UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Alteration of poly(adp-ribose) polymerase activity influences the mode of oxidant-induced endothelial… Walisser, Jacqueline Anne 1999

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

Item Metadata

Download

Media
831-ubc_1999-389952.pdf [ 16.8MB ]
Metadata
JSON: 831-1.0089331.json
JSON-LD: 831-1.0089331-ld.json
RDF/XML (Pretty): 831-1.0089331-rdf.xml
RDF/JSON: 831-1.0089331-rdf.json
Turtle: 831-1.0089331-turtle.txt
N-Triples: 831-1.0089331-rdf-ntriples.txt
Original Record: 831-1.0089331-source.json
Full Text
831-1.0089331-fulltext.txt
Citation
831-1.0089331.ris

Full Text

ALTERATION OF POLY(ADP-RIBOSE) POLYMERASE ACTIVITY INFLUENCES T H E MODE OF OXIDANT-INDUCED ENDOTHELIAL C E L L DEATH  by JACQUELINE ANNE WALISSER B . S c , Simon Fraser University, 1988  A THESIS SUBMITTED IN P A R T I A L F U L F I L L M E N T OF THE REQUIREMENTS FOR T H E D E G R E E OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES The Faculty of Pharmaceutical Sciences Division of Pharmacology and Toxicology  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH C O L U M B I A April 1999  © Jacqueline Anne Walisser, 1999  In  presenting  degree freely  at  the  available  copying  of  department publication  this  of  in  partial  fulfilment  University  of  British  Columbia,  for  this or  thesis  reference  thesis by  this  for  his thesis  and  study.  scholarly  or  her  for  of I  I further  purposes  gain  shall  requirements  agree  that  agree  may  representatives.  financial  the  be  It not  is  of  PhMm(XCtLVh'COlI  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  r^0AA^^J f ?9 C C  Library  be  JUMICQO  an  granted  by  allowed  advanced  shall  permission for  understood  permission.  Department  that  the  for  the that  without  make  it  extensive  head  of  my  copying  or  my  written  ABSTRACT Poly(ADP-ribose) polymerase (PARP) is a zinc-finger enzyme activated by D N A strand scissions, as occur during oxidative stress. Extensive PARP activation causes depletion of its substrate, N A D , and a consequent decline in cellular energy, which contributes to oncotic cell +  death. Cell death is categorized either as oncosis, a passive form of death, or apoptosis, an active energy-dependent form of death. This thesis examines the effect of altering PARP activity on the cellular response to oxidative injury and the influence that P A R P activity has on the process of endothelial cell (EC) death. Fluorescence microscopy reveals hallmark nuclear features consistent with apoptosis in oxidant-stressed ECs when PARP is inhibited with  1,5-  isoquinolinediol and oncosis when exposed to oxidant alone. Caspase-3-like activity, a marker of apoptosis, is negligible in oxidant-treated ECs but increases when P A R P activity is inhibited. These results demonstrate that the process of oxidant-induced E C death is switched from oncosis to apoptosis by pharmacological inhibition of PARP activity. In accordance with this switch to an active form of death, PARP inhibition maintains energy-dependent cellular processes, such as retention of low concentrations of intracellular free calcium and lysosomal uptake of acridine orange. Metal replacement in the PARP zinc fingers is a potential mechanism of altering PARP activity that may have important consequences in metal toxicology. Because zinc is essential for both the ability of PARP to bind D N A and the activation of its catalytic activity, the potential for altering the DNA-binding capacity of P A R P through metal replacement in the zinc-finger motif is investigated. Purified PARP, derived from a baculovirus expression system, is employed for these studies. Various mass spectrometric methods, including laser ablation-ICPMS, M A L D I TOF MS and ESI-MS, are evaluated for their effectiveness in the detection of metal replacement in purified PARP. Analysis of the native PARP is hindered at neutral pH due to suppression of  ii  ionization and protein adsorption to instrument components. However, analysis of the 24 kDa caspase-3-digested PARP fragment containing the 2 zinc fingers is accomplished by ESI-MS using  a  solvent  composed  of  20%  acetonitrile  containing  53  mM  1,1,1,3,3,3-  hexafluoroisopropanol pH 5.1. Therefore, the groundwork is laid to examine metal replacement in PARP as a potential mechanism for alteration of PARP activity. In conclusion, alteration of PARP activity, either via its catalytic site or at the zinc-finger site, may have important toxicological consequences in a variety of processes including metal toxicity and oxidantinduced cell death.  