophoresis PCR Polymerase chain reaction PDB Pro tein Data Bank PEG Polyethylene glycol PELDOR Pulsed electron double resonance PES Polyethersulfone Phe (F) Phenylalanine P i Normalized population probability PI Protease inhibitor pI Isolectric point PMPR Pentamutated protease Pol Polymerase ppm par ts per million PR Protease Pro (P) Proline PROXYL 2,2,5,5 tetramethylpyrrolidine N oxyl r Pearson product moment correlation coefficient r AB Inter spin distance

PAGE 20

20 R Gas constant (8.3144621 J/mol.K) R 1 Longitudinal relaxation rate (reciprocal of T 1 ) R 2 Tran sverse relaxation rate (reciprocal of T 2 ) R 2 Coefficient of determination Rev Anti repression transactivator protein RNA Ribonucleic acid RT Reverse transcriptase RTV Ritonavir SDS Sodium dodecyl sulfate SDSL Site directed spin labeling Ser (S) Seri ne SIV Simian immunodeficiency virus SQV Saquinavir SNR Signal to noise ratio SU Surface T Thymine T Temperature T 1 Longitudinal relaxation tim T 2 Transverse relaxation time Tat Transactivating regulatory protein TCEP T ris(2 carboxyethyl)phosphine TEM PO 2,2,6,6 tetramethylpiperidine 1 oxyl Thr (T) Threonine TKR Tikhonov regularization TM Transmembrane

PAGE 21

21 T m Melting temperature T m Phase memory time TOAC 2,2,6,6 tetramethyl N oxyl 4 amino 4 carboxylic acid TOF Time of flight TPV Tipranavir Trp (W) Tryp tophan Tyr (Y) Tyrosine URF Unique recombinant form UV Ultraviolet Val (V) Valine Vif Virion infectivity factor Vis Visible Vpr Viral protein R Vpu Viral protein U 18 WT Wild type x c Zero time y i exp Expected value y i TKR Data point in Tikhonov regul arization fit

PAGE 22

22 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy INSIGHTS INTO THE MECHANISMS OF HIV 1 PROTEASE DRUG RE SISTANCE FROM PULSED ELECTRON PARAMAGNETIC RESONANCE AND NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY By Ian Mitchelle Sayo de Vera December 2012 Chair: Gail Elizabeth Fanucci Major: Chemistry The hum an immunodeficiency virus type 1 protease (HIV 1 PR) is a n attractive target in AIDS therapies that utilize protease inhibitors (PIs) because this enzyme is required for viral maturation. However, the rapid emergence of drug pressure selected mutations in the HIV genome after exposure to PIs oft en results in d rug resistance. The understanding of the molecular mechanism by which the accumulation of mutations impart s resistance to PIs is important for future drug design. W e used double electron electron resonance (DEER) spectroscopy to illuminate how amino aci d substitutions in the apo protease and in the presence of inhibitors result in perturbations of the equilibrium fractional occupancies of HIV 1 PR conformation al populations We found that drug resistance emerges from combinations of mutations that stabi lize open like conformations while destabilizing the closed state and retaining the semi open population seen in native protease. Changes in protease flap conformations were also monitored in the presence of inhibitors in the multi drug resistant HIV 1 PR variant MDR769 Specifically variations in inhibitor IC 50 values compared to the native enzyme are related to the relative change

PAGE 23

2 3 in the inhibitor induced shift to the closed state | C |. A linear correlation was found between | C | and the fold change i n IC 50 when inhibitor binding is not too weak. In 1 H 15 N heteronuclear single quantum coherence ( HSQC ) titration experiments for MDR769 asymmetric i nhibitors that have longer residence time in the protease binding pocket showed peak splitting of sever al HSQC resonances because of slow exchange in the NMR timescale Peak splitting coincided with strong induction of flap closure based on DEER results On the other hand, several resonances we re missing in the HSQC spectra when an inhibitor undergo es inte rmediate exchange in the NMR timescale with the binding cleft Peak disappearances due to broadened line widths of several HSQC peaks were consistent with weak inhibitor induced flap closure. Finally, we compared the residue specific backbone flexibility of MDR769 with subtype B protease by measuring 15 N spin relaxation parameters, namely T 1 T 2 and NOE Results suggest that the flap elbow and T80 loop of MDR769 are more rigid relative to subtype B.

PAGE 24

24 CHAPTER 1 INTRODUCTION AND RESEARCH OVERVIE W Human Immunodeficiency Viru s and the AIDS Pandemic Approximately 60 million people have been infected by the human immuno deficiency vi rus (HIV) from the time acquired immunodeficiency syndrome (AIDS) has been identified as a disease in 1981, resulting in 25 million deaths 1 AIDS is a severe HIV related immunological disorder characterized by the increased susceptibility to opportunistic infections cer tain rare cancers and neurological disorders. 2 HIV is a member of genus Lentivirus of the Retroviridae family. 3 Lentiviruses are known for their long incubation periods, and being a retrovirus, these are characterized by their single stranded positive sense ribonucleic acid (RNA) genome that is reverse transcribed into deoxyribonucleic acid (DNA) for integration into the genome of the host cell HIV 1 vs. HIV 2 The most virulent, pandemic HIV strain called HIV 1 was found to be almost identical to a type of simian immunodeficiency viru s (SIV) namely SIVcpz a strain identified from a subgroup of chimpanzees known as Pan troglodytes troglodytes which were once common in west central Africa. 4 HIV 1 is believed to have been introduced into humans in Central Africa between 1884 and 1924. 5 Meanwhile, t he significantly less infectious and rarer HIV 2 virus is thought to come from SIV in sooty mangabeys, and believed to have been transmitted to humans around 1940 in Guinea Bissau. 6 The DNA sequences of HIV 1 an d HIV 2 genomes, although similar, differ by 55%. Genetic Variability of HIV 1 HIV 1 has great genetic diversity, being classified into groups, subtypes, circulating recombinant forms (CRF) and unique reco mbinant forms (URF) 7, 8 as shown

PAGE 25

25 in t he phylogenetic tree in Figure 1 1 The most common group is Group M (major), which is further subdivided to subtypes A D, F H, J, K and several CRFs such as CRF01 AE and CRF02 AG. The subtypes are taxonomic classifications within a particular lineage whereas CRFs are from different recombinant forms of the virus that arise from genetic combination of viral subtypes. Meanwhile, other viral groups are extremely rare, such as G roup O viral strains that can be found in west central Africa and Groups N and P from Cameroon. 9 This genetic variability could be attributed to the lack of proofreading ability of the viral reverse transcriptase, resul ting in a substantially high mutation rate in HIV 1 (3.4 x 10 5 mutations/bp/replication). 10 Other factors accounting for the genetic variation include high recombination frequency and high in vivo replication rate. 9 Figure 1 1. Phylogenetic tree of HIV 1 Although s ubtype B is the predominant form of HIV 1 found in the A mericas, Western Europe and Australia it accounts for < 10% of the world wide infections. 7

PAGE 26

26 This subtype has been widely studied in clinical trials for screening new drugs and for susceptibility evaluation. Th e consensus sequen ce for subtype B is derived from the LAI isolate, 11, 12 designated as B.FR.83.HXB2 in the Los Alamos HIV database The subtype that accounts for the most number of infections worldwide is Subtype C from sub Saharan Africa. 1, 9 The circulating recombinant forms are mostly genetic mosaics of subtype s A and E, with CRF01 AE an d CRF01 AG being comm on in East Asia and West Africa, respectively. 13 URFs, on the other hand, are un ique sequences obtained from individual s that differ from earlier classifications The geographical distributio n of the different HIV 1 variants is shown in Figure 1 2. 14 Figure 1 2. Geographical distribution of HIV 1 protease variants (modified from Chan and Kantor). 14

PAGE 27

27 The HIV 1 Genome The HIV 1 genome (Figure 1 3 ) is approximately 9.8 kb and consists of several open reading frames (ORF ) that are transcribed into viral proteins 3 The three largest genes, namely gag, pol and env are translated into polyprotein chains that are even tually processed into the individual proteins by the viral protease. The group specific antigen ( gag ) gene is the gen omic region encoding the structural proteins. The p55 myristoylated protein precursor called assemblin associates with the plasma membrane and is processed by the viral protease to p17 (MAtrix), p7 (NucleoCapsid) p24 (CApsid), and p6 proteins. Meanwhil e, the polymerase ( pol ) gene encodes the viral enzymes, consisting of protease (PR), reverse transcriptase (RT) and integrase. These enzymes are produced as a Gag Pol precursor polyprotein gag that gets proce ssed by the viral protease. 11 Figure 1 3. Schematic diagram of the HIV 1 genome ( adapted from Levy). 3 The third largest gene, the envelope gene ( env ) encodes proteins involved in host cell recogniti on and binding, namely surface (SU) and transmembrane (TM) proteins.

PAGE 28

28 The viral glycoprotein gp160 is produced as a precursor, and is processed into a non covalent complex of the external glycoprotein gp120 and transmembrane glycoprotein gp41. The mature gp120 gp41 proteins are held noncovalently and are associated as a trimer on the cell surface. The binding site for t he CD4 receptor and seven trans membrane chemokine co receptors of HIV 1 are in gp120. Meanwhile, tat and rev are genes that code for reg ulatory proteins involved in upregulating viral replication and expression, respectively. The negative factor protein that downregulates CD4 and MHC Class I molecules is encoded by Nef. The genes, vif vpr and vpu are responsible for the production of a ccessory proteins that promote viral replication, assembly and budding; CD4 cell degradation; virulence; and immune response suppression of the host cell. Structure of the HIV 1 Virion The spherical HIV 1 virions (Figure 1 4) have a diameter of 100 120 nm 15 and have two main com ponents: (1) the cone shaped core tha t has a shell made up of multiple copies of the viral capsid (CA) protein and (2) the envelope that consists of viral surface proteins embedded on the lipid bilayer Within the core, two identical RNA strands are closely associated with reverse transcrip tase (RT) and nucleocapsid (NC) proteins. 3 The NC proteins sterically block access to the RNA, thereby helping with the maintenance of nucleic acid integrity. Other viral enzymes, such as protease (PR) and integrase (IN) are also localized within the co re. The lipid bilayer is embedded with up to 72 copies of gp41 transmembrane (TM) protein s, 16 and each is noncovalently linked to three gp120 external surface (SU) protein s These spiky microscopy tomography images. 17 The inner bilayer leaflet of the virion consists of

PAGE 29

29 myristoylated matrix (MA) proteins. 18 This modification provides a 14 carbon fatty acid region that conveniently anchors MA to the bilayer. Immature and mature virions can be differ entiated based on their interior organization. Since the polyprotein precursors, gag and gag pol have not yet been processed by the viral protease, immature virions do not have a core. However, the envelope structures are essentially the same for immatur e and mature virus particles because gp120 and gp41 are proteolytically processed prior to viral budding. Figure 1 4. Structural assembly of viral proteins in an HIV 1 virion (adapted from Levy). 11 The HIV 1 Life Cycle The life cycle of HIV 1 is shown in Figure 1 5. 19 HIV binds through its gp120 protein onto T cells that possess CD4 antigen on their surface. After CCR5 and CXCR4 recruitment, the virus enters host cells by fusion and endocytosis. Once inside the cell,

PAGE 30

30 viral uncoa ting takes place, releasin g the virion contents into the host cell. Successful uncoating generates the reverse transcription complex, which reverse transcribes viral RNA into double stranded DNA. Figure 1 5. The HIV viral life cycle. Figure adapted from the U.S. Department o f Health and Human Services. After completing reverse transcription, the viral DNA enters the cell nucleus, where it is integrated with the genomic DNA of the host cell by a virally encoded enzyme called integrase. The integrated DNA remains dormant unti l the region is activated for transcription to messenger RNA (mRNA), which is then translated into viral proteins. The viral RNA and protein s assemble at the cell membrane and bud from the cell as new immature and non infectious viral particles. The vira l protease then cleaves itself from the polypeptide chain and cleaves the rest of the polyprotein precursor into individual mature proteins, which rearrange within

PAGE 31

31 the virion. Polypeptide precursor cleavage and rearrangement of proteins is the maturation step the last step in the HIV life cycle. HIV 1 Therapeutic Regimens One of the first attempts to treat HIV started with the introduction of zidovudine or AZT ( D (+) azido deoxythimidine) in 1987. 20 Howeve r, the drug did not seem to give durable efficacy, leading to the development of other antiretroviral drugs. Currently, there are 25 FDA approved antiretroviral drugs, and these are classified based on their target step in the HIV life cycle: (1) fusion inhibitors and coreceptor antagonists that prevent viral entry, (2) nucleoside/nucleotide and non nucleoside reverse transcriptase inhibitors that block reverse transcription, (3) integrase inhibitors that deter viral DNA integration into the genom e of the host cell, and (4) pro tease inhibitors that block viral maturation. 21 The use of highly active antiretroviral therapy (HAART) at the end of 1995 in US A and Western Europe led to massive improvements in the treatment of HIV infection. 22 HAART uses a combination of three drugs belonging to the nucle oside/nucleotide reverse transcriptase inhibitors (NRTIs), non nucleoside reverse transcrip t ase inhibitors (NNRTIs) and/or aspartic protease inhibitors (PIs). Protease inhibitors (PI) specifically target HIV 1 protease ( HIV 1 PR ), which block HIV 1 replic ation and deter post translational processing, consequently suppress ing viral load. 23 HIV 1 Protease Protease i nhibitors target the human immunodeficiency vir us type 1 protease (HIV 1 PR), a 99 amino acid homodimeric aspa rtic protease that cleaves G ag and Gag P ol polyproteins at specific sites to liberate the mature structural (capsid, nucleo capsid and matrix) a nd functional (reverse transcriptase, RNase H and integrase)

PAGE 32

32 proteins important for viral maturation and propagation. HIV 1 PR catalyzes its own release from the Gag P ol polyprotein by cleaving its N and C termini to produce the mature protease domain. S tructure of HIV 1 Protease HIV 1 protease (EC 3.4.23.16), a member of the aspartic protease family, 24 is a C2 symmetric homodimer composed of two 99 resid ue polypeptides. 9 The dimer is made up of structurally identical monomers stabilized by inter monomer interactions The m aj ority of noncovalent interactions occur in the dimerization domain, where 12 hydrogen bo nds stabilize the four intercalated terminal strands from each monomer. The dimer is further stabilized by inter monomer interactions in the active site. Like most r etroviral protease s HIV 1 PR is homologous to several cellular aspartic proteases, suc h as pepsin. 25 The general template for aspartic proteases is mapped onto the LAI amino acid sequence and structure of HIV 1 PR in Figure 1 6. 26 29 The structural components are labeled alphabetically from the N to C terminus. The outer edges of the four sheet dimerization domain are residues 1 6 ( strand A) at the N terminus The fulcrum consists of residu es 9 15 ( strand B) and 18 24 ( strand C). The active si te triad (Asp25, Thr26, and Gly 27) can be found in a loop between strands C and D (residues 30 35). In most cellular proteases, strand D is followed by a helix E. This structural element is s ubstituted by a disordered loop E ( residues 36 42) in viral protea ses, which is called an elbow in HIV 1 PR. The flap consists of residues 43 49 ( strand F) and 52 66 (half of strand G). Access to the active site floor formed by the loop between s trands C and D is only

PAGE 33

33 Figure 1 6. HIV 1 PR topology. A) LAI amino acid sequence and corresponding s chematic diagram of secondary structural elements based on the aspartic protease template. The barrel and arrows are the helix and strands, respect ively. B) HIV 1 PR ribbon diagram (PDB ID 1HHP) that is color coded in a manner that matches t he aspartic protease template. C) Top view of the HIV 1 PR ribbon diagram. Structures are rendered in PyMOL. gained upon opening of the glycine rich flaps Th e cantilever is made up of residues 69 78 ( strand H) and the other half of strand G. Residues 83 85 ( strand I) form a portion of the active site wall that leads to helix J (residues 86 94). Finally, the dimer interface is completed by strand K (r esidues 95 99) at the C terminus and is situated between the A strands.

PAGE 34

34 Se quence homology comparison of HIV 1 PR with other cellular aspartic proteases reve al that this enzyme contains active site catalytic triad 30 Asp25, Thr26 and Gly 2 7 with a highly conserved geometry The binding cleft of the protease is formed by residues 25 32, 47 5 3, and 80 84, where Asp25/Asp25 are the major catalytic residues. Note that residues of the other monomer are designated with the prime ( ) symbol. In Asp25, Thr26, and Gly27 (Figure 1 7) Carboxylate O 1 atoms in Asp25/Asp25 closely interact in a co planar geometry. Both the backbone carbonyl of Leu24 and backbone 1 atom of each Thr26 is protona ted by the amide group of Thr26 and deprotonated by the Leu24 hydrogen bonds between the backbone carbonyl of Asp25 and the backbone amide of Gly27. The hydrogen bond network results in a geometrically conserved active site, where the catalytic Asp residues are strategically positio ned to allow proteolytic activity. Like all retroviral proteases, HIV 1 PR is only ac tive as a homodimer. Each mono mer contributes one aspartyl (Asp25) residue that lies at the bottom of the cavity. The pH rat e profile of this enzyme suggests that one of the active site aspartic acid residues (pKa = 3.1 and 5.2) 31 is protonated in the active pH range. 32 Each subunit of the active PR homodimer has a g lycine rich extended hair pin re gions called flaps. These two flexible antiparallel strands act as a gate that restricts substrate or inhibitor access to the active site. Inhibitor bound crystal structures reveal that flap residues have multiple interactions with the ligand tha t contribute to 50% of the inhibitor protease interactions. 33

PAGE 35

35 Figure 1 7. A) Top view of t 1 PR active site. B) Front Atoms are color coded such that car bon, oxygen, and nitrogen are gray, red and blue, respectively. To distinguish the carbon atoms of Leu24 and Thr26, these are shaded cyan for Leu24 and green for Thr26. Note that residues of the other monomer are designated with the prime ( ) symbol. St ructures are rendered in P yMOL. The flaps are locked in toward the active site due to van der Waal contacts and hydrogen bonding with the inhibitor. In this position, the flaps are in the closed conformation and access to the active site is blocked Mean while, i n the apoenzyme, the flaps are typically positioned slightly farther from the bin ding pocket into a semi open conformation, which is the most thermodynamically stable conformation in solution in the absence of a ligand. 27 Different techniques have been employed to study the flap conformational sampling and stability of HIV 1 PR, including nuclear magnetic resonance (NMR) spectroscopy, isothermal titration calorimetry (ITC), differential scanning calorimetry (DSC) molecular dynamics (MD) studies and pulsed electron paramagne tic resonance (EPR) experiments 34 52 Substrate access is blocked in the semi open conformation as shown in the space filling model in Figure 1 8. Clearly, in the closed conformation, there will likewise be no sufficient space for ligand entry. The inaccessibility of the active site pocket in both

PAGE 36

36 closed and semi open conformations implies that a large scale flap opening is neces sary to allow substrate or inhibitor entry. Figure 1 9 show s the three predominant HIV 1 PR conformations from results of molecular dynamics (MD) simulations. 46 MD results revea l small motions within the active site pocket in the course of flap conformational changes, where symmetry partners move closer upon flap closing and slightly farther when the flap opens. Flap opening involves synchronized movements of the fulcrum, elbow and cantilever (Figure 1 8). Figure 1 8. Structure of HIV 1 PR (PDB ID 1HHP). A) Space filling model of HIV 1 PR showing the restricted access to the binding pocket even when the flaps sit slightly farther away from the active site in the semi open conformation. B) Ribbon diagram showing the structural anatomy of the protease. As the flaps open, the cantilever and elbows move downwards, while the fulcrum pivots around the protease core, thereby accommodating the cantilever shift. In contrast to the outer edges, the binding pocket remains relatively static. 46, 53 In the closed state, the flap handedness is reversed relative to the semi open conformation.

PAGE 37

37 Figure 1 9. Predominant flap conformations of H IV 1 protease. A) Closed conformation (PDB ID 1OHR). B) Semi open conformation (PDB ID 1HHP). C) Wide open conformation. D to F are the corresponding top view images of A to C, sho wing the relative horizontal flap displacement and how ligand access is con trolled by the flap conformation Note that the flap handedness is reversed in the semi open conformation when compared to the closed state Figures A, B, D and E are rendered in PyMOL while C and F are modified from Hornak et al. 46 The semi open state has been proven to be the most stable conformation for the free protease based on NMR, x ray crystallography and M D studies. N uclear O ver hauser effect (NOE) and 15 N spin relaxation experiments have revealed flap flexibility in the apoenzyme, where conformational exchange occur s as indicated by high amplitude backbone motions in the s to ms timescale. 38 Moreover, the flap tips (residues 48 52) displayed sub ns fluctuations indicating the existence of an ensemble of semi open conformations along with minor populations of the closed and wide open states. 37 NMR results have been corroborated by MD simulations that reveal presence of the closed, semi open, wide open and curled/tucked states in the free protease 46, 49, 54 The curled/tucked conformation is an alternative flap opening whe re the flap tips curl in toward the binding pocket, burying several hydrophobic residues into the binding cleft. 54

PAGE 38

38 Depending on the MD study invoked, the curled/tucked conformation which involves horizontal or sideways flap displacement may or may not allow substrate or inhibitor entry 49, 54 Horizontal flap opening has been observed in sev eral x ray crystal structures of HIV 1 PR. For instance, the x ray crystal structure of the multi drug resistant HIV 1 PR variant MDR769 revealed an expanded active site pocket and horizontally opened flaps 55 (Fig ure 1 10) as opposed to the vertical flap opening observed in MD studies. 46 Horizontal flap opening is also observed i n HIV 1 PR crystallized with an inhibitor in the presence of bulky inorganic molecules. 56 Figure 1 10. Apoenzyme X ray crystal structure of MDR769 overlaid with wild type (LAI). A) Front view of MDR769 (PDB ID 1TW7) ribbon diagram shown in orange overlaid with BRU/LAI (PDB ID 1HHP) rendered in cyan. B) Side view clearly shows the horizontal flap opening of MDR769 where the distance between the tips increased by 6 relative to LAI. Structures are rendered in P yMOL modeling software. HIV 1 Protease Inhibitors The design of HIV 1 PR inhibitors wa s originally based on peptidomimetic inhibitor models of other aspartic proteases, such as rennin. 57 Many peptide mimetics cont ain various tetrahedral intermediates substituted for the scissile amide bond. 58, 59

PAGE 39

39 Tetrahedral intermediates containing a hydroxyethylene isostere 60 (Figure 1 11A) are fo und to be more superior over other analogues that contain phosphinates, fluoro ketones or reduced amide groups (Figure 1 11B). The structure of efficient transition state analogues slightly varies among inhibitors but almost all are based on a hydroxy ethylene or hydroxyethylamine (Figure 1 11C) scaffold. 59 Current ly, 10 protease inhibitors (PIs) are approved by the United States Food and Drug Administration (FDA) (Figure 1 11D). Figure 1 11. Protease inhi bitor scaffolds and structures. (A) Hydroxyethylene scaffold. (B) Hydroxyethylamine scaffold. (C) Reduced amide scaffold. (D) FDA approved protease inhibitors: saquinavir (SQV), indinavir (IDV), ritonavir (RTV), nelfinavir (NFV), fosamprenavir (FPV ), lopinavir (LPV), darunavir (DRV), amprenavir (APV), atazanavir (ATV) and tipranavir (TPV). The first PI to get FDA approval in 1995 wa s saquinavir (SQV) 61 R itonavir (RTV) and indinavir (IDV) were available a year later. Other inhibitors availab le in the market

PAGE 40

40 and the corres ponding year of FDA approval are as follows: nelfinavir (NFV; 1997), amprenavir (APV; 1999), lopinavir (LPV, 2000), fosamprenavir (FPV; 2003), atazanavir (ATV; 2003), tipranavir (TPV; 200 5), and darunavir (DRV; 2006). 62 64 The chemical structures for these inhibitors are shown in Figure 1 11D. All of these PIs are competitive inhibitors with HIV 1 PR bindin g affinities ranging from low nanomolar to picomolar 62 TPV is the only PI that does not have a hydroxy ethylene or hydroxyethylamine scaffold, and hence the only non peptidomimitic inhibitor. Instead, TPV contains a coumarin scaffold, which is base d on the lead compound, phenpro coumon. 65 HIV 1 Protease Subtype Polymorphisms Recent data from the Los Alamos HIV database reveal high genetic variability among clinical isolates o f HIV 1 protease 66 The prototype sequence for wild type HIV 1 protease 67 is unknown, although HXB2 has been used as a consensus sequence based on historical precedence. 68 HIV 1 protease sequences that arise from natural polymorphisms are classified into groups, subtypes and circulating recombinant forms as shown earlier in the phylogenetic tree in Figure 1 1. The taxonomic groups refer to the viral lineage while subtypes refer to sequences from a common progenitor that differs by more than 15% and 20% in the gag and env genes, respectively. 11 G ene tic variability in HIV 1 PR is attributed to the high mutation rate of the encoding gene, which arises from the lack of proofreading ability in the HIV reverse transcriptase. Other factors, including variant accumulatio n during the course of infection, high recombination frequency and high replication rate have been cited to contribute to the genetic differences among clinical isolates. 9

PAGE 41

41 N atural polymorphisms within a subgro up are not random but are found in specific regions of HIV 1 PR. Critical regions to protease structure or function are either conserved or show no significant variation. Drug pressure Selected Mutations in HIV 1 Protease The high mutation frequency in th e HIV genome plays a crucial role in drug resistance development and evasion of the immune system by providing a mechanism for evolving adaptation to constantly changing environments M ore than 50% of the amino acid substitutions in drug resistant HIV 1 p rotease variants arise from the accumulation of multi ple mutations, resulting in reduced affinity to protease inhibitors while maintaining turnover rates for natural substrates. 69 During several generations of viral replication in the presence of an inhibitor, viruses that contain random mutations that favor viral propagation will survive. These variants have increased viral fitn ess over the generations and have developed drug resistance. Amino acid substitutions in HIV 1 PR that result from mutations in the HIV genome are classified into two categories: prima ry and compensatory (Figure 1 12 ). The efficacy of currently availab le PIs is limited by the rapid emergence of mutations, where changes in at least 38 out of 99 amino acid residues occur under the selective pressure of PI therapy. 70 Primary mutations usually occur within the activ e site pocket and alter the shape of the b inding cavity. Structural evidence clearly shows that primary mutations lead to loss of crucial H bonding or van der Waals interactions with the inhibitor. 21, 63, 69, 71, 72 However, these mutations lead to viral replication impairment due to inefficient natural substrate processing.

PAGE 42

42 Figure 1 12 Sites of amino acid substitutions in HIV 1 protease A) Front view of the ribbon diagram (1OHR) shows primary and seconda ry mutations rendered in red and blue, respectively. B) Top view of the ribbon diagram. Figure adapted from Weber and Agniswamy. 70 Table 1 1 lists 15 primary mutations associated with PI resistance. The most co mmon mutations that emerge from therapy using FDA approved PIs are those in positions 10, 46, 54, 82 and 90 (Figure 1 13) Mutations that occur with highest frequency in drug resistant variants include D30N, G48V, I50V/L, V82A, I84V and L90M. These mutat ions usually appear in combinations. For instance, studies show that the I50L/V and I84V emerge after treatments using atazanavir, fosamprenavir and darunavir. 72 The I84V mutation coul d co evolve with L90M as primary mutations involved in therapy that uses ritonavir, indinavir, amprenavir and saquinavir. 72, 73 In addition, the G48V/L90M pair i s common ly associated with saquinavir resistance 21 while the D30N/L90M pair emerge after nelfinavir treatment. 74 Ritonavir is often employed as a pharma cological booster for other PIs and when used in combination with lopinavir and indinavir, results in the emergence of the V82A mutation. 73, 75 As shown i n Table 1 1, the selected mutations are the same with or without the ritonavir booster for treatments that utilize atazanavir 72

PAGE 43

43 Table 1 1. Mutations associated with protease inhibitor treatments. 72 Inhibitor Primary m utations Compensatory m utations Saquinavir a G48V, L90M L10I/R/V, L24I, I54V/L, I62V, A71V/T, G73S, V77I, V82A/F/T/S, I84V N elfinavir D30N, L90M L10F/I, M36I, M46I/L, A71V/T, V77I, V82A/F/T/S, I84V, N88D/S Indinavir a M46I/L, V82A/F/T, I84V L10I/R/V, K20M/R, L24I, V32I, M36I, I54V, A71V/T, G73S/A, L76V, V77I, L90M Atazanavir b I50L, I84V, N88S L10I/F/V/C, G16E, K20R/M/I /T/V, L24I, V32I, L33I/F/V, E34Q, M36I/L/V, M46I/L, G48V, F53L/Y, I54L/V/M/T/A, D60E, I62V, I64L/M/V, A71V/I/T/L, G73C/S/T/A, V82A/T/F/I, I85V, L90M, I93L/M Fosamprenavir a I50V, I84V L10F/I/R/V, V32I, M46I/L, I47V, I54L/V/M, G73S, L76V, V82A/F/S/T, L9 0M Darunavir a I47V, I50V, I54L/M, L76V, I84V V11I, V32I, L33F, T74P, L89V Lopinavir a V32I, I47V/A, L76V, V82A/F/T/S L10F/I/R/V, K20M/R, L24I, L33F, M46I/L, I50V, F53L, I54V/L/A/M/T/S, L63P, A71V/T, G73S, I84V, N88D/S Tipranavir a I47V, Q58E, T74P, V82L/T, N83D, I84V L10V, L33F, M36I/L/V, K43T, M46L, I54A/M/V, H69K/R, L89I/M/V a) with ritonavir b) with or without ritonavir On the other hand, c ompensatory or secondary substitutions appear at distal positions and compensate for the viral replica tion impairment due to the primary mutation or through natural polymorphisms prior to PI exposure. 76 For example, the N88D secondary mutation emerges in nelfinavir treatment to recover viral fitness that is impaire d by the primary D30N/L90M pair. Moreover, secondary mutations also appear as polymorphisms before therapy, such as those found in positions 10, 36, 46, 63, 71,

PAGE 44

44 77 and 82. 77 These secondary polymorphisms are found to influence inhibitor binding but are typically not located in regions of the protein that make physical contact with the PIs. 76, 78 80 The mechanism by which secondary mutations transmit their effects to the active site pocket and confer drug resistance remains unclear, but these non active site mutations have been implicated in altering flap dynamics and flexibility of HIV 1 PR throug h the hydrophobic sliding mechanism 81 or by either restricting flap opening for inhibitor uptake or preventing flap closure to allow inhibitor geometry optimization in the active site for high affinity bind ing. 82 These mutations have also been implicated in the restoration of protein stability. 83 Understandin g the role of these mutations in flap conformational sampling and inhibitor binding is essential for future drug design. Figure 1 13. Subtype B HIV 1 PR mutation prevalence in PI nave and treated patient isolates.