iii  TABLE OF CONTENTS Page  ABSTRACT  ii  T A B L E OF CONTENTS  iv  LIST OF T A B L E S  ix  LIST OF FIGURES  x  LIST OF A B B R E V I A T I O N S  xiv  ACKNOWLEDGEMENTS  xvi  DEDICATION  xviii  /. PROLOGUE  1  2. INTRODUCTION  2  2.1.  2  Oxidative Stress 2.1.1. 2.1.2. 2.1.3. 2.1.4.  Chemistry of Oxidative Stress Biological Sources of Reactive Oxygen Species Antioxidant Defense Mechanisms Oxidative Stress and the Endothelium  2 4 5 6  2.2.  Metal Toxicity  7  2.3.  Cell Death  9  2.4.  2.5.  The 2.4.1. 2.4.2. 2.4.3. 2.4.4. 2.4.5.  Processes of Cell Death Oncosis Apoptosis The Role of Caspases in the Execution of Apoptosis Determination of the Mode of Cell Death in vitro The Fate of a Cell: Apoptosis versus Oncosis  Poly(ADP-ribose) Polymerase 2.5.1. PARP Functional Domains 2.5.2. Cellular Function of PARP 2.5.3. Poly(ADP-ribose) Polymerase and Oxidant-Induced Cell Death 2.5.3.1. PARP: A Caspase-3 Substrate in Apoptosis  2.6.  Research Hypothesis 2.6.1. P A R P Activation and the Turning-Point Between Oncosis and Apoptosis  11 11 12 14 16 19 20 21 25 26 27 29 29  iv  2.7.  Modulation of PARP Activity 2.7.1. Inhibition of P A R P Catalytic Activity 2.7.2. Modulation of P A R P DNA-Binding Capacity 2.7.2.1. Evidence for Metal Replacement in Zinc-finger Motifs 2.7.2.2. Evidence for Metal Replacement in the Zinc Fingers of P A R P 2.7.3. Mass Spectrometric Analysis of Metal-Protein Complexes 2.7.3.1. Matrix-Assisted Laser Desorption Ionization - Time of Flight Mass Spectrometry 2.7.3.2. Electrospray Ionization Mass Spectrometry  2.8.  Thesis Objectives  31 33 34 35 36 37 37 38 42  3. EXPERIMENTAL  44  3.1.  Materials  44  3.2.  Cell Culture Techniques  45  3.3.  3.2.1. Cell Culture of Bo vine Pulmonary Artery Endothelial Cells  45  3.2.2. Cell Culture of Spodoptera frugiperda (S/9) Cells  45  Recombinant PARP Production Using the Baculovirus Expression System  46  3.3.1. AcPARP Virus Propagation 3.3.2. Assessment of Viral Titre 3.3.3. Effect of MOI on Extent and Time-Course of Recombinant P A R P Expression 3.3.4. Routine Expression of Recombinant P A R P 3.4. Purification of Recombinant PARP  46 47 47 48 48  3.4.1. Preparation of a Crude Cell Extract 3.4.2. Preparation of the 3AB-Affigel Affinity Chromatography Resin 3.4.3. Final Purification of P A R P by Affinity Chromatography  48 50 50  3.5.  Protein Quantitation  51  3.6.  Protein Analysis by SDS-Polyacrylamide Gel Electrophoresis  52  3.7.  PARP Enzyme Activity Assay  53  3.8.  Proteolytic Digestion of PARP  55  3.9.  3.8.1. Trypsin Digestion  55  3.8.2. Caspase-3 Digestion  55  Separation of Caspase-3-Digested PARP Fragments 3.9.1. 3.9.2. 3.9.3. 3.9.4.  Centrifugal Ultrafiltration Under Non-Denaturing Conditions Centrifugal Ultrafiltration Under Denaturing Conditions Affinity Chromatography H P L C Separation of P A R P Fragments Under Acidic Conditions  56 56 57 58 61  v  3.10. Mass Spectrometric Analysis of Native PARP and PARP Fragments 3.10.1. Laser Ablation-Inductively Coupled Plasma Mass Spectrometry (LA-ICPMS) of PARP Bands on Acrylamide Gels 3.10.2. Inductively Coupled Plasma Mass Spectrometry (ICPMS) of PARP in Solution 3.10.3. Protein Analysis by Matrix Assisted Laser Desorption Ionization Time-of-Flight Mass Spectrometry (MALDI-TOF MS) 3.10.4. Protein Analysis by Electrospray Ionization Mass Spectrometry (ESI-MS) 3.11. Endothelial Cell Injury Experiments 3.11.1. Preparation of EC for Cell Injury Experimentation 3.11.2. Cell Injury Assessment 3.11.3. Cell Injury Data Analysis  64 64 65 67 68 71 71 72 72  3.12. Assessment of Caspase Activity in Oxidant-Stressed Endothelial Cells  73  3.13. Immunoblot Analysis of PARP Cleavage by Caspase  74  3.14. Fluorescence Microscopy and Morphologic Assessment  75  3.15. Assessment of Intracellular Free Calcium in Oxidant-Stressed Endothelial Cells  76  3.16. Data Analysis  77  4. RESULTS  78  4.1.  78  4.2.  Recombinant PARP Production Using the Baculovirus Expression System 4.1.1. AcPARP Virus Propagation and Viral Titre Determination  78  4.1.2. Effect of M O I on Time-Course of PARP Expression  78  Purification of Recombinant PARP  4.2.1. Preparation of a Crude Cell Extract 4.2.2. Affinity Chromatography for PARP Purification 4.2.2.1. Modification of the Elution Buffer Composition 4.2.2.2. Direct Application of the Cell Extract to 3AB-Affinity Column 4.2.3. Yield of P A R P from the Purification Procedure 4.3. PARP Enzyme Assay Development 4.3.1. Effect of Temperature on P A R P Activity 4.3.2. Effect of D N A Sonication Time on P A R P Activity 4.3.3. Effect of the Presence or Absence of Histones on PARP Activity 4.3.4. Effect of Different Preparations of Histones on PARP Activity 4.3.5. Effect of Histone Concentration on P A R P Activity 4.3.6. Effect of the Presence or Absence of D N A on P A R P Activity 4.3.7. Effect of the Amount of Histones and D N A on P A R P Activity 4.3.8. Effect of Increasing Amounts of PARP on PARP Activity 4.3.9. The Kinetics of PARP Activity 4.3.10. Substrate Saturation Curve for PARP Activity  81 81 85 85 86 89 90 90 90 94 95 95 98 99 101 102 103  4.3.11. Standardized Assay for P A R P Activity 4.4.  4.5.  103  Inhibition of PARP Activity  105  4.4.1. Reaction Product Inhibition by Nicotinamide  105  4.4.2. Inhibition of P A R P Activity with Known and Novel Inhibitors  105  Proteolytic Digestion of PARP  108  4.5.1. P A R P Digestion by Trypsin 4.5.2. P A R P Digestion by Caspase-3 4.6. Separation of Caspase-3-Digested PARP Fragments  108 109 112  4.6.1. Centrifugal Ultrafiltration of P A R P Fragments 112 4.6.1.1. Centrifugal Ultrafiltration Under Non-Denaturing Conditions 112 4.6.1.2. Centrifugal Ultrafiltration Under Denaturing Conditions 113 4.6.2. Affinity Chromatography for Separation of P A R P Fragments 116 4.6.3. Separation of Caspase-3-Digested P A R P Fragments by H P L C 120 4.6.3.1. H P L C of Basic Proteins Under Acidic Conditions 120 4.6.3.2. Development of an H P L C Method for Separation of P A R P Fragments Under Neutral Conditions 124 4.6.3.3. Choice of H P L C Column for Protein Separation 126 4.7.  Mass Spectrometric Analysis of Native PARP and PARP Fragments  127  4.7.1. Development of a LA-ICPMS Method for Analysis of P A R P Metal Content 4.7.1.1. L A - I C P M S Analysis of Zinc Isotopes in P A R P Bands on a Polyacrylamide Gel 4.7.1.2. L A - I C P M S Analysis of Sulfur Isotopes in P A R P Bands on a Polyacrylamide Gel 4.7.1.3. Analysis of Zinc in a PARP Solution by ICPMS 4.7.2. M A L D I - T O F Mass Spectrometric Analysis of Intact P A R P and P A R P Fragments 4.7.3. ESI-MS Analysis of Native P A R P and P A R P Fragments 4.7.3.1. ESI-MS Analysis of Basic Proteins Under Acidic Conditions 4.7.3.2. Comparison Between Positive and Negative Ion Electrospray M S 4.7.3.3. Development of an ESI-MS Method for Analysis of Basic Proteins Under Neutral Conditions 4.7.3.4. Protein Adsorbance to the Instrument Components 4.8.  127 127 127 130 131 136 136 141 147 154  Endothelial Cell Injury  160  4.8.1. Effect of H 0 on the Time-Course and Extent of E C Death 4.8.2. Effect of the Combination of P A R P and Protein Synthesis Inhibition on H 0 Induced E C Death 4.8.3. Effect of P A R P and Endonuclease Inhibition on H 0 -Induced E C Death 4.8.4. Effect of Staurosporine on the Time-Course and Extent of E C Death 2  2  2  2  4.9.  2  Caspase Activation in Endothelial Cells 4.9.1. Caspase-3-like Activity in Oxidant-Stressed E C 4.9.2. Caspase-3-like Activity in E C Treated with Staurosporine  160  2  160 163 165 166 166 171  vii  4.10. Caspase-Induced Cleavage of PARP in Oxidant-Injured E C 4.10.1. Determination of Cell Number Required for Western Blot Analysis of PARP 4.10.2. Western Blot of PARP Cleavage in Oxidant-Injured E C  174 174 174  4.11. Morphological Assessment of Apoptosis in E C  176  4.12. The Effect of PARP Inhibition on [Ca ]i in Oxidant-Stressed E C  180  4.13. The Effect of DIQ is Not a Result of Antioxidant Activity  182  5. DISCUSSION  186  5.1.  Zinc Finger Metal Replacement and Modulation of PARP Activity  186  5.2.  Biological Consequences of Pharmacological Inhibition of PARP Activity  201  2+  6. SUMMARY 7. FUTURE  & CONCLUSIONS DIRECTIONS  8. REFERENCES  208 211 214  LIST OF TABLES Page  Table 1. Summary of H P L C conditions used in the development of a method to separate PARP fragments under acidic and non-acidic conditions 63 Table 2. The operating conditions for the V G Q U A T T R O mass spectrometer used for positive ion ESI-MS of proteins 69 Table 3. The solvent systems evaluated for the ESI-MS analysis of proteins under acidic and neutral conditions Table 4. Standardized assay components and incubation time for the P A R P enzyme assay  70 103  Table 5. Summary of results from the development of an H P L C assay to separate caspase-3digested PARP fragments under acidic and neutral conditions 121 Table 6. Comparison of molecular weights and PI of model proteins used for ESI-MS method development 137  ix  LIST OF FIGURES Page  Fig. 1. Scheme showing the