PAGE 45

45 A comparison of the HIV 1 subtype B sequences of the various H I V 1 protease s from the Stanford HIV database (updated June 2012) that were isolated from 17165 PI nave patients and 8231 patients who have received PI therapy, reveal that several residues in the protease flaps, active site and dimerization re gion are conserved (Figure 1 13). Drug Resistance Mechanisms The binding affinity of PIs can be severely decreased by natural polymorphisms and drug pressure selected mutations in the HIV 1 protease. This can be attr ibuted to the inherent flexibility of HIV 1 PR, where mutations could result to structural adjust ments throughout the protease that could transmit their effects to the binding pocket and alter crucial protein inhibitor interactions. 63, 69, 80, 81, 84, 85 The substrate envelope hypothesis proposed by Schiffer and co workers 86, 87 explains that an inhibitor that protrudes beyond the substrate envelope causes amino acid sub stitutions in regions of inhibitor contact, including but not limited to positions 30, 47, 48, 50, 82 and 84. Secondary mutations often found on the periphery of the protease, are also believed to confer drug resistance by favoring open flap conformati ons and simultaneously decreasing the rate of flap closure so that the protease would prefer substrate uptake while escaping inhibitor s 82 T he hydrophobic sliding mechanism 81 corroborates this hypothesis, where molecular dynamics (MD) simulations have revealed that 19 residues in the hydrophobic core, which includes 7 isoleucines, slide past one another by exchanging van der Waal s contacts while undergoing conformati onal transitions, which could be correlated with changes in protease dynamics. Moreover, these hydrophobic contact exchanges have minimal associated energy penalties, where the isoleucine residues have several

PAGE 46

46 rotameric states that facilitate the slidin g mechanism. Therefore, conformational heterogeneity alterations conferred by non active site compensatory mutations can be explained by the inherent plasticity of the hydrophobic core, where extensive mutation induced core rearrangements could be detrim ental to PI binding. Multi drug Resistant Patient Isolate MDR769 The multi drug resistant variant MDR769 has been isolated from a patient who has been failing treatment after long term antiretroviral therapy using saquinavir (SQV), nelfinavir (NFV), amp renavir (APV) and indinavir (IDV) Relative to the LAI consensus sequence, the MDR769 variant contains L10I, M36V, S37N, M46L, I54V, I62V, L63P, A71V, V82A, I84V and L90M (Figure 1 14). Figure 1 14. Mutation sites in patient isolate MDR769 (1TW7) rel ative to the LAI consensus sequence. The ribbon diagram is colored by subunit where sites of mutation are rendered as spheres on one monomer and as capped sticks on another. MDR769 exhibits higher level of resistance to m ost FDA approved inhibitors, ex cep t darunavir (DRV), tipranavir (TPV) and lopinavir (LPV). 88 Crystal structures of this variant reveal an expanded biding pocket in the apoenzyme, substrate bo und and inhi bitor bound forms. 55, 88 90 The expanded active site cavity is attributed to the primary

PAGE 47

47 mutations at positions 82 and 84, where substitutions of V82 and I84 with smaller hydro phobic residues A and V, respectively, afforded more space in the binding pocket Identical Transition States in Native and Drug resistant HIV 1 Protease Enzymatic transition structures for native and drug resistant I84V HIV 1 PR variants were established by Schramm and co workers 91 via kinetic isotope effect (KIE) measurements, im plying tha t protease inhibitors, as true tetrahedral transition state analogues should be effective for both native and drug resistant enzymes. However, the drug resistant variant I84V displayed up to 32 fold reduction in inhibitor binding affinity 92 suggesting that drug resistance in this variant arises from alterations in the enzyme distant from transition state interactions. Therefore, when discussing drug resistance in HIV 1 protease, other contributing factors such a s flap conformational sampling, backbone dynamics, and shape or volume of the binding pocket have to be considered. Research Overview Challenges to the Anal ysis of Conformational Sampling Various techniques are suitable for characterizing protein conformat ions, ensembles and ensemble shifts. However, most of these techniques have limited sensitivity and are incapable of detecting sparsely populated states. In addition, most techniques are applicable to a limited range of timescales and can only analyze particular energy landscape regions. For instance, fast timescale motions are difficult to characterize and can only be discussed in terms of statistical distributions because there are too many states present that cannot be individually trapped. On the other hand, motions in the slow timescale undergo slow interconversion, making them amenable to direct observation The h igher energy states present in slow timescale

PAGE 48

48 motions can be trapped via ligand binding or by choosing a set of conditions that stabi lize these states, such as pH, salt concentration and temperature. Structure determination techniques such as X ray crystallography and NMR spectroscopy, can be employed to study the high energy states. However, most of these techniques are limited by p articular technical requirements. For example, the state of interest needs to be amenable to crystallization to be suitable for X ray crystallography. On the other hand, isotopic labeling is required in NMR studies, thereby ruling out proteins isolated f rom natural sources. Moreover, NMR is limited by protein size, where analysis of proteins with molecular weight (MWt) >70 kDa is extremely difficult M oder n NMR techniques may be needed to approach this MWt limit, including: (1) ultra high field magnets (i.e. 800 1000 MHz ), (2) advanced pulse sequences and NMR electr onics, (3) isotopic labeling schemes, and (4) the highest sensitivity cryop robes P ulsed EPR methods, particularly double electron electron resonance (DEER) spectroscopy ha ve several advant ages over most techni ques for conformational studies. DEER spectroscopy (1) enable s simultaneous sampling of all the protein conformations; (2) can observe sparsely populated conformational states ; and (3) has no molecular weight upper limit for proteins that can be analyzed. However, acquisition of DEER data requires (1) incorporation of an EPR active spin label, (2) the addition of cryoprotectants (e.g., glycerol), and (3) sample freezing. Consequently, questions arise as to whether sample freezing is fast enough t o trap the conformations at physiological temperature. Moreover, there are also concerns whether the introduction of the spin label and addition of co solutes could perturb the conformational sampling. Concerns

PAGE 49

49 regarding sample freezing are addressed here by acquiring solution NMR data to corroborate the puls ed EPR results. Meanwhile, conformational perturbations due to spin labeling were minimized because solvent exposed sites on the HIV 1 PR flaps were used in DEER distance measurements. Finally, the addition of co solute could perturb conformational sampling depending on the amount and type of additive used. Previous pulsed EPR distance measurements 93 have shown that the addition of up to 40% glycerol did not alter the flap conformations of HIV 1 PR In this study, only 30% glycerol was added to the samples for D EER experiments. Specific Aims In this dissertation, the HIV 1 PR flap conforma tional sampling and protease inhibitor interaction dynamics were investigated using pulsed EPR and NMR spectroscopy. Specifically, the aims of this work were (1) to determine the effect of individual and combined drug pressure selected mutations to the HIV 1 PR flap conformational sampling, (2) to monitor alterations in inhibitor induced flap closure as a result of single and combined mutations, (3) to statistically correlat e the flap conformational sampling and inhibitor induced conformational sampling shifts to published enzyme kinetics parameters (i.e, k cat and K m ), inhibit ion constants ( K i ) and half maximal inhibitory concentration (IC 50 ) of inhibitors to the HIV virus, a nd ( 4 ) to compare the backbone flexibility of a multi drug resistant HIV 1 protease variant to that of a native protease. Scope s and Limitations T he conformational sampling and flexibility of HIV 1 PR were determined for subtype B HIV 1 protease variants with drug pressure selected D30N, M36I and A71V mutations using double electron electron resonance (DEER) pulsed EPR spectroscopy

PAGE 50

50 D30N is a primary mutation that emerges in response to nelfinavir treatment, 70, 94 while M36I and A71V are compensatory amino acid substitutions associated with treatments using various protease inhibitors. 70, 94, 95 The effects of the combined D30N, M36I and A71V mutations to conformational s ampling were correlated to the previously determined enzymatic parameters, namely k cat K m and k cat /K m 76 and inhibition constants ( K i ) for nelfinavir (NFV), ritonavir (RTV) and indinavir (IDV). The inhibitor RTV often used as a therapeutic booster in HIV treatment, and the non hydrolyzable substrate mimic, CA p2 were also added to the protease variants D30N, M36I, A71V and D30N/M36I/A71V to look into alterations in inhibitor induced flap conformational shifts in HIV 1 PR as a result of single and combined mutations Aside from the effect of single point mutations t o flap conformational sampling, the conformational heterogeneity of the multi drug resistant HIV 1 PR patient isolate MDR769 was monitored in the apoenzyme and in the presence of 9 FDA approved inhibitors. The conformational sampling shifts upon add ing inhibitors to MDR769 were compared to the degree of flap closure in subtype B PR, as previously determined. 34 The degree of flap closure was correlated w ith published IC 50 data for PIs. To corroborate DEER results acquired at cryogenic conditions (60 K) for MDR769 solution NMR data were also acquired at 20 C. In particular, protease inhibitor interaction dynamics were investigated by acquiring 1 H 15 N he teronuclear single quantum coherence (HSQC) spectra for samples titrated with increasing concentration of PIs (up to 1:1 protease to inhibitor ratio) Moreover, 15 N spin relaxation and nuclear Overhauser effect (NOE) parameters were determined in MDR769 t o look into protein backbone flexibility and internal mobility. Results of relaxation measurements for

PAGE 51

51 MDR769 were compare d to that of subtype B protease, which illuminate d the effects of drug pressure selected mutation s to the backbone dynamics of this c linical isolate To retard autoproteolysis, subtype B PR variants with stabilizing mutations Q7K, L33I and L63I were employed in pulsed EPR and NMR experiments 96, 97 However, stabilizing mutations were not incor porated in MDR769 PR. For pulsed EPR samples, t he naturally occurring cysteine residues (C67A and C95A) were substituted with alanine, while K55 was mutated to cysteine for chemical modification with a methane thiosulfonate (MTSL) nitroxide label. Moreov er, the active site D25N substitution was incorporated in all variants for pulsed EPR and NMR experiments to impart sample stability and homogeneity. 98, 99

PAGE 52

52 CHAPTER 2 PULSED EPR DISTANCE MEASUREMENTS IN SOLU BLE P ROTEINS BY SITE DIRECTED SPIN LABELING Introduction Distance measurements by site directed spin labeling ( SDSL ) electron paramagnetic resonance ( EPR ) are based on the dependence of the electron dipole dipole couplings, w hich scale as 1/ r 3 where r is the distance between unpaired spins There are four common EPR techniques utilized to interrogate specific distanc e ranges: exchange EPR for short distances in the 4 8 range 100 continuous wave (CW) EPR for distances of 8 25 101 103 an d two pulsed methods that cover distan ces from 15 80 : double electron electron resonance (DEER), also known as pulsed electron double resonance (PELDOR) 104, 105 and double quantum coherence (DQC ). 106, 107 The first technique, exchange EPR, rel ies on the overlap of two unpaired electron orbitals and dipole dipole broadening between the electron spins. Meanwhile, in CW EPR methods, distances are obtained by analyzing the l ine width increases that result from the dipole dipole interactions 101, 108 The distances are then obtained from a Fourier transform deconvolution to extract the Pake broadening function, which is simulated in ter ms of the distances and distance distributions that generated the Pake function. The upper limit of 20 25 for traditional nitroxide spin labeling experiments is set by the inherent inhomogeneous broadening of the nitroxide spectrum. This limit can be e xtended by deuterating the spin label or by using a spin label with a narrower line shape. Typically pulsed EPR techniques are utilized when measuring distances in the 20 80 range 106, 109 113 The four pulse D EER methodology determines the dipolar coupling between spins in the form of a modulation of the spin echo amplitude, with great sensitivity in the range of 20 80 achieving a precision of 0.3 for the lower end

PAGE 53

53 of this range 35, 112, 114 In DQC another pulsed EPR technique, the dipole interactions between spins generates double quantum coherences, where the rate of coherence formation directly reports on the strength of the dipolar interaction 106, 113 Either extrinsic or intrinsic EPR active species can be utilized in distance measurements. The former are artificially introduced, such as nitroxide radicals or metal centers chelated at targeted specific sites, whereas the latter are unpaired electrons that exist naturally in protein s such as amino acid radicals, paramagnetic metals and radical cofactors DEER has been utilized in a variety of biomolecules, such as peptides 115 proteins with tyrosyl radicals 116 soluble protein s 117 integral membrane protein s 118 and nucleic acids 119, 120 Distance measurements with DEE R are not limited to nitroxides D istances between metal centers, such as Cu 2+ Cu 2+ 121 between iron paramagnets in FeS clusters 122 Gd 3+ Gd 3+ 123, 124 and between metal and spin label, such as Cu 2+ nitroxide 125, 126 and iron sulfur center nitroxide 127 pairs have also been measured u sing DEER This list serves to provide select examples and is in no way inclusive of the numerous papers published in this discipline. Although distances up to 60 80 can be achieved for soluble proteins with matrix deuteration, the limit falls to 4 0 4 5 for membrane proteins, as fast relaxation times often do not allow for dipola r evolution times exceeding 1.8 2.0 s 128 In addition to using alternative spin markers for protein labeling, such as Gd 3+ sensitivity of the DEER ex periment on conventional nitroxide radicals can potentially be enhanced by going to higher frequencies. Assuming that the experiment can be done under otherwise identical conditions pulse d EPR sensitivity scales with the square of the resonance frequency 129 implying about a 13 and 100 fold sensitivity gain whe n going from X to Q

PAGE 54

54 and W band s, respectively Consequently, for Q band, a 169 fold decrease in data acquisition time has been achieved 130 However, such enhancement in absolute sensiti vity does not necessarily translate to better concentration sensitivity due to smaller sample volumes and lower excitation power at higher frequency. This chapter discusses the important experimental considerations for DEER data collection and analysis ut ilized for nitroxide based DEER studies of HIV 1 protease, which is compatible with distance measurements in o ther soluble proteins in the 20 50 range. Site directed Spin labeling Background Information Spin labeling of m uscle f ibers Early EPR studies focused on the role of the orientation and rotational motion of myosin heads in muscle contraction paved the way to the spin labeling of glycerinated rabbit muscle fibers. These experiments used spin labels with maleimide (MSL) and iodoacetamide (IASL) reactivities that target thiol groups on myosin heads 131 135 Another study on rabbit skeletal muscles attached the spin label 3 (5 fluoro 2,4 dinitro anilino)PROXYL (FDNA) that is specific to the amino group on a lysine residue in G actin 136 Spin labeling conditions, such as pH, vary depending on the spin label used. For instance, labeling of the thiol groups in myosin using MSL and IA SL was done at pH 7.0 131, 132, 135 while FDNA labeling of lysine in G actin was accomplished at a more basic condition (pH 8.0) 136

PAGE 55

55 S ynthetic and orthogonal labeling strategies for incorporating unnatural amino acid spin labels Aside from site directed spin labeling (SDSL), unnatural nitroxide amino acid residues such as the conformationally constrained 2,2,6,6 tetramethyl N oxyl 4 am ino 4 carboxylic acid ( TOAC ) label can be added to specific sites of the protein during peptide synthesis 137 Novel nitroxide labels can also be incor porated into the recombinant pro tein via the u nique amber stop codon in an orthogonal labeling strategy 138, 139 These advances are important for consideration of protein systems where cysteine cannot be utilized as a labeling site, such as those that have fu nctionally important native cysteine residues and disulfide bridges. Nitroxide Spin Labels By definition, a spin label is any molecule containing an unpaired electron and reactive moiety for binding to another molecule. Most of the spin labels employed for studying structure and flexibility of biomolecules are nitroxide radicals because their simple line shape is highly sensitive to motion, making them ideal for such investigations 140 The nitroxide radicals for spin labeling applications are persistent because the electron is prote cted by bulky methyl groups that sterically prevent collisions and hence limit the reactivity of the radical. The geometry of the radical and surrounding methyl groups is typically preserved by inclusion in a 5 membered pyrrole ring or 6 member ed piperidi ne ring. Pyrrole rings with an unsaturated bond have lower flexibility. Many versions of nitroxide spin labels have been generated by Kalman Heidig and the popularity of these labels has led to their commercial availability from Toronto Research Chemical s

PAGE 56

56 Commercially available spin labels vary in both their ring structure and in their reaction chemistry for attachment to the protein. Figure 2 1 shows the structures of 4 common nitroxide radicals with thiol based linkers as well as the structures of the modified cysteine after reacting w ith the spin label. Unlike other nitroxide spin labels, the unnatural, conformationally restricted nitroxide amino acid TOAC readily adopts helical backbone torsion angles and is incorporated as a rigid spin label d uring peptide synthesis 137, 141 Biradical nitroxide labels with linkers of different flexibility have also been used in membrane protein studies 142 Recently, the use of a triarylmethyl based spin label with a relatively long relaxation time enabled pulsed EPR distance measurements in a protein immobilized to a solid support 143 Figure 2 1. Chemical s tructure s of four common nitroxide spin labels before and af ter reacting to a cysteine side chain. A) and B) MTSL: (1 o xyl 2,2,5,5 t etramethyl p yrroline 3 m ethyl) m ethanethiosulfonate; C) and D) IAP : 3 (2 iodoaceta mido) PROXYL ; E) and F) MSL: 4 m aleimido TEMPO; and G) and H) IASL: 4 (2 i odoacetamido) TEMPO. The r ectangular box represents the protein backbo ne. Selection of Labeling Sites The choice of labeling sites within biological systems is not trivial. Obviously, the site should be chosen to report on the relevant aspect of the system under study. However, the introduction of the spin label should not alter the structure stability or function of the system and functional assays of the modified proteins are essential

PAGE 57

57 components to any SDSL EPR study For distance measurements, a solvent exposed reporter s ite is preferred over a buried site because it is less likely to produce structural perturbation due to its presence. T he extent of solvent exposure at the selected site ca n also impact the distance distribution profile 144 For solvent ex posed sites, the spin label has conformational rotational freedom dictated by the structure of the spin label, which often has averaged conformational sampling. Oftentimes, spin labels in buried sites or site of tertiary contact adopt multiple conformations, wh ich can complicate analysis of distance distribution profiles. A solvent exposed site with minimal contacts is often considered an optimum labeling site for distance measurements because the spin label adopts a limited number of rotamers 144 An important complement to any DEER experiment is the use of computational simulations, both for protein conformational sampling 145 and likely spin label rotamers at the chosen site in a protein 146 These studies are often necessary when building a biologi cal model around the experimentally observed distances and distance changes, and determining favorable spin labeling sites 147 An open source software MMM 2011.2, available from the Swiss Federal Institute of Technology Zurich website ( http://www.epr. ethz.ch ), all ows in silico incorporation of spin labels and model based fitting of confor mational changes in proteins. With this program, provided that the initial structural conditions and a set of probable distance constraints for the chosen spin label are known, p redictions for the structure following labeling can be made. Commonly, nitroxide spin labels (SL) are covalently linked to proteins at specific amino acid residues. The functional group attached to the nitroxide provides specificity to the labeling reac tion. In particular, cysteine residues are targeted by iodoacetamide,

PAGE 58

58 maleimide, indanedione and ketone functional groups attached to a nitroxide moiety, while activated esters are specific to lysines. Nitroxide probes are often used as extrinsic spi n labels in SDSL 148 Here native cysteines are usually replaced with another amino acid, such as alanine or serine residues, depending upon the relative hydrophobicity of the local site, and new cysteine residues are engineered at specific position s. Sulfhydryl specific spin labels a re then attached to the Cys residues. The linker by which nitroxide probes conjugate to the protein confers an intrinsic conformational flexibility. To reduce probe flexibility, MTSL has been modified in the 4 posit ion of the pyrrole ri ng to include bulky substituents such as phenyl 149 or bromo moieties. The bromo deriv ative of MTSL is commercially available 108, 150, 151 Additional restriction of spin label motion can be incorporated by bifunctional SL that enable s cross linking of two target sites on a peptide or protein 1 52 Alternatively, rigid unnatural amino acid spin labels can be incorporated into the backbone at specific sites of the poly peptide chain during peptide synthesis, in the case of TOAC 137 or by usin g an elegant method of unnatural amino acid mutagenesis 138 Site directed S pin labeling of Soluble Proteins The term SDSL has been coined to describe site directed mutagenesis combined with spin labeling. The most common site chosen for attaching a spin label is throu gh the reactive thiol functional group afforded by the amino acid cysteine. Given that cysteine occurs in most proteins with relatively low abun dance, it is typically straight forward to utilize modern molecular biology protocols to alter the DNA that enc odes for the addition or removal of cysteine in a protein. In general, the DNA encoding the protein of interest is altered such that the codon specific to an amino acid is mutated to

PAGE 59

59 encode for a cysteine residue. The mutant DNA is used to express the cy steine variant in a recombinant expression system such as E. coli or S. cerevisiae The protein is purified and labeled with a thiol reactive spin label. To ensure that the cysteine thiol group is in the reduced f orm during spin labeling, treat ment with reducing agents such as mercaptoethanol, t ris(2 carboxyethyl)phosphine (TCEP) or dithiothreitol (DTT ) may be necessary. The development of rapid site directed mutagenesis protocols allowed researchers to routinely introduce an engineered labeling site a t a desired position while allowing for the remov al of all native labeling sites; thus greatly enhancing the feasibility of SDSL EPR for proteins 153 There are three common chemistries for label attachment to cysteine side chains for SDSL of a soluble prote in; namely disulfide exchan ge via thiosulfonate, iodoace ta mide and male i mide reactivities (Figure 2 1) Distance Measurements via DEER Figure 2 2 A illustrates the pulse sequence used for four pulse DEER 105 In DEER experiments microwave pulses are used to excite two distinct spin populations (generally referred to as spins A and B). Figure 2 2 A ) Four pulse DEER. Each pulse delay labeled with remains constant while spacing labeled with T is incremented. B) Pump and observe frequencies for nitroxide labels at X band.

PAGE 60

60 Applying a pulse flips the B spins, which perturbs the coupled A spins r esul ting in a modulation on the A spins. The sequence begins with a two pulse Hahn echo sequence on mw A After the appearance of the Hahn echo, a pump pulse is applied on mw B with a varying time delay after the echo. At 2 the echo is refocused by an ad ditional pulse on mw A Again, the echo intensity is recorded as a function of the time delay between the first echo and the pump pulse. T he pump pulse flip s the B spins at time T which alters the effective magnetic field of the A spins that are coupl ed to the B spins. This change in the magnetic field changes the precession frequency of the coupled spins via electron electron coupling ( ee ), which results in the magnetization being out of phase by the angle ee = ee T Thus, ee can be determined by integrating the echo intensity as a function of T Equation 2 1 defines ee (2 1) where r AB is the distance between the spi ns, AB is the angle subtended by the static field B o and the vector between the spins, J is the exchange coupling, and dd is the dipolar coupling between the electrons. Equation 2 1 is valid as long as the positions of the electron spins are relatively well defined in relation to the distance between them, such that the point dipole approximation holds 154 This restrict ion is easily met for spins more than 15 apart, which is the lower limit for a DEER experiment. The J coupling is significant only at short distances and considered negligible for distances greater than 20 111 In four pulse DEER, mw A is the obse rve frequency (corresponding to A spins) and mw B is the pump frequency (B spins). Both frequencies are chosen such that there is

PAGE 61

61 no overlap (or minimal overlap) between the excitation windows of the pulses and that the greatest number of spins is excited. For nitroxide labels at X band the low and center field transitions are ~26 G apart (which corresponds to ~72 MHz) as shown in Figure 2 2B. Typically, the pump frequency is chosen to correspond to the center field transition because it is the most populated region o f the spectrum. M eanwhile, the low field manifold, which is the second most populated region, is often selected as the observe frequency. However, any position along the powder spectrum can be chosen for either the pump or probe pulses as long as the frequencies are well separated, so as to prevent exciting A spins with B pulses and vice versa. For split ring or loop gap resonators, it could be advantageous to situate the pump pulse to the resonator dip and to increase the observed frequency so as to compensate for the l imited pumping power 155 Note that experiments performed at higher frequencies (i.e. Q band or W band) follow this same procedure 123, 124, 142 However, as the field strength increases, the anisotropy of the powder spectrum is altered. For nitroxide labels, at X band, the anisotropy of the hype rfine tensor dominates the shape of the powder spectrum; whereas at W band the g tensor anisotropy dominates. The frequency offsets chosen will be dictated by the frequency of the experiment. Additionally, these offsets and powder spectra will differ wh en metals are used as the spin label, as g tensor anisotropy may dominate even at X band 121, 124 The DEER method is intrinsically insensitive. The DEER pulse sequence utilizes frequencies that are selective to g iven orientations of spin labels in the sample. The motional rearrangement of the molecules during the time of the experiment. Because

PAGE 62

62 the pulses his is tantamount to a small f raction of the total sample being excited by the pulse s Furthermore, given the random p owder orientation of the sam ple, only a fraction of the protein molecule s in the sample will have the magnetic components of the nitroxi de radical that is oriented so they match the orientation selection of the A and B spins dictated by the pulse sequence fr equency offsets Inspection of the echo detected powder pattern spectrum in Figure 2 2 B reveals that majority of spins are neither A nor B spins (portions of spectrum not highlighted). Additionally, the practical concern of incomplete sample spin labeling could further reduce the signal intensity Although the selective pulses lead to inherently lower signals, one benefit of the frequ ency offsets is the ability to perform orientation selection for spin labels, which has advantages in spin labeling of nucleic acids and distance determination in metals 156 The total DEER signal detected arises from both the intramolecular distance of interest as well as contributions from signals of all other random intermolecular intera c tions. Given a frozen protein sample, each of the desired A spins may have an intra m olecular B spin for interrogation, but it will also be surrounded by other spins on neighboring proteins, some of which will also be B spins. These intermolecular interactions give rise to a background signal, which comes from a random distribution of large distances, and thus can usually be m odeled by an exponential decay 110 T he total signal, which is a combin ation of both intra and inter molecular distances takes the form of a damped oscillation, as shown in Figure 2 3. In this illustrat ion, t he raw dipolar evolution curve is shown as a solid gray line. This signal is usually designated as V (t). The background contribution is plotted as a dashed blue line and is represented by B (t).

PAGE 63

63 The background corrected signal, F (t) is plotted as a solid black line. The relationship between the signal, the background, and the background corrected signal is given by Equation 2 2 The decay time for the oscillations in F (t), decay and the m aximum dipolar evolution time, max are also shown in Figu re 2 3. The modulation depth is a correction factor that compensates for the incomplete excitation of all B spins by the pump pulse. (2 2) Figure 2 3 Sample dipolar evolution curve before (gray solid line) and after ( black solid line) applying the background subtraction function (blue dashed line). The red line is the regenerated echo curve from data analysis Preparing DEER Samples The most obvious choice to improve signal to noise ratio (SNR) in a DEER experiment is to increase sample concentration and to ensure optimum labeling efficiency. However, increasing the spin concentration does not always translate to greater sensitivity. One reason is that at higher concentration, the distance between neighboring molecule s is reduced, leading to increased intermolecular dipolar coupling contributions to the signal. Additionally, phase memory loss associated

PAGE 64

64 with instantaneous diffusion at high concentrations also limits sample concentrations 111 The optimal concentration ( c opt ) based on the instant aneous diffusion restriction is given by Equation 2 3, where r AB is the inter spin distance and N A constant 111 (2 3) Figure 2 4 shows the plots of concentration limits as a function of the inter spin distance for both the intermolecular distance restrictions (dashed line) and the instanta neous diffusion restr ictions (solid line). The intermolecular distance restriction assumes that the spin labels are on solvent exposed sites of a protein with 60 diameter. Figure 2 4. Plots of the maximum spin concentration as a function of the inter spin distance. Sol id line corresponds to the restriction imposed by instantaneous diffusion, while the dashed line corresponds to the restriction imposed by intermolecular distances. Figure adapted from Blackburn. 157 Another experimental consideration in DEER experiments that impacts the SNR is the chosen dipolar evolution time ( max ) in the pulse sequence. In practice, longer dipolar evolution times may be required to resolve the detail ed shape of distance distributions and to accurately measure longer distances. A way to increase both

PAGE 65

65 the SNR and the longest distances interrogated is to prolong the phase memory time, T m of the sys tem, which is usually accomplished by replacing neighb oring nuclei with nuclei that have smaller magnetic moments such as deuterium that minimizes the effect of nuclear spin diffusion on T m Nuclear spin diffusion results from the flipping of nuclear spins, which are coupled to the electron spins. Deuterium replacement can be accomplished by using deuterated solvent (i.e., D 2 O) and deuterated co solute 158 and/or deuterating the protein 159 Eaton and co workers 160 demonstrated that r eplace ment of solvent protons with deuterons or removal of solvent protons ( i.e. salts containing methyl groups) in a samp le could extend the phase memory time ( T m ) by up to three fold for solvent exposed spin labeling sites on carbonic anhydrase Because DEER experiments are performed at cryogenic temperatures (<100K), cryoprotectants are required to prevent ice crystal f ormation and protein aggregation where protein aggregation and crystallization leads to dramatic decreases in T m thus compromising SNR. Co solutes that act as glassing agents are often utilized to over come these problems. A glassing agent reduces t he glass transition temperature so that the sample remains in a glass state throughout the experiment, which prevents protein aggregation during sample freezing. Glycerol is by far the most common glassing agent and cryoprotectant used, although many other c o solutes can be used, including sucrose, Ficoll polymers, ethylene glycol and polyethylene glycol (PEG). However, the effect of the solute to protein conformational equilibria or nitr oxide orientation and motion, has to be understood 161, 162 For instance, perturbations in the X band EPR line shape for common nitroxide labels have been observed in HIV 1 protease when PEG3000 or glycerol are added,

PAGE 66

66 due to changes in preferential interactions with the co solute tha t alter spin label mobility. However, adding up to 40% glycerol did not alt er the flap conformations o f HIV 1 PR based on pulsed EPR distance measurements. 93 In another study, rapid freezing of T4 lysozyme samples reveal that lower concentration of glycerol ( 10%) preserve the protein properties and at the same time enable acquis ition of high quality DEER signals. 150 On the other hand, addition of osmolytes has been shown to modulate conformational change in a membrane protein 161 The e xclusion of the co solute from the protein surface requires energy that is proportional to the surface area. Consequently, the energy is minimize d when there is a decrease in surface area thereby favoring the most compact form of the protein. For insta nce, sucrose was shown to increase the surface tension of water and has a stabilizing effect on protein structure 163 Although surface tension may play a role in the effect of solutes on the protein, it is not a reliable predictor of the effect. For example, glycerol reduces the surface tension but also stabilizes protein structure 164 In fact, the stabilizing effect of glycerol is attributed to the mechanism of preferential hydration. Addition of proteins to a glycerol water mixture increases the chemical potential of glycerol resulting in its unfavorable interaction with the protein surface, which has a stabilizing effect to the native structure of globular proteins. Reduction of aqueous chemical potential results in a decrease of water activity and an increase in the osmotic pressure 165 The availability of bulk water in a protein solute solution may affect the balance of protein water s olute interactions, which can lead to preferential hydration. Finally, solutes can also alter solution viscosity, which has an effect on the translational and rotational diffusion of proteins 166, 167 Note that

PAGE 67

67 co solutes are seldom used in DEER sample preparation of membrane proteins since the membranes themselves appear to lower the glass temperature of the system. DEER Experimental Considerations In an EPR spectrometer, the resonator (or cavity) is responsib le for converting the microwave power into the B 1 field necessary for flipping the spins. The commercially available dielectric and split ring resonators are currently the most popular. The dielectric resonators, EN4118X MD4/EN4118X MD5 from Bruker Biospi n, offer a larg e filling factor and variable Q or ratio of microwave power stored in the resonator to power lost via heat absorption, which provides a high degree of sensitivity and adaptability for various experiments. The split ring resonators, ER 4118X MS 2 /4118X MS3/4118X MS 5 also from Bruker Biospin, generate the highest B 1 fields and have the largest bandwidths among the commercially available resonators Both dielectric and split ring resonators are suitable for DEER experiment s, although each off ers distinct advantage s. The sample volume required for DEER vary depending on the resonator used 155 Generally, the SNR 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 EPR tube, this corresponds to 100 L of sample. The 2 mm split ring resonator, ER 4118X MS 2, is often utilized for membrane protein samples. An important experimental parameter in any pulsed magnetic resonance experiment is the time required for the magnetization to return to thermal equilibrium. Spin relaxation is characterized by two time constants: T 1 or spin lattice relaxation time,

PAGE 68

68 which is the relaxation of bulk magnetization along the z axis; and T 2 or spin spin relaxation time, which is the time constant for relaxation in the x y plane. Meanwhile, the spin echo dephasing time, T m comprises all processes that lead to the loss of electron spin phase coherence (which includes T 2 ). 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 the SRT is set to be 1.26* T 1 for optimal SNR. The effec tive T 1 can be measured via an inversion recovery ex periment, although saturation recovery may also be employed 168 The max of the DEER experiment is limite d by the T m which in turn is strongly dependent upon temperature. The ideal temperature is the one at which T m is dominated by the spin diffusion of nuclear spins, as opposed to the modulation of the hyperfine or g tensor resulting from the molecular reo rientation 114 For nitroxide radicals in aqueous buf fers, this point is usually at or below 80 K. For most DEER experiments, temperatures between 55 K and 80 K are suitable. Although the T m will continue to increase at lower temperature, the T 1 will also increase, which lengthens the necessary experiment time. Also note that liquid helium used to cool down the system is a limited and expensive resource. Typically, a 100 L helium Dewar could last at least a week if D EER experiments are done at 50 70 K. More helium is use d if lower temperature is needed Due to current costs and limitations in liquid helium, DEER experiments are also performed at 80 K with liquid nitrogen. Other pulsed EPR experiments, such as electron spin echo envelope modulation ( ESEEM ) can be performed at 80 K using liquid nitrogen 169 To circumvent the need for utilizing cryogen in pulsed EPR experiments, a cryogen free cryostat can be a strategic alternative. This system

PAGE 69

69 fr om Oxford Instruments cools the sample down to cryogenic temperatures using helium gas instead of liquid helium or nitrogen. DEER Data Acquisition The steps outlined in this section for DEER data acquisition summarizes the more detailed procedure from Thus, this section will focus on the crucial steps and troubleshooting procedures when problems arise. For more information, the reader is encouraged to refer to the Bruker First the liquid nitrogen for the temperature reference is set up by filling a Dewar with liquid nitrogen and immersing the refere nce the rmocouple into the Dewar. After that, the cavity is pumped down to ~10 6 Torr and continuous flow cryostat is set up at th e chosen temperature. The Bruker continuous flow cryostat (ER 4112HV for X band; ER 4118CF for Q band) and transfer line for liquid helium (HTL 438A B2 series from CRYO Industries) can work with both liquid helium and liquid nitrogen. After the magnet a nd console is turned on the EPR tube with protein sample is inserted into a suitable adapter, followed by flash freezing of the sample by immersing the tube into a liquid nitrogen bath for 1 min. T he EPR tube is adjusted such that the sample is situated within the active region of the resonator once the sample is loaded. When flash freezing, the EPR tube has to be tilted at an acute angle before immersing into liquid nitrogen to prevent breaking the tube. Alternative methods that provide for faster fre ezing rates exist. The typical method utilized is the slower nitrogen dunking method. However if faster freezing rates are required for trapping conformations, the liquid isopentane method is a suitable alternative 150 At this point, the valves leading to

PAGE 70

70 the helium pump has to be turned off After pressure goes down to zero, the sample tube holder is loosened and the EPR tube is immediately inserted into the resonator. The vacuum has to be turned off when the cavity is open because moist air will be sucked into the cold cavity where ice can form. Ice formation can freeze the EPR tube in place and could also interfere with the cryogen cooling system by obstructing gas flow. While waiting for the press ure to drop to zero, the EPR tube has to remain submerged in liquid nitrogen. By inserting the sample as quickly as possible, the introduction of moisture into the resonator is minimized. The next step is to s et up the spectrometer and to tune the cavi ty. After that, the defense p ulse has to be checked along with the cavity ring down The ring down may occur after the pulse and cover the echo signal. After the preliminary checks a field swept spectrum is collected (Figure 2 5A). The Hahn echo inten sity scales with the intensity of the CW line shape, so the frequency (or field) can be varied and the Hahn echo intensity can be used to generate a derivative spectrum, which will enable determining the center field. This spectrum can also be used for p ump and observe fre quency selection After coll ecting a field swept spectrum, the phase memory time ( T m ) of the sample has to be measured by performing an echo decay experiment (Figure 2 5B). T he curve is fitted to a single exponential decay function to d etermine T m Here, the exponential decay of the echo intensity is obtained as a function of time between pulses. The oscillations present in the earlier part of the curve are ESEEM modulations of the spin labels interacting with deuterium in the solvent.

PAGE 71

71 Figure 2 5. Sample preliminary experiments prior DEER experiment set up. A) Field swept spe ctrum to determine center field. B) Echo decay experiment to measure phase memory time, T m C) DEER echo acquisition to d etermine d0 and gate parameters. D) F ield swept ech o detected spectrum to identify pump and observe frequency positions. All data were collected at 65K. After measuring T m the DEER echo is acquired (Figure 2 5C) by using 8 step phase cycling to eradicate the stimulated echoes, then d0 and ga te parameters are determined The former is the time domain start position of the echo while the latter corresponds to the difference between the start and end positions (echo width). The integrated gate width influences SNR and has to be measured carefu lly. ELDOR = current B*2.83 MHz/G (2 4) The last preliminary experiment prior to DEER experimental set up is the acquisition of a field swept echo detected spectrum using two step phase cycling to

PAGE 72

72 determine pump and observe frequencies (Figu re 2 5D). T he current frequency center field and low field of the spectrum have to be noted T he ELDOR frequency ( ELDOR ) is calculated using Equation 2 4, where B is the difference between the center and low fields. Figure 2 6. The four pulse DEER sequence with the pulse spacings labeled according to Bruker Xepr software package nomenclature. In reference to the four pulse DEER sequence in Figure 2 2A d1 = 1 and d2 = 2 (or max ). The time increment parameter dx is equal to d2/N, where N is the number of real data points. The start position of the echo signal is designated as d0 while d3 is a critical delay parameter that prevents overlapping of the second observe pulse and pump pulse. Once the four preliminary experiments have been completed, th e DEER experi ment can now be set up Figure 2 6 shows the four pulse DEER sequence, with pulse spacings labeled using the Bruker Xepr software p ackage nomenclature. T he DEER dipolar m odulation curve is acquired using two step phase cycling with sufficie nt number of scans to generate a n SNR of at lea st 15. SNR can be measured by obtaining the ratio of the modulation depth to that of peak to peak noise in the last 0.5 s of the dipolar evolution curve. 34 Analysis of DEER Data For a complete understanding of how to properly analyze data from DEER experiments, the reader is directed to a number of excellent sources 109, 114, 170, 171

PAGE 73

73 including the DeerAnalysis2008 www.epr.ethz.ch ) if data analysis using this software is desired In this section, the process of converting a dipolar evolution curve into a distance profile will be outlined. The dipolar modulation curve is t he manifestation of additi onal modulation imposed upon A spins by their coupling with the B spins. The frequency of this additional modulation can be determined in a variety of ways. The simplest method is to obtain the Fourier transform of the dipolar modulation curve so as to ge nerate a Pake pattern in the frequency domain, where the splitting between the singularities is proportional to 1/ r 3 where r is the inter spin distance (Figure 2 7, Method #1) T he Pake pattern needs to be simulated to obtain the distance profiles. A com mon approach to obtain distance profile s, particularly well suited for noisy data sets, is the use of curve fitting to optimize the solution (Figure 2 7, Method #2) M any variations of this method exist but all include a modeling of the distance profile based on the current information on the system and generating the correspon ding theoretical dipolar evolution curve for comparison with the experimental data. The process involves changing the distance profile to optimize the fit between the theoretical a nd experimental dipolar evolution curves. One variation of this approach utilizes Monte Carlo (MC) methods to generate a distance profile with an assumed form, such as Gaussian or Lorentzian shape by combining random function s 172 The third method, Tikhonov regularization (TKR) 170 is a mathematical method used for solving an ill posed problem by introducing a penalty for smoothness (Figure 2 7, Method #3) TKR uses the function in Eq uation 2 5 ( 2 5)

PAGE 74

74 Figure 2 7. Schematic diagram of three methods utilized for obtaining distance information from the background subtracted dipolar evolution curve. TKR balance s the quality of fit to the experimental data (first term) with the smoothness of the solution (second term) by varying the magnitude of the regularization parameter ( ), where P is the probability distribution of the inter spin distance, K is the operator that maps the function P onto the experimental data vector S and L is usually a second derivative operator. The visual result of the TKR process is a plot of log ( ) against log ( ). Equations 2 6 and 2 7 give the corresponding L curve equations. (2 6 ) (2 7 ) Figure 2 8 illustrates the effect of non optimal values to the distance profile and the TKR fit. T he dist ance profile in Figure 2 8 B is under smoothed and corresponds to an value that is too low. The corresponding dipo lar evolution curve in Figure 2 8 C is

PAGE 75

75 over fit such that the theoretical dipolar evolution curve was fit to some of the noise. The distance profile in Figure 2 8 D is qualitatively similar to the distance profile in B in that the most probable distance is within the breadth of the peak in D. However the distance profile in D is smoother and corresponds to the optimal value and dipolar evol ution curve in Figure 2 8 E. The distance profile in Figure 2 8 F is over smoothed and thus overly broad. The corresponding dipolar evolution curve in Figure 2 8 G is under fit, such that some of the oscillations in the signal are neglected in the TKR fit. In the process of converting dipolar evolution curves to distance profiles (Figure 2 9) DeerAnalysis software uses a combination of shell factorization to simulate the dipolar evolution curves and Tikhonov regularizat ion (TKR) to optimize the solution. This software contains a variety of data analysis tools, such as approximat e Pake transformation (APT) model fitting to a Gaussian or other user defined functions and options for background corrections. Figure 2 8. Example of an L curve (A) and the cor responding distance profiles and dipolar modulation curves for low (B and C), optimal (D and E) and high (F and G) regularization parameters ( ).