process of ADP-ribosylation of an acceptor protein  21  Fig. 2. Scheme showing the functional domains of PARP  22  Fig. 3. Amino acid sequence of the zinc fingers of PARP  23  Fig. 4. Scheme illustrating the proposed role of PARP in the modulation of oxidant-stressed endothelial cell death from oncosis to apoptosis  30  Fig. 5. Scheme illustrating the role of PARP mediating the process of cell death following exposure to a D N A damaging agent  32  Fig. 6. Scheme depicting the process of laser ablation - inductively coupled plasma mass spectrometry (LA-ICPMS) of PARP protein bands localized on a polyacrylamide gel by SDS-PAGE  66  Fig. 7. Effect of different MOI's on the time-course of protein expression in AcPARPinfected 5/9 cells  80  Fig. 8. Purification of rePARP from AcPARP-infected SJ9 cells  83  Fig. 9. Concentration of DIQ required to elute PARP from the 3AB-affinity column  87  Fig. 10. Purification of rePARP from a partially purified AcPARP-infected 5/9 cell extract by 3AB-Affigel affinity chromatography  88  Fig. 11. Effect of incubation temperature on rePARP enzyme activity  92  Fig. 12. Effect of duration of D N A sonication on rePARP activity  93  Fig. 13. Effect of the presence or absence of histones on PARP activity  94  Fig. 14. Comparison of Gibco histone H I and Worthington lysine-rich histones on PARP activity  96  Fig. 15. Effect of histone amount on PARP activity  97  Fig. 16. Effect of the presence or absence of D N A on PARP activity  98  Fig. 17. Effect of various amounts of histone H I and D N A on PARP activity  100  Fig. 18. Effect of increasing PARP amount on activity..  101  Fig. 19. Kinetics of PARP activity  102 x  Fig. 20. Substrate saturation curve for N A D in the PARP enzyme assay  104  Fig. 21. Inhibition of PARP activity by nicotinamide  106  +  Fig. 22. Inhibition of PARP activity by the known PARP inhibitors, DIQ and 3AB, and allopurinol  107  Fig. 23. Time-course of PARP digestion with trypsin  110  Fig. 24. Time-course of P A R P digestion by caspase-3  Ill  Fig. 25. Separation of PARP fragments by centrifugal ultrafiltration  114  Fig. 26. Separation of PARP fragments by centrifugal ultrafiltration in the presence of 8 M urea Fig. 27. Differential binding affinity of 3AB-Affigel for the 89 kDa PARP fragment under  115  different KC1 concentrations  117  Fig. 28. Binding affinity of two different preparations of 3AB-Affigel for digested P A R P  118  Fig. 29. Binding affinity of two different preparations of 3AB-Affigel for native P A R P  119  Fig. 30. H P L C separation and SDS-PAGE identification of caspase-3digested PARP fragments Fig. 31. Abundance of zinc isotopes in PARP protein bands isolated by SDS-PAGE and in the blank gel measured by LA-ICPMS  122 128  Fig. 32. Abundance of sulfur isotopes in PARP protein bands isolated by SDS-PAGE and in the blank gel measured by LA-ICPMS  129  Fig. 33. Analysis of purified PARP in solution by ICPMS  130  Fig. 34. M A L D I - T O F MS analysis of native PARP  133  Fig. 35. M A L D I - T O F MS analysis of caspase-3-digested PARP  134  Fig. 36. M A L D I - T O F MS analysis of caspase-3-digested PARP without T F A present in the matrix mixture  135  Fig. 37. ESI-MS analysis of trypsinogen under acidic conditions  138  Fig. 38. ESI-MS analysis of cytochrome c under acidic conditions  139  Fig. 39. ESI-MS analysis of the 24 kDa PARP fragment under acidic conditions  140  xi  Fig. 40. Comparison between the raw spectra obtained following ESI-MS analysis of myoglobin under acidic positive electrospray conditions, neutral positive electrospray conditions, and neutral negative electrospray conditions  143  Fig. 41. Comparison between the MaxEnt transformed spectra obtained following ESI-MS analysis of myoglobin under acidic positive electrospray conditions, neutral positive electrospray conditions, and neutral negative electrospray conditions  145  Fig. 42. Relative abundance of trypsinogen in the total ion chromatogram following ESI-MS analysis under acidic and neutral conditions 150 Fig. 43. ESI-MS analysis of cytochrome c under neutral conditions using hexafluoroisopropanol  151  Fig. 44. ESI-MS analysis of myoglobin under neutral conditions using hexafluoroisopropanol  152  Fig. 45. ESI-MS analysis of the purified 24 kDa P A R P fragment under neutral conditions using hexafluoroisopropanol  153  Fig. 46. Adsorption of cytochrome c to components of the H P L C and/or mass spectrometer and its release following injection of formic acid  157  Fig. 47. Adsorption of the 24 kDa P A R P fragment to components of the H P L C and/or mass spectrometer and its release following injection of formic acid 158 Fig. 48. The time-course of irreversible E C injury after exposure to increasing concentrations of H2O2  161  Fig. 49. The effect of P A R P and protein synthesis inhibition on the time-course of H 0 -induced E C death  162  Fig. 50. The effect of P A R P and endonuclease inhibition on the time-course of H 0 -induced E C death  164  Fig. 51. The time-course of irreversible E C injury after exposure to increasing concentrations of staurosporine  165  Fig. 52. Comparison between the time-course of caspase-3-like activation and loss of cell membrane integrity in PARP-inhibited E C treated with H 0  168  Fig. 53. The relationship between H 0 concentration and total caspase-3-like activity in PARP-inhibited and non-PARP-inhibited E C  170  Fig. 54. Comparison between the time-course of caspase-3-like activation and loss of cell membrane integrity in E C treated with staurosporine  172  Fig. 55. The concentration dependence of total caspase-3-like activity in E C treated with staurosporine  173  2  2  2  2  2  2  2  2  Fig. 56. Determination of the number of EC required for immunoblot analysis of P A R P  174  Fig. 57. PARP cleavage in oxidant-stressed E C treated with a P A R P inhibitor  175  Fig. 58. The morphological assessment for apoptosis and oncosis in oxidant-stressed E C following P A R P inhibition  178  Fig. 59. Intracellular free calcium concentration ([Ca ]i) in oxidant-stressed E C following PARP inhibition  181  Fig. 60. Effect of PARP inhibition and antioxidant treatment on reversing the oxidantinduced rise in [Ca ]; and loss of cell viability  184  Fig. 61. The effect of addition of a PARP inhibitor at 0 h or 2 h after the initial oxidant exposure on caspase-3-like activity  185  2+  2+  LIST OF ABBREVIATIONS 3AB  3-arninobenzamide  AcMNPV  Autographa californica multicapsid nucleopolyhedrosis virus  AcPARP  recombinant A c M N P V containing the full-length gene for human P A R P  ACN  acetonitrile  AO  acridine orange  AP  allopurinol  ATA  aurintricarboxylic acid  P-ME  (3-mercaptoethanol  [Ca ]i  intracellular free calcium  CC  cytochrome c  CHAPS  3-[(3-cholamidopropyl)dimethylammonio]-l-propane sulfonate  CHX  cycloheximide  Da  dalton  DEVD-AFC  (Asp-Glu-Val-Asp)-7-amino-4-trifluoromethyl coumarin  DIQ  1,5-isoquinolinediol  DMSO  dimethyl sulfoxide  DTT  dithiothreitol  EC  endothelial cell  EDTA  ethylenediamine tetraacetic acid  ESI-MS  electrospray ionization mass spectrometry  HBSS  Hanks balanced salt solution  HEPES  A^-(2-hydroxyethyl)piperizine-A '-2-ethanesulfonic acid  2+  HFIPA  /  1,1,1,3,3,3-hexafluoro-2-propanol  HPLC  high-performance liquid chromatography  HO*  hydroxyl free radical  ICPMS  inductively coupled plasma mass spectrometry  IC50  50% inhibitory concentration  I.D.  internal diameter  kDa  kilodalton  LA-ICPMS  laser ablation - inductively coupled plasma mass spectrometry  MG  myoglobin  xiv  MALDI-TOF MS  matrix assisted laser desorption time-of-flight mass spectrometry  MOI  multiplicity of infection  MW  molecular weight  NAD  +  nicotinamide adenine dinucleotide (oxidized form)  NMWL  nominal molecular weight limit  m/z  mass to charge ratio  PAGE  polyacrylamide gel electrophoresis  PARP  poly(ADP-ribose) polymerase  PBS  phosphate buffered saline  pfu  plaque forming units  PI  propidium iodide  PMSF  phenylmethylsulfonylfluoride  r  coefficient of determination  2  rePARP  recombinant poly(ADP-ribose) polymerase  SD  standard deviation  SDS  sodium dodecyl sulfate  SEM  standard error of the mean  S/9  Spodoptera frugiperda  oss  superoxide anion  TEMED  A-A-A-A'-tetramethylethylenediamine  TCA  trichloroacetic acid  TCID50  tissue culture infection dose  TFA  trifluoroacetic acid  TG  trypsinogen  UV  ultraviolet  wt  wild-type  YO-PRO  YO-PRO-1-iodide  2  cells  staurosporine  XV  ACKNOWLEDGEMENTS I would like to sincerely thank my advisor, Dr. Robert Thies, for his guidance, perseverance, understanding and, especially, for his friendship. His interest in research and his enthusiasm helped me stay motivated. Thank you for teaching me that the most important lesson of graduate studies was to have fun (and to "get out"!). I would also like to thank the members of my Research Committee, Drs. Frank Abbott, Anne Autor, Gail Bellward and Ron Reid for their continuous interest and valuable contributions to this work. It was a pleasure and an honor to work with each of you. There are a number of people in the