PAGE 76

76 In the analysis of DEER spectra, DeerAnalysis2008 is executed in MATLAB environ ment, and the DEER spectrum is loaded T he appropriate zero time in the raw dipolar modulation curve is selected Data collected from 0 to max actually start at small negative time so that the data can be corrected for any discrepancy between the instr umental and actual zero time. Because the echo intensity data is set to start collecting before the top of the curve, the true zero time must be selected after data acquisition. A Gaussian function (Equation 2 8) can be fitted using a plotting software such as Origin 8.5 to data poin ts from 300 to 300 ns. The center of the curve is assigned as the zero point. ( 2 8) I nitial background subtraction is performed using an exponential function corres pon ding to a t hree dimensional homogeneous background, and digital long pass filter is applied to the dipolar evolution curve. The initial attempt is not necessarily the corr ect level of background subtraction. Digital long is applied to the dipolar evolutio n curve prior to extracting the distance profile so as to remove artifacts attributed to high frequency noise and nuclear modulations 171 In DeerAnalysis, Tikhonov regularization is selected and the L curve is generated. T he distance distribution profile that corresponds to the regularization parameter situate d at the corner of the L curve (optimum ) is selected The file that stores L curve regularization parameters ( ) can be altered to decrease increments, which confers more resolution around the L curve corner, enabling more accurate determination of optimal for longer distances and broader distance distribution profiles.

PAGE 77

77 Figure 2 9. Self consistent procedure for determining the appropriate level of back ground subtraction. After saving the selected distance prof ile DeerSim is opened in MATLAB environment In DeerSim, t he TKR distance profile is regenerated using a linear combination of Gaussian shaped functions with definite center positions, full width at half maxima (FWHM) and relative population percentages. The initial guesses for these populations can be determined from software analysis such as those available in Origin8.5. T he Gaussian regenerated dipolar modulation curve and the original echo curve with an initial background subtraction can be visually compared in Origin 8.5 If t he two curves do not overlay, the process is repeated by using a new background subtraction level for each attempt until the two echo curves are perfectly overlaid. At this point, the level of background subtraction has been verified in a self consistent manner and the

PAGE 78

78 distance profile has been deconstructed to the corresponding Gaussian populations (Figure 2 9). For HIV 1 protease, these Gaussian populations correspond to various protein conformations that have been modeled by combining results from DEE R, X ray and molecular dynamic (MD) simulations. The meaning of the Gaussian populations will vary from system to system. The original use for the Gaussian reconstruction was a simple mathematical way to generate a theoretical echo curve in a MATLAB envir on ment for comparison to the TKR distance profile. This method was developed as a self consistent method for background subtraction. Critical DEER Parameters and Troubleshooting As illustrated in Figure 2 10, when considering only intramolecular distanc e for F (t), the frequency and decay rate of the oscillations depend on the length of the most probable distance and the breadth of the distance distribution, respectively. By varying the breadth of a distance profile centered at 36 (Figure 2 10 A) from 1 to 10 and generating the theoretical di polar evolution curves (Figure 2 10 B), the narrowest distributions are found to have the most well defined oscillations, corresponding to the longest decay rates. The frequency of oscillations can likewise be il lustrated by comparing the dipolar evolution curves (Figure 2 10 D) corresponding to distance profiles (Figure 2 10 C) that have the same breadth (7 ) and vary in the most probable distance from 18 to 78 These dipolar evolution curves have different dec ay rates, but because the frequency of the oscillations also changes, the curves have the same number of oscillations before being completel y damped. The inset in Figure 2 10 D highlights the dipolar evolution curves that decay within the first 2 s, a com mon data acquisition parameter reported in

PAGE 79

79 literature. Curves with solid lines correspond to 36 The dipolar evolution curves corresponding to center distances larger than 36 are plotted as dashed lines The distinction is made to show whi ch distances of this breadth have echo curves containing two full oscillations, which is required for accurate distance and distance breadth determination within 2 s 111 Figure 2 10. Effect of the breadth and most probable distance on the dipolar modulation curves. A) Varying FWH M of distance profiles centered at 36 B) The corresponding theoretical dipolar evolution curves with different decay rates. C) Profiles with different center distances but same FWHM. D) The corresponding theoretical dipolar evolution curves with differe nt oscillation frequencies. Inset highlights dipolar modulations that decay within the first 2 s. Figure adapted from Blackburn. 157 A practical consideration when set ting up a DEER experiment is to properly choose the data acquisition time, 2 or max for the longest distance under consideration. Unfortunately, one of the difficulties with detecting longer distances with DEER is the increased pulse length time, where the overall signal intensity is dictated by the phase memory time, T m of the system. As max increases, the spacing between the Hahn

PAGE 80

80 echo pulses increases and the echo intensity decreases as dictated by spin relaxation. Consequently, data acquisition tim e should be increased so as to attain the desired SNR. Data has been collected on protein samples for distances greater than 60 ; however, the acquisition times are typically long and SNR are lowered. 111, 171, 17 3 Thus, in this situation, experiments performed at Q band frequencies is beneficial, 142 as well as using deuterated protein s or matrices. 158, 159 Figure 2 11 demonstrates the effects of the chosen value of max on proper rege neration of the breadths of DEER distance distributions. A distance of 48 is chosen for this demonstration. F igure 2 11A shows Gaussian shaped distance distribution profiles centered at 48 with FWHM of 1 and 7. The theoretical dipolar evolution curves generated are shown in Figure 2 11B. Given the theoretical echo curves do not contain any noise, the nois e generator in Origin 8.0 was utilized to introduce white noise to an SNR of 40, a value that is twice of that obtained in HIV 1 protease experiments. The data sets were then truncated to max of 2 s (Figure 2 11C). Tikhonov Regularization (TKR) methods in DeerAnalysis were employed to generate distance profiles, shown in Figure 2 11 D. The most probable distance for the curves is the same (48.00 0.01 ), and the broader signal is accurately regenerated with a breadth of 7.0 0.1. However, for max o f 2 s, the breadth of the 1 FWHM profile was broadened to 6.8 If instead the echo curves are truncated to max values of 3 4 s, the breath of the original distance distri bution profiles are accurately regenerated. This exer cise demonstrates that t he accu racy of the breadth for the distance profile strongly relies on a sufficiently long max when collecting the dipolar evolution curve.

PAGE 81

81 Figure 2 11. The influence of max length to the corresponding distance profile. A) Gaussian distance profile s with most probable distance of 48 with FWHM of 1 (solid) and 7 (dashed). B) The corresponding theoretical dipolar evolution curves. Solid vertical line corresponds to max of 2 s C) The same dipolar evolution curves in (B) but max is shortened to 2 s. D) The corresponding TKR distance profiles for (C) analyzed using DeerAnalysis software. Figure adapted from Blackburn. 157 Figure 2 12. DEER data for s ubty pe B HIV 1 protease acquired at variable max (A) Background subtracted dipolar evolution curves and the corresponding TKR fits (gray solid line); (B) The corresponding TKR distance profiles. The vertical dashed line marks the position for a minor popul ation centered at 40 Data are vertically offset for clarity. Figure adapted from Blackburn. 157

PAGE 82

82 To further illustrate the importance of collecting data at a sufficie ntly long max experimental dipolar evolution curves are obtained for subtype B HIV 1 protease in deuterated matrix (Figure 2 12A), collected as a function of or length of the dipolar evolution curve, and the corresponding distance profiles (Figure 2 12 B) are derived. These data were collected sequentially without sample thawing to ensure reproducibi lity of data acquisition. Dipolar evolution curves were analyzed with DeerAnalysis in a manner such that the same background subtraction level was used. Data analysis reveals that t he most probable distance is independent of max However, a minor population centered at changes in intensity depending on the max used. Convergence of the shape of the distance profile resulted for max > 2.5 s suggesting that long acquisition times are necessary when information about the shape of the distance profile is required. When only the most probable distance is of interest, shorter values of max can be emp loyed, resulting in stronger signals and faster acquisition times In addition to choosing a long enough 2 such as for long distance measurements or accurate distance profile shape, long acquisition times are also needed for proper background subtraction. B ecause the acquired signal, V (t), contains contributions from both the intramolecular signal F (t) and the background signal B (t), data needs to be collected for long enough times ( max > decay ) for accurate separation of B (t) from V (t ). Typically B (t) can be modeled by an exponential fu nction of the form in Equation 2 9 (2 9) where D is the dimensionality of the background, typically three dimensions for soluble proteins and ln [ B (t)] can be described as a simple low order polynomial. If V (t) is co ted

PAGE 83

83 such that 2 is too short for the oscillations to be dampened, it becomes difficult to determine the proper background subtraction. W hen only a single distance, such as the most probable distance, is of interest, small errors in the background subtraction do not significantly alter the value of the most probable dis tance obtained. However, when information from the shape of the distance profile or the presence of minor populations is desired, improper background subtraction can inadvertently affect the distance profiles obtained. Note that the background subtractio n for membrane protein samples can be troublesome. Given that the proteins are contained within the lipid leaflets of vesicles, which can be described as a 2D sheet, the 3D background may not be an appropriate model. Additionally, the concentration enhan cement of integral proteins in lipid bilayers may completely abolish the oscillations in the raw data. To overcome this problem, novel membrane mimetic system of nanolipoprotein phospholipid bilayers, most commonly referred to as nanodiscs, have been pione ered and shown to be successful 174 The distance limitations for DEER experiments are frequently cited as being from 15 to 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 15 T he upper limit, on the other hand, is limited by the phase memory time ( T m ), the inherent sensitivity of the system, sample concentration and contribution from intermolecular interactions For biological molecules, the solvent is typically restricted to a queous solutions. Moreover, the biological molecule typically contains methyl protons, which contribute to shorter T m These restrictions typically limit the T m

PAGE 84

84 to 5 s, al though 3 4 s is more common. These T m values set the time limit on the spacing between pulses, which corresponds to the evolution time in the echo curve. As shown above, this impacts the ability to accurately determine distances and breadths of distance profiles. In practice, T m values of 3 4 s correspond to the ability to measure distances of 60 as a practical limit. To increase the upper limit for distance measurements, T m can be prolonged by deuteron replacement of protons in the sample. This can be accomplished by replacing neighboring nuclei, usually from solvent and cryo protectant with deuterium or by deuterating the protein 159 D istance measurements up to 70 was achieved on the histone core particle by using deuterated solvents and 50% deuterated glycerol as a glassing agent 158 Figure 2 13 shows the T m cu rves in echo decay experiments for HIV 1 protease in protonated and deuterated matrices, revealing the dependence of T m upon the solvent Compared to other matrices used, T m is longest when the protein is in a matrix of D 2 O and 30% deuterated glycerol. L onger T m enables longer distance measurements and improved SNR of the corresponding dipolar modulation curves; hence faster data acquisition. Unfortunately, for buried spin labels without access to the solvent, also the case for some membrane proteins, s olvent deuteration does not always lead to increases in T m 160, 175 Other methods to overcome limitations of T m include alternative pulse sequences, such as variable time DEER 171 5 pulse DEER 176 and 6 pulse DQC ESR. 113 Although variable time DEER confers increased sensitivity in most cases, it is not recommended for membrane proteins rec onstituted in liposomes because of inhomogeneous spatial distribution in three dimensions and dominant instantaneous diffusion mechanism. 105

PAGE 85

85 Aside from variable time DEER, a 5 pulse version based on the 4 pulse DEER that provides longer acquisition time has been introduced. This sequence provides maximum suppression of spin diffusion and an extra pump pulse that allows up to twice as much dipolar evolution time and thus, longer distance measurement. 176 DQC, another pulsed EPR technique may also be employed. Partial suppression of the nuclear spin diffusion effects on electron spin T m has been achieved using a 6 pulse variant of DQC, also known as double quantum filtered refocused electron spin echoes (DQFR ESE), allowing accurate distance measurements up to 70 113 Figure 2 13. Intensity normalized T m curves and corresponding exponential decay fits for MTSL labeled HIV 1 protease in 2 mM NaOAc buffer pH 5.0 with H 2 O and 30% glycerol (dotted line), 2 mM NaOAc buffer pH 5.0 with H 2 O and 30% deuterated glycerol (dashed line), 2 mM NaOAc buffer pH 5.0 with D 2 O and 30% deuterated glycerol (solid line). The oscillations in the deuterated sol vents originate from ESEEM e ffects between the deuterons and spin labels. The vertical dashed line marks = 3 s. Figure adapted from Blackburn. 157 Attempts have also been made to measure longer distances by going to higher frequencies such as Q and W bands. The enhanced sensitivity observed at higher frequency stems from an increase in Boltzmann population difference between the spin states at the higher Zeeman energy, which contributes to sign al improvement and increased sensitivity of the resonator at the higher frequency 177 I n addition, distance

PAGE 86

86 range increase for a membrane protein was reported at Q band by employing discoidal nanoscale lipoprotein bound bilayers 174 U tilization of Gd 3+ markers can be advantage ous due to the high spin quantum number for these nuclei, enabling distance measurements in the 60 100 range at Ka and W band frequencies 176 Also note that several model systems, such as rod like shape persistent rigid biradicals, allowed distance measurements from 50 to 75 171 Before acquiring any DEER data, it is important to make sure that the protein samples are homogeneous, f ree of contaminants, and properly spin labeled. Given that non labeled protein contaminants do not directly affect the DEER signal, oftentimes these aspects of sample purity are overlooked. However, in cases where sample homogeneity is important, all ste ps must be taken to ensure sample quality. Typically, combining size exclusion chromatography with analytical high pressure liquid chromatography (HPLC) is a suitable means to demonstrate a homogeneous protein sample. The spin labeled protein samples can be analyzed by electrospray ionization time of flight mass spectrometry (ESI TOF MS) to validate homogeneous spin labeling If necessary, tandem mass spectrometry (MS MS) methods can also be utilized to confir m the location of the spin label, such as HPL C ESI ion trap MS n and matrix assisted laser desorption ionization (MALDI) TOF tandem MS of the protein tryptic digest. Results from TKR analysis often generate distance profiles with more than one peak. Questions arise as to the significance of smalle r peaks. The Gaussian self consist ent method will oftentimes regenerate these smaller signals. Depending on the SNR of the raw dipolar modulation curve, minor peaks with relative population

PAGE 87

87 percentage <5% may or may not be real These peaks may be omitte d or suppressed when regenerating the dipolar evolution curve. If the model curve is statistically comparable to the experimental curve by 2 analysis (Equation s 2 10 and 2 11) 178 then the suppresse d population is an artifact at the confidence level used, usually set to 95% 2 analysis uses Equation 2 10 where a data point in the TKR curve ( y i TKR ) is assumed to be the expected value ( y i exp ). The calculated model echo curve which corresponds to one or more suppressed Gaussian populations, is represented by y i cal The noise parameter, y is given in Equation 2 11 where y i echo is a point in the echo curve y i TKR is a data point in the TKR fit, and n is the number of data points employed. (2 10 ) (2 11) Anticip ated Results Figure 2 1 4 shows a sample DEER result for HIV 1 protease. The Tikhonov regularization (TKR) dis tance distribution profiles are obtained from the background subtracted and long pass filtered dipolar modulation curves using DeerAnalysis2008 T he optimum regularization parameter, selected right at the corner of the L curve, was used for TKR. Gaussian shaped functions were used to regenerate the TKR distance profile. Populations with relative percentage of 5% were validated by suppress ing them individually or in combination and regenerating a dipolar modulation model. If the model curve is comparable to the original dipolar evolution curve using a 2 criteria, then the questionable minor populations are regarded as artifacts of noise. The min or peaks

PAGE 88

88 Figure 2 14. DEER data processing for HIV 1 protease. A) Determination of zero time (x c ) by fitting a Gaussian function to the 300 to 30 0 ns region of the echo curve. B) Raw dipolar echo curve and the expone ntial decay function (red line) corresponding to a homogeneous three dimensional distribution that is employed for background s ubtraction. C) Long pass filtered and background subtracted dipolar modulation curve with Tikhonov regu larization (TKR) fit (red line) overlaid with Gaussian re constructed dipolar modulation (blue line). D) The L curves derived from TKR analysis that helps determine the optimal regularization parameter ( ). E) TKR distance profile overlaid with the summation of Gaussian populations (red dashed line). Peaks labe led with asterisk indicate populations 5% that are not statistically significant at 95% confidence level based on the 2 criteria. F) The Gaussian populations employed to regene rate the TKR distance profile. G) The Pake dipolar pattern that results from the Fourier transformation of the background subtracted dipolar modulation curve and the correspon ding fit (red dashed line). H) Table of values for the most probable or center distances, full width at half maxima (FWHM), and relative percentage of conf ormational populations (Pop. %) used for generating the Gaussian reconstructed distance profile in E (red dashed line).

PAGE 89

89 in Figure 2 1 4E that are not statistically signifi cant at the 95% confidence level are marked with an asterisk Time Considerations D EER data acquisition at 65 K using an X band spectrometer, on average, takes 4 to 16 hours for a homogeneously spin labeled sample and 2 days for a poorly labeled sample For longer distance measurements and narrower distance profiles, dipolar evolution t ime needs to be prolonged and acquisition time will increase. Note that the maximum dipolar evolution time ( max ) should be chosen by the anticipated distance and distance distribution profile, but its maximum value is limited by the phase memory time ( T m ) of the sample. Thus, for longer distances, using a deuterated matrix or protein deuteration may be necessary for longer T m and higher SNR. Data acquisition on higher field spectrometers significantly lower the acquisition time and extend the possible co ncentration range for DEER investigations.

PAGE 90

90 CHAPTER 3 UNRAVELING THE RELAT IONSHIP BETWEEN CONF ORMATIONAL SAMPLING AND DRUG RESISTANCE IN HIV 1 PROTEASE Introduction The i nhibition of enzymes through small molecules that compete with a substrate for the ac tive site is a common clinical method for effective treatment of a disease. However, drug resistance develops in proliferating cells or pathogenic organisms through selective pressure, where the incorporation of random genetic mutations generates an enzym e with amino acid substitutions that render s drug molecule s less effective 70, 179 The emergence of primary mutation s often results in a change in an amino acid whose structure interacts less favorably with the inh ibitor because of steric hindrance or remov es essential molecular interactions such as charge stabilization or van der Waals contacts 70, 179 In the case of competitive inhibition, these primary mutations also ten d to alter the interactions of the enzyme with the substrate or product, thus negatively impacting enzyme efficiency and compromising fitness. The observed pattern of continued evolutionary mutations shows that secondary (or compensatory) mutations recove r fitness, while maintaining drug resistance 94 An understanding of the molecular mechanism by which the accumulation of mutation s impart these effects is important for the rational design of future generati ons of drugs. 78 In some enzymes, a clear rationale is illuminated through structural changes induced b y the pattern of mutations while in others indirect effects such as changes in enzyme dynamics or protein ligand dynamics are evoked 179 In this chapter an indi how shifts in equilibrium conformational sampling via drug pressu re accumulated mutations can alter enzyme kinetics and inhibitor susceptibility S ome proteins sample multiple conformations, where interaction with a

PAGE 91

91 ligand or an inhibitor simply shifts the population to an already accessible state 180 182 Specifically this chapter for drug resistance in HIV 1 protease (HIV 1 PR). HIV 1 PR an aspartic protease that processes the G ag and Gag P ol viral poly peptides, is an attractive target for AIDS antiviral therapy 1 because of its central role in viral maturation 183 Protease inhibitors (PIs) that target HIV 1 PR prevent the formation of infectious virions by block ing viral replication. PIs bind in the protease active site where the two flexible hairpin turns (aka the flaps ) are folded over the inhibitor, giving a stable complex that prevents substrate processing 184 The efficacy of currently available PIs is time limited by the rapid emergence o f mutations in HIV 1 PR where changes in at least 3 8 out of 99 amino acid residues occur under the selective pressure of PI therapy, leading to lowered drug susceptibility. 95 In HIV 1 PR, structural evidence clearly explains the effects of primary mutations that occur within the active site pocket. 71 Secondary substitutions, on the other hand, appear at distal positions and function to compensate for the viral replication impairment due to the primary m utation 70 or thr ough natural polymorphisms prior to PI exposure 76 These secondary polymorphisms are found to influence inhibitor binding but are typically not located in regions of the protein that make physical contact with the PIs 76, 78 80 The mechanisms by which distal mutations transmit their effects to the active site pocket and confer drug resistance are unclear, but they have been implicated in altering protein flexibility 85 through the hydrophobic sliding mechanism 81 or in restoring protein stability. 83 Recently, the transition states of native and drug resistant HIV 1 PR we re shown to be identical 91 Because both primary and secondary

PAGE 92

92 mutations are present in multidrug resistant HIV 1 PR variants previously analyzed by double electron ele ctron resonance ( DEER ) spectroscopy, 35, 52 where flap conformation and flexibility are altered, these mutations are hypothesized to individually and in com binations modify the HIV 1 PR flap conformational sampling. Site directed spin labeling (SDSL) DEER spectroscopy is a pulsed electron paramagnetic resonance (EPR) technique that measures the strength of the dipole interaction between unpaired electrons 111, 112, 114 This method has been applied in dis tance measurements to study conformational chan ges in biomolecules 119, 152, 153, 185 187 SDSL DEER have been previously utilized to monitor flap conformational sampling in HIV 1 PR 34, 35, 51, 52, 93, 153, 188, 189 Because HIV 1 PR is a homo dimer, labeling an amino acid residue in the polypeptide incorporates two spin labels into the holoenzyme Specifically, a site specific cysteine residu e is incorporated into the aqueous solvent active spin probe 153 (Figure 3 1B), wher e inter spin distance in the 20 60 range can readily be studied by DEER 104, 114 Previous DEER studies on HIV 1 PR have shown sampling of flap conformers consistent with conformational ensembles described as closed, semi open, curled/tucked, and wide open 34, 35 52, 188 Both X ray diffraction and molecular dynamics (MD) simulation models 46, 48, 50, 54, 190, 191 were used when assigning the conformational populations observed in DEER distance profiles. For PI nave pro tein sequence s ( i.e., from patients infected with HIV 1 that have not taken any PI therapy ), MD simulations show a predominant semi open flap confor mation for the HIV 1 PR apo enzyme with only a small percentage of the conformers found in either closed or wide open state 46, 48, 49 Ligand bound protein is shown to

PAGE 93

93 adopt the closed state consistent with enzyme inhibitor crystal complexes 78, 192 194 Additional MD studies of apo protein reveal a curled/tucked conformation of the flaps, which is proposed to be a conformational trigger for flap opening. 49, 54 X ray structures have also revealed curled flap conformations. 50 In these conformations, the flap tips curl in towards the active site, and depending on the study may or may not limit subs trate and inhibitor access to the binding pocket, revealing a high degree of variability in this conformation. B ased upon these studies and analysis presented here the curled/ tucked state detect ed in DEER distance profiles of HIV 1 PR is hypothesized to be a state whereby inhibitor can escape from the pocket. conformers where inhibitors ar e likely able to escape from the binding cleft. Thus, both the wide open and curled/tucked conformers are collectively referred state. Both DEER studies and MD simulations of two drug resistant variants MDR769 and V6 HIV 1 PR, have sh own that the conformational sampling and average structure of the flaps are altered relative to the native enzyme 35 Additional experiments have reported HIV 1 PR conformat ional samplin g shifts due to subtype polymor phisms 52 The effects of individ ual and combined amino acid substitutions as well as the addition of inhibitor or substrate mimic, on flap conformation and flexibility are described herein In particular, SDSL DEER is utilized to understand the effects of the acc umulation of primary, D30N, and secondary mutations M36I and A71V on the flap conformational sampling of subtype B HIV 1 PR D30N occurs specifically in response to nelfinavir treatment 70, 94 whereas M36I and A71 V, along with other non active site substitutions,

PAGE 94

94 appear as a result of selective pressure of treatments using various protease inhibitors 70, 94, 95 The location s of these sites in HIV 1 PR are shown in Figure 3 1 A Figure 3 1. A) Ribbon diagram of HIV 1 PR (PDBID: 1HHP) rendered using PyMol modeling software. The spin probes (K55R1) are incorporated in silico via MMM 2011.1 and shown as capped sticks. Mutation sites D30, M36 and A71 are shown as red, gree n and blue spheres; respectively. B) The sulfhydryl specific spin labeling reaction yielding the disulfide linked MTSL spin label at K55C, referred to as K55R1 after labeling. The effects that these combined mutations have upon the enzymatic parameters ( k cat K m k cat / K m ) had been investigated previously 76 and serve as the basis for the correlation studies. DEER data analyses show that secondary mutations alter protein

PAGE 95

95 conformational sampling profiles. By correl ating the relative percentages of various conformational states with enzyme kinetic parameters and inhibition constants, drug resistance is found to emerge when mutations combine to stabilize the open like states at the expense of the closed state. In th is chapter a direct link is shown between equilibrium conformational sampling to enzyme kinetic and inhibition parameters in HIV 1 PR, and forms the basis of a hypothesis for a possible mechanism of how secon dary mutations combine to elici t drug resistanc e. Namely, secondary mutations combine to shift conformational sampling state, while retaining a sufficiently high population of the semi open conformation for the enzyme to maintain viral fitness. The c onformational shift s that result after the addition of an inhibitor or substrate mimic to HIV 1 PR variants wi th single and combined amino acid substitutions are also described herein. Moreover, a possible relationship bet ween inhibitor induced conformational shift and drug resistance is investigated Materials and Methods Nomenclature for HIV 1 Protease Variants The seven subtype B HIV 1 protease (HIV 1 PR ) variants examined in this work; D30N, M36I, A71V, D30N/M36I, D30 N/A71V, M36I/A71V, and D30N/M36I/A71V are LAI variants with three stabilizing amino acid substitutions (Q7K, L33I, and L 63I) that retard auto proteolysis 97 To ensure that spin labeling is site specific, the two native cysteine residues were chan ged to alani ne (C67A and C95A). These five residue substitutions on HIV 1 PR have been utilized previously in NMR and X ray studies where the variants have been referred to as pentamutated protease ( PMPR). 38, 195 For the EPR investigations, Lys 55 on the solvent exposed side of the protease flaps was

PAGE 96

96 chosen as our reporter site fo r cysteine substitution and site directed spin la beling 153 with methanethiosulfonate (MTSL). The inactivating D25N mutation is also incorporated to abol ish autocatalytic self cleavage, thus aiding in sample handli ng during purification and spin labeling. Figure 3 2 shows the amino acid sequence information for the stabilized and inactive subtype B (B si ) HIV 1 PR proteins investi gated here. The B si const ruct is treated as wild type (WT) in this work. Figure 3 2. Protein sequence for subtype B HIV 1 PR (B si ) construct. The following residues different from the LAI consensus sequence are in red: stabilizing mutations (Q7K, L33I and L63I), the cysteine t o alanine substitutions, inactivating D25N mut ation, and the engineered labeling site (K55C) Cloning and Site directed Mutagenesis DNA that encodes E. coli codon optimized subtype B HIV 1 PR (DNA 2.0) was cloned into a pET 23a vector (Novagen Gibbstown NJ ) under the control of T7 promoter. Seven s tabilized (Q7K, L33I, L63I) and i nactive (D25N) constructs (B si ) with engineered labeling sites (K55C) were made using the QuikChange site directed muta genesis kit ( Stratagene ) : D30N, M36I, A71V, D30N/M36I, D30N/A71V, M36I/A71V, and triple mutant D30N/M36I/A71V Primers were designed using PrimerX ( http://www bioinformatics.org/primerx/) Note that this proced ure renders all mutations symmet rically applied to both subunits of th e homodimer. Moreover, natural cysteine residues (C6 7 and C95) in these constructs we re mutated to alanine to prevent non specific disulfide bridge formation and to ensure site specific labeling at C ys 55. The C67A and C95A mutation s have been utilized in numerous X ray crystallography studies and do

PAGE 97

97 not alter kinetic parameters, protein stability or dimer dissociation compared to the unmutated sequence 195, 196 The fidelity of HIV 1 PR DNA sequences w as confirmed by Sanger DNA sequencing (ICBR Genomics Faci lity, University of Florida ). Protein Expressi on, Purification and Spin labeling The HIV 1 PR was expressed into inclusion bodies using BL21*(DE3) pLysS E. coli cells (Invitrogen, Carlsbad, CA). Cells were gr own at 37 C with 250 rpm shaking in a 2.8 L Fernback flask containing 1 L of Luria Bertani (LB) media, supplemented with 100 g/mL of ampicillin and 25 g/mL of chloramphenicol After approx. 2 3 h of incubation at 37 C, p rotein expression was induced when the optical density at 600 nm (OD 600 ) is 1.0 by adding IPTG (isopropyl D thiogalactoside) to a final concentration of 1 mM Expression was al lowed to co ntinue at 37 C for 5 6 h to a final OD 600 of 1.5 1.7. Cells were pelleted via centrifugation at 7500 x g for 20 mins at 4 C, and then resuspended in 3 0 mL of 20 mM Tr is, 1 mM EDTA, 10 mM BME pH 7.5 buffer Cells were lysed via sonication for 2 minutes with on off cycle at 5 second intervals to prev ent sample heating. The cells we re lysed fur ther by three times pass a ge through a 35 mL French pressure cell (Thermo Scientific). The inclusion bodies and cellular debris underwent centrifugation at 18500 x g for 30 mins at 4 C. The inclusion body pellet was resuspended in 40 mL of wash buffer 1 (25 mM TrisHCl, 2.5 mM EDTA, 0.5 M NaCl, 1 mM diGly, 50 mM BME, pH 7.0). The inclusion bodies were sonicated as bef ore, homogenized using a 40 mL D ounce tissue homogenizer, and re pelleted via centrifugation. The inclusion bodies were subsequently washed with buffer 2 ( 25 mM TrisHCl, 2.5 mM EDTA, 0.5 M NaCl, 1 mM diGly, 50 mM BME,

PAGE 98

98 1 M u rea pH 7.0 ) and buffer 3 ( 25 mM TrisHCl, 1 mM EDTA, 1 mM diGly, 50 mM BME, pH 7.0). The inclusion bodies were solubilized in a resuspension buffer that contains 9 M urea, 25 mM TrisHCl, 5 mM NaCl, 1 mM EDTA, 1 mM diGly and 50 mM BME, ad jus ted to a 0.5 pH units lower than the pI of the corresponding HIV 1 PR construct to minimize protease binding to the anion exchange Q column. T he buffer pH used for wild type (WT) subtype B ( B si ) D30N, M36I, A71V, D30N/M36I, D30N/A71V, M36I/A71V, and D30N/M36I/A71V, respectively are as follows : 8.85, 9.00, 8.82, 8.80, 8.95, 8.98, 8.85, and 8.88. The solubilized inclusion bodies were passed through two 5 mL Q columns (Amersham Bio sciences) using the KTAprime TM liquid chromatography system The column flow through was acidified by the addition of formic acid to a final concentration of 25 mM and stored at 4C for 12 h to maximize precipitation of contaminating proteins. The precip itate was removed by centrifugatio n at 38500 x g for 30 min at 25 C The supernatant was added drop wise to 10 mM formic acid (pH ~ 3.0) solution on ice in a 10 fold dilution for optimal refolding of HIV 1 PR. The pr otein solution pH was adjusted to 3.8 by drop wise addition of 2.5 M sodium acetate (pH 5.5) and was allo wed to equilibrate to 30 C for an hour. Afterwards the pH was adjusted to 5.0 by adding 2.5 M sodium acetate (NaOAc), pH 5.5 until the protein precipitates. After 30 min the turbid prote in solution w as centrifuged at 18500 x g for 30 mins at 23 C to remove precipitated contaminants. The HI V 1 PR was concentrated until absorbance at 280 nm (A 280 ) is equal to 0.5 using an Amicon 8200 stirred cell with a polyethersulfone (PES) membrane (Mil lipore),

PAGE 99

99 and with molecular weight cut off of 10 kDa, and was buffer exchanged into 10 mM Tris buffer pH 6.9 using a 53 mL HiPrep desalting column (Amersham Bi osciences). The sample was con centrated back to A 280 of 0.5. Methanethiosulfonate (MTSL) spin la bel (Toronto Research Chemicals) previously dissolved in absolute ethanol, was added in 3 to 4 fold molar excess to 8 M HIV 1 PR homodimer and the reaction is allowed to proceed in the dark for 12 h at 25 C, with 150 rpm shaking For increased stabil ity and homogeneity, t he l abeled protein was desalted against 2 mM NaOAc buffer pH 5.0 and concentrated to A 280 > 1.0 using a stirred cell with a PES membrane. The sample was subsequently stored at 20C until ready for analysis Confirmation of Homogeneou s Spin labeling by ESI TOF MS Homogeneous spin labeling was verified via electrospray ionization time of flight mass spectrometry ( ESI TOF MS ). MTSL labeled subtype B variants with D30N, M36I and A71V amino acid substitutions reconstituted in 2 mM sodium acetate were concentrated to 50 M homodimer prior to ESI TOF MS on an Agilent 6210 instrument. MS characterization of spin labeled HIV 1 PR is described in more detail in Chapter 6. Secondary Structure Characterization by Circular Dic h r orism (CD) Spectr oscopy To ensure that the spin labeled protease had proper secondary structure in deuterated matrix buffer exchanged samples were diluted to 30 M, suitable for CD data collection 5 on an Aviv 400 spectr ometer. CD spectra were acquired by scanning at 1 nm steps from 200 to 250 nm at constant temperature (25 o C) and compar ed to a reference spectrum for s ubtype B HIV 1 PR in 2 mM NaOAc, pH 5.0.

PAGE 100

100 Sample Preparation and DEER Data Acquisition Protein samples were prepared as 100 M HIV 1 PR homodimer in 20 mM D 3 NaOAc/D 2 O, pH 5.0, 30% D 8 glycerol (Cambridge Isotope Laboratories). For samples with ligand, th e inhibitor or substrate mimic wa s added in 4 fold molar excess to the protein. To allow sufficient time for substrate or inhib itor binding, the sample w as incubated at room temperature for at least 45 min and transferred to a 4 mm quartz EPR tube The tube wa s flash frozen in liquid nitrogen before insertion into the resonator. All pulsed EPR data were collected in a Bruker EleXsys E580 spectrometer eq uipped with the ER 4118X MD 5 dielectric ring resonator at 65 K using a four pulse DEER sequence 104 described in detail previously 35 DEER Data Processing The DEER dipolar modulation curves were background subtracted, long pass filtered, and converted to distance distribution profiles via Tikhonov regularization (TKR) using De erAnalysis2008 170, 197 a free software from the Swiss Federal Institute of Tech nology Zurich website ( http://www.epr.ethz.ch/software/index ). Background subt raction level was determined using a self consistent analysis procedure 34 The optimal regula rization parameter was used for the conversion of the dipolar modulation curve to a TKR distance profile. Zero time was determined by fitting the 300 to 300 ns region of the dipolar modulation curve with a Gaussian function where the center of the Ga ussian fit wa s equal to the zero time. A s eries of Gaussian shaped populations representing the nominal conformations of HIV 1 PR 34, 52 with estimated relative percentage, fu ll width a t half maximum (FWHM) and most probable distance were summed to reconstruct the distance profile via DeerSim. Using this softwar e, the dipolar evolution curve wa s regenerated from the

PAGE 101

101 summed Gaussian profile for comparison to the experimental background s ubtracted data and TKR fit. DeerSim is a MATLAB based software created by the Fanucci laboratory Pearson Correlation Pearson correlation coefficient calculation and the criteria for its interpretation was previously established. 198 The following criteria was used: 0.8 1.0 for strong; 0.5 0.8 for moderate; 0.2 0.5 for weak; 0.0 0.2 for no asso ciation. The same criteria held true for neg ative correlations. 2 Error Analysis E rrors associated with percent relative populations were determined using 2 error analysis 178 as described in Chap ter 2 2 error analysis 50 was performed for populations <20% by sequentially suppressing these populations and their linear combinations. The regenerated echo curve after population suppression was compared to the TKR fit, and the 2 value was calculate d. When 2 was less than the critical value at P =0.05 ( 2 0.95 ) for certain degrees of freedom (df), the questionable populations were discarded. This method is based on the 2 analysis previously utilized to evaluate simulation uncertainty 178 E cho in tensity for 300 time points within = 0 3 s we re used for calculating 2 using Equations 2 10 and 2 11 in Chapter 2 If the 2 value is less than 2 0.95 then the model that contains the least number of populations is accepted at 95% confidence level ( P = 0.05). In this manner, smaller populations are validated or rejected. The value used for degrees of freedom (df) is calculated by subtracting the number of independent variables (e.g., distance, FWH M, population percentage) from the number

PAGE 102

102 of data points used. The value of 2 0.95 is determined using the 2 calculator available at http://ww w.fourmilab.ch/rpkp/experiments / analysis/chiCalc.html Results of the error analysis are shown in Appendix B. Results and Discussion Expression, Purification, Spin labeling and C haracterization of HIV 1 Protease The sodium dodecylsulfate polyacrylamide gel electrophoresis ( SDS PAGE ) of the ~11 kDa M36I HIV 1 PR monomer (Figure 3 3 ) illustrates the removal of contaminat ing proteins from the sample after each purification step The purified and spin labeled HIV 1 PR samples were estimated to be > 98% pure Figure 3 3 SDS PAGE of HIV 1 protease The reducing Tris Tricine gel illustrates the purification of HIV 1 PR subtype B M36I ( 11 kDa monomer marked with an arrow ) Lane s 1 10 and A : Broad range mole cular weight marker. Lane 2: Total cell extract. Lanes 3 4: Homogenate (H) and supernatant (S) of first pellet wash, respectively. Lanes 5 6: H and S of second pellet wash. Lanes 7 8 : H and S of third pellet wash. Lane 9: Inclusion bodies resuspension in 9 M urea. Lanes 11 18 : F ractions 4 11 (at 4 mL/fraction) from the anion exchange (Q) column. Lane s B and C : MTSL labeled M36I in 10 mM Tris HCl, pH 6.9 and 2 mM NaOAc, pH 5.0 repectively To ensure that the spin labeled protease had the proper secondary fold, circular dichroism (CD) spectra were acquired. Figure 3 4 shows the CD spectra for MTSL labeled D30N, M36I, A71V, D30N/M36I, D30N/A71V, M36I/A71V and D30N/M36I/A71V overlaid with a reference spectrum for subtype B PR.

PAGE 103

103 Conformational Sampling of Point Mutation Variants via DEER Spectroscopy The effects that accumulated primary and secondary mutations have on flap conformational sampling were determined from a series of single, double and triple mutant variants containing amino acid substitutions D30N, M36I and A71V. Figure 3 4 Circular dichroism spectra for MTSL lab eled HIV 1 PR variants in 2 mM NaOAc pH 5.0 overlaid w ith the reference spectrum for s ubtype B (b lack). Spectra are consistent with the predominantly strand secondary fold of this protease. Distance profiles between the two spin probes incorporated in to the flaps at site were determined from DEER spectroscopy (also referred to as PELDOR). The time domain, background subtracted dipolar modulation curves for wild type (WT) HIV 1 PR ( also referred to as subtype B), and for the seven variants are shown in Figure 3 5 A. The rate of decay and frequency of the oscillations seen in the dipolar modulation curves are directly related to the distance and breadth of the distance profiles 110, 197 The presence of secondary mutation, A71V alone or in combination with either D30N or M36I ( i.e., A71V, D30N/A71V, and M36I/A71V), shows steeper echo decay at <0.50 s

PAGE 104

104 (solid line) indicating that shorter distances consistent with a closed conformation dominate in the c orresponding distance profiles. Figure 3 5. A) Background subtracted and long pass filtered dipolar modulation curves in the time domain for WT subtype B and for variants with D30N, M36I, and A71V single and combined mutations overlain with Tikhonov r egularization (TKR) fits. The solid and dashed vertical lines mark the local minimum and maximum of th e WT echo curve, respectively. B) Stack plot of the corres ponding TKR distance profiles. T he dashed line corresponds to the semi open conformation and the solid line indicates the peak position of the closed state. Single asterisk indicate s the increase in the curled/tucked population while two asterisks signify an increase in the wide open population.

PAGE 105

105 For the other variants investigated, the slope of the initial echo decay is similar to WT, signifying most probable distances consistent with a predominant semi open conformation or comparable closed and semi open populations (e.g., D30N/M36I) Figure 3 5 B shows the distance profiles obtained from Tikhon ov regularization (TKR) of the background subtracted time domain dipolar modulation echo curves. For WT, D30N, M36I and D30N/M36I/A71V, the most probable inter spin distance occurs near 35 37 (dashed line), corresponding to a predominant semi open confo rmation This distance decreases to ~33 (solid line) for A71V, D30N/M36I, D30N /A71V, and M36I/A71V; indicating that a distance consistent with a closed like or closed confor mation is favored. Previously, this distance was only observed in DEER measure ments when inhibitor was added. 34, 188 Hence, here, we distinguish between the closed conformation when the protease is ligand bound and a closed like state in the absence of a substrate or inhibitor. Additional ly, these variants show a marked increase of a population located within 26 31 (asterisk). This distance is suggestive of a curled/tucked conformation of the flaps 49, 54 hairpin would be arranged in a manner that the spin labels would point towards one another, leading to a shorter distance between the spin probes, but where the backbone conformation may produce an opening to the active site pocket allowing fo r inhibitor to escape. For A71V, D30N/M36I, M36I/A71V and D30N/M36I/A71V, a small but distinct peak in the range of 40 45 ( two asterisk s ) wa s observed. This distance wa s assigned to a wide open flap conformation 46, 49, 199, 200

PAGE 106

106 Gaussian Reconstruction Profiles and Hypothetical Folding Funnels In a self consistent analy tical technique for background subtraction of the dipolar modulated echo curves, a series of Gaussian shaped populations is used to regenerate the TKR distance profile and fit to the time domain data 34, 52 A representative Gaussian reconstruction of the distance profile for M36I/A71V is shown in Figure 3 6 A. Combining the results of these reconstruction s with those from MD simulations and X ray models, we assign four predominant types of conformational populations observed in the SDSL DEER distance profiles for HIV 1 PR 46, 49, 50, 54, 190 193 Gaussian populati ons with average flap distances in the range of 26 31 are defined as curled /tucked ensem bles. Meanwhile, populations centered near 33 are defined as closed like conformations, whereas those near 35 37 are referred to as semi open conformations. Fin ally, distances within 40 45 are described as wide open populations The bar graph in Figure 3 6 B summarizes the results from the Gaussian reconstruction analyses. The changes in SDSL DEER distance distribution prof iles are summarized in Table 3 1 inc luding information on peak center position, full width at half maximum (FWHM) and relative population percentage of th e individual Gaussian shaped po pulations used to regenerate the TKR distance profile. Meanwhile, a summary of distance parameters, such a s range, average and most probable distance for each construct and for the individual conformations are shown in Tables 3 2 and 3 3. The com plete DEER data analysis for subtype B variants is shown in Appendix A.

PAGE 107

107 Figure 3 6 A) Gaussian reconstr uction profile for M36I/A71V. B) Results from Gaussian reconstruction analysis of TKR distance profiles for the various variants Complete DEER data analysis is shown in Appendix A. Several conclusions regarding the effects of sequential mutations on the confo r mational sampling can be made. When compared to WT, the incorporation of the D30N primary mutation does not significantly alter the conformational sampling profile, where roughly 65 70% of the conformers are in a semi open state. Strikingly, when D30N i s combined with either A71V or M36I, a closed like conformational ensemble dominates, with less than 35% of the population in the semi open conformation. The closed like conformation also dominates with A71V alone or in combination with M36I. The dramati c change in flap conformational sampling upon incorporation of the non polar mutation of A71 to valine is consistent with the results of earlier X ray crystal structure and MD investigations of this polymorphism 201 where it was shown that this substitution requires the local structure adjustment in the cantilever region, particularly amino acid residues 67 71 to accommodate the bulkier valine side chain 201 Because the loo p formed by residues 68 71 is highly mobile during flap opening and closing upon substrate/inhibitor binding 85 it is not surprising that this single mutation can markedly shift the flap conformational sampling profile

PAGE 108

108 Table 3 1 Changes i n flap di stance distribution profile in s ubtype B variants deter mined by SDSL DEER. Data is presented as distribution center, full width at half maximum (FWHM), and relative population in percent. Construct Center ( 0.2 ) FWHM ( 0.3 ) Population ( % ) Pea k Assignment Subtype B a 28.5 4.0 5 3 Curled/Tucked 33.3 3.9 24 3 Closed 36.1 5.2 67 3 Semi o pen 40.4 2.8 4 3 Wide o pen D30N 27.0 5.5 11 6 Curled/Tucked 33.0 3.0 20 6 Closed 36.7 5.1 69 6 Semi o pen M36I 30.0 5.6 11 3 Curled/Tucked 33.5 3.7 29 3 Closed 36.4 4.8 60 3 Semi o pen A71V 30.6 2.6 17 6 Curled/Tucked 33.0 3.0 67 6 Closed 36.0 2.8 7 6 Semi o pen 42.8 3.0 9 6 Wide o pen D30N/M36I 26.1 2.9 18 4 Curled/Tucked 32.8 3.1 4 2 4 Closed 35.6 4.3 35 4 Semi o pen 45.7 2.9 5 4 Wide o pen D30N/A71V 25.8 2.6 21 4 Curled/Tucked 32.9 3.3 51 4 Closed 36.3 4.3 23 4 Semi o pen 40.5 3.3 5 4 Wide o pen M36I/A71V 28.2 2.6 11 5 Curled/Tucked 33.0 3.0 65 5 Closed 35.2 2.3 9 5 Semi o pen 40.0 3.5 15 5 Wide o pen D30N/M36I/A71V 27.6 4.0 13 4 Curled/Tucked 32.7 4.8 12 4 Closed 35.1 5.4 63 4 Semi o pen 43.0 4.4 12 4 Wide o pen a) Data obtained from Kear et al. 52 error for relative % population was recalculated.

PAGE 109

109 Tab le 3 2 Summary o f DEER distance parameters for s ubtype B HIV 1 PR variants Span is defined as the difference between the farthest and shortest distance between the K55R1 flap sites. Construct Range (span) Most Prob. Dist. Avg. Dist. ( 1 ) ( 0.2 ) ( 0.2 ) Subtype B 24 43 (19) 35.1 35.2 D30N 22 42 (20) 36.6 34.9 M36I 25 41 (16) 35.0 34.9 A71V 28 46 (18) 32.9 33.7 D30N/M36I 23 49 (26) 33.3 33.2 D30N/A71V 23 44 (21) 33.0 32.6 M36I/A71V 26 44 (18) 33.1 3 3.7 D30N/M36I/A71V 24 47 (23) 34.9 34.8 Table 3 3 Average distance and range for HIV 1 PR conformations determined for seven subtype B variants with D30N, M36I, and A71V mutations. Given that each distance distribution represents an ensembl e, it is not surprising that the different variants have slight changes in the average distances of each conformational ensemble. Construct Range () Avg. Dist. () curled/tucked 26 31 28 2 C losed 32.7 33.5 33.0 0.3 semi open 35.0 37.0 35.9 0.5 wide open 40 46 42 2 R ecent MD simulations show that A71V has a stabilizing effect that may be essential to offset other mutations that alter hairpin conformation and destabilize the dimer interfac e. 202 Additional work has upheld the MD results, showing an increase in protein stability after combining A71V to a primary mutation. 83 The stabilizing effect of A71V is consistent with a more compact structure conferred by flap clo sure, indicated by an increase in closed like populations. The M36 substitution with isoleucine is the most common mutation in the hydro phobic core 203 and is proposed to alter flap conformation via the hydropho bic sliding mechanism 81 Additionally, MD simulations have shown that M36I decreases the

PAGE 110

110 binding cavity volume 204 H owever, we find herein that the M36I mutation alone has only a minimal effect on the conformational sampling ensemble compared to WT. The population of the semi open ensemble slightly dec reases with an increase of the closed and curled conformations. These findings are consistent with insights from MD simulations, suggesting that the decrease in the binding cavity volume results from the increase of closed like population When M36I and A 71V secondary polymorphisms are present along with primary mutation D30N, protease conformational sampling reverts to the semi open ensemble as the predominant form with its relative percentage similar to WT. Also, compared to WT, the triple muta tion cons truct shows simultaneous increase in the percentage of the wide open and curled conformations and a decrease in closed population percentage Dunn and co workers 76 reported that the D30N/M36I/A71V construct has w eaker inhi bition (higher K i ) for nelfinavir indinavir, and ritonavir, and better catalytic efficiency with respect to WT The recovery of major semi open population via accumulation of primary and secondary mutations appears to be a mechanism by which HIV 1 PR main t ain s catalytic efficiency while escaping inhibitor binding. Figures 3 7 A and 3 7 B summarize these results graphically in the form of hypothetical protein folding funnels In these figures, the relative energ y of the semi open states is alig ned. The probability distribution profiles are directly converted to the thermodynamic energies via the relationship G = RT ln P i where T was taken as 300 K and P i determined from the normalized population probability. Because curling of the flaps is co nsidered an alternative flap opening mechanism 54 the curled/tucked and wide open ensembles are combined into the single open like state for these energy

PAGE 111

111 landscapes. Note that the energy surfaces are called hypoth etical because from DEER data alone the heights of the barriers between the states are unknown Only the relative free energies of the states can be obtained from the population analyses. Figure 3 7 H ypothetic al protein folding funnels for WT subtyp e B (A) ; and the D30N/M36I/A71V variants (B) For each, the energy of the semi open conformations are set equ al and the relative free energy differences are calculated with respect to the semi open conformation. Free energies were calculated directly fro m the relative percentages given in B where open like percentage is the sum of the wide open and curled/tucked states. The temperature used to calculate G is assumed to be constant at 300 K, assuming that the freezing technique is fast enough to trap th e conformations at this temperature. However, recognizing the possibility of temperature fluctuations and that flash freezing in liquid N 2 is not rapid enough, the parameter G T1 / G T2 can also be calculated where G T1 and G T2 are the changes in Gibbs fr ee energy at temperatures

PAGE 112

112 T 1 and T 2 respectively. This parameter could provide information on conformational sampling alterations attributed to mutations in a temperature independent manner. The calculated G values between conformational states indic ate that relative to the semi open conformation, the combined mutations act to stabilize the open like states and destabilize the closed like conformation w hen compared to the WT variant This effect can be seen clearly in Figure 3 8 where the relative pe rcen tage change in the conformational populations for each variant is compared to wild type. Clearly, most of the variants have a destabilized semi open conformation, but have increased curled/ tucked and closed like populations. The triple mutant vari ant however, has a destabilized closed state, a semi open population similar to WT, and increased populations of both the curled/tucked and wide open conformational states. Figure 3 8 Relative percentage change in the population of each conformation f or the variants studied. The relative percentage change is calculated with respect to WT subtype B, % (P i P WT ).

PAGE 113

113 Pearson Correlation of DEER Results with Enzymatic and Inhibition Parameters Statistical correlations, via Pearson product moment correlation 205 were made between the populations of conformations observed in the DEER distance profiles, and kinetic and inhibition data Note that protease variants used in EPR experiments differ from those u tilized in en zymatic investigations. DEER samples contain the K55R1 spin label site and the catalytic residue substitution D25N. This substitution is often used in spectroscopic studies 37, 39, 98, 99, 206, 207 because of the rela tive ease of sample preparation, and sample stability and homogeneity. The D25N mutation does not significantly change the protein structure, altho ugh, it does increase the dissociation constant ( K d ) 98 by up t o a factor of ~ 10 6 Nevertheless, correlations among the data were observed. Pearson product moment correlation coefficient ( r ) values of 0.84, and 0.19 were obtained for the relative percentage closed conformation (%C) with k cat and K m respectively r is > 0.80, the correlation tailed critical value for r at P = 0.05 for a sample size of 8 is 0.707. 208 As shown in Figure 3 9 trend ( r = 0.84) is obtained between %C and reported values for k cat 76 m eaning, a higher percentage closed like population coincides with lower catalytic rate The strong negative correlation of %C with k cat appears reasonable given that substrate entry requires flap opening, and mutations that stabilize a closed like confor mation could have reduced turnover rates. NMR studies of substrate binding in HIV 1 PR suggest that in order for substrate uptake to occur, the flaps would have to open first, followed by slow flap closure and concomitant positioning of the substrate int o a reactive geometry in the binding cleft. 207 Thus, if HIV 1 PR is initially in the closed like

PAGE 114

114 state, the protease would need to adopt an open conformation for substrate entry, then a rate limiting flap closure and concurrent positioning of substrate in the active site would e nsue prior to catalysis, which explains the lowered catalytic rate in variants that have high closed like population. Meanwhile, F igure 3 9B reveals K m ( r = 0.19), thus explaining the slightly weaker correlation ( r = 0.73) seen between %C and k cat / K m. The lack of correlation between %C and K m is not surprising given that the combined mutations are not expected to significantly alter the shape of the substrate binding pocket in a systematic manner. Figures 3 9 C and percentage semi open conformation ( %SO ) with k cat ( r = K m ( r = 0.22). Again the correlation with k cat / K m ( r = between catalytic rate and the semi open conformation can be understood by consi dering that the flexibility and entropy of the semi open conformational ensemble may be needed for maintaining the turnover rate and catalytic efficiency. This finding is consistent with MD investigations that reveal a unique flap handedness in the semi open state of the protease which is necessary for subsequent flap reversal, flap opening and substrate uptake steps of the HIV 1 PR catalytic cycle. 20 9 The strongest correlation s w ere observed when values of inhibition constants ( K i ) for nelfinavir (NFV), ritonavir (RTV) and indinavir (IDV) were plotted against the ratio of the closed like conformation to the open like states ( Figure 3 10 ). The states, consists of combine d wide open conformers ( distance ~43 ) and curled or tucked populations (distances ~25 30 ). If only a single population was considered or

PAGE 115

115 when a different conformational ratio was used lower correlation coef f icients we re obtained (Table 3 4 ) Figure 3 9 C orrelation plots for trends of closed and semi open conformation percen tages with kinetic parameters The Pearson product moment coefficient ( r ) is reported for each plot. A) Percentage of closed population is inv ersely cor r elated with k cat B) Percentage closed population shows poor correlation with K m C) Percentage of semi open popu lation is directly correlated with k cat D) Percentage semi open population shows poor correlation with K m As shown in Figur e 3 10, s trong Pearson correlation coefficients were calculated between the confor mational ratio, %(closed like /open like) and the K i for NFV ( r = 0.98), RTV ( r = 0.91), and IDV ( r = 0.95) These findings suggest that an increase open

PAGE 116

116 decrease in the population of the closed state, where inhibitor binding is stabilized, leads to larger values of inhibition constants. Higher K i i ndica tes we aker inhibitor binding strength, and thus, drug resistance. Table 3 4 Summary of Pearson product moment coefficients ( r ) for DEER relative percentages correlated to inhibition constants ( K i ) a Relative percentage of conformation from DEER R K i (NFV) K i (RTV) K i (IDV) % semi open 0 .39 0.22 0.39 % wide open 0.30 0.46 0.04 % open like 0.28 0.44 0.24 %curled/tucked 0.12 0.17 0.32 %(semi open/open like) 0.06 0.23 0.04 %(closed like/semi open) 0.43 0.31 0.55 %(closed like/wide open) 0.52 0.55 0.35 % closed like 0.59 0.44 0.58 %(closed like/open like) 0.98 0.91 0.95 a) Clemente et al., 2003 76 Furthermore, these findings show that enzymatic activity is maintained only when the combination of mutations does not significantly alter the population of the semi open co nformation relative to WT subtype B protease The correlations between inhibitor binding and conformational sampling also show that alternative flap opening modes that could promote weaker inhibit or interactions like the curled and tucked states need to be considered when discussing drug resistance mechanisms This point was recently demonstrated via X ray crystallography of an extremely drug resistant variant found in an open like conformation 210 Therefore drug resistance appears when co e volving mutations destabilize the closed state in favor of open like conformations to e lude inhibitor binding, while maintaining a semi open population similar to wild type so as to sustain catalytic efficiency

PAGE 117

117 Figure 3 10 Strong negative correlati ons are observed from Pearson product correlation plots for K i values of the inhibitors nelfinavir (top), ritonavir (middle), and indinavir (bottom) against the ratio of the percentage closed to the percentage of open like conformations. The Pearson produ ct moment correlation coefficient ( r ) is reported for each plot.

PAGE 118

118 Inhibitor bound DEER Data for HIV 1 Protease with Point Mutations To monitor the effect of inhibitor binding to the flap conformational sampling of HIV 1 protease contain ing single point muta tions, 4 fold molar excess of inhibitor or substrate mimic is added to MTSL labeled D30N, M36I, A71V and D30N/M36I/A71V su btype B HIV 1 PR variants F igu re 3 11 show s the background sub tracted dipolar evolution curves and corresponding TKR fits for subtyp e B variants bound to ritonavir (RTV) and the substrate mimic CA p2. T the echo curve s for both D30N and M36I bou nd to RTV or CA p2 (Figure 3 11 A to D) show less dampening of the oscillations when compared to the corresponding dipolar evolution curves fo r the apoenzyme, indicating shorter flap distances in the presence of inhibitor or substrate. Meanwhile, for A71V, the frequency of oscillations is not substantially altered in the presence of either CA p2 or RTV, but the strength of oscillations is incre ase d for both cases. Interestingly, for D30N/M36I/A71V, the oscillation frequency increased with respect to the apo echo curv e when CA p2 was added. In contrast when RTV wa s added, the di polar evolution curve obtained wa s strikingly similar to apo, with a slight increase in modulation depth. The DEER dipolar evolution curves we re consistent to the derived TKR distance di stribution profiles (Figure 3 12 ). Summary of the Gaussian reconstruction analysis of ligand bound DEER data for s ubtype B variants are shown in Tables 3 5 to 3 7 The area normalized distance profiles in the presence of the non hydrolyzable substrate mimic ( CA p2 ) and ritonavir (RTV) overlaid with the apoenzyme (apo) reveal a shift of the most probable distance from 35 37 to 33 in D 30N and M36I variants (Figures 3 12 A and B), indicating a conformational shi ft in favor of the closed state. The bar graphs in the figure insets show 60 66% increase of the closed population percentage

PAGE 119

119 for CA p2 or RTV bound D30 N and M36I Moreover, th e distance distribution widths we re narrower after adding the substrate mimic or inhibitor. By contrast, the most probable distance for the A71V DEER distance profile (Figure 3 12 C) wa s unchanged at 33 even in the presence of substrate mimic or inhibitor suggesting that t he closed like or closed state wa s the major population for both the apo and the CA p2 or RTV bound protease. Noticeably, th e distance distribution widths we re narrower for A71V in the presence of substrate mimic or inhibitor, with a 33 % and 27% increase in the percentage occupancy of the closed state for CA p2 and RTV, respectively. The triple mutation variant, D30N/M36I/A71V, demonstrated flap closure when CA p2 was added, with a shift in most probable distance from 35 to 33 si milar to D30N and M36I variants bound to CA p2. However, addition of RTV did not induce flap closur e in D30N/M36I/A71V (Figure 3 12 D) and the most probable distance remained at 35 similar to apo. Not e that the curled/tucked and wide open populations s imultaneously disappeared after adding CA p2 or RT V to the triple mutation variant suggesting a relationship between these populations consistent with the assumption that curling of the flaps may be an alternative flap opening mechanism. T he ligand boun d DEER distance profiles suggest that both CA p2 and ritonavir could efficiently close the flaps of the D30N, M36I and A71V single point mutation variants Unlike D30N and M36I, the A71V construct has a predominant closed like population in the apo state. The same flap distance of 33 observed for the closed like state in the apoenzyme and the closed conformation upon inhibitor binding

PAGE 120

120 suggests that these states are indistinguishable and validates that the incorporation of the single point A71V mutatio n promotes the closed flap conformation in the apo state Figure 3 11 Background subtracted dipolar evolution curves and TKR fits for MTSL labeled subtype B variants D30N, M36I, A71V and D3 0N/M36I/A71V bound to CA p2 or RTV (red) overlaid with apo (blue)

PAGE 121

121 Figure 3 12 Area normalized TKR distance distribution profiles for subtype B variants A) D30N, B) M36I, C) A71V and D) D30N/M36I/A71V for apo (black) CA p2 bound (red) and RTV bound (blue) protease. Inset figures show the corresponding bar g raphs summarizing the DEER percentage of the closed population (DEER % closed) for the apoenzyme and ligand bound protease derived from the Gaussian reconstruction analysis of the distance profiles. The open like states (i.e., curled/tucked and wide open populations) almost dis appeared completely after adding a ligand to all single point mutation variants ( 5%), suggesting that inhibitor or substrate bi nding imposes a steric strain to the flaps, making open like conformations less stable relative to the closed and semi open states Also note that compared to apo, the dis tance distribution widths we re narrower after ligand binding, as previously observed, 34, 211 implying that binding to the substrate mimic or inhi bitor confers conformational rigidity as the flaps are locked in onto the ligand in the binding cleft.

PAGE 122

122 Table 3 5 Ligand induced c hanges in flap distance distribution profile in s ubtype B variants determined by SDSL DEER. Data are presented as distribu tion center, full width at half maximum (FWHM), and relative population in percent. Construct Center ( 0.2 ) FWHM ( 0.3 ) Population ( % ) Peak Assignment D30N + CA p2* 33.9 3.5 80 3 C losed 35.5 3.8 15 3 semi open 39.5 2.5 5 3 wide open D30N + RTV 33.1 2.9 80 4 C losed 35.0 2.0 15 4 semi open 38.2 2.0 5 4 wide open M36I + CA p2 33.2 3.4 95 3 C losed 37.2 1.6 5 3 semi open M36I + RTV 33.4 3.3 95 3 C losed 39.5 2.1 5 3 wide open A71V + CA p2 33 .1 2.6 100 4 C losed A71V + RTV 33.2 3.4 94 4 C losed 37.2 1.6 6 4 semi open D30N/M36I/A71V + CA p2 33.8 4.0 90 3 C losed 39.3 2.7 10 3 wide open D30N/M36I/A71V + RTV 33.8 3.8 10 3 C losed 35.3 5.5 90 3 semi open Interestingly, for the triple mutant construct, D30N/M36I/A71V, the substrate mimic CA p2 induced flap closure, but not ritonavir (RTV). Since this construct is known to be resistant to RTV based on K i 76 measure ments, it is no t surprising that flap closure wa s not observed after adding this inh ibitor. Although flap closure wa s not observed, presence of the ligand prevented the flaps to adopt open like conformational states, and promote d the semi open conformati on.

PAGE 123

123 Table 3 6 Summary o f DEER distance parameters for D30N, M36I, A71V and D30N/M36I/A71V with inhibitor or substrate mimic Span is defined as the difference between the farthest and shortest distance between the K55R1 flap sites. Sample Range (span) Most Probable Distance Average Dist. ( 1 ) ( 0.2 ) ( 0.2 ) D30N + CA p2 30 43 (13) 33.9 34.4 D30N + RTV 30 40 (10) 33.1 33.6 M36I + CA p2 30 39 (9) 33.2 33.4 M36I + RTV 30 43 (13) 33.4 33.7 A71V + CA p2 30 36 (6) 33.1 33.1 A71V + RTV 30 39 (9) 33.2 33.4 D30N/M36I/A71V + CA p2 30 42 (12) 33.8 34.2 D30N/M36I/A71V + RTV 30 41 (11) 35.3 35.2 Table 3 7 Average distance and range for HIV 1 PR conformations determined for D30N, M36I, A71V and D30N/M36I/A71V with inhibi tor or substrate mimic. Given that each distance distribution represents an ensemble, it is not surprising that the different variants have slight changes in the average distances for each population Population Range () Average Distance () c losed 33. 1 33.9 33.4 0.3 semi open 35 37 36 1 wide open 38.2 39.5 39.1 0.8 Conclusions Both primary and secondary mutations alter the conformational sampling in HIV 1 PR based on DEER results The secondary mutation A71V promotes a closed li ke conformation of the protease except w hen both D30N and M36I substitutions are also present. The recovery of a predominant semi open population in the D30N/M36I/A71V variant after the step positive Pearson correlations for the percentage of the semi open population vs. k cat suggests the importance of the semi open conformation in maintaining turnover rate and catalytic

PAGE 124

124 efficiency. negative inverse Pearson correlation for the percentage of closed conformation vs. k cat implies that certain mutations that promote a closed like conformation (i.e., A71V) could hamper substrate entry, resulting in lower ed catalytic rate However, a more closed like conformation may also sta bilize protease inhibitor interactions. The accumulation of additional mutations is required (i.e., D30N and M36I) to recover a major semi open population and to restore catalytic activity. negative correlations among the ratio o f closed like to open like fractional occupancy with K i values for NFV, RTV, and IDV implies that drug like conformation while main taining a semi open population similar to WT. The correlations are corroborated by the lack of flap closure upon adding ritonavir to D30N/M36I/A71V, a construct that exhibited resistance to this inhibitor based on K i measurements. Correlation measurements with k cat and K i show that by accumulating mutations, HIV 1 PR could recover close to WT catalytic efficiency while eluding inhibitors. These results indicate a possible mechanism for drug resistance where the population of a thermodynamic state is altere d. Alternative hypotheses for how drug resistance is affected in HIV 1 PR also evoke the idea that protein dynamics, hydrophobic core flexibility 212 or t he exchange rates among conformational states 213 are altered by secondary mutations, which in turn modulates enzyme function. I n fact, numerous MD simulations provide insights into hypotheses where protein backbone dynamics and flexibility are altered 212 214

PAGE 125

125 From an experimental approach, NMR investigations of protein conformational exch ange and dynamics have a strong hold 215, 216 Although the backbone dynamics of the native subtype B construct have been explored extensively, 37, 99, 206, 207 217 the exchange rates in native HIV 1 PR are difficult to characterize due to the fast dynamics of the flaps and relatively slower time scale of the exchange process. 214 Recently, exchange dynamics were measured on a synthetic tethered HIV 1 PR construct revealing that mutations may alter the rate of exchange among t he states, thus impacting catalysis. 189 A combination of effects are likely at play to generate drug resistance, and possibly, a combination of all these aspects o f the protein are changing as drug pressure selected mutations arise. DEER results for the apoenzyme are corroborated by acquiring DEER data for D30N, M36I, A71V; and D30N/M36I/A71V variants with CA p2 or RTV The presence of CA p2 and RTV was shown to i nduce flap closure in D30N and M36I, and to reduce the distance distribution widths in D30N, M36I and A71V. Furthermore, the most probable distance of 33 measured for A71V apo is unchanged after adding inhibitor or substrate mimic, validating that the c losed like conformation observed for A71V apo is indistinguishable from the closed conformation observed in the inhibitor bound protease. However, conformational flexibility is lim ited in the inhibitor bound protease resulting in decreased distance distr ibution width. Finally, the absence of flap closure in D30N/M36I/A71V after adding RTV coincides with the drug resistance exhibited by this variant to RTV. The disappearance of wide open and curled/tucked populations after adding CA p2 or RTV suggests t hat these states are simultaneously destabi lized by inhibitor or substrate binding.