Faculty of Pharmaceutical Sciences, including Drs. Stelvio Bandiera, Keith McErlane and Ron Reid, that I would like to acknowledge for generously allowing me to use their laboratories and equipment to complete my research. A special thank you goes to Dr. David Theilmann, Department of Plant Science, U B C , for allowing me to work in his lab and use the baculovirus expression system (all in exchange for racking pipet tips!). I am gratefully indebted to Mr. Roland Burton for his hard work and assistance in performing the protein mass spectrometry and especially for showing me the wonderful mountains of British Columbia and the Rockies. Thanks to Alan, Barry, Dag, David, Deanna, Eileen, Linda, Lynn and Terry for your support and for your constant reminders of the important things in life. To Dr. Lilian Clohs and Joan Cosar thank-you for your continuous friendship. I would especially like to thank Dr. Virginia Borges for her willingness to read this thesis. The opportunity to meet a person who will be a life-long friend does not come along very often, and I am grateful, Virginia, that we have met. To Donna and Lee Nicol, Barbara McPhate, Irene Hsu and Melinda Marshall friendships that stand the test of time are more valuable than anything, and I am privileged to  xvi  have you each as my friend. To my friend Richard Poulter, for his love, constant belief in my abilities and for sharing some wonderful travelling experiences - terima kasih. To the Baculovirus Queen, Dr. Cindy Shippam, thanks for sharing both the joy and misery of grad school and life in general. I look forward to celebrating our Ph.D.'s with a bottle of ice-wine! Thank you to my family, Sharon, Brian, Colin and Denise, for your support over the years and for giving me a place to belong. To my Aunt Berna Schwartz, thank-you for watching over me I hope I have made you proud. M y doctoral training was not possible without the financial support of the Natural Sciences and Engineering Research Council of Canada (NSERC), the University of British Columbia and my father, Mr. Alphonse Walisser. Finally, to all the people I have come to know over the course of my studies - it is the experiences that I have shared, both good and bad, that have had shaped the person I am today and left me with indelible memories of my time here.  xvii  DEDICATION  To my mother and my friend,  Jessie Alexander Walisser  Thank you for being the one I could always rely on.  Ring the bells that still can ring. Forget your perfect offering. There is a crack in everything. That's how the light gets in.  ...Leonard Cohen  xix  1.  PROLOGUE  Poly(ADP-ribose) Polymerase (PARP) is a nuclear zinc-finger metalloprotein, which is increasingly of interest across a broad range of areas in toxicology and pathology. When activated, P A R P catalyzes the post-translational modification of a variety of nuclear proteins (Benjamin and Gill, 1980; Hayaishi and Ueda, 1982; Althaus and Richter, 1987). Although the biological function of P A R P remains unclear, the post-translational modification by P A R P has been implicated in an array of cellular processes that include modulation of histone and chromatin architecture (de Murcia et al, 1988; Realini and Althaus, 1992), genome surveillance and integrity (Satoh and Lindahl, 1992) and D N A repair (Molinete et al, 1993). P A R P activity is also implicated in the mechanism of cell death following D N A damage from reactive oxygen species (Schraufstatter et al, 1986; Thies and Autor, 1991; Kirkland, 1991; Monti et al, 1994; Nosseri et al, 1994). Reactive oxygen species are common initiators of injury shared by many chemicals and pathological processes. In addition, PARP, by virtue of being a zinc-finger protein, may be of importance in the area of metal toxicology. A basic premise of metal toxicology is the concept of metal substitution in metalloproteins, which may disrupt activity and consequent cellular function of the protein. Metal exposure and toxicity is an occupational and environmental health issue, and therefore,a major concern in environmental toxicology. It was the broad toxicological implications of P A R P in oxidant-induced cell death and the potential for alteration of its activity by metals that formed the impetus to investigate the following: 1) the role of P A R P inhibition in modulating oxidant-induced cell death and 2) the potential for metal replacement in P A R P as a mechanism of altering activity and contributing to metal toxicity.  1  2.  2.1.  