PAGE 126

126 CHAPTER 4 CORRELATING CONFORMA TIONAL SHIFT INDUCTI ON WITH ALTERED INHB ITOR POTENCY IN A MULTI DRUG RESISTANT HIV 1 PROTEASE VARIANT Introduction The c ross resistance of H IV 1 protease (HIV 1 PR) to several protease inhibitors (PIs) could result from primary and secondary amino acid changes 218 that occur because of natural polymor phisms 219 drug pressure selected mutations 70, 76 or combinations of both. These substitutions have been shown to alter flap conformational sampling 35, 52, 188 and dynamics 85 in the free HIV 1 protease Many residue substitutions outside the active site are often associated with drug resistance and has an effect to substrate or inhibitor binding. 76, 78 80, 220 Mutation of the se residues facilitate the conformational changes that occur in HIV 1 PR to allow ligand binding. Of the 99 residues in this protease, 40 are hydrophobic and some have close van der Waal s contacts with the ligand at the active site or are located in the f lap region. Amino acid substitutions especially in the hydrophobic core, could potentially increase HIV 1 PR flexibility, which in turn could decrease inhibitor or substrate binding because of higher protease inhibitor dissociation rate. 81 T his chapter describes the inhibitor induced conformational sampling shifts in the multi drug resistant patient isolate MDR769 using double electron electron resonance (DEER) spectroscopy A parameter defined as | C | or the change in the inhibitor induced conformational shift to the closed state is correlated to the change in half maximal inhibitory concentration (IC 50 ) of a PI against the HIV virus. The MDR769 Patient Isolate The failure of HIV/AIDS treatment parti cularly Highly Active A ntiretroviral Therapy (HAART) is attributed to the appearance of drug pressure selected mutations in the HIV

PAGE 127

127 genome after PI exposure 70, 76 One particular multi drug resistant HIV 1 PR vari ant is the patient isolate MDR769 (Figure 4 1). MDR769 is resistant to various inhibitors 221 and exhibits higher level of resistance to most FDA approved PIs, with the exception of DRV, TPV and LPV. 88 Crystal structures of this variant in the apo, substrate bound and inhibitor bound forms 55, 88 90 illuminate an expanded active site pocket As a result, ther e are fewer hydrogen bonding and van der Waals interactions result between inhibitors and the binding cavity contributing to drug resistance in MDR769 55 Figure 4 1. Ribbon diagram of MDR769 (PDB file 1TW7) pr otease colored by subunit, rendered in PyMol 1.3. Primary mutations relative to wild type (LAI) are shown as red spheres, while compensatory mutation sites are rendered as yellow spheres. MTSL spin probes (K55R1) are incorporated in silico via MMM 2011.1 222 and shown as gray capped sticks. Definition of the | C | Parameter T he relationship between changes in flap closure and dru g potency were investigated by defining a parameter | C |, which is the magnitude of the difference in inhibitor induced conformational shift to the closed state between two HIV 1 protease variants. For instance, if a given protease inhibitor induced X% an d Y% of the closed

PAGE 128

128 population (%closed) in variants A and B, respectively, | C | can be calculated by using Eq uation 4 1 (4 1) This parameter is employed to calculate the inhibitor induced percentage change of flap closure between subtype B and MDR769. The values of | C | are then compared to drug potency using previo usly reported half maximal inhibitory concentration s (IC 50 ) Materials and Methods Details on the MDR769 Variant The MDR769 construct used in pulsed EPR experiments do not contain the stabilizing mutations (Q7K, L33I and L63I) 97 present in subtype B variants The native cysteince mutations to alanine (C67A and C95A) and the K55C substitution that creates the spin labeling site are present. Moreover, the inactivating D25N mutation is also incorporated for improved sample stability and homogeneity. 98 Because of these mutations that would guarantee site specific labeling at site C55, the type of MDR769 variant used is abbreviated SDSL i which stands for i nactive protease for s ite d irected s pin l abeling Details of the amino acid differences of the MDR769 construct relative to native enzy me (LAI) are summarized in Figure 4 2 Figure 4 2. Amino acid sequence alignment of subtype B and MDR769 HIV 1 PR. The amino acid substitutions relative to s ubtype B BRU (LAI) that emerge in MDR 769 are drug pressure selected mutations (in red). The s ubstitutions of native cysteine residues to alanine, t he active site D25N mutation and the substitution that creates the labeling site (K55C) are in green

PAGE 129

129 Cloning, Protein Expression S pin labeling and Purification The methodology for cloning, protein expression, spin labeling and purification of MDR769 is the same for that employed in subtype B variants as described in detail in Chapter 3 but with the following exception: the pH of th e buffer used to solubilize the inclusion bodies is adjusted to 8.80. Substrates and Inhibitors Most of the protease inhibitors (PIs) were acquired through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH. These PIs are as f ollows: indinavir sulfate, nelfinavir, atazanavir sulfate, saquinavir, ritonavir, lopinavir, tipranavir and amprenavir. Darunavir was obtained from Tibotec Pharmaceuticals. The peptide substrate mimic, CA p2 (H Arg Val Leu r Phe Glu Ala Nle NH2; Nle=nor l eu cine, r=reduced), was purchased from the University of Florida Protein Chemistry Core Facility. DEER Experiments and Data Processing Samples were prepared as described in Chapter 3. The DEER dipolar modulation curves were converted to Tikhonov regulari zation (TKR) distance distribution profiles using the DeerAnalysis2008 software package. 109 Gaussian reconstruction was performed via DeerSim software and any population that has a fractional occupancy of less than 20% were validated by 2 error analysis as described in Chapter 2. Results of the error analysis are summarized in Appendix B. Results and Discussion Ligand induced Flap Closure in MDR769 Ligand i nduced confor mational shifts in MDR769 211 were determined from DEER measurements, with methanethiosulfonate (MTSL) spin probes incor porated at sites

PAGE 130

130 K5 5R1/ Figure 4 3 shows select DEER data and distance profiles (comple te data in Appendix A ). The effects of inhibitors on the average flap distance can clearly be seen in the DEER echo curves in Figure 4 3 A. LPV and TPV shift the frequency with less d ampening of the oscillations; generating distance profiles with most probable distances of 33 and narrower breadths. In Figure 4 3 B, the solid line indicates a distance of ~37 which coincides with that expected for the semi open conformational state, where the distance profile for apo MDR769 indicates a longer most probable flap distance (37 versus 36 ) relative to subtype B, as seen previously 35 The dashed line at 33 marks the distance expected for the closed conformation 34, 35 induced by an inhibitor Figure 4 3 A) Background subtracted DEER dipolar evolution curves with fits from Tikhonov regularization (TKR) analysis for MDR769. Vertical dashed line ma rks the local minimum for apo. B) Stack plot of distance profiles for free MDR769 and with FDA approved inhibitors or substrate mimic, CA p2. Semi open and closed populations have flap distances of 37 ( solid line) and 33 ( dashed line), respectively. The minor population at 26 30 corresponds to the curled/tucked flap conformation (asterisk), whereas those at 40 45 are assigned to t he wide open populations (two asterisk s ).

PAGE 131

131 Fractional Occupancies of Putative Confo rmations from Gaussian Reconstruction T o ensure proper background subtraction, the Tikhonov regularization (TKR) distance profiles from DeerAnalysis2008 are deconstructed to a linear combination of Gaussian populations. As mentioned in Chapter 3 four Ga ussian populations are required for sufficient regeneration of the TKR data, a nd these are assigned to curled or tucked, closed, semi open, and wide open HIV 1 PR conformational states 35, 52 Population assignmen ts are based on assessing MTSL distances in HIV 1 PR models from molecular dynamics simulation and X ray studies 46, 48, 190, 191 Minor populations located at 26 30 and 40 45 are assigned to the curled or tuc ked 54 and wide open states 46 respectively. Results of the Gaussian recons truct ion are summarized in Tables 4 1 to 4 3 For apo subtype B HIV 1 PR, previous DEER studies reveal a distance profile containing a predominant semi open conformation, where inhibitor binding shifts the conformational ensemble to increasing population s of the closed state 34 For subtype B, the fractional occupancy of the closed state was >60% for 7 out of 10 inhibitor s with a concomitant shift in the most probable distance from 36 to ~33 (Figure s 4 4A and 4 5 A ) These PIs are ritonavir (RTV), saqu inavir (SQV), amprenavir (APV), lopinavir (LPV), darunavir (DRV) tipranavir (TPV), and Ca P2 substrate mimic. As shown in Figure 4 5 A, f or the remaining three inhibitors, atazanavir (ATV) has ~40% closed population, whereas only <15% closed population is seen for nelfinavir (NFV) and indinavir (IDV).

PAGE 132

132 Table 4 1 Inhibitor induced changes in MDR769 flap distance distribution profile Data is presented as peak center, full width at half maximum (FWHM), and relative population percentage. Construct Cent er ( 0.2 ) FWHM ( 0.3 ) Population ( % ) Peak Assignment MDR769 apo 33.2 3.1 18 5 Closed 36.5 4.4 75 5 Semi o pen 45.1 3.4 7 5 Wide o pen Indinavir 30.8 4.5 22 4 Curled/Tucked 33.9 3.5 18 4 Closed 36.9 4.5 43 4 Semi o pen 40.1 4.0 17 4 Wide open Nelfinavir 29.5 5.6 18 5 Curled/Tucked 33.3 4.5 20 5 Closed 36.4 4.6 56 5 Semi o pen 40.9 4.0 6 5 Wide o pen Atazanavir 33.2 3.3 11 4 Closed 35.9 4.5 74 4 Semi o pen 39.0 4.5 15 4 Wide o pen S aquinavir 30.0 4.0 7 5 Curled/Tucked 33.9 3.7 32 5 Closed 36.6 3.5 44 5 Semi o pen 39.5 4.6 17 5 Wide o pen Amprenavir 30.1 4.1 13 5 Curled/Tucked 33.1 3.5 36 5 Closed 36.0 4.0 51 5 Semi open Ritonavir 34.0 3.3 31 5 Closed 37.2 3.4 69 5 Semi open Darunavir 33.6 2.5 68 5 Closed 35.2 2.5 18 5 Semi open 39.0 2.9 14 5 Wide open Lopinavir 29.2 1.8 7 4 Curled/Tucked 33.1 3.0 71 4 Closed 37.0 3.0 22 4 Semi open Tipranavir 30.8 1 .2 6 4 Curled/Tucked 33.1 2.5 84 4 Closed 45.8 1.5 10 4 Wide open CA p2 25.8 1.3 6 4 Curled/Tucked 33.4 3.0 88 4 Closed 37.0 1.5 6 4 Semi Open

PAGE 133

133 Table 4 2 Summary of DEER distance parameters for free and inhibitor bound MDR 769 sam ples Span is defined as the difference between the farthest and shortest distance between the K55R1 flap sites. Sample Range (S pan) Most Probable D istance ( 1 ) ( 0.2 ) MDR769 apo 33 45 (12) 36.7 Indinavir 31 40 (9) 36.9 Nelfinavi r 30 41 (11) 36.0 Atazanavir 33 39 (6) 35.9 Saquinavir 30 40 (10) 36.3 Amprenavir 30 36 (6) 35.3 Ritonavir 34 37 (3) 36.3 Darunavir 34 39 (5) 33.8 Lopinavir 29 37 (8) 33.1 Tipranavir 31 46 (15) 33.1 CA p2 26 37 (11) 33.3 Table 4 3 Average distance and range for HIV 1 PR conformations determined for MDR769 apo and in the presence of inhibitors. Conformation Range () Average Distance () c urled/tucked 26 31 28 2 c losed 33.1 34.0 33.4 0 .3 semi open 35.2 37.2 36.5 0.6 wide open 39 46 42 3

PAGE 134

1 34 Figure 4 4 A) S ubtype B distance distribution profiles in the free and ligand bound state. B) MDR769 distance distribution profiles in the free and ligand bound state Semi open and c losed populations have flap distances of 37 ( solid line) and 33 ( dashed line), respectively. Highlighted in gray are inhibitors that can no longer efficiently close the flaps of MDR769. For MDR769, only 3 out of 10 inhibitors, LPV, DRV, and TPV, shi ft the population to >60% closed (Figure 4 4 B and 4 5B ) Six inhibitors retain the most probable distance at 35 37 (Figure 4 4 B). For APV, RTV or SQV, only 31 36% closed p o pulation is observed (Figure 4 5 B ), which is 40 60% less than that seen previou sly for subt ype B (Figure 4 5 C ). Likewise, ATV exhibited a 30% decrease in the shift to the closed population. Because MDR769 represents a clinical isolate from a patient failing extensive antiretro viral therapy, it is not surprising that many of the in hibitors lose the ability to induce flap closure. Within error, IDV and NFV induced similar degrees of flap closure in subtype B and MDR769 ( 20%). CA p2, a substrate mimic that serves as positive control, was found to have comparable flap closure to sub type B. Complete DEER data analysis for MDR769 is shown in Appendix A.

PAGE 135

135 Figure 4 5. A) Degree of flap closure measured as percentage occupancy of the closed state (% closed) in subtype B PR 34 B) Degree of flap closure measured as percentage occupancy of th e closed state (% closed) in MDR769. C) The difference in flap closure between MDR769 and subtype B PR Correlation of the | C | Parameter with IC 50 The relationship between induced conformational shifts detected with DEER spectroscopy and underlying biol ogical implications were examined by plotting previously reported half maximal inhibitory concentrations (IC 50 ) for MDR769 88 in logarithmic scale, against the percentage of closed state (Figure 4 6 A ). The correla tion between induced conformational shifts and IC 50 measurements agrees with X ray models 55, 88 where drugs that bind the HIV 1 PR active site pocket and inhibit viral propagation have multiple interactions with HIV 1 PR that stabilize the closed conformation. Results show that inhibitors with percentage closed (%closed) 68%, namely LPV, DRV and TPV, have excellent drug potency (IC 50 0.7 nM) to the HIV virus. DEER populations indicated that the distance distri bution widths are also narrower for these PIs, suggesting conformationally rigid flaps that are locked in onto the PI in the active site pocket, consistent with inhibitor bound MDR769 PR crystal structures that reveal multiple inhibitor flap interactions 55, 88 In contrast most PIs with %closed 32% have IC 50 in the 60 300 nM range. Currently, there is no available

PAGE 136

136 crystal structure for MDR769 bound to inhibitors NFV, IDV, RTV, SQV, APV or ATV; interestingly the D EER data for these inhibitors reveal conformational flexibility, which may complicate crystallographic resolution or even interfere with crystallization. Figure 4 6 A) Logarithmic plot of half maximal inhibitory concentration (IC 50 ) 88 against DEER % closed for MDR769. Inhibitors clustered into two distinct groups but ATV a nd APV appear to be outliers. B ) Plot of the log IC 50 fold change 88 versus the magnitude of % di fference in DEER closed population (| C |) between MDR769 and subtype B reveals a linear correlation Inhibitors that exhibited low % closed population in both MDR769 and subtype B PR (IDV and NFV) are excluded from analysis. Even though APV and ATV have I C 50 of 3 5 nM, DEER data suggest that these inhibitors may bind to MDR769 without promoting substantial flap closure. The presence of out l iers, APV and ATV, in Figure 4 6 A may arise because of the active site mutation, D25N that was introduced in the DEER variants This substitution is often incorporated in spectroscopic investigations 98, 99 because it imparts sample stability and homogeneity. Without inhibitor bound crystal structures, it is difficult to unders tand why APV and ATV are affected by the D25N substitution, but we speculate that the mode of binding may be altered by this substitution. Nevertheless, when the logarithmic fold change of IC 50 (MDR769 relative to wild type) a measure of drug resistance is plotted against the magnitude of | C |, defined

PAGE 137

137 earlier as % change in flap closure, ATV and APV now fit within a linear trend (Figure 4 6 B ) with a coefficient of determination (R 2 ) equal to 0.91. Here, however, NFV and IDV were excluded because of their weak ability to induce flap closure in both subtype B and MDR769. The weak effects of these two inhibitors is not surprising, given they have the largest wild type K d values, 60 67 times the K d of DRV (10 pM) 34 Finall y the linear trend in Figure 4 6 B sugg ests that aside from IC 50 the parameter | C | can be used to evaluate inhibitor effectiveness against emerging drug resistant variants or the effectiveness of novel inhibitors for drug potency against various HIV 1 PR variants. Other drug resistant varian ts or inhibitors can be used to further test and validate the proposed method. Conclusion s These results demonstrate the applicability of the | C | parameter as a criteria in the potency evaluation of new PIs and existing FDA approved inhibitors in the tre atment of patients infected with other HIV 1 PR subtypes or drug resistant variants. Interestingly, a linear correlation with the logarithmic fold change in IC 50 has been demonstrated for the MDR769 variant, implying that if the proposed method is furthe r validated to be applicable to other drug re sistant variants and subtypes, the | C | parameter could be utilized to complement established biological assay techniques, such as IC 50 Moreover, if handling high risk pathogens such as the HIV virus needs to be avoided, using the pulsed EPR approach could be a safer and less expensive alternative.

PAGE 138

138 CHAPTER 5 ANALYSIS OF HIV 1 PROTEASE BACKBONE FLEXIBILITY AND PROTEIN LIGAND INTERACTION DYNAMICS BY NMR SPECTROSCOPY Introduction The i nformation about local fle xibility and internal mobility of HIV 1 PR can be derived by nuclear magnetic resonance (NMR) spectroscopic methods. 223 In particular, these techniques employ spin systems belonging to proton heteronucleus bonds (e.g. 1 H 15 N, 1 H 13 C) as local probes of protein motion. The m ost common starting point for protein NMR is a 1 H 15 N heteronu clear single quantum coherence (HSQC) experiment wherein each amide proton pair in a protein appears as a peak in a two dimensional spectrum. If the protein is folded, the p eaks are usually well dispersed and distinct ; t h us the number of peaks in the spe ctrum match es the number of residues in the protein except when proline residues are present In this chapter, HSQC will be utilized to determine backbone chemical shift perturbations in HIV 1 PR as a result of acquiring mutations 39 or binding with a ligand 207, 224 Moreover, 15 N spin relaxation and nuclear Overhauser effect parameters will be employed to compare the flexibility and internal mobility of the multi drug resistant patient isolate MDR769 to subtype B PR. Protein Backbone Chemical Sh ift Assignment via Triple Resonance Experiments T he HSQC spectrum cannot be assigned to the corresponding amino acid residues in a protein without initially performing triple resonance experiments (3D NMR) 225 These experiments require that the protease is isotopically labeled with 15 N and 13 C. To perf orm backbone assignment, HNCA and HN(CO)CA spectra were usually acquired. Meanwhile, side chain information were often derived from

PAGE 139

139 CBCA(CO)NH (or HN(CO)CACB) and HNCACB (Figure 5 1 ) experiments. HNCACB is usually preferred over CBCA (CO) NH since the form er is 4.5 times more sensitive than the latter. Figure 5 1 Triple resonance experiments. The naming of the experiments describes the magnetization transfer, with the the nuclei chemical shifts not being mapped in paren theses The experiments are run i n pairs with one giving rise to correlations to both the residue itself (residue i) and the previous residue (residue i 1) and its partner experiment giving only the inter residue (i 1) correlation. Any experiment that contains the magnetization transfer s tep N C a gives rise to the two sets of correlations because of the similar size of the intra and inter residue coupling. Figure adapted from Cavanagh et al. 225

PAGE 140

140 15 N Spin Relaxation and Nuclear Overhauser Effect After the HSQC peak assignment is completed relaxation can then be measured for each prot on nitrogen pair. Relaxation refers to the process where non equilibrium spin order conditions caused by local magnetic field are allowed to return to the equilibrium Zeeman magnetization state. A c hange in NH bond axis orientation is known to cause rela xation of the amide 15 N spin. 223 In the absence of unpaired electrons (paramagneti sm) and quadrupole relaxation due to I > nuclei, there are two relaxation mechanisms : (1) spin lattice relaxation, where the system returns to its initial condition by transferring energy to neighboring molecules or the surroundings, and (2) spin spin relaxation, where the system returns to its initial condition by transferring their energy to surrounding nuclei. 223 Longitudinal relaxation time (T 1 ), wh ich is a measure by which the z magnetization (M z ) increases back to its equilibrium value M o is primarily determined by the spin lattice relaxation mechanism, but not the spin spin mechanism. Meanwhile, transverse relaxation time (T 2 ), which determines the rate at which magnetization component in the xy plane (M xy ) decays to its initial value of zero, is determined by both spin lattice and spin spin mechanisms. Another important rel axation parameter in protein NMR is the nuclear Overhauser enhancement (NOE), 226 which results from the transfer of spin polarization from one spin to another via cross relaxation. 15 N spin relaxation measurements (T 1 T 2 and NOE) were done previously for HIV 1 PR to acquire data for mo tions in the fast (picoseconds to nanoseconds) and intermediate (microseconds to milliseconds) time scales. 37, 207 In these studies, T 1 T 2 and NOE gave qualitative information about local flexibility and internal mobility of HIV 1 PR in the free and and ligand bound states

PAGE 141

141 The f ollowing is an overview of NMR relaxation measurements as described by Barbar and co workers 227, 228 Briefly local anisotropic motion in disorder ed segments of proteins can lead to longer T 1 and T 2 Lowest values of T 1 and T 2 can be obtained for residues with most restricted local mobility in non exchanging systems. Meanwhile, the steady state heteronuclear NOE can be measured by saturating the pro ton signal and monitoring the intensity change of the 15 N signal. This NOE intensity change depends on the heteronuclear cross relaxation rate. Aside from the NOE intensity, its magnitude can provide information on the mobility of the N H vector. Negative zero, and small positive values is tantamount to high internal mobility. In summary, when comparing picosecond to nanosecond motions of different conformations for the same residue or of different residues in the same conformation, greater flexibility leads to higher T 1 higher T 2 and lower NOE. Meanwhile, motions in the microsecond to millisecond regime (i.e. intermediate chemical exchange) give rise to smaller T 2 values. To evaluate the extent of exchange, T 2 values must be measured at two different field strengths. If exchange is present, T 2 would be shorter at higher field. Previous NMR Relaxation Studies on HIV 1 Protease NMR studies of the free HIV 1 PR showed that the flap region can move on a sub microsecond time scale, while the glycine rich flap tips are more flexible, moving on sub nanosecond time scale. 36 These NMR experiments have established the flexibility of the flap region, suggesting that closed, semi open and wide open conformations of the protease are in dynamic equil ibrium, with the semi open being prevalent in the apo state. 36, 38

PAGE 142

142 Katoh and co workers 207 conducted NMR studies on the binding of a substrate or inhibitor to inactive HIV 1 PR. Their 15 N spin relaxation measurements and subsequent model free analysis reveal restricted backbone motion in the flap region on the sub nanosecond timescale for the substrate bound protease. Meanw hile, residues 50 and 51 at the flap tips undergo slow dynamics on the conformational exchange timescale (millisecond to microsecond). Moreover, there is clear evidence that asymmetric ligands pack more tightly with one flap than the other. Protein Ligand Exchange Dynamics The a mide backbone chemical shifts are sensitive to changes in the local environment, such as inhib itor binding to a protein. Thus, t he exchange dynamics of ligands with proteins can be assessed by monitoring shifts and broadening of the HSQC resonances. The type of change in the HSQC resonances, i.e., shifts or broadenings, is dictated by the exchan ge rate (k ex ) between unbound and bound protein species relative to the differences in chemical shifts for the free and ligand bound s tate s As illustrated in Figure 5 2, 229 for 1:1 free to bound protein species, residues under slow exchange appear as two distinct resonan ces that represent the free and ligand b ound states. Meanwhile, intermediate exchange with a ligand results in linewidth broadening, leading to decreased resonance intensity and possible peak disappearance. Finally, residues under fast exchange have a resonance that appears midway of the ini ti al and final frequencies ; where the free and bound states merge into a single HSQC cross peak. One unique attribute of homodimeric proteins, such as HIV 1 PR, is the chemical shift degeneracy of residues for each monomer due to the C2 symmetry in the apo enzyme. However, upon binding with an asymmetric ligand, the residues for each

PAGE 143

143 monomer will sense disparate environments, leading to peak splitting for bound species in the HSQC spectra. In the case of a slow exchange coupled to the lost degeneracy for the bound protease, the HSQC cross peak for the affected residue can be split into as many as three separate resonances (two for each bound monomer and a degeneracy signal for the free protease). Figure 5 2. The effect of ligand exchange rate (k ex ) on reso nance intensity and chemical shift When the difference in the final ( f ) and initial ( i ) resonance frequencies is larger than the exchange rate (k ex << f i ) the ligand protein interaction dynamics described by the equilibrium P + L PL is slow re lative to the NMR timescale. Arrows indicate progress of ligand titration. For slow exchange, the intensity of the peak corresponding to a residue in the free (P) protein (with frequency i ) decreases while that for the ligand bound (PL) protein (with fr equency f ) increases. For intermediate exchange, k ex is comparable in magnitude to the frequency difference (k ex f i ) and linewidth ( ex ) broadening results because ex is a function of ( f i ) 2 Finally, for a ligand in fast exchange (k ex >> f i ) a peak shifts from the initial to the final frequency with almost unchanged intensity during the course of ligand titration. When 50% of the protein species is in the bound state, into the P and P L resonances For intermediate exchange, peaks disappear due to linewidth broadening. Lastly for fast exchange, the peak shifts to a frequency that is midway the P and PL frequencies ( i and f respectively).

PAGE 144

144 Materials and Methods HIV 1 Protease E xpression in M9 Media and NMR Sample Preparation DNA encoding E. coli codon optimized HIV 1 PR amino acid sequence of MDR769 lacking the K55C substitution ( i.e., K55) were cloned into pET 23a vector (Novagen, Madison, WI) under the control of the T7 promot er. The vector was transformed in BL21*(DE3)pLysS E. coli cells (Invitrogen, Carlsbad, CA) and grown in modified M9 minimal media with 15 NH 4 Cl (Sigma Aldrich, St. Louis, MO) as the sole nitrogen source. For NMR samples used in triple resonance experiment s, 13 C glucose is utilized as the only carbon source. Overexpression of HIV 1 PR was induced when optical density of the culture is 0.8 (measured as absorbance at 600 nm), by adding isopropyl D thiogalac toside (IPTG) to a final media concentration of 1 mM. Induction was al lowed to proceed at 37 C for 5 6 hours. HIV 1 PR was purified from inclusion bodies as described in Chapters 3 and 4. [U 15 N] MDR769 for HSQC titration experiments and [U 15 N, 13 C] MDR769 HIV 1 PR for triple resonance NMR were prepare d at 40 M homodimer in 2 mM D 3 NaOAc buffer at pH 5.0 with 10% D 2 O and 0.1 mM DSS as internal reference. NMR Data Acquisition and Processing All NMR spectr a for MDR769 were obtained at 29 3 K with a Bruker 5 mm TXI Cryoprobe Avance II system operating at 600 MHz at the University of Florida AMRIS Facil i ty. DSS was used as an internal standard for referencing all proton chemical shifts and as an external standard for referencing nitrogen and carbon chemical shifts. The NMR data were processed using NMRP ip e 230 and analyzed with Sparky (Goddard and Kneller, Sparky 3, UCSF, San Francisco).

PAGE 145

145 1 H 15 N HSQC Inhibitor Titration Experiments Stepwise add ition of inhibitors or substrate into 40 45 M HIV 1 PR to a final concentration of 1:1.protease to inhibitor ratio was performed. Protease to inhibitor ratio at 1:1 was chosen for peak counting. At this condition, comparable amounts of free and bound pr otease species are present, thereby giving a more observable peak pattern change. Backbone Chemical Shift A ssignments Backbone resonance assignments were carried out using 2D 1 H 15 N HSQC and 3D HNCACB, CBCA ( CO ) NH, HNCA and HN(CO)CA experiments. Relaxati on Experiments 15 N T 1 values were measured using relaxation delays of 16, 64 128, 256, 384, 512, 640 768, 896 and 1024 ms at 20C, whereas T 2 measured using standard Carr Purcell Meiboom Gill (CPMG) pulse train with relaxation delays of 8, 17, 35, 52, 69, 86, 104 and 121 ms. 15 N { 1 H} NOE experiments were performed using a water flip back sequence 231 NOE values were measured by taking the ratio of peak intensities fr om experiments performed with and without 1 H presaturation. Errors associated with NOE values are determined by measuring the standard deviation of triplicate runs. Peak heights measured in the processed spectra from T 1 and T 2 experiments were fitted with a two parameter exponential function to extract relaxation rates R 1 and R 2 by using r elax software 232 234 available from http://www.nmr relax.com/ Errors in R 1 and R 2 we re determined from in Monte Carlo simulations 235

PAGE 146

146 Results and Discussion 1 H 15 N HSQC of Subtype B Single point Mutation Variants To assess the effect of single point m utations to the local environment of individual HIV 1 PR residues, 1 H 15 N HSQC spectra were obtained for D30N, M36I and A71V (Figure 5 3 ) Chemical shift perturbations were measured using Equation 5 1, where 1 H and 15 N are the changes (in ppm) along the proton and nitrogen axi s, respectively. (5 1) The perturbation plots show that the primary D30N substitution did not result in substantial chemical shift changes of distal residues. This indicates that such a mutation in the active site pocket do not propagate to other regions of the protease. The largest chemical shift change other than the mutation site is observed in the adjacent D29 residue (1.3 ppm), as expected. The chemical shift perturbation for most residu es in D30N is less than the average change ( 0.4 ppm). In contrast, secondary mutations such as M36I and A71V exhibited chemical shift change in more residues than the primary D30N substitution. M36I, a substitution in the hydrophobic core near the flap h inge, resulted in chemical shift perturbations larger than 0.5 ppm in residues situated at the flap and cantilever regions of HIV 1 PR. Meanwhile, the A71V mutation at the cantilever transmitted its effects to several residues in the active site pocket an d to the dimerization region. The largest average chemical shift change due to the A71V substitution is observed in residues 65 to 75 in the cantilever (0.6 to 1.4 ppm) and several residues in the helix dimer interface, particularly those at positions 2, 93, and 98 (0.6 to 0.9 ppm).

PAGE 147

147 Figure 5 3 1 H 15 N HSQC spectra for D30N (A), M36I (B), and A71V (C) in blue overlaid with wild type subtype B (red); and corresponding plots showing chemical shift perturbation (D to F) as a function of residue position.

PAGE 148

148 Positions 36 and 71 are situated in protein regions that are highly dynamic during flap opening and closing. It is not surprising that single point mutation at these positions could have a dramatic effect on the local magnetic field environment of neighb oring residues and could transmit their effect to residues that are distal from the mutation site. Backbone Chemical Shift Assignment of MDR769 For subtype B single point mutation variants such as D30N, M36I and A71V in the previous section, assigning th e HSQC crosspeaks to the corresponding HIV 1 PR residues could be accomplished by comparing the spectrum to that of published wild type (WT) peak assignment This is quite straightworward since most of the peaks have not substantially shifted relative to WT resonances, and could be easily identified. However, for a different HIV 1 PR variant such as MDR769, where there are 17 residue substitutions with respect to published subtype B HSQC spectrum. 38 Thus, peak a ssignment by simple comparison of MDR769 to the subtype B reference spectrum is inaccurate because of large chemical shift perturbations attributed to the numerous residue substitutions. In such situation, performing triple resonance experiments for backbo ne 1 H, 15 N and 13 C chemical shift assignment of a protein is needed. For MDR769, 3D HNCACB, CBCA ( CO ) NH, HNCA and HN(CO)CA experiments were performed to accomplish this task. These experiments come in pairs, where one gives only inter residue correlations while the other give both intra and inter residue correlations. For instance, data from CBCA(CO)NH reveal correlations of C and C of a residue and the amide of the succeeding amino acid showing two peaks in one of the 2D slices in the 3D data Meanw hile, HNCACB gives the same inter residue information as CBCA(CO)NH, but

PAGE 149

149 also provides additional intra residue correlations of the C and C with its own amide, showing a total of four crosspeaks in the 2D slice. This process leads to the identification of succeeding residues along the polypeptide chain; hence, complete backbone chemical shift assignment can be accomplished. The backbone chemical shift assignment was made for 88 out of the 92 HSQC resonances in MDR769 236 This variant has seven proline residues, which do not have amide resonances, while 4 residues (L5, I15, E21 and A82) were not assigned due to spectral overlap. Backbone c hemical shift as signments for MDR769 were deposited in the BMRB database at Madison, WI (http://www.bmrb.wisc.edu/) with accession number 18138 Figure 5 4 shows the 1 H 15 N HSQC of MDR769 with peak assignments. To show the chemical shift perturbations due to the mutation s in MDR769, its HSQC spectrum is overlaid to that of subtype B (Figure 5 5A). The corresponding chemical shift perturbation plot is shown in Figure 5 5B and the residue positions wi th largest chemical shift changes were highlighted in the HIV 1 PR ribbon diagram in Figure 5 5C. 1 H 15 N HSQC Ligand Titration Experiments for MDR769 In Chapter 4, inhibitor induced flap closure using 9 FDA approved inhibitors and a substrate mimic have been illuminated using pulsed EPR technique called DEER spectrosocpy Howev er, this method requires that the sample be frozen to cryogenic conditions to prolong phase memory time, which is required to acquire sufficient DEER echo signal before the dipolar evolution curve completely decays. It is assumed that the initial flash fr eezing of the sample in liquid nitrogen is fast enough to trap the conforma tions present at room temperature. To validate the assumption that the conformational populations in the glass state are the same as those present at room temperature

PAGE 150

150 ligand titr ation was performed using the same inhibitors and substrate mimic u tilized in DEER experiments and 1 H 15 N HSQC spectra were acquired at 0.5:1 and 1:1 inhibitor to protein ratio. All HSQC titration spectra of MDR769 are shown in Appendix C. Figure 5 4. 1 H 15 N HSQC spectrum of HIV 1 PR MDR769 at 6 00 MHz and 29 3 K. Peaks are labeled according to one letter amino acid codes and to position in the protein sequence. 88 out of 92 p eaks have been assigned. Resonances corresponding to L5, I15, E21 and A82 were not assigned probably due to spectral overlap Cro ss peaks assigned with an asterisk (*) are side chain amide resonances

PAGE 151

151 Fig ure 5 5 A) Overlay of 1 H 15 N HSQC spectra of subtype B (blue) and MDR769 (red). B) D ifferences in the backbone amide chem ical shifts of HIV 1 PR ( MDR769 versus s ubtype B ) are the differences in chemical shifts (in ppm ) for the individua l 1 H and 15 N atoms, respectively Mutation sites are shown as red open circle s C) The protease molecule is colored according to the chemical shifts differences in B Mutation sites are rendere d as spheres.

PAGE 152

152 After completing the 1 H 15 N HSQC peak assignment for MDR769, as discussed in the preceding section, residue specific changes in the HSQC spectra can be monitored at increasing co ncentrations of inhibitor or substrate during the course of the titration In particular, the changes in the spectra reflect the type of ligand exchange dynamics with the protein. There are two types of protein ligand interaction dynamics observed: slow exchange and inter mediate/fast exchange. The dynamics for slow exchange is exhibited by ligands with relatively long residence time in the binding pocket of the protease and is consistent with tight binding inhibitors. For MDR769, tipranavir (TPV), lo pinavir (LPV) and darunavir (DRV) belong to this classification. Representative spectra for TPV and DRV are shown in Figure 5 6. Because all of the FDA approved protease inhibitors utilized in our investigations are asymmetric, inhibitor binding to the p rotease active site pocket abolishes the homo dimeric chemical shift degeneracy of residues in the free protease. Presence of the inhibitor in the binding cleft situates homodimeric residue pairs in disparate magnetic field environments, thereby resulting in peak splitting into two resonances. Moreover, the equilibrium condition associated with a 1:1 protease to inhibitor concentration ratio favors a mixture of inhibitor bound and free protease in solution. Thus, a third peak representing the degenerate u nbound state can also be observed. Due to peak splitting, inhibitors that exhibit slow exchange dynamics have HSQC spectra peak count that is higher than apo. On the other hand, intermediate/fast exchange is exhibited by ligands with a relatively short residence time in the binding pocket; hence, weakly binding inhibitors. For MDR769, this is demonstrated by the inhibitors, indinavir (IDV), nelfinavir (NFV),

PAGE 153

153 atazanavir (ATV), amprenavir (APV), saquinavir (SQV), and ritonavir (RTV). Note that in subtype B, SQV and RTV ligand binding dynamics is classified as slow exchange (Table 5 1). This suggests that drug pressure selected mutations present in MDR769 altered the ligand exchange dynamics for these inhibitors. Figure 5 6 shows the HSQC spectra for NF V and ATV. In the intermediate exchange limit, the linewidth of resonances is a function of ( f i ) 2 where f and i are the frequencies of the initial and final positions of a crosspeak respectively Peak broadening results in substantial intensity decrease until several resonances disappear in the bound HSQC relative to the free protease spectrum. Because of peak disappearances, the HSQC spectrum peak count for HIV 1 PR in the presence of weakly binding inhibitors is less than that in the apoenzy me. Table 5 1 summarizes the ligand exchange dynamics of the 9 FDA approved inhibitors utilized previously for subtype B and in this study. Comparison of NMR and Pulsed EPR Results The main motivation for thet inhibitor titration and solution NMR experim ents was to corroborate pulsed EPR results shown in Chapter 4, where conformational shifts in favor of the closed state is observed after adding inhibitors or substrate mimic to HIV 1 protease. Table 5 2 summarizes NMR and pulsed EPR results for MDR769 an d sub type B proteases, where HSQC spectra peak count and percentage occupancy of the closed state (% closed) are reported. From the data presented, it is apparent that if an inhibitor efficiently induces flap closure, more crosspeaks are observed in the HSQC spectra. This implies that if there is an increase in % closed or % closed change in the inhibitor bound protease relative to apoenzyme, the inhibitor likely undergoes slow exchange in the binding cleft.

PAGE 154

154 Figure 5 6. Representative 1 H 15 N HSQC sp ectra for 1:1 inhibitor to protease samples illuminating two types of exchange dyn amics in HIV 1 PR: inter mediate/fast a nd slow Inhibitors that have relatively short residence time in the binding cleft exhibit linewidth broadening due to intermediate lig and exchange. Consequently, there are missing resonances for nelfinavir and atazanavir bound protease (A and B) due to the decreased peak intensity. In contrast, inhibitors with relatively long residence time in the binding pocket such as tipranavir an d darunavir (C and D) undergo slow exchange relative to the ligand structure that breaks the homodimer chemical shift degenerac y in the bound state Note that the third crosspeak is f rom the unbound protease species. Table 5 1. Comparison of ligand exchange dynamics in MDR769 and subtype B PR HIV 1 PR variant Exchange dynamics Fast/Intermediate exchange Slow exchange Subtype B IDV, NFV, ATV, APV RTV, SQV, LPV, DRV, TPV MDR769 ID V, NFV, ATV, APV, RTV, SQV LPV, DRV, TPV

PAGE 155

155 The NMR results are consistent to earlier puls ed EPR investigations because an inhibitor will have longer residence time in the binding pocket and exhibit slow exchange dynamics (higher HSQC peak count) if there i s a higher population of the closed state (Figure 5 7) From a structural viewpoint, flap closure prevents inhibitor escape from the binding pocket, which explains the longer residence time. This delineates the possibility that multiple hydrogen bonding a nd/or van der Waals interactions may be present between the inhibitor and the binding pocket that promotes the closed state and consequently favors longer inhibitor residence time in the binding cleft. In fact, this is supported by the X ray crystal struc tures of MDR769 complexed with TPV, LPV and DRV. 55, 88 Table 5 2 Summary of HSQC peak count and DEER percentage of closed population for HIV 1 PR variants with inhibitors. HSQC peaks are initially selected by us ing the automatic p ick peaking function in Sparky. Only peaks with S/N ratio > 5 are counted. Peak counting uncertainty results from partial peak overlap. HIV 1 PR Sample HSQC Peak Count DEER % Closed Subtype B MDR769 Subtype B MDR769 Apo 87 87 3 18 IDV 73 65 14 18 NFV 73 61 14 20 ATV 77 63 53 11 APV 65 83 76 36 LPV 109 9 3 84 71 DRV 96 101 87 68 RTV 106 7 3 90 31 TPV 110 1 21 91 84 SQV 126 80 93 32 CA p2 115 111 80 88

PAGE 156

156 Figure 5 7. P lots of HSQC peak count against DEER% closed In subtype B, strong flap closure is demonstrated by SQV and RTV; and moderate for ATV. For the MDR769 variant the inhibitors SQV, RTV ATV and APV exhib ited weak flap closure Error bars for DEER% were estimated to be 3 5%, and 3 for HSQC peak number. The horiz ontal dashed line marks the number of HSQC peaks for the subtype B apo protease, and serves as a de marcation line exhibit <20% and >50% flap closure (vertical lines), respectively. In contr ast, inhibitors that exhibit poor flap closure efficiency (low % closed) such as NFV, IDV, RTV, SQV, ATV, and APV, demonstrated intermediate/fast exchange dynamics. If the degree of inhibitor induced flap closure is low, there is higher probability tha t the inhibitor can escape from the binding pocket and its residence time will be short. Consequently, the exchange dynamics of these inhibitors would be in

PAGE 157

157 the intermediate/fast regime relative to the NMR timescale. Moreover, since the inhibitor could e asily escape from the binding cavity, it is possible that critical interactions that favors inhibitor binding and flap closure are lost, consistent to the expanded active site model of MDR769. 55 Relaxation Measure ments for MDR769 To understand the dynamics of free MDR769 in solution and how drug pressure selected mutations alter backbone dynamics and consequently ligand binding, 15 N spin relaxation parameters R 1 R 2 and NOE were derived for this HIV 1 PR variant. Note that R 1 and R 2 (in ms 1 ) are the reciprocals of spin lattice relaxation time (T 1 ) and spin spin relaxation time (T 2 ), respectively. Figure 5 8 shows the residue specific R 1 R 2 and NOE parameters for MDR769 at 600 MHz, 20 C. Results show conform ational rigidity of the flap elbow and around the T80s loop in MDR769 relative to subtype B. This is evident from the higher 15 N NOE for several residues within these regions, particularly G40 and T80. However, the NOE values for these residues are sti ll substantially below the average, indicating that internal motions in the sub nanosecond timescale are still present. In contrast to the NOE data, the R 1 values for MDR769 are similar to that of subtype B. F or MDR769, the measured R 1 values indicate t hat flap motions on the sub nanosecond timescale are comparable to the T 1 values that arise from the overall tumbling in the 10 ns timescale, as previously observed in subtype B PR. 38 Meanwhile, R 2 values for flap elbow and T80 residues are lower than average but slightly higher than subtype B The R 2 measurements for other residues in MDR769 are comparable to that of subtype B.

PAGE 158

158 Note that the information obtained from these relaxation measurements provides limi ted information on protein dynamics unless model free analysis 237, 238 is performed for MDR769. This would require obtaining another R 1 R 2 and NOE data set at a higher field, say 700 MHz. Figure 5 8. 15 N spin relaxation parameters R 1 R 2 and NOE for MDR769 (green squares ) and subtype B (red circles) apo at 600 MHz, 20 C (A C), and their corresponding change (D F) in MDR769 with respect to subtype B as a function of residue position Residues situated at the flap, elbow and T80 loop are marked with a vertical dashed lin e (blue)

PAGE 159

159 Conclusions From the NMR s tudies, we have found th at secondary mutations M36I and A71V on the periphery of HIV 1 PR give rise to substantial chemical shift perturbation of more residues when compared to the primary mutation D30N. Results from 1 H 15 N HSQC inhibitor titration experiments indicate that inhibitors utilized in our investigations exhibit two types of ligand exchange dynamics, namely slow and intermediate/fast exchange. Slow exchange is characterized by peak splitting in the HSQC spect ra due to longer inhibitor residence time in the binding pocket relative to the NMR timescale, and coincides with higher inhibitor induced flap closure efficiency from pulsed EPR results. In contrast, intermediate/fast exchange manifests as peak disappea rances In the HSQC spectrum due to short inhibitor residence time and coincides with poor inhibitor induced flap closure. This evokes the hypothesis that flap closure induction is coupled to protease inhibitor interaction dynamics, whereby multiple interact ions of the inhibitor with the binding pocket stabilize the closed state and prolongs inhibitor residence time in the binding cleft Finally, preliminary 15 N spin relaxation measurements performed on MDR769 indicate conformational rigidity of the flap e lbows and the T80 loop. Model free analysis has to be performed, which requires obtaining an additional relaxation data set at a higher field, to derive residue specific order parameters.

PAGE 160

160 CHAPTER 6 CHARACTERIZATION OF HIV 1 PROTEASE BY MASS S PECTROMET RY Introduction Protein purity and identity are often determined by SDS PAGE, reverse phase high performance liquid chromatography (HPLC) and N terminal Edman sequencing. However, these conventional techniques are incapable of providing accurate molecular weight. Analysis of the intact protein by mass spectrometry (MS) can be accomplished either by electrospray ionization (ESI) or linear mode matrix assisted laser desorption ionization (MALDI) MS. The degree of protonation, which depends on the ionizatio n process or type of mass analyzer utilized, dictates whether singly or multi charged ions are formed. MALDI often generates singly charged ions while ESI produces multi charged species. The detemination of intact protein molecular weight is not enough to differentiate isobaric proteins. Traditional approach for protein characterization involves enzymatic cleavage, HPLC separation and subsequent N terminal Edman sequencing of the peptides. 239 However, these techniques a re labor intensive and incompatible to proteins with N terminal modification. Peptide mass fingerprinting by MS analysis can be performed after enzymatic or chemical digestion of a protein. Trypsin is often the enzyme of choice for protein digestion becau se of its high specificity in terms of peptide bond hydrolysis. The observed masses can be compared with in silico protein digest that can be generated from NCBI 240 and SwissProt protein 241 knowledgebases. MALDI time of flight (TOF) MS in the reflectron mode is a common method of choice to analyze the tryptic peptides. Unlik e the linear mode employed for intact

PAGE 161

161 proteins, the reflectron mode utilizes ion mirrors 242 to allow the delayed arrival of faster ions into the detector. As a result, initial ion velocity variation s are compensated, allowing ions with the same mass to charge ratio ( m/z ) to reach the detector within a narrow time window. Furthermore, the reflectron mode increases the effective flight length and enhances TOF MS instrument resolution. Consequently, m onoisotopic peaks up to m/z 3000 can be identified. Peptides that have the same amino acid residues but different sequences cannot be distinguished by MALDI TOF MS, unless multiple enzymes with high specificity are used. Validation of the amino acid seque nces can be accomplished using on line liquid chromatography MS (LC MS n ). Peptide MS/MS fragments are generated from a n isolated parent peptide ion using collision induced dissociation (CID). Fragmentation for ion t rap mass analyzers is that for triple quadrupoles is performed After separating the peptide fragments by HPLC, the eluent is vaporized into fine spray droplets upon applying 3 5 kV onto the electrospray needle. Uneven fission of original droplets via coloumbic repulsion coupled to inert drying gas (e.g., nitrogen) desolvation yield s smaller droplets. The fission desolvation process continues until or multi roduced via a heated capillary into the mass spectrometer and are led to the ion trap after passing through lenses and skimmers. Detection of charged species is accomplished by gradually increasing the radio frequency (RF), resulting in the ejection of ion s in the order of increasing m/z Ions then l eave through the end cap towards the dynode and electron multiplier.

PAGE 162

162 MS/MS or product ion spectra are acquired by sequentially ejecting ions, except the parent ion, out of the trap via RF amplification and reso nant excitation. CID is utilized to fragment the isolated ions at optimum voltage, followed by detection through RF scanning. The acquisition method is usually programmed such that MS and MS 2 modes occur in succession. Spectra from the latter are collec ted either in selected reaction monitoring (SRM) or data dependent (DD) mode. SRM is performed using a pre determined list of parent ions and is often used for quantitative analyses. 243 Meanwhile, the DD mode enables selection of the two most intense ions in the full MS scan. The individual ions are isolated and fragmented in the succeeding two scans in the three scan cycle that is iterated throughout the chromatogram. CID fragmentation through amide bond cleavage is most common in ion traps. Usually, the y ions have higher intensities than the b ions. The latter frequently undergo structural rearran gement or dehydration (i.e. water elimination). The amino acid sequence of the polypeptide can be deduced from the series of ions detected. Fortunately, proteomics programs are commercially available or from the internet thereby facilitating de novo se quencing or sequence comparison against in silico data for a known peptide. This chapter focuses on the analysis of MTSL labeled protease in aqueous sodium acetate solutions via direct inj ection electrospray ionization time of flight mass spectrometry ( ESI TOF MS) 244 and MALDI TOF/TOF MS. 245 The analysis of intact nitroxide labeled HIV 1 PR by ESI TOF MS was found to be an essential step for verifying homogeneous spin labeling of protein samples and molecular we ight deter mination. Me anwhile, the HIV 1 PR variants were also analyzed for autoproteolytic

PAGE 163

163 fragments using MALDI TOF/TOF MS to establish protein shelf life that is needed for establishing the repeatability of the DEER conformational studies. A mino aci d substitutions and site specific labeling of HIV 1 protease were confirmed via tandem mass spectrometry. MALDI TOF/TOF was performed to verify D30N, M36I and A71V site directed mu tations, while LC ESI I ontrap MS n was done to confirm the attachment of the MTSL label at the C55 position. Materials and Methods Determination of Intact Protein Average Molecular Weight by ESI TOF MS MTSL labeled subtype B variants with D30N, M36I and A71V amino acid substitutions were reconstituted in 2 mM sodium acetate and c oncentrated to 50 M homodimer prior to positive mode ESI TOF MS on an Agilent 6210 instrument. Some of the spin labeled variants were re analyzed after several months to check the stability of the MTSL label. Data analysis was performed using MassHunter software (Agilent). Theoretical masses for the H IV 1 PR variants were calculated at 1000 resolution. Meanwhile, [U 15 N, 13 C] MDR769 PR used in 3D NMR experiments was reconstituted in 20 mM deuterated NaOAc and concentrated to 80 M homodimer before colle cting ESI TOF MS. In solution Trypsin Digestion and MALDI TOF MS Analysis In solution t rypsin digestion was done in a 3 0 C water bath for 3 h using enzyme from New England Biolabs Aliquots were collected every hour for an active (D25) subtype B K55 cons truct and MTSL labeled inactive (D25N) subtype B variants with point mutations D30N, M36I and A71V Control samples were heated without the enzyme to determine thermal degradation products. The peptides were mixed with

PAGE 164

164 cyano 4 hydroxycinnamic acid (CH CA) and trifluoroacetic acid (TFA), spotted on a plate, and analyzed using an AB Sciex TOF/TOF 5800 instrument in positive (+) and negative ( ) modes. Ions of interest underwent TOF/TOF MS and de novo sequencing. Data processing was performed using Data Explorer software (AB Sciex). HPLC ESI Ion trap MS Analysis of Tryptic Digests HPLC ESI ion trap MS n analysis of tryptic peptides was performed on a Thermo Finnigan LCQ Classic equipped with a regular electrospray, and operated in positive mode by Dr. Jod ie Johnson. Briefly, r everse phase chromatography was done on a Waters XTerra C18 column (3.5 mm x 2.1 mm x 150 mm) at a flow rate of 0.15 mL/min, using 0.5% acetic acid in water as mobile phase A and 0.5 % acetic acid in methanol as mobile phase B. Dat a were processed using Xcalibur software (ThermoScientific). All HPLC grade solvents were from Honeywell Burdick & Jackson, while the ACS grade acetic acid was purchased from Fisher Scientific. Results and Discussion Determination of MTSL Spin labeling Ef ficiency by ESI TOF MS Continuous wave electron paramagnetic resonance spectroscopy (CW EPR) was often used to confirm successful spin labeling before analyzing the protease variants for conformational sampling by double electron ele ctron resonance (DEER ) spectros copy. However, CW EPR does not give an approximate ratio of labeled and unlabeled protease. To monitor the degree of MTSL labeling in HIV 1 PR, intact protein samples were analyzed via ESI TOF MS. Figure 6 1 shows the optimization of conditions for the MTSL labeling of subtype B HIV 1 PR. In panel A, spin labeling is poor as shown by ion masses corresponding to the unlabeled mon omer and disulfide linked dimer

PAGE 165

165 The formation of a disulfide linked dimer is expected for a poorly labeled protein b ecause the labile cysteine resid ues eventually form a disulfide bridge in solution in the absence of a reducing agent Increasing the incubation temperature to 25 C (see Figure 6 1 B and C) improved the labeling efficiency, as exhibited by lower ion i ntensities for the unlabeled monomer and the disap pearance of ions for the disulfide linked dimer. Figure 6 1. ESI TOF MS reveals optimized conditions to achieve homogeneous spin labeling of HIV 1 protea se. A) MTSL labeling for 12 h and 4 C using 2 fol d molar excess of MTSL showed ions for the unlabeled HIV 1 PR monomer (red asterisk) and disulfide l i nked dimer (blue asterisk). B C) Incubation of HIV 1 PR with two and four fold mo lar excess of MTSL for 12 h at 25 C

PAGE 166

166 Average Molecular Weight by ESI T OF MS and Sequence Coverage by MALDI TOF MS Multiple charging of proteins in ESI TOF MS result ed in the presence of different charge states for the same protein. Th e average molecular weight of proteins were derived from each charge state (n = 6 14) and then the average molecular weight with standard deviation are reported Meanwhile, MALDI TOF MS and subsequent tandem MS analysis of tryptic digests verif ied peptide sequences via de novo sequencing and provided sequence coverage for each protein. A sum mary of the average molecular weight via ESI TOF MS, and sequence coverage from MALDI TOF MS and tandem MS analysis of tryptic peptides is shown in Table 6 1 E SI TOF MS gives average molec ular weights within 0.2 Da of theoretical values; thus, intact pr oteins that are 1 Da apart are easily differentiated. M eanwhile, using MALDI TOF MS in both positive and negative mode s increased amino acid percent coverage (%AA) of trypsin digested samples to 94 %. Table 6 1 Molecular weights of intact protease via ESI TOF MS and percent amino acid coverage of trypsin digests via MALDI TOF MS Construct Molecular Weight %AA Coverage Theoretical Observed Positive Mode Negative Mode Total Subtype B 25/55 a 10728.69 10728.68 0.11 86 81 98 Subtype B b 10886.95 10887 .00 0.03 88 90 100 D30N b 10885.96 10885.92 0.12 88 73 100 M36I b 10868.90 10868.72 0.11 88 79 100 A71V b 10914.98 10915.00 0.12 66 87 94 D30N/M36I b 10867.91 10867.81 0.10 88 73 100 D30N/A71V b 10914.00 10914.03 0.15 88 73 100 M36I/A71V b 10896.94 10897.05 0.05 82 66 94 D30N/M36I/A71V b 10895.96 10895.91 0.11 88 73 100 MDR769 K55 c 10637.47 10637.50 0.09 86 65 98 15 N, 13 C MDR769 c 11248.91 11242.46 0.15 86 65 98 a) D25/K55, active construct with no labeling site; b) K55C with MTSL; c) D25N /K55 inactive construct with no labeling site

PAGE 167

167 Figure 6 2. Sequence coverage of subtype B HIV 1 protease variants Highlighted in yellow and orange are peptide sequences confirmed via MALDI (+) and MALDI ( ), respectively. Peptide fragments identified in both modes are shaded in blue Missing fragments are not highlighted Shelf Life Studies of MTSL labeled HIV 1 Protease The s tability evaluation of MTSL labeled HIV 1 PR is important in pulsed EPR studies that utilize active protease variants or when repeat measurements of the same sample are needed after long storage in the freezer. It was also used in this study for error calculations and repeatability assessment s via ESI TOF MS of MTSL labeled HIV 1 PR samples that were stored at 20 C over the span of 8 months, and at 4 C for 3 months. The ESI TOF MS r esults show that t he nitroxide label was intact for all inactive (D25N) subtype B variants tested even after three months at 4 o C. Only one out of eight variants showed MTSL label cleavage after 8 months of storage at 20 C (Figure 6 3). These results suggest that certain amino acid substitutions could render the disulfide linkage to the nitroxide label less stable perhaps by making the disulfide bond susceptible to reduction. The stability of a ctive (D25) sub type B HIV 1 protease variant with stabilizing mutations Q7K, L33I and L63I (B sa ) was also evaluated using MALDI TOF MS

PAGE 168

168 Mildner and co workers 97 reported three cleavage sites in HIV 1PR: between (1) Leu5 and Trp6, (2) Ile33 and Glu34, and (3) Ile63 and Ile64. In Figure 6 4, MALDI TOF MS showed s ubstantial amount of the expected autolysis fragments for B sa that was stored at 20 C for a week. Ions at m/z 3449 and 3713 correspond to Glu34 Ile63 and Ile 64 Phe 99 peptides respectively The observed i ons below m/z 2000, after incubation at 30 C for at least 2 h were attributed to thermal degradation Figure 6 3 Deconvoluted ESI TOF MS spectrum of the triple mutant, after 8 months of storage at 20 o C, shows disulfide linked dimer form ation ( peaks marked with an asterisk) after the MTSL label fell off. The results suggest that the active protease samples have to be analyzed immediately after isolation. Moreover, the a ctive enzyme may not be compatible for DEER analysis in the apo for m because thermal degradation could occur even at room temperature during spin labeling. A possible solution to this problem is to perform the nitroxide labeling at 4 C in order to slow down autoproteolysis However, as shown in Figure 6 1, homogeneous s pin labeling could not be achieved at low temperature even after overnight incubation.

PAGE 169

169 Figure 6 4. MALDI TO F MS of active subtype B (D25) variant after storage at 20 C for a week ( c ontrol) and after heating at 30 C for up to 3 h Tracking Amino Aci d Substitutions by MALDI TOF MS L ocating the sites of a mino acid substitutions in HIV 1 PR required trypsin digestion and sequencing of the modified peptides. For instance, t he MALDI TOF MS spectra in Figure 6 5 illustrate the change in pe ptide mass due t o two mutations, D30N and M36I, with respect to subtype B (WT). The Glu21 Arg 41 peptide at m/z 2230.9 in WT was reduced to m/z 2213.1 in the M36I construct and to m/z 2212.1 in D30N/M36I.

PAGE 170

170 In contrast, all other peptide mass to charge ratios remained the same with variable relative abundances compared to the base peak at m/z 1677.0. Also shown in Figure 6 5 is the labeled peptide with the most intense ion at m/z 1407.6, corresponding to [M+2H] + due to the protonation of the nitroxide radical in the pres ence of TFA. For those variants that contain the A71V amino acid substitution, the m/z 1677 .0 base peak for the Ala71 Arg87 WT peptide increased to m/z 1705.0 (data not shown). Figure 6 5. The MALDI TOF MS spectra of WT subtype B M36I and D 30N/M36I tr yptic digests all show the same isotope pattern for the (M46 R57)+K55R1 peptide beginning at m/z 1406.6 ( ). The insets help emphasize the change in m/z after nitroxide labeling and amino acid substitutions with respect to E21 R41 in subtype B

PAGE 171

171 De Novo S equencing of Modified Peptides by MALDI TOF MS 2 The positions of the modified amino acids were confirmed via tandem MS. As shown in Figure 6 6 the MALDI TOF MS 2 spectra of m/z 2213.0 and 2211.9 peptides showed y 12 to y 20 ions that support the D30N mutat ion in D30N/M36I. The product ions below m/z 1250 are the same for the two variants No significant b ions were observed because the proton from the [M+H] + ion was sequestered by the arginine residue at the C terminus. These tandem MS results validate t he chemical shift perturbations in 1 H 15 N HSQC NMR spectra and changes in inter flap DEER distance profiles that were attributed to the M36I and D30N substitutions. Figure 6 6 MALDI TOF MS 2 spectra of m/z 2213.0 [M+H] + ion from M36I and m/z 2211.9 f rom D30N/M36I showed product ions that differ after the 10 th amino acid. The ions below m/z 1250 were identical for both peptides but differ in rel ative abundance Mutations in the peptides were incorporated via Phyre v2.0 and the models are rendered in PyMOL v1.5.

PAGE 172

172 Confirmation of Site specific Spin labeling in HIV 1 Protease via ESI Ion trap MS n In this section, the validation of site specific MTSL labeling at C55 is described. The addition of MTSL label (R1) increased the molecular weight of the unlabe led 1222 Da peptide by 184 Da This increase in mass was not observed in other tryptic peptides. The ESI Ion trap MS n spectra (Figure 6 7) of the corresponding m/z 1406.6 [M+H] + ion showed cleavage of the nitroxide label before and after the disulfide b ond. De novo sequencing was done using the MS/MS spectrum of the m/z 704.1 [M+2H] 2+ ion. As shown in Figure 6 8 the y 3 to y 11 ions verified the MIGGIGGFIXVR sequence and the position of Cys+R1. In contrast, the a and b ions did not differentiate the labeled and unlabeled peptides. Figure 6 7 The ESI TOF MS spectrum (top) shows the [M+H] + ion of the nitroxide labeled MIGGIGGFIXVR peptide at m/z 1406.6 together with several sodium adducts. The tandem mass spectra (middle and bottom) show loss of R1 with and without the disulfide bond at m/z 1188.9 and 1253.6, respectively

PAGE 173

173 Figure 6 8 The ESI TOF MS spectrum (top panel ) shows the doubly charged ion of the nitroxide labeled MIGGIGGFIXVR peptide at m/z 704.1 together with several sodium ad ducts The tandem mass spectrum (bottom panel ) shows product ions that confirmed the presence of amino acid X (Cys+R1). Conclusion s This chapter shows that ESI and MALDI TOF MS n are essential in establishing the integrity of nitroxide lab eled HIV 1 protease samp les prior to EPR investigations. Aside from its ability to estimate spin labeling efficiency, ESI T OF MS also provide d information regarding the presence o f contaminants in the sample. Me a nwhile, the MALDI TOF mass spectra showed that the MTSL labeled in active subtype B variants were stable at 20 o C, with the exception of the D30N/M36I/A71V variant that formed the disulfide linked dimer after 8 months. T he active subtype B enzyme showed some autoproteolytic fragments at 20 C. The d egradation of activ e protease was accelerated at 30 C based on MALDI TOF MS data Substantial auto proteolytic fragments and thermal degradation products were observed in the mass spectra after at least 2 h incubation

PAGE 174

174 The p ercent AA coverage of trypsin digests improved to at least 94% by collecting MALDI TOF MS in positive and negative modes. Meanwhile, t he full mass spectra of trypsin digests conf irmed the presence of amino acid substitution in peptides via change in m/z but not the exact positions of the amino acid subs titutions. The b and y ions from tandem mass spectra enabled de novo peptide sequencing and confirmation of AA substitution in each variant The specificity of the MTS L labeling towards cysteine was established through ESI TOF MS n of the MTSL labeled pe ptide. Tandem MS of the corresponding singly charged, protonated ion showed cleavage of the nitroxide group before and after the disulfide bond Meanwhile MS/MS of the doubly charged ion resulted in b and y product ions for de novo sequencing and con firmation of the MTSL site directed spin labeling at Cys55 These results verified the presence of amino acid substitutions and/or site specific labeling in HIV 1 protease variants making subsequent pulsed EPR experiments more reliable.

PAGE 175

175 CHAPTER 7 FUT URE DIRECTIONS Model free Analysis of MDR769 As mentioned in Chapter 5, 15 N spin relaxation data need to be acquired at a higher field to complement the current R 1 R 2 and NOE data set for MDR769. After testing for data consistency, model free analysis 237, 238 needs to be pe rformed to derive order parameters. Relaxation Meas urements for Single point Mutation Variants 15 N spin relaxation parameters need to be obtained for single point mutants, particularly those that show increased percentage of the closed state in pulsed EP R experiments, such as A71V. Similar to MDR769, order parameters need to be derived for these variants using model free analysis DEER An alysis of Active HIV 1 Protease T he use of the D25N mutation for the structural studies introduces a major am biguity in to the interpretation of correlations with kinetics and inhibition parameters since the effects of the D25N mutation may or may not be energetically additive 246 with the effects of the mutations of interest. If the HIV protease mutations of interest are not energetically additive with D25N, the effect of the second mutation on, e.g., k cat would differ in the wild type and the D25N backgrou nds, conceivably by up to one or two orders of magnitude (for energetic coupling of 1 3 kcal/mol between mutations). Such differences might have the effect of obscuring true structure function correlations Therefore it is highly recommended that DEER an alysis of subtype B variants with D30N, M36I and A71V mutations be repeated using active enzyme.

PAGE 176

176 Testing other Drug resistant HIV 1 Protease Variants and Inhibitors in | C | Correlations with IC 50 We have recently proposed that the parameter | C|, defined as the difference in the inhibitor induced percentage occupancy of closed state between two HIV 1 protease variants, correlates linearly with the fold change in half maximal inhibitory concentrations (IC50) of the inhibitors utilized 211 This method needs to be verified by testing other drug resistant variants and protease inhibitors.

PAGE 177

177 APPENDIX A SUPPLEMENTAL INFORMA TION FOR DEER DATA A NA LYSIS Figure A 1 DEER data processing for the D30N variant (A) Determination of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background sub traction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks l abeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H ) Results table.

PAGE 178

178 Figure A 2 DEER data proce ssing for D30N with CA p2 (A) Determination of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically sig nificant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table.

PAGE 179

179 Figure A 3 DEER data processing for D30N with ritonavir (RTV). (A) Determination of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipol ar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table.

PAGE 180

180 Figure A 4 DEER data processing for the M36I variant (A) Determination of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit ove rlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table.

PAGE 181

181 Figure A 5 DEER data processing for M36I with CA p2 (A) Determination of zero time. (B) Raw dipolar echo curve and the ex ponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employe d to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table.

PAGE 182

182 Figure A 6 DEER data processing for M36I with ritonavir (RTV) (A) Determination of zero time. (B) Raw dipolar echo curve and the exponential decay function co rresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR dis tance profile. (G ) Frequency domain spectrum (H) Results table

PAGE 183

183 Figure A 7 DEER data processing for the A71V variant (A) Determination of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dim ensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile ove rlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain s pectrum (H) Results table.

PAGE 184

184 Figure A 8 DEER data processing fo r A71V with CA p2. (A) Determi nation of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian p opulations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table.

PAGE 185

185 Figure A 9 DEER data processing for A71V wit h ritonavir (RTV). (A) Determi nation of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled wit h are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table.

PAGE 186

186 Figure A 10 DEER data processing for t he D3 0N/M36I variant (A) Determi nation of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and backg round subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically signific ant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table.

PAGE 187

187 Figure A 11 DEER data processing for the D30N/A71V variant (A) Deter mination of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulat ion curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level base d on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table.

PAGE 188

188 Figure A 12 DEER data processing for the M36I/ A71V variant (A) Determination of zero time. (B) Raw d ipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table.

PAGE 189

189 Figure A 13 DEER data processing for the D30N/M36I/ A71V variant (A) Det er mination of zero time. (B) Raw dipolar echo curve and the ex ponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employe d to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table.

PAGE 190

190 Figure A 14 DEER data processing for D30N/M36I/A71V with CA p2 (A) Determi nation of zero time. (B) Raw dipolar echo curve and the exponential decay function corr esponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR dista nce profile. (G ) Frequency domain spectrum (H) Results table.

PAGE 191

191 Figure A 15 DEER data processing for D30N/M36I/A71V with ritonavir (RTV). (A) Determination of zero time. (B ) Raw dipolar echo curve and exponential decay function corresponding to a homogen eous three dimensional distri bution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR dista nce profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Freq uency domain spectrum (H) Results table.