INTRODUCTION  Oxidative Stress Oxygen is an essential component for living organisms. However, the ubiquity of  molecular oxygen (O2) in aerobic organisms and its ability to readily accept electrons, resulting in reactive forms of oxygen, has made oxygen a common initiator of toxic injury. There are numerous intracellular sources of reactive oxygen species including many cellular enzymes and electron transport processes where one-electron transfers yield free radical intermediates. A l l aerobic cells generate a constitutive flux of reactive oxygen species. In response to this generation, biochemical defense mechanisms have evolved to protect organisms from oxidant damage. However, when the rate of reactive oxygen formation is increased and/or antioxidant defenses of cells are weakened, oxidative damage occurs.  While virtually any cellular  constituents are potential targets, cellular macromolecules, such as unsaturated lipids, proteins and D N A , are of major concern and can lead to critical structural and functional alterations in the cell. 2.1.1.  Chemistry of Oxidative Stress Although molecular oxygen contains an even number of electrons, it has two unpaired  electrons in its molecular orbitals and is said to be in a triplet ground state, meaning that these two electrons have the same spin (Cadenas, 1989). Therefore, the kinetic restrictions that are imposed by these two unpaired spin-parallel electrons limit the oxidative potential of molecular oxygen (Imlay and Linn, 1988). Oxygen is normally completely reduced by the concerted addition of four electrons. Due to energy restrictions, catalysts are required for this reduction. However, under certain circumstances, primarily in the presence of transition metals (e.g., Fe),  2  the catalytic action of the reduced metal can facilitate the univalent and sequential reduction of oxygen. The biochemistry of oxygen toxicity and formation of reactive oxygen species have been reviewed extensively (Imlay and Linn, 1988; Cadenas, 1989; Basaga, 1990; Farber et al, 1990). Molecular oxygen can undergo a one electron reduction to form the superoxide anion, 0~". The 2  superoxide anion can undergo a spontaneous dismutation reaction under aqueou neutral pH conditions to yield hydrogen peroxide (H2O2) (Reaction 1). This reaction is also catalyzed by the enzyme superoxide dismutase. Reaction 1  0 '* + 0 '' + 2H -> H 0 +  2  2  2  2  + 0  2  H 0 can undergo a further one electron reduction to form the hydroxyl radical, HO*, a reaction 2  2  known as a Fenton reaction (Reaction 2). This reaction requires a transition metal, such as iron, as a catalyst to add one electron to H 0 . The superoxide anion, 0"", is one source of reducing 2  2  2  power to provide the electron to form the reduced transition metal in the Fenton reaction (Reaction 3). Reaction 2  Fe  Reaction 3  0 *+ Fe  2 +  + H 0 2  3 +  2  2  ->Fe  -> F e  2 +  3+  +  + HO" + HO' 0  2  The net balance of Reaction 2 and Reaction 3 is referred to as an iron-catalyzed Haber-Weiss reaction (Reaction 4). Fe  Reaction 4  0 "' + H 0 2  2  2  -> 0 + HO' + HO" 2  Superoxide anion and hydrogen peroxide alone are, relatively, not very reactive.  However,  when these reactive oxygen species are in the presence of a transition metal, the hydroxyl radical, HO", can be formed, as depicted in Reaction 4. The hydroxyl radical is a powerful oxidizing agent that reacts at its site of formation by removing an electron from a variety of  3  macromolecules, such as D N A and lipids, thus causing structural damage and biological dysfunction (Pryor, 1986). 2.1.2.  Biological Sources of Reactive Oxygen Species There are several intracellular and extracellular ways in which an organism can be exposed  to reactive oxygen. Subcellular organelles, structural components and cytoplasmic contents all contribute to generation of a variety of reactive oxygen species. Soluble cytoplasmic enzymes, such as xanthine oxidase and aldehyde oxidase, generate 0{' during their catalytic cycling. The rates of reactive oxygen production by these enzymes can be influenced by certain metabolic states, such as ischemia, as in the case of xanthine oxidase (Hille and Nishino, 1995). Electron transport systems in the cell are capable of generating both 0 ~' and H2O2. The percentage of 2  oxygen reduced in this way is small but measurable. In the case of mitochondrial electron transport, the ubiquinone-cytochrome b region is the major site of O2" production (Freeman and Crapo, 1982). In the endoplasmic reticulum and nuclear membrane, cytochromes P450 and bs are