PAGE 192

192 Figure A 16 DEER data processing for MDR769 apo. (A) Determination of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table

PAGE 193

193 Figure A 17 DEER data processing for MDR769 with indinavir ( IDV ). (A) Determination of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for backgroun d subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Pe aks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table

PAGE 194

194 Figure A 18 DEER data processing for MDR769 with nelfinavir ( NFV ). (A) Determina tion of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table

PAGE 195

195 Figure A 19 DEER data processing for MDR769 wi th atazanavir ( ATV ) (A) Determi nation of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and back ground subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically signifi cant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table

PAGE 196

196 Figure A 20 DEER data processing for MDR769 with saquinavir ( SQV ) ( A) Determi nation of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipo lar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table

PAGE 197

197 Figure A 21 DEER data processing for MDR769 with amprenavir ( APV ) (A) Determi nation of z ero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve wi th TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 erro r analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table

PAGE 198

198 Figure A 22 DEER data processing for MDR769 with ritonavir ( RTV ) (A) Determi nation of zero time. (B) Raw dipol ar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with (TKR fit overlaid wi th Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Ga ussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table

PAGE 199

199 Figure A 23 DEER data processing for MDR769 with darunavir ( DRV ) (A) Determi nation of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with (TKR fit overlaid with Gaussian reco nstruc ted dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations emp loyed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table

PAGE 200

200 Figure A 24 DEER data processing for MDR769 with lopinavir ( LPV ) (A) Determi nation of zero time. (B) Raw dipolar echo curve and the exponential decay func tion corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance profile. (G ) Frequency domain spectrum (H) Results table

PAGE 201

201 Figure A 25 DEER data processing for MDR769 bound to tipranavir ( TPV ) (A) Deter mination of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimensional distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with (TKR fit overlaid with Gaussian reco nstructed dipolar modulation. (D ) The L curve. (E ) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F ) The Gaussian populations employed to regener ate the TKR distance pro file. (G ) Frequency domain spectrum (H) Results table

PAGE 202

202 Figure A 26 DEER data processing for MDR769 with CA p2 (A) Determination of zero time. (B) Raw dipolar echo curve and the exponential decay function corresponding to a homogeneous three dimension al distribution that is employed for background subtraction. (C) Long pass filtered and background subtracted dipolar modulation curve with (TKR fit overlaid with Gaussian reconstructed dipolar modulation. (D) The L curve. (E) TKR distance profile overlaid with the summation of Gaussian populations. Peaks labeled with are not statistically significant at 95% confidence level based on 2 error analysis. (F) The Gaussian populations employed to regenerate the TKR distance profile. (G) Frequency domain spectru m (H) Results table

PAGE 203

203 APPENDIX B 2 ERROR ANALYSIS FOR POPULATION VALIDATIO N Table B 1. Values of 2 for Gaussian regenerated echo curves for subtype B variant s with suppressed populations I f 2 is less than the critical value at P = 0.05 ( 2 0.95 ), a pop ulation is regarded noise artifact. Construct Suppressed Population(s)* D f 2 0.95 2 Artifact Peaks D30N wide open (wo) 91 114.3 78 curled/tucked (ct) 91 114.3 1087 wo (wo) and (ct) 94 117.6 1082 M36I unassigned, 21.1 (u) 88 110.9 61 wi de open (wo) 88 110.9 15 curled/tucked (ct) 88 110.9 705 (u) and (wo) 91 114.3 67 u, wo (u) and (ct) 91 114.3 1150 (wo) and (ct) 91 114.3 646 (u), (wo), and (ct) 94 117.6 1084 A71V wide open (wo) 91 114.3 1113 semi open (so) 91 11 4.3 141 curled/tucked (ct) 91 114.3 861 (wo) and (so) 94 117.6 1966 none (wo) and (ct) 94 117.6 1668 (so) and (ct) 94 117.6 540 (wo), (so), and (ct) 97 121.0 2194 D30N/M36I unassigned, 18.5 (u1) 85 107.5 8 unassigned, 21.3 (u2 ) 85 107.5 7 wide open (wo) 85 107.5 393 curled/tucked (ct) 85 107.5 1963 (u1) and (u2) 88 110.9 29 (u1) and (wo) 88 110.9 2487 (u2) and (wo) 88 110.9 389 (ct) and (u1) 88 110.9 2241 u1, u2 (ct) and (u2) 88 110.9 2250 (ct) and (wo) 8 8 110.9 2198 (ct), (u1) and (u2) 91 114.3 2562 (ct), (u1) and (wo) 91 114.3 2487 (ct), (u2) and (wo) 91 114.3 2496 (u1), (u2), and (wo) 91 114.3 401 (ct), (u1), (u2), and (wo) 94 117.6 2823 *wo = wide open; c = closed; so = semi open; curl ed/tucked = ct; u = unassigned

PAGE 204

204 Table B 1. Continued Construct Suppressed Population(s)* Df 2 0.95 2 Artifact Peaks D30N/A71V wide open (wo) 91 114.3 140 curled/tucked (ct) 91 114.3 2138 none (wo) and (ct) 94 117.6 2034 M36I/A71V semi ope n (so) 91 114.3 223 curled/tucked (ct) 91 114.3 1423 wide open (wo) 91 114.3 4326 (ct) and (so) 94 117.6 1261 none (ct) and (wo) 94 117.6 5026 (so) and (wo) 94 117.6 6608 (ct), (so) and (wo) 97 121.0 6958 D30N/M36I/A71V wide open (wo) 91 114.3 2502 closed (c) 91 114.3 280 curled/tucked (ct) 91 114.3 1801 (wo) and (c) 94 117.6 1866 none (wo) and (ct) 94 117.6 3216 (c) and (ct) 94 117.6 3798 (wo), (c), and (ct) 97 121.0 3794 *wo = wide open; c = closed; so = semi o pen; curled/tucked = ct; u = unassigned Table B 2. 2 error analysis for s ubtype B variants with or without ligands summarizing the v alues of 2 for Gaussia n regenerated echo curves with suppressed populations. A p opulation is considered an artifact of noise at 95% confidence level ( P = 0.05) if 2 < 2 0.95 Sample Suppressed Population(s)* Df 2 0.95 2 Artifact Peaks D30N + CA p2 wide open (wo) 94 117.6 356 semi open ( so ) 94 117.6 13 87 none (wo) and ( so ) 97 121.0 2084 D30N + RTV wide ope n (wo) 94 117.6 208 semi open ( so ) 94 117.6 945 none (wo) and ( so ) 97 121.0 1128 M36I + CA p2 semi open (so) 97 121.0 605 none M36I + RTV wide open (wo) 97 121.0 567 none A71V + RTV semi open (so) 97 121.0 337 none D30 N/M36I/A71V + CA p2 unknown, 43 .1 (u) 94 117.6 108 semi open (s o) 94 117.6 758 u (u) and (so) 97 121.0 801 D30N/M36I/A71V + RTV closed (c ) 97 121.0 827 none *wo = wide open; c = closed; so = semi open; curled/tucked = ct; u = unassigned

PAGE 205

205 Table B 3 Values of 2 for Gaussian regenerated echo curves for MDR 769 with or without ligands I f 2 is less than the critical value at P = 0.05 ( 2 0.95 ), a population is regarded noise artifact. Construct Suppressed Population(s)* D f 2 0.95 2 Artifa ct Peaks MDR 769 apo curled/tucked (ct) 291 331.8 163 closed (c) 291 331.8 667 wide open (wo) 291 331.8 551 (ct) and (c) 294 335.0 1436 ct (ct) and (wo) 294 335.0 657 (c) and (wo) 294 335.0 893 (ct), (c) and (wo) 297 338.1 1616 ID V closed (c) 291 331.8 1477 wide open (wo) 291 331.8 3756 none (c) and (wo) 294 335.0 2638 NFV unassigned, 44.3 (u) 288 328.6 178 curled/tucked (ct) 288 328.6 611 wide open (wo) 288 328.6 372 (u) and (ct) 291 331.8 658 u (u) and (wo) 291 331.8 872 (ct) and (wo) 291 331.8 525 (u), (ct) and (wo) 294 335.0 1056 ATV closed (c) 294 335.0 835 wide open (wo) 294 335.0 1803 none (c) and (wo) 297 338.1 1445 SQV curled/tucked (ct) 291 331.8 260 closed (c) 291 331.8 392 wide open (wo) 291 331.8 2175 (ct) and (c) 294 335.0 1533 ct (ct) and (wo) 294 335.0 1958 (c) and (wo) 294 335.0 1803 (ct), (c) and (wo) 297 338.1 2535 APV curled/tucked (ct) 291 331.8 754 wide open (wo) 291 331.8 276 w o (ct) and (wo) 294 335.0 827 RTV unassigned, 44.3 (u) 288 328.6 84 curled/tucked (ct) 288 328.6 15 wide open (wo) 288 328.6 158 (u) and (ct) 291 331.8 163 u, ct wo (u) and (wo) 291 331.8 240 (ct) and (wo) 291 331.8 168 (u), (ct) and (wo) 294 335.0 315 DRV semi open (so) 294 335.0 587 wide open (wo) 294 335.0 868 none (so) and (wo) 297 338.1 1797 *wo = wide open; c = closed; so = semi open; curled/tucked = ct; u = unassigned

PAGE 206

206 Table B 3 Continued Construct S uppressed Population(s)* df 2 0.95 2 Artifact Peaks LPV curled/tucked (ct) 294 335.0 487 semi open (so) 294 335.0 5280 none (ct) and (so) 297 338.1 5678 TPV unassigned, 24.0 (u) 288 331.8 80 curled/tucked (ct) 288 331.8 497 semi ope n (so) 288 331.8 35 wide open (wo) 288 335.0 1899 (u) and (ct) 291 335.0 351 (u) and (so) 291 335.0 108 (u) and (wo) 291 338.1 2024 (ct) and (so) 291 338.1 484 u, so (ct) and (wo) 291 338.1 2007 (so) and (wo) 291 338.1 2281 (u), (ct) and (so) 294 335.0 348 (u), (ct) and (wo) 294 335.0 2268 (u), (so) and (wo) 294 335.0 2419 (ct), (so) and (wo) 294 335.0 2417 (u), (ct), (so) and (wo) 297 338.1 2700 CA p2 unassigned, 20.5 (u) 288 331.8 36 curled/tucked (ct) 288 331.8 352 semi open (so) 288 331.8 349 wide open (wo) 288 335.0 235 (u) and (ct) 291 335.0 478 (u) and (so) 291 335.0 350 (u) and (wo) 291 338.1 266 (ct) and (so) 291 338.1 423 u, wo (ct) and (wo) 291 338.1 470 (so) and (wo) 291 338. 1 896 (u), (ct) and (so) 294 335.0 651 (u), (ct) and (wo) 294 335.0 698 (u), (so) and (wo) 294 335.0 707 (ct), (so) and (wo) 294 335.0 896 (u), (ct), (so) and (wo) 297 338.1 1144 *wo = wide open; c = closed; so = semi open; curled/tucked = ct; u = unassigned

PAGE 207

207 APPENDIX C 1 H 15 N HSQC TITRATION FIG URES FOR MDR769 Figure C 1. 1 H 15 N HSQC spectra MDR769 (red) titrated with amprenavir (APV) at 0.5:1 (A) and 1:1 (B) inhibitor to protein concentration ratio, overlaid with apo (blue).

PAGE 208

208 Figu re C 2 1 H 15 N HSQC spectra MDR769 (red) titrated with atazanavir (AT V) at 0.5:1 (A) and 1:1 (B) inhibitor to protein concentration ratio, overlaid with apo (blue).

PAGE 209

209 Figure C 3. 1 H 15 N HSQC spectra MDR769 (red) titrated with darunavir (DRV) at 0.5:1 (A) and 1:1 (B) inhibitor to protein concentration ratio, overlaid with apo (blue).

PAGE 210

210 Figure C 4. 1 H 15 N HSQC spectra MDR769 (red) titrated with saquinavir (SQV) at 0.5:1 (A) and 1:1 (B) inhibitor to protein concentration ratio, overlaid with apo (blue ).

PAGE 211

211 Figure C 5. 1 H 15 N HSQC spectra MDR769 (red) titrated with lopinavir (LPV) at 0.5:1 (A) and 1:1 (B) inhibitor to protein concentration ratio, overlaid with apo (blue).

PAGE 212

212 Figure C 6 1 H 15 N HSQC spectra MDR769 (red) titrated with nelfinavir (NFV) a t 0.5:1 (A) and 1:1 (B) inhibitor to protein concentration ratio, overlaid with apo (blue).

PAGE 213

213 Figure C 7 1 H 15 N HSQC spectra MDR769 (red) titrated with indinavir (IDV) at 0.5:1 (A) and 1:1 (B) inhibitor to protein concentration ratio, overlaid with a po (blue).

PAGE 214

214 Figure C 8 1 H 15 N HSQC spectra MDR769 (red) titrated with ritonavir (RTV) at 0.5:1 (A) and 1:1 (B) inhibitor to protein concentration ratio, overlaid with apo (blue).

PAGE 215

215 Figure C 9. 1 H 15 N HSQC spectra MDR769 (red) titrated with tipranavir (TPV) at 0.5:1 (A) and 1:1 (B) inhibitor to protein concentration ratio, overlaid with apo (blue).

PAGE 216

216 Figure C 10. 1 H 15 N HSQC spectra MDR769 (red) titrated with the non hydrolyzable substrate mimic, CA p2 to a ratio of 1:1 substrate to protein, overlai d with apo (blue).

PAGE 217

217 LIST OF REFERENCES 1. UNAIDS Report on the Global AIDS Epidemic (2010) pp 19 57, Joint United Nations Programme on HIV/AIDS (UNAIDS). 2. Sepkowitz, K. A. (2001) AIDS -the first 20 years, N Engl J Med 344 1764 1772. 3. Levy, J. A. (1993) Pathogenesis of human immunodeficiency virus infection, Microbiol Rev 57 183 289. 4. Bailes, E., Gao, F., Bibollet Ruche, F., Courgnaud, V., Peeters, M., Marx, P. A., Hahn, B. H., and Sharp, P. M. (2003) Hybrid origin o f SIV in chimpanzees, Science 300 1713. 5. Worobey, M., Gemmel, M., Teuwen, D. E., Haselkorn, T., Kunstman, K., Bunce, M., Muyembe, J. J., Kabongo, J. M., Kalengayi, R. M., Van Marck, E., Gilbert, M. T., and Wolinsky, S. M. (2008) Direct evidence of exte nsive diversity of HIV 1 in Kinshasa by 1960, Nature 455 661 664. 6. Lemey, P., Pybus, O. G., Wang, B., Saksena, N. K., Salemi, M., and Vandamme, A. M. (2003) Tracing the origin and history of the HIV 2 epidemic, Proc Natl Acad Sci U S A 100 6588 6592. 7. Santos, A. F., and Soares, M. A. (2010) HIV Genetic Diversity and Drug Resistance, Viruses 2 503 531. 8. Wainberg, M. A., and Brenner, B. G. (2010) Role of HIV Subtype Diversity in the Development of Resistance to Antiviral Drugs, Viruses 2 2493 250 8. 9. Shafer, R. W., Stevenson, D., and Chan, B. (1999) Human Immunodeficiency Virus Reverse Transcriptase and Protease Sequence Database, Nucleic Acids Res 27 348 352. 10. Mansky, L. M., and Temin, H. M. (1995) Lower in vivo mutation rate of human immu nodeficiency virus type 1 than that predicted from the fidelity of purified reverse transcriptase, J Virol 69 5087 5094. 11. Levy, J. A. (2007) HIV and the Pathogenesis of AIDS 3 ed., ASM Press, Washington DC. 12. Wong Staal, F. (1985) Name for AIDS vi rus, Nature 314 574. 13. Clemente, J. C., Coman, R. M., Thiaville, M. M., Janka, L. K., Jeung, J. A., Nukoolkarn, S., Govindasamy, L., Agbandje McKenna, M., McKenna, R., Leelamanit, W., Goodenow, M. M., and Dunn, B. M. (2006) Analysis of HIV 1 CRF_01 A/E protease inhibitor resistance: structural determinants for maintaining sensitivity and developing resistance to atazanavir, Biochemistry 45 5468 5477.

PAGE 218

218 14. Chan, P. A., and Kantor, R. (2009) Transmitted drug resistance in nonsubtype B HIV 1 infection, HI V Ther 3 447 465. 15. Kuznetsov, Y. G., Victoria, J. G., Robinson, W. E., Jr., and McPherson, A. (2003) Atomic force microscopy investigation of human immunodeficiency virus (HIV) and HIV infected lymphocytes, J Virol 77 11896 11909. 16. Gelderblom, H. R., Ozel, M., Hausmann, E. H. S., Winkel, T., Pauli, G., and Koch, M. A. (1988) Fine Structure of Human Immunodeficiency Virus (Hiv), Immunolocalization of Structural Proteins and Virus Cell Relation, Micron and Microscopica Acta 19 41 60. 17. Zhu, P., Liu, J., Bess, J., Chertova, E., Lifson, J. D., Grise, H., Ofek, G. A., Taylor, K. A., and Roux, K. H. (2006) Distribution and three dimensional structure of AIDS virus envelope spikes, Nature 441 847 852. 18. Resh, M. D. (2004) A myristoyl switch regula tes membrane binding of HIV 1 Gag, Proc Natl Acad Sci U S A 101 417 418. 19. Greene, W. C., and Peterlin, B. M. (2002) Charting HIV's remarkable voyage through the cell: Basic science as a passport to future therapy, Nat Med 8 673 680. 20. Yarchoan, R. Klecker, R. W., Weinhold, K. J., Markham, P. D., Lyerly, H. K., Durack, D. T., Gelmann, E., Lehrman, S. N., Blum, R. M., Barry, D. W., and et al. (1986) Administration of 3' azido 3' deoxythymidine, an inhibitor of HTLV III/LAV replication, to patients w ith AIDS or AIDS related complex, Lancet 1 575 580. 21. Menendez Arias, L. (2010) Molecular basis of human immunodeficiency virus drug resistance: An update, Antivir Res 85 210 231. 22. Mastrolorenzo, A., Rusconi, S., Scozzafava, A., Barbaro, G., and S upuran, C. T. (2007) Inhibitors of HIV 1 protease: current state of the art 10 years after their introduction. From antiretroviral drugs to antifungal, antibacterial and antitumor agents based on aspartic protease inhibitors, Curr Med Chem 14 2734 2748. 23. Ashorn, P., McQuade, T. J., Thaisrivongs, S., Tomasselli, A. G., Tarpley, W. G., and Moss, B. (1990) An inhibitor of the protease blocks maturation of human and simian immunodeficiency viruses and spread of infection, Proc Natl Acad Sci U S A 87 7472 7476. 24. Barrett, A. J., Rawlings, N. D., and Woessner, J. F. (2003) The Handbook of Proteolytic Enzymes Academic Press, London.

PAGE 219

21 9 25. Toh, H., Ono, M., Saigo, K., and Miyata, T. (1985) Retroviral Protease Like Sequence in the Yeast Transposon Ty1, Natur e 315 691 691. 26. Dunn, B. M., Goodenow, M. M., Gustchina, A., and Wlodawer, A. (2002) Retroviral proteases, Genome Biol 3 REVIEWS3006. 27. Wlodawer, A., and Gustchina, A. (2000) Structural and biochemical studies of retroviral proteases, Biochim Biop hys Acta 1477 16 34. 28. Andreeva, N. (1991) A consensus template of the aspartic proteinase fold, in Structure and Function of Aspartic Proteinases (Dunn, B. M., Ed.), pp 559 572, Plenum Press, New York. 29. Andreeva, N. S., and Gustchina, A. E. (1979) On the supersecondary structure of acid proteases, Biochem Biophys Res Commun 87 32 42. 30. Cairns, J., Overbaugh, J., and Miller, S. (1988) The origin of mutants, Nature 335 142 145. 31. Hyland, L. J., Tomaszek, T. A., Jr., and Meek, T. D. (1991) Hum an immunodeficiency virus 1 protease. 2. Use of pH rate studies and solvent kinetic isotope effects to elucidate details of chemical mechanism, Biochemistry 30 8454 8463. 32. Brik, A., and Wong, C. H. (2003) HIV 1 protease: mechanism and drug discovery, Org Biomol Chem 1 5 14. 33. Gustchina, A., and Weber, I. T. (1990) Comparison of inhibitor binding in HIV 1 protease and in non viral aspartic proteases: the role of the flap, FEBS Lett 269 269 272. 34. Blackburn, M. E., Veloro, A. M., and Fanucci, G. E. (2009) Monitoring inhibitor induced conformational population shifts in HIV 1 protease by pulsed EPR spectroscopy, Biochemistry 48 8765 8767. 35. Galiano, L., Ding, F., Veloro, A. M., Blackburn, M. E., Simmerling, C., and Fanucci, G. E. (2009) Drug pr essure selected mutations in HIV 1 protease alter flap conformations, J Am Chem Soc 131 430 431. 36. Nicholson, L. K., Yamazaki, T., Torchia, D. A., Grzesiek, S., Bax, A., Stahl, S. J., Kaufman, J. D., Wingfield, P. T., Lam, P. Y., Jadhav, P. K., and et al. (1995) Flexibility and function in HIV 1 protease, Nat Struct Biol 2 274 280. 37. Ishima, R., Freedberg, D. I., Wang, Y. X., Louis, J. M., and Torchia, D. A. (1999) Flap opening and dimer interface flexibility in the free and inhibitor bound HIV prot ease, and their implications for function, Structure 7 1047 1055.

PAGE 220

220 38. Freedberg, D. I., Ishima, R., Jacob, J., Wang, Y. X., Kustanovich, I., Louis, J. M., and 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 Sci 11 221 232. 39. Ishima, R., Torchia, D. A., and Louis, J. M. (2007) Mutational and structural studies aimed at characterizing the monomer of HIV 1 p rotease and its precursor, J Biol Chem 282 17190 17199. 40. Ishima, R., Torchia, D. A., Lynch, S. M., Gronenborn, A. M., and Louis, J. M. (2003) Solution structure of the mature HIV 1 protease monomer: insight into the tertiary fold and stability of a pr ecursor, J Biol Chem 278 43311 43319. 41. Louis, J. M., Ishima, R., Nesheiwat, I., Pannell, L. K., Lynch, S. M., Torchia, D. A., and Gronenborn, A. M. (2003) Revisiting monomeric HIV 1 protease. Characterization and redesign for improved properties, J Bi ol Chem 278 6085 6092. 42. Todd, M. J., Semo, N., and Freire, E. (1998) The structural stability of the HIV 1 protease, J Mol Biol 283 475 488. 43. Todd, M. J., and Freire, E. (1999) The effect of inhibitor binding on the structural stability and coope rativity of the HIV 1 protease, Proteins 36 147 156. 44. Todd, M. J., Luque, I., Velazquez Campoy, A., and Freire, E. (2000) Thermodynamic basis of resistance to HIV 1 protease inhibition: calorimetric analysis of the V82F/I84V active site resistant muta nt, Biochemistry 39 11876 11883. 45. Velazquez Campoy, A., Todd, M. J., and Freire, E. (2000) HIV 1 protease inhibitors: enthalpic versus entropic optimization of the binding affinity, Biochemistry 39 2201 2207. 46. Hornak, V., Okur, A., Rizzo, R. C., and Simmerling, C. (2006) HIV 1 protease flaps spontaneously open and reclose in molecular dynamics simulations, Proc Natl Acad Sci U S A 103 915 920. 47. Collins, J. R., Burt, S. K., and Erickson, J. W. (1995) Flap Opening in Hiv 1 Protease Simulated by Activated Molecular Dynamics, Nat Struct Biol 2 334 338. 48. Ding, F., Layten, M., and Simmerling, C. (2008) Solution structure of HIV 1 protease flaps probed by comparison of molecular dynamics simulation ensembles and EPR experiments, J Am Chem Soc 13 0 7184 7185.

PAGE 221

221 49. Toth, G., and Borics, A. (2006) Flap opening mechanism of HIV 1 protease, J Mol Graph Model 24 465 474. 50. Heaslet, H., Rosenfeld, R., Giffin, M., Lin, Y. C., Tam, K., Torbett, B. E., Elder, J. H., McRee, D. E., and Stout, C. D. (2007 ) Conformational flexibility in the flap domains of ligand free HIV protease, Acta Crystallogr D Biol Crystallogr 63 866 875. 51. Galiano, L., Bonora, M., and Fanucci, G. E. (2007) Inter flap distances in HIV 1 Protease determined by pulsed EPR measureme nts, J Am Chem Soc 129 11004 11005. 52. Kear, J. L., Blackburn, M. E., Veloro, A. M., Dunn, B. M., and Fanucci, G. E. (2009) Subtype polymorphisms among HIV 1 protease variants confer altered flap conformations and flexibility, J Am Chem Soc 131 14650 1 4651. 53. Harte, W. E., Jr., Swaminathan, S., Mansuri, M. M., Martin, J. C., Rosenberg, I. E., and Beveridge, D. L. (1990) Domain communication in the dynamical structure of human immunodeficiency virus 1 protease, Proc Natl Acad Sci U S A 87 8864 8868. 54. Scott, W. R., and 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. 55. Logsdon, B. C., Vickrey, J. F., Martin, P., Proteasa, G., Koepke, J. I., Terlecky, S. R., Wawrzak, Z., Winters, M. A., Merigan, T. C., and Kovari, L. C. (2004) Crystal structures of a multidrug resistant human immunodeficiency virus type 1 protease reveal an expanded active site cavity, J Virol 78 3123 3132. 56. Cigler, P., K ozisek, M., Rezacova, P., Brynda, J., Otwinowski, Z., Pokorna, J., Plesek, J., Gruner, B., Doleckova Maresova, L., Masa, M., Sedlacek, J., Bodem, J., Krausslich, H. G., Kral, V., and Konvalinka, J. (2005) From nonpeptide toward noncarbon protease inhibitor s: metallacarboranes as specific and potent inhibitors of HIV protease, Proc Natl Acad Sci U S A 102 15394 15399. 57. McQuade, T. J., Tomasselli, A. G., Liu, L., Karacostas, V., Moss, B., Sawyer, T. K., Heinrikson, R. L., and Tarpley, W. G. (1990) A synt hetic HIV 1 protease inhibitor with antiviral activity arrests HIV like particle maturation, Science 247 454 456. 58. Richards, A. D., Roberts, R., Dunn, B. M., Graves, M. C., and Kay, J. (1989) Effective blocking of HIV 1 proteinase activity by characte ristic inhibitors of aspartic proteinases, FEBS Lett 247 113 117.

PAGE 222

222 59. Dreyer, G. B., Metcalf, B. W., Tomaszek, T. A., Jr., Carr, T. J., Chandler, A. C., 3rd, Hyland, L., Fakhoury, S. A., Magaard, V. W., Moore, M. L., Strickler, J. E., and et al. (1989) I nhibition of human immunodeficiency virus 1 protease in vitro: rational design of substrate analogue inhibitors, Proc Natl Acad Sci U S A 86 9752 9756. 60. Tomasselli, A. G., Olsen, M. K., Hui, J. O., Staples, D. J., Sawyer, T. K., Heinrikson, R. L., and Tomich, C. S. (1990) Substrate analogue inhibition and active site titration of purified recombinant HIV 1 protease, Biochemistry 29 264 269. 61. la Porte, C. J. L. (2009) Saquinavir, the pioneer antiretroviral protease inhibitor, Expert Opinion on Drug Metabolism & Toxicology 5 1313 1322. 62. Anderson, J., Schiffer, C. A., Lee, S. K., and Swanstrom, R. (2009) Viral Protease Inhibitors, in Handbook of Experimental Pharmacology (Krausslich, H. G., and Bartenschlager, R., Eds.), Springer Verlag Berlin, H eidelberg. 63. Wensing, A. M., van Maarseveen, N. M., and Nijhuis, M. (2010) Fifteen years of HIV Protease Inhibitors: raising the barrier to resistance, Antiviral Res 85 59 74. 64. Tomasselli, A. G., and Heinrikson, R. L. (2000) Targeting the HIV prote ase in AIDS therapy: a current clinical perspective, Biochim Biophys Acta 1477 189 214. 65. 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., Bohanon, M. J., Horn g, M. M., Lynn, J. C., Chong, K. T., Hinshaw, R. R., Watenpaugh, K. D., Janakiraman, M. N., and Thaisrivongs, S. (1998) Tipranavir (PNU 140690): a potent, orally bioavailable nonpeptidic HIV protease inhibitor of the 5,6 dihydro 4 hydroxy 2 pyrone sulfonam ide class, J Med Chem 41 3467 3476. 66. Los Alamos HIV Database accessed October 6, 2012 67. Barrie, K. A., Perez, E. E., Lamers, S. L., Farmerie, W. G., Dunn, B. M., Sleasman, J. W., and Goodenow, M. M. (1996) Natural variation in HIV 1 protease, Gag p7 and p6, and protease cleavage sites within gag/pol polyproteins: amino acid substitutions in the absence of protease inhibitors in mothers and children infected by human immunodeficiency virus type 1, Virology 219 407 416. 68. Shafer, R. W., Rhee, S. Y., Pillay, D., Miller, V., Sandstrom, P., Schapiro, J. M., Kuritzkes, D. R., and Bennett, D. (2007) HIV 1 protease and reverse transcriptase mutations for drug resistance surveillance, AIDS 21 215 223.

PAGE 223

223 69. Ali, A., Bandaranayake, R. M., Cai, Y., King, N. M., Kolli, M., Mittal, S., Murzycki, J. F., Nalam, M. N., Nalivaika, E. A., Ozen, A., Prabu Jeyabalan, M. M., Thayer, K., and Schiffer, C. A. (2010) Molecular Basis for Drug Resistance in HIV 1 Protease, Viruses 2 2509 2535. 70. Weber, I. T., and Agni swamy, J. (2009) HIV 1 Protease: Structural Perspectives on Drug Resistance, Viruses 1 1110 1136. 71. Wilson, S. I., Phylip, L. H., Mills, J. S., Gulnik, S. V., Erickson, J. W., Dunn, B. M., and Kay, J. (1997) Escape mutants of HIV 1 proteinase: enzymic efficiency and susceptibility to inhibition, Biochim Biophys Acta 1339 113 125. 72. Johnson, V. A., Brun Vezinet, F., Clotet, B., Gunthard, H. F., Kuritzkes, D. R., Pillay, D., Schapiro, J. M., and Richman, D. D. (2010) Update of the drug resistance muta tions in HIV 1: December 2010, Top HIV Med 18 156 163. 73. Tie, Y., Boross, P. I., Wang, Y. F., Gaddis, L., Liu, F., Chen, X., Tozser, J., Harrison, R. W., and Weber, I. T. (2005) Molecular basis for substrate recognition and drug resistance from 1.1 to 1.6 angstroms resolution crystal structures of HIV 1 protease mutants with substrate analogs, FEBS J 272 5265 5277. 74. Sugiura, W., Matsuda, Z., Yokomaku, Y., Hertogs, K., Larder, B., Oishi, T., Okano, A., Shiino, T., Tatsumi, M., Matsuda, M., Abumi, H. Takata, N., Shirahata, S., Yamada, K., Yoshikura, H., and Nagai, Y. (2002) Interference between D30N and L90M in selection and development of protease inhibitor resistant human immunodeficiency virus type 1, Antimicrob Agents Chemother 46 708 715. 75. 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., Leonard, J. M., Norbeck, D. W., and Kempf, D. J. (1996) Ordered accum ulation of mutations in HIV protease confers resistance to ritonavir, Nat Med 2 760 766. 76. Clemente, J. C., Hemrajani, R., Blum, L. E., Goodenow, M. M., and Dunn, B. M. (2003) Secondary mutations M36I and A71V in the human immunodeficiency virus type 1 protease can provide an advantage for the emergence of the primary mutation D30N, Biochemistry 42 15029 15035. 77. Kozal, M. J., Shah, N., Shen, N., Yang, R., Fucini, R., Merigan, T. C., Richman, D. D., Morris, D., Hubbell, E., Chee, M., and Gingeras, T R. (1996) Extensive polymorphisms observed in HIV 1 clade B protease gene using high density oligonucleotide arrays, Nat Med 2 753 759.

PAGE 224

224 78. Mahalingam, B., Boross, P., Wang, Y. F., Louis, J. M., Fischer, C. C., Tozser, J., Harrison, R. W., and Weber, I T. (2002) Combining mutations in HIV 1 protease to understand mechanisms of resistance, Proteins 48 107 116. 79. Mo, H., Parkin, N., Stewart, K. D., Lu, L., Dekhtyar, T., Kempf, D. J., and Molla, A. (2007) Identification and structural characterization of I84C and I84A mutations that are associated with high level resistance to human immunodeficiency virus protease inhibitors and impair viral replication, Antimicrob Agents Chemother 51 732 735. 80. Clemente, J. C., Moose, R. E., Hemrajani, R., Whitfor d, L. R., Govindasamy, L., Reutzel, R., McKenna, R., Agbandje McKenna, M., Goodenow, M. M., and Dunn, B. M. (2004) Comparing the accumulation of active and nonactive site mutations in the HIV 1 protease, Biochemistry 43 12141 12151. 81. Foulkes Murzycki J. E., Scott, W. R., and Schiffer, C. A. (2007) Hydrophobic sliding: a possible mechanism for drug resistance in human immunodeficiency virus type 1 protease, Structure 15 225 233. 82. Rose, R. B., Craik, C. S., and Stroud, R. M. (1998) Domain flexibil ity in retroviral proteases: structural implications for drug resistant mutations, Biochemistry 37 2607 2621. 83. Chang, M. W., and Torbett, B. E. (2011) Accessory mutations maintain stability in drug resistant HIV 1 protease, J Mol Biol 410 756 760. 8 4. Prabu Jeyabalan, M., Nalivaika, E. A., King, N. M., and Schiffer, C. A. (2003) Viability of a drug resistant human immunodeficiency virus type 1 protease variant: structural insights for better antiviral therapy, J Virol 77 1306 1315. 85. Piana, S., C arloni, P., and Rothlisberger, U. (2002) Drug resistance in HIV 1 protease: Flexibility assisted mechanism of compensatory mutations, Protein Sci 11 2393 2402. 86. Prabu Jeyabalan, M., King, N. M., Nalivaika, E. A., Heilek Snyder, G., Cammack, N., and Sc hiffer, C. A. (2006) Substrate envelope and drug resistance: crystal structure of RO1 in complex with wild type human immunodeficiency virus type 1 protease, Antimicrob Agents Chemother 50 1518 1521. 87. Nalam, M. N., and Schiffer, C. A. (2008) New appro aches to HIV protease inhibitor drug design II: testing the substrate envelope hypothesis to avoid drug resistance and discover robust inhibitors, Curr Opin HIV AIDS 3 642 646. 88. Wang, Y., Liu, Z., Brunzelle, J. S., Kovari, I. A., Dewdney, T. G., Reite r, S. J., and Kovari, L. C. (2011) The higher barrier of darunavir and tipranavir resistance for HIV 1 protease, Biochem Biophys Res Commun 412 737 742.

PAGE 225

225 89. Martin, P., Vickrey, J. F., Proteasa, G., Jimenez, Y. L., Wawrzak, Z., Winters, M. A., Merigan, T C., and Kovari, L. C. (2005) "Wide open" 1.3 A structure of a multidrug resistant HIV 1 protease as a drug target, Structure 13 1887 1895. 90. Liu, Z., Wang, Y., Brunzelle, J., Kovari, I. A., and Kovari, L. C. (2011) Nine crystal structures determine t he substrate envelope of the MDR HIV 1 protease, Protein J 30 173 183. 91. Kipp, D. R., Hirschi, J. S., Wakata, A., Goldstein, H., and Schramm, V. L. (2012) Transition states of native and drug resistant HIV 1 protease are the same, Proc Natl Acad Sci U S A 109 6543 6548. 92. Gulnik, S. V., Suvorov, L. I., Liu, B., Yu, B., Anderson, B., Mitsuya, H., and Erickson, J. W. (1995) Kinetic characterization and cross resistance patterns of HIV 1 protease mutants selected under drug pressure, Biochemistry 34 9 282 9287. 93. Galiano, L., Blackburn, M. E., Veloro, A. M., Bonora, M., and Fanucci, G. E. (2009) Solute effects on spin labels at an aqueous exposed site in the flap region of HIV 1 protease, J Phys Chem B 113 1673 1680. 94. Kantor, R., Fessel, W. J., Zolopa, A. R., Israelski, D., Shulman, N., Montoya, J. G., Harbour, M., Schapiro, J. M., and Shafer, R. W. (2002) Evolution of primary protease inhibitor resistance mutations during protease inhibitor salvage therapy, Antimicrob Agents Chemother 46 1086 1 092. 95. Martinez Cajas, J. L., and Wainberg, M. A. (2007) Protease inhibitor resistance in HIV infected patients: molecular and clinical perspectives, Antiviral Res 76 203 221. 96. Rose, J. R., Salto, R., and Craik, C. S. (1993) Regulation of autoprote olysis of the HIV 1 and HIV 2 proteases with engineered amino acid substitutions, J Biol Chem 268 11939 11945. 97. Mildner, A. M., Rothrock, D. J., Leone, J. W., Bannow, C. A., Lull, J. M., Reardon, I. M., Sarcich, J. L., Howe, W. J., Tomich, C. S., Smit h, C. W., and et al. (1994) The HIV 1 protease as enzyme and substrate: mutagenesis of autolysis sites and generation of a stable mutant with retained kinetic properties, Biochemistry 33 9405 9413. 98. Sayer, J. M., Liu, F., Ishima, R., Weber, I. T., and Louis, J. M. (2008) Effect of the active site D25N mutation on the structure, stability, and ligand binding of the mature HIV 1 protease, J Biol Chem 283 13459 13470.

PAGE 226

226 99. Louis, J. M., Ishima, R., Torchia, D. A., and Weber, I. T. (2007) HIV 1 protease: structure, dynamics, and inhibition, Adv Pharmacol 55 261 298. 100. Miick, S. M., Martinez, G. V., Fiori, W. R., Todd, A. P., and Millhauser, G. L. (1992) Short alanine based peptides may form 3(10) helices and not alpha helices in aqueous solution, Natu re 359 653 655. 101. Rabenstein, M. D., and Shin, Y. K. (1995) Determination of the distance between two spin labels attached to a macromolecule, Proc Natl Acad Sci U S A 92 8239 8243. 102. McHaourab, H. S., Oh, K. J., Fang, C. J., and Hubbell, W. L. ( 1997) Conformation of T4 lysozyme in solution. Hinge bending motion and the substrate induced conformational transition studied by site directed spin labeling, Biochemistry 36 307 316. 103. Cooke, J. A., and Brown, L. J. (2011) Distance measurements by c ontinuous wave EPR spectroscopy to monitor protein folding, Methods Mol Biol 752 73 96. 104. Tsvetkov, Y. D., and Grishin, Y. A. (2009) Techniques for EPR Spectroscopy of Pulsed Electron Double Resonance (PELDOR): A Review, Instrum Exp Tech 52 615 636. 105. Jeschke, G. (2012) DEER distance measurements on proteins, Annu Rev Phys Chem 63 419 446. 106. Borbat, P. P., Mchaourab, H., and Freed, J. H. (2002) Protein structure determination using long distance constraints from double quantum coherence ESR: Study of T4 Lysozyme, Bioph ys J 82 360A 360A. 107. Misra, S. K., Borbat, P. P., and Freed, J. H. (2009) Calculation of Double Quantum Coherence Two dimensional Spectra: Distance Measurements and Orientational Correlations, Appl Magn Reson 36 237 258. 1 08. Altenbach, C., Oh, K. J., Trabanino, R. J., Hideg, K., and Hubbell, W. L. (2001) Estimation of inter residue distances in spin labeled proteins at physiological temperatures: experimental strategies and practical limitations, Biochemistry 40 15471 154 82. 109. Jeschke, G., Chechik, V., Ionita, P., Godt, A., Zimmermann, H., Banham, J., Timmel, C. R., Hilger, D., and Jung, H. (2006) DeerAnalysis2006 A Comprehensive Software Package for Analyzing Pulsed ELDOR Data, Appl Magn Reson 30 473 498.

PAGE 227

227 110. Jes chke, G., Panek, G., Godt, A., Bender, A., and Paulsen, H. (2004) Data analysis procedures for pulse ELDOR measurements of broad distance distributions, Applied Magnetic Resonance 26 223 244. 111. Jeschke, G., and Polyhach, Y. (2007) Distance measurement s on spin labelled biomacromolecules by pulsed electron paramagnetic resonance, Phys Chem Chem Phys 9 1895 1910. 112. Pannier, M., Veit, S., Godt, A., Jeschke, G., and Spiess, H. W. (2000) Dead time free measurement of dipole dipole interactions between electron spins, J Magn Reson 142 331 340. 113. Borbat, P. P., Davis, J. H., Butcher, S. E., and Freed, J. H. (2004) Measurement of large distances in biomolecules using double quantum filtered refocused electron spin echoes, J Am Chem Soc 126 7746 7747. 114. Jeschke, G. (2002) Distance measurements in the nanometer range by pulse EPR, Chemphyschem 3 927 932. 115. Milov, A. D., Maryasov, A. G., Tsvetkov, Y. D., and Raap, J. (1999) Pulsed ELDOR in spin labeled polypeptides, Chem Phys Lett 303 135 143. 116. Bennati, M., Weber, A., Antonic, J., Perlstein, D. L., Robblee, J., and Stubbe, J. (2003) Pulsed ELDOR spectroscopy measures the distance between the two tyrosyl radicals in the R2 subunit of the E. coli ribonucleotide reductase, J Am Chem Soc 125 1 4988 14989. 117. Persson, M., Harbridge, J. R., Hammarstrom, P., Mitri, R., Martensson, L. G., Carlsson, U., Eaton, G. R., and Eaton, S. S. (2001) Comparison of electron paramagnetic resonance methods to determine distances between spin labels on human ca rbonic anhydrase II, Biophys J 80 2886 2897. 118. Jeschke, G., Wegener, C., Nietschke, M., Jung, H., and Steinhoff, H. J. (2004) Interresidual distance determination by four pulse double electron electron resonance in an integral membrane protein: the Na + /proline transporter PutP of Escherichia coli, Biophys J 86 2551 2557. 119. Cai, Q., Kusnetzow, A. K., Hideg, K., Price, E. A., Haworth, I. S., and Qin, P. Z. (2007) Nanometer distance measurements in RNA using site directed spin labeling, Biophys J 93 2110 2117. 120. Kim, N. K., Bowman, M. K., and DeRose, V. J. (2010) Precise mapping of RNA tertiary structure via nanometer distance measurements with double electron electron resonance spectroscopy, J Am Chem Soc 132 8882 8884.

PAGE 228

228 121. Yang, Z., Becker, J., and Saxena, S. (2007) On Cu(II) Cu(II) distance measurements using pulsed electron electron double resonance, J Magn Reson 188 337 343. 122. Roessler, M. M., King, M. S., Robinson, A. J., Armstrong, F. A., Harmer, J., and Hirst, J. (2010) Direct assi gnment of EPR spectra to structurally defined iron sulfur clusters in complex I by double electron electron resonance, Proc Natl Acad Sci U S A 107 1930 1935. 123. Raitsimring, A. M., Gunanathan, C., Potapov, A., Efremenko, I., Martin, J. M., Milstein, D ., and Goldfarb, D. (2007) Gd 3+ complexes as potential spin labels for high field pulsed EPR distance measurements, J Am Chem Soc 129 14138 14139. 124. Potapov, A., Yagi, H., Huber, T., Jergic, S., Dixon, N. E., Otting, G., and Goldfarb, D. (2010) Nanome ter scale distance measurements in proteins using Gd 3+ spin labeling, J Am Chem Soc 132 9040 9048. 125. Narr, E., Godt, A., and Jeschke, G. (2002) Selective measurements of a nitroxide nitroxide separation of 5 nm and a nitroxide copper separation of 2.5 nm in a terpyridine based copper(II) complex by pulse EPR spectroscopy, Angew Chem Int Ed Engl 41 3907 3910. 126. Yang, Z., Kise, D., and Saxena, S. (2010) An approach towards the measurement of nanometer range distances based on Cu 2+ ions and ESR, J Ph ys Chem B 114 6165 6174. 127. Lovett, J. E., Bowen, A. M., Timmel, C. R., Jones, M. W., Dilworth, J. R., Caprotti, D., Bell, S. G., Wong, L. L., and Harmer, J. (2009) Structural information from orientationally selective DEER spectroscopy, Phys Chem Chem Phys 11 6840 6848. 128. Dastvan, R., Bode, B. E., Karuppiah, M. P., Marko, A., Lyubenova, S., Schwalbe, H., and Prisner, T. F. (2010) Optimization of transversal relaxation of nitroxides for pulsed electron electron double resonance spectroscopy in phos pholipid membranes, J Phys Chem B 114 13507 13516. 129. Borbat, P. P., Crepeau, R. H., and Freed, J. H. (1997) Multifrequency two dimensional Fourier transform ESR: an X/Ku band spectrometer, J Magn Reson 127 155 167. 130. Ghimire, H., McCarrick, R. M. Budil, D. E., and Lorigan, G. A. (2009) Significantly improved sensitivity of Q band PELDOR/DEER experiments relative to X band is observed in measuring the intercoil distance of a leucine zipper motif peptide (GCN4 LZ), Biochemistry 48 5782 5784.

PAGE 229

229 131. Thomas, D. D., Ishiwata, S., Seidel, J. C., and Gergely, J. (1980) Submillisecond rotational dynamics of spin labeled myosin heads in myofibrils, Biophys J 32 873 889. 132. Thomas, D. D., and Cooke, R. (1980) Orientation of spin labeled myosin heads in glycerinated muscle fibers, Biophys J 32 891 906. 133. Roopnarine, O., and Thomas, D. D. (1994) A spin label that binds to myosin heads in muscle fibers with its principal axis parallel to the fiber axis, Biophys J 67 1634 1645. 134. Fajer, P. G. (1994 ) Determination of spin label orientation within the myosin head, Proc Natl Acad Sci U S A 91 937 941. 135. Fajer, P. G., Fajer, E. A., Brunsvold, N. J., and Thomas, D. D. (1988) Effects of AMPPNP on the orientation and rotational dynamics of spin labele d muscle cross bridges, Biophys J 53 513 524. 136. Waring, A. J., and Cooke, R. (1987) The molecular dynamics of actin measured by a spin probe attached to lysine, Arch Biochem Biophys 252 197 205. 137. Hanson, P., Anderson, D. J., Martinez, G., Millha user, G., Formaggio, F., Crisma, M., Toniolo, C., and Vita, C. (1998) Electron spin resonance and structural analysis of water soluble, alanine rich peptides incorporating TOAC, Mol Phys 95 957 966. 138. Cornish, V. W., Benson, D. R., Altenbach, C. A., H ideg, K., Hubbell, W. L., and Schultz, P. G. (1994) Site specific incorporation of biophysical probes into proteins, Proc Natl Acad Sci U S A 91 2910 2914. 139. Fleissner, M. R., Brustad, E. M., Kalai, T., Altenbach, C., Cascio, D., Peters, F. B., Hideg, K., Peuker, S., Schultz, P. G., and Hubbell, W. L. (2009) Site directed spin labeling of a genetically encoded unnatural amino acid, Proc Natl Acad Sci U S A 106 21637 21642. 140. Fajer, P. (2000) Electron Spin Resonance Spectroscopy Labeling in Peptide and Protein Analysis, in Encyclopedia of Analytical Chemistry (Meyers, R. A., Ed.), pp 5725 5761, John Wiley and Sons, Ltd. 141. Schreier, S., Bozelli, J. C., Jr., Marin, N., Vieira, R. F., and Nakaie, C. R. (2012) The spin label amino acid TOAC and its uses in studies of peptides: chemical, physicochemical, spectroscopic, conformational aspects, Biophys Rev 4 45 66. 142. Polyhach, Y., Bordignon, E., Tschaggelar, R., Gandra, S., Godt, A., and Jeschke, G. (2012) High sensitivity and versatility of the DE ER experiment on nitroxide radical pairs at Q band frequencies, Phys Chem Chem Phys

PAGE 230

230 143. Yang, Z., Liu, Y., Borbat, P., Zweier, J. L., Freed, J. H., and Hubbell, W. L. (2012) Pulsed ESR Dipolar Spectroscopy for Distance Measurements in Immobilized Spin L abeled Proteins in Liquid Solution, J Am Chem Soc 134 9950 9952. 144. Altenbach, C., Kusnetzow, A. K., Ernst, O. P., Hofmann, K. P., and Hubbell, W. L. (2008) High resolution distance mapping in rhodopsin reveals the pattern of helix movement due to acti vation, Proc Natl Acad Sci U S A 105 7439 7444. 145. Durrant, J. D., and McCammon, J. A. (2011) Molecular dynamics simulations and drug discovery, BMC Biol 9 71. 146. Polyhach, Y., Bordignon, E., and Jeschke, G. (2011) Rotamer libraries of spin labelle d cysteines for protein studies, Phys Chem Chem Phys 13 2356 2366. 147. Polyhach, Y., and Jeschke, G. (2010) Prediction of favourable sites for spin labelling of proteins, Spectrosc Int J 24 651 659. 148. Hubbell, W. L., Gross, A., Langen, R., and Liet zow, M. A. (1998) Recent advances in site directed spin labeling of proteins, Curr Op Struct Biol 8 649 656. 149. Fawzi, N. L., Fleissner, M. R., Anthis, N. J., Kalai, T., Hideg, K., Hubbell, W. L., and Clore, G. M. (2011) A rigid disulfide linked nitrox ide side chain simplifies the quantitative analysis of PRE data, J Biomol NMR 51 105 114. 150. Georgieva, E. R., Roy, A. S., Grigoryants, V. M., Borbat, P. P., Earle, K. A., Scholes, C. P., and Freed, J. H. (2012) Effect of freezing conditions on distanc es and their distributions derived from Double Electron Electron Resonance (DEER): a study of doubly spin labeled T4 lysozyme, J Magn Reson 216 69 77. 151. Pistolesi, S., Pogni, R., and Feix, J. B. (2007) Membrane insertion and bilayer perturbation by an timicrobial peptide CM15, Biophys J 93 1651 1660. 152. Fleissner, M. R., Bridges, M. D., Brooks, E. K., Cascio, D., Kalai, T., Hideg, K., and Hubbell, W. L. (2011) Structure and dynamics of a conformationally constrained nitroxide side chain and applicat ions in EPR spectroscopy, Proc Natl Acad Sci U S A 108 16241 16246. 153. Fanucci, G. E., and Cafiso, D. S. (2006) Recent advances and applications of site directed spin labeling, Curr Opin Struct Biol 16 644 653. 154. Riplinger, C., Kao, J. P., Rosen, G. M., Kathirvelu, V., Eaton, G. R., Eaton, S. S., Kutateladze, A., and Neese, F. (2009) Interaction of radical pairs through bond and through space: scope and limitations of the point dipole approximation in

PAGE 231

231 electron paramagnetic resonance spectroscopy, J Am Chem Soc 131 10092 10106. 155. Fajer, P., Brown, L., and Song, L. (2006) Practical Pulsed Dipolar ESR (DEER), Spin Labeling 1 95 128. 156. Schiemann, O., Cekan, P., Margraf, D., Prisner, T. F., and Sigurdsson, S. T. (2009) Relative Orientation of R igid Nitroxides by PELDOR: Beyond Distance Measurements in Nucleic Acids, Angew Chem Int Edit 48 3292 3295. 157. Blackburn, M. E. (2010) Monitoring Polymorphism and Inhibitor induced Conformational Ensemble Shifts in HIV 1 Protease via Pulsed Electron Pa ramagnetic Resonance, in Chemistry p 305, Unviersity of Florida, Gainesville. 158. Ward, R., Bowman, A., El Mkami, H., Owen Hughes, T., and Norman, D. G. (2009) Long distance PELDOR measurements on the histone core particle, J Am Chem Soc 131 1348 1349. 159. Ward, R., Bowman, A., Sozudogru, E., El Mkami, H., Owen Hughes, T., and Norman, D. G. (2010) EPR distance measurements in deuterated proteins, J Magn Reson 207 164 167. 160. Huber, M., Lindgren, M., Hammarstrom, P., Martensson, L. G., Carlsson, U. Eaton, G. R., and Eaton, S. S. (2001) Phase memory relaxation times of spin labels in human carbonic anhydrase II: pulsed EPR to determine spin label location, Biophysical Chemistry 94 245 256. 161. Kim, M., Xu, Q., Fanucci, G. E., and Cafiso, D. S. (2 006) Solutes modify a conformational transition in a membrane transport protein, Biophys J 90 2922 2929. 162. Lopez, C. J., Fleissner, M. R., Guo, Z., Kusnetzow, A. K., and Hubbell, W. L. (2009) Osmolyte perturbation reveals conformational equilibria in spin labeled proteins, Protein Sci 18 1637 1652. 163. Lee, J. C., and Timasheff, S. N. (1981) The stabilization of proteins by sucrose, J Biol Chem 256 7193 7201. 164. Gekko, K., and Timasheff, S. N. (1981) Mechanism of protein stabilization by glycero l: preferential hydration in glycerol water mixtures, Biochemistry 20 4667 4676. 165. Timasheff, S. N. (2002) Protein hydration, thermodynamic binding, and preferential hydration, Biochemistry 41 13473 13482.

PAGE 232

232 166. Lavalette, D., Hink, M. A., Tourbez, M ., Tetreau, C., and Visser, A. J. (2006) Proteins as micro viscosimeters: Brownian motion revisited, Eur Biophys J 35 517 522. 167. Kuttner, Y. Y., Kozer, N., Segal, E., Schreiber, G., and Haran, G. (2005) Separating the contribution of translational and rotational diffusion to protein association, J Am Chem Soc 127 15138 15144. 168. Berliner, L. J., Eaton, S. S., and Eaton, G. E. (2000) Distance Measurements in Biological Systems by EPR Vol. 19, Kluwer Academic/Plenum Publishers, New York. 169. Mao, Q., Schleidt, S., Zimmermann, H., and Jeschke, G. (2008) A pulsed EPR study of surfactant layer structure in composites of a synthetic layered silicate with polystyrene and polycaprolactone, Phys Chem Chem Phys 10 1156 1167. 170. Chiang, Y. W., Borbat, P P., and Freed, J. H. (2005) The determination of pair distance distributions by pulsed ESR using Tikhonov regularization, J Magn Reson 172 279 295. 171. Jeschke, G., Bender, A., Paulsen, H., Zimmermann, H., and Godt, A. (2004) Sensitivity enhancement i n pulse EPR distance measurements, J Magn Reson 169 1 12. 172. Sen, K. I., Logan, T. M., and Fajer, P. G. (2007) Protein dynamics and monomer monomer interactions in AntR activation by electron paramagnetic resonance and double electron electron resonanc e, Biochemistry 46 11639 11649. 173. McHaourab, H. S., Steed, P. R., and Kazmier, K. (2011) Toward the fourth dimension of membrane protein structure: insight into dynamics from spin labeling EPR spectroscopy, Structure 19 1549 1561. 174. Zou, P., and McHaourab, H. S. (2010) Increased sensitivity and extended range of distance measurements in spin labeled membrane proteins: Q band double electron electron resonance and nanoscale bilayers, Biophys J 98 L18 20. 175. Volkov, A., Dockter, C., Bund, T., Pa ulsen, H., and Jeschke, G. (2009) Pulsed EPR determination of water accessibility to spin labeled amino acid residues in LHCIIb, Biophys J 96 1124 1141. 176. Song, Y., Meade, T. J., Astashkin, A. V., Klein, E. L., Enemark, J. H., and Raitsimring, A. (201 1) Pulsed dipolar spectroscopy distance measurements in biomacromolecules labeled with Gd(III) markers, J Magn Reson 210 59 68. 177. Prisner, T., Rohrer, M., and MacMillan, F. (2001) Pulsed EPR spectroscopy: biological applications, Annu Rev Phys Chem 52 279 313.

PAGE 233

233 178. Gabrys, C. M., Yang, J., and Weliky, D. P. (2003) Analysis of local conformation of membrane bound and polycrystalline peptides by two dimensional slow spinning rotor synchronized MAS exchange spectroscopy, J Biomol NMR 26 49 68. 179. An derson, A. C. (2012) Winning the arms race by improving drug discovery against mutating targets, ACS Chem Biol 7 278 288. 180. Csermely, P., Palotai, R., and Nussinov, R. (2010) Induced fit, conformational selection and independent dynamic segments: an e xtended view of binding events, Trends Biochem Sci 35 539 546. 181. Ma, B., and Nussinov, R. (2010) Enzyme dynamics point to stepwise conformational selection in catalysis, Curr Opin Chem Biol 14 652 659. 182. Miller, Y., Ma, B., and Nussinov, R. (2010 ) Polymorphism in Alzheimer Abeta amyloid organization reflects conformational selection in a rugged energy landscape, Chem Rev 110 4820 4838. 183. Huff, J. R. (1991) Hiv Protease a Novel Chemotherapeutic Target for AIDS J Med Chem 34 2305 2314. 184 Wlodawer, A., and Vondrasek, J. (1998) Inhibitors of HIV 1 protease: a major success of structure assisted drug design, Annu Rev Biophys Biomol Struct 27 249 284. 185. Finiguerra, M. G., Prudencio, M., Ubbink, M., and Huber, M. (2008) Accurate long ran ge distance measurements in a doubly spin labeled protein by a four pulse, double electron electron resonance method, Magn Reson Chem 46 1096 1101. 186. Gruene, T., Cho, M. K., Karyagina, I., Kim, H. Y., Grosse, C., Giller, K., Zweckstetter, M., and Beck er, S. (2011) Integrated analysis of the conformation of a protein linked spin label by crystallography, EPR and NMR spectroscopy, J Biomol NMR 49 111 119. 187. Swanson, M. A., Kathirvelu, V., Majtan, T., Frerman, F. E., Eaton, G. R., and Eaton, S. S. (2 009) DEER distance measurement between a spin label and a native FAD semiquinone in electron transfer flavoprotein, J Am Chem Soc 131 15978 15979. 188. Torbeev, V. Y., Raghuraman, H., Mandal, K., Senapati, S., Perozo, E., and Kent, S. B. (2009) Dynamics of "flap" structures in three HIV 1 protease/inhibitor complexes probed by total chemical synthesis and pulse EPR spectroscopy, J Am Chem Soc 131 884 885.

PAGE 234

234 189. Torbeev, V. Y., Raghuraman, H., Hamelberg, D., Tonelli, M., Westler, W. M., Perozo, E., and Ke nt, S. B. (2012) Protein conformational dynamics in the mechanism of HIV 1 protease catalysis, Proc Natl Acad Sci U S A 108 20982 20987. 190. Spinelli, S., Liu, Q. Z., Alzari, P. M., Hirel, P. H., and Poljak, R. J. (1991) The three dimensional structure of the aspartyl protease from the HIV 1 isolate BRU, Biochimie 73 1391 1396. 191. Rick, S. W., Erickson, J. W., and Burt, S. K. (1998) Reaction path and free energy calculations of the transition between alternate conformations of HIV 1 protease, Protein s 32 7 16. 192. Bandaranayake, R. M., Kolli, M., King, N. M., Nalivaika, E. A., Heroux, A., Kakizawa, J., Sugiura, W., and Schiffer, C. A. The effect of clade specific sequence polymorphisms on HIV 1 protease activity and inhibitor resistance pathways, J Virol 84 9995 10003. 193. Klei, H. E., Kish, K., Lin, P. F., Guo, Q., Friborg, J., Rose, R. E., Zhang, Y., Goldfarb, V., Langley, D. R., Wittekind, M., and Sheriff, S. (2007) X ray crystal structures of human immunodeficiency virus type 1 protease mutan ts complexed with atazanavir, J Virol 81 9525 9535. 194. Shen, C. H., Wang, Y. F., Kovalevsky, A. Y., Harrison, R. W., and Weber, I. T. Amprenavir complexes with HIV 1 protease and its drug resistant mutants altering hydrophobic clusters, FEBS J 277 369 9 3714. 195. Liu, F., Kovalevsky, A. Y., Tie, Y., Ghosh, A. K., Harrison, R. W., and Weber, I. T. (2008) Effect of flap mutations on structure of HIV 1 protease and inhibition by saquinavir and darunavir, J Mol Biol 381 102 115. 196. Louis, J. M., Clore G. M., and Gronenborn, A. M. (1999) Autoprocessing of HIV 1 protease is tightly coupled to protein folding, Nat Struct Biol 6 868 875. 197. Jeschke, G., Chechik, V., Ionita, P., Godt, A., Zimmermann, H., Banham, J., Timmel, C. R., Hilger, D., and Jung, H. (2006) DeerAnalysis2006 a comprehensive software package for analyzing pulsed ELDOR data, Appl Magn Reson 30 473 498. 198. Zou, K. H., Tuncali, K., and Silverman, S. G. (2003) Correlation and simple linear regression, Radiology 227 617 622. 199. Sadiq, S. K., and De Fabritiis, G. Explicit solvent dynamics and energetics of HIV 1 protease flap opening and closing, Proteins 78 2873 2885.

PAGE 235

235 200. Tozzini, V., Trylska, J., Chang, C. E., and McCammon, J. A. (2007) Flap opening dynamics in HIV 1 protease explored with a coarse grained model, J Struct Biol 157 606 615. 201. Kozisek, M., Bray, J., Rezacova, P., Saskova, K., Brynda, J., Pokorna, J., Mammano, F., Rulisek, L., and Konvalinka, J. (2007) Molecular analysis of the HIV 1 resistance development: enzymatic activities, crystal structures, and thermodynamics of nelfinavir resistant HIV protease mutants, J Mol Biol 374 1005 1016. 202. Wright, D. W., and Coveney, P. V. (2011) Resolution of discordant HIV 1 protease resistance rankings using molecular dynamics simulations, J Chem Inf Model 51 2636 2649. 203. Rhee, S. Y., Fessel, W. J., Zolopa, A. R., Hurley, L., Liu, T., Taylor, J., Nguyen, D. P., Slome, S., Klein, D., Horberg, M., Flamm, J., Follansbee, S., Schapiro, J. M., and Shafer, R. W. (2005) HIV 1 Protease and reverse transcriptase mutations: correlations with antiretroviral therapy in subtype B isolates and implications for drug resistance surveillance, J Infect Dis 192 456 465. 204. Ode, H., Matsuyama, S., Hata, M., Neya, S., Kakizawa, J., Sugiura, W., and Hoshino, T. (2007) Computational characterization of structural role of the non active site mutation M36I of human immunodeficiency virus type 1 protease, J Mol Biol 370 598 607. 205. Weinberg, S. L., and Abramowitz, S. K., (Eds.) (2008 ) Statistics using SPSS: an Integrative Approach 206. Ishima, R., and Torchia, D. A. (2003) Extending the range of amide proton relaxation dispersion experiments in proteins using a constant time relaxation compensated CPMG approach, J Biomol NMR 25 243 248. 207. Katoh, E., Louis, J. M., Yamazaki, T., Gronenborn, A. M., Torchia, D. A., and Ishima, R. (2003) A solution NMR study of the binding kinetics and the internal dynamics of an HIV 1 protease substrate complex, Protein Sci 12 1376 1385. 208. Shes kin, D. J. (2007) Handbook of Parametric and Nonparametric Statistical Procedures 4 ed., Chapman and Hall/CRC, New York. 209. Karthik, S., and Senapati, S. (2011) Dynamic flaps in HIV 1 protease adopt unique ordering at different stages in the catalytic cycle, Proteins 79 1830 1840. 210. Agniswamy, J., Shen, C. H., Aniana, A., Sayer, J. M., Louis, J. M., and Weber, I. T. ( 2012) HIV 1 Protease with 20 Mutations Exhibits Extreme Resistance to Clinical Inhibitors through Coordinated Structural Rearrangements, Biochemistry 51 2819 2828.

PAGE 236

236 211. de Vera, I. M. S., Blackburn, M. E., and Fanucci, G. E. (2012) Correlating conformatio nal shift induction with altered inhibitor potency in a multidrug resistant HIV 1 protease variant, Biochemistry 51 7813 7815. 212. Mittal, S., Cai, Y., Nalam, M. N., Bolon, D. N., and Schiffer, C. A. (2012) Hydrophobic core flexibility modulates enzyme activity in HIV 1 protease, J Am Chem Soc 134 4163 4168. 213. Perryman, A. L., Lin, J. H., and McCammon, J. A. (2004) HIV 1 protease molecular dynamics of a wild type and of the V82F/I84V mutant: possible contributions to drug resistance and a potential new target site for drugs, Protein Sci 13 1108 1123. 214. Cai, Y., Yilmaz, N., Myint, W., Ishima, R., and Schiffer, C. A. (2012) Differential Flap Dynamics in Wild Type and a Drug Resistant Variant of HIV 1 Protease Revealed by Molecular Dynamics and NMR Relaxation., J Chem Theory Comput 215. Kempf, J. G., and Loria, J. P. (2004) Measurement of intermediate exchange phenomena, Methods Mol Biol 278 185 231. 216. Loria, J. P., Berlow, R. B., and Watt, E. D. (2008) Characterization of enzyme motions by s olution NMR relaxation dispersion, Acc Chem Res 41 214 221. 217. Myint, W., Cai, Y., Schiffer, C. A., and Ishima, R. (2012) Quantitative comparison of errors in (15)N transverse relaxation rates measured using various CPMG phasing schemes, J Biomol NMR 218. Rhee, S. Y., Taylor, J., Fessel, W. J., Kaufman, D., Towner, W., Troia, P., Ruane, P., Hellinger, J., Shirvani, V., Zolopa, A., and Shafer, R. W. (2010) HIV 1 protease mutations and protease inhibitor cross resistance, Antimicrob Agents Chemother 54 4253 4261. 219. Bandaranayake, R. M., Kolli, M., King, N. M., Nalivaika, E. A., Heroux, A., Kakizawa, J., Sugiura, W., and Schiffer, C. A. (2010) The effect of clade specific sequence polymorphisms on HIV 1 protease activity and inhibitor resistance path ways, J Virol 84 9995 10003. 220. Olsen, D. B., Stahlhut, M. W., Rutkowski, C. A., Schock, H. B., vanOlden, A. L., and Kuo, L. C. (1999) Non active site changes elicit broad based cross resistance of the HIV 1 protease to inhibitors, J Biol Chem 274 236 99 23701. 221. Palmer, S., Shafer, R. W., and Merigan, T. C. (1999) Highly drug resistant HIV 1 clinical isolates are cross resistant to many antiretroviral compounds in current clinical development, Aids 13 661 667.

PAGE 237

237 222. Polyhach, Y., Bordignon, E., an d Jeschke, G. Rotamer libraries of spin labelled cysteines for protein studies, Phys Chem Chem Phys 13 2356 2366. 223. Peng, J. W., and Wagner, G., (Eds.) (1994) Nuclear magnetic resonance probes of molecular dynamics Kluwer Academic Publishers, Norwell 224. Ishima, R., Louis, J. M., and Torchia, D. A. (1999) Transverse 1H cross relaxation in 1H 15N correlated 1H CPMG experiments, J Magn Reson 137 289 292. 225. Cavanagh, J., Fairbrother, W. J., Palmer, A. G., Skelton, N. J., and Rance, M. (2007) Prot ein NMR spectroscopy 2 ed., Elsevier Academic Press, Burlington. 226. Wasylishen, R. E. (1996) NMR relaxation and dynamics, in NMR Spectroscopy T echniques (Bruch, M. D., Ed.) 2 ed., pp 105 144, Marcel Dekker, Inc., New York. 227. Barbar, E., Hare, M., D aragan, V., Barany, G., and Woodward, G. (1998) Dynamics of the conformational ensemble of partially folded bovine pancreatic trypsin inhibitor, Biochemistry 37 7822 7833. 228. Barbar, E., Hare, M., Daragan, V., Barany, G., and Woodward, C. (1998) Slow a nd fast dynamics of a fluctuating ensemble of partially folded BPTI conformations, Biophys J 74 A11 A11. 229. McKay, R. T., Tripet, B. P., Hodges, R. S., and Sykes, B. D. (1997) Interaction of the second binding region of troponin I with the regulatory d omain of skeletal muscle troponin C as determined by NMR spectroscopy, J Biol Chem 272 28494 28500. 230. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J., and Bax, A. (1995) Nmrpipe a Multidimensional Spectral Processing System Based on Unix Pipes, J Biomol NMR 6 277 293. 231. Grzesiek, S., and Bax, A. (1993) The Importance of Not Saturating H 2 O in Protein NMR Application to Sensitivity Enhancement and NOE Measurements, J Am Chem Soc 115 12593 12594. 232. d'Auvergne, E. J., and Goo ley, P. R. (2008) Optimisation of NMR dynamic models II. A new methodology for the dual optimisation of the model free parameters and the Brownian rotational diffusion tensor, J Biomol NMR 40 121 133. 233. d'Auvergne, E. J., and Gooley, P. R. (2008) Opti misation of NMR dynamic models I. Minimisation algorithms and their performance within the model free and Brownian rotational diffusion spaces, J Biomol NMR 40 107 119.

PAGE 238

238 234. Bieri, M., d'Auvergne, E. J., and Gooley, P. R. relaxGUI: a new software for fas t and simple NMR relaxation data analysis and calculation of ps ns and mus motion of proteins, J Biomol NMR 50 147 155. 235. Kamath, U., and Shriver, J. W. (1989) Characterization of thermotropic state changes in myosin subfragment 1 and heavy meromyosin by UV difference spectroscopy, J Biol Chem 264 5586 5592. 236. Huang, X., De Vera, I. M. S., Veloro, A. M., Rocca, J. R., Simmerling, C., Dunn, B. M., and Fanucci, G. E. (2012) Backbone 1H, 13C and 15N Chemical Shift Assignments for HIV 1 Protease Varia nts, Biomol NMR Assign, DOI 10.1007/s12104 012 9409 7 237. Lipari, G., and Szabo, A. (1982) Analysis of Nmr Relaxation Data on Macromolecules Using the Model Free Approach, Biophys J 37 A380 A380. 238. Lipari, G., and Szabo, A. (1982) Model Free Approa ch to the Interpretation of Nuclear Magnetic Resonance Relaxation in Macromolecules .2. Analysis of Experimental Results, J Am Chem Soc 104 4559 4570. 239. Reim, F., and Speicher, D. W. (2004) Unit 11.3, in Current Protocols in Protein Science (Coligan, J. E., Dunn, B. M., Speicher, D. W., and Wingfield, P. T., Eds.), John Wiley and Sons, Inc. 240. Geer, L. Y., Markey, S. P., Kowalak, J. A., Wagner, L., Xu, M., Maynard, D. M., Yang, X., Shi, W., and Bryant, S. H. (2004) Open mass spectrometry search algo rithm, J Proteome Res 3 958 964. 241. Boeckmann, B., Bairoch, A., Apweiler, R., Blatter, M. C., Estreicher, A., Gasteiger, E., Martin, M. J., Michoud, K., O'Donovan, C., Phan, I., Pilbout, S., and Schneider, M. (2003) The SWISS PROT protein knowledgebase and its supplement TrEMBL in 2003, Nucleic Acids Res 31 365 370. 242. Glish, G. L., and Vachet, R. W. (2003) The basics of mass spectrometry in the twenty first century, Nat Rev Drug Discov 2 140 150. 243. Kirkpatrick, D. S., Gerber, S. A., and Gygi, S. P. (2005) The absolute quantification strategy: a general procedure for the quantification of proteins and post translational modifications, Methods 35 265 273. 244. Banerjee, S., and Mazumdar, S. Electrospray ionization mass spectrometry: a technique to access the information beyond the molecular weight of the analyte, Int J Anal Chem 2012 282574.

PAGE 239

239 245. El Aneed, A., Cohen, A., and Banoub, J. (2009) Mass Spectrometry, Review of the Basics: Electrospray, MALDI, and Commonly Used Mass Analyzers, Appl S pec trosc Rev 44 210 230. 246. Schreiber, G., and Fersht, A. R. (1995) Energetics of Protein Protein Interactions Analysis of the Barnase Barstar Interface by Single Mutations and Double Mutant Cycles, J Mol Biol 248 478 486.

PAGE 240

240 BIOGRAPHICAL SKETCH Ian Mitchelle Sayo de Vera was born in 1982 in Manila, Philippines. He obtained his Bachelor of Science degree in c hemistry at the Ateneo de Manila University in March 2003. In October of 2005, he obtained his second Bachelor of Science degree in Computer Engineering from the same institution. After graduation, he worked as an NMR spectroscopist at the National Chemistry Instrumentation Center (NCIC) which houses one of the few NMR facilities in the Philippines. While employed in NCIC, he pursued Master of Science degree in Chemistry and wor ked as an Assistant Instructor at Ateneo de Manila University. In August of 2008, he applied to graduate school and got accepted in the Department of Chemistry at the University of Florida. A month later, he joined D r. He received his Ph.D. from the University of Florida in the fall of 2012.