sites of uncoupling of electron transport and generation of O2"" and H2O2. Peroxisomes are potent sources of cellular H2O2 due to the high concentrations of oxidases, such as amino acid oxidase and urate oxidase (Freeman and Crapo, 1982). Reactive oxygen species generated extracellularly must cross the plasma membrane and, in the process, may cause damage to the membrane components, such as unsaturated fatty acids and transmembrane proteins. Reactive oxygen damage of membrane components can lead to lipid peroxidation, amino acid oxidation, protein-protein cross-linking, as well as protein scission. Hydrogen peroxide can diffuse freely across membranes, while superoxide is thought to enter cells via transmembrane anion channels (Freeman and Crapo, 1982). Therefore, in contrast to  4  HO', the less reactive oxygen species, H2O2 and O2", may be capable of reacting distally from their sites of generation. In addition to these intracellular sources of reactive oxygen, there are a variety of extracellular sources. The generation of partially reduced oxygen is an important component of the bactericidal action of neutrophils. The cell and tissue injury to the host tissue associated with acute and chronic inflammation is attributable, in part, to the generation of superoxide anions by N A D P H oxidase of activated phagocytic cells, and the subsequent dismutation of O2"" to hydrogen peroxide (Weiss and LoBuglio, 1982). Quinones and quinone-containing compounds, such as the antineoplastic agent daunorubicin and doxorubicin, generate reactive oxygen species (Kalyanaraman et al, 1980). These agents have been designed to cause D N A damage to rapidly proliferating malignant cells. In addition, a wide variety of environmental toxic xenobiotics, including photochemical air pollutants, pesticides, tobacco smoke, solvents and aromatic hydrocarbons, can also cause free radical damage to cells (Mason and Chignell, 1981). Therefore, the diverse sources of oxidative stress suggest that reactive oxygen species are important common mediators of cell injury and cell death. 2.1.3.  Antioxidant Defense Mechanisms The biochemical defense mechanisms that have evolved to protect organisms from free  radical-induced damage include both low molecular weight free radical scavengers and complex enzyme systems (Freeman and Crapo, 1982; Cadenas, 1989). The intracellular small molecular weight free radical scavengers include ascorbate, tocopherol, P-carotene and glutathione. Oxidation of these compounds by reactive oxygen species results in the formation of less toxic compounds. However, any molecule that reacts with a free radical can be termed a "scavenger",  5  and therefore,cellular components such as sugars, unsaturated amino acids and unsaturated fatty acids can also scavenge free radicals. A family of enzymes, including glutathione peroxidase, catalase and superoxide dismutase, has evolved to catalytically detoxify and remove superoxide, hydrogen peroxide and lipid peroxides. Superoxide dismutases catalytically dismute superoxide anion to hydrogen peroxide at a rate approximately 10 times faster than the spontaneous dismutation reaction (Reaction 1) 9  (McCord and Fridovich, 1978). Superoxide dismutase is thought to decrease oxidative damage by decreasing the steady state level of O2"* and thus limiting the iron-catalyzed Haber-Weiss reaction. The hydrogen peroxide produced by the dismutation of superoxide anions can be a precursor of more potent radical species, namely the hydroxyl free radical, an extremely powerful oxidizing agent. Catalase is present in high concentrations in the peroxisomes and is responsible for catalyzing the conversion of H2O2 to water. In addition, the glutathione peroxidase/reductase enzyme system in the presence of N A D P H and G S H can reduce H2O2 and lipid peroxides to either water or the conjugate alcohol. These antioxidant enzyme systems are thought to protect the cell against reactive oxygen damage.  However, in the event that an  oxidative stress occurs in excess of the cell's antioxidant defense mechanisms, cellular injury and death can occur. 2.1.4.  Oxidative Stress and the Endothelium The endothelium is recognized as being more than just a barrier between the blood and the  extracellular space. The endothelium is, in a sense, a separate organ that supplies and communicates with every tissue and organ of the body. Endothelial cell injury and cell death are implicated in the pathogenesis of a variety of disease states and pathological conditions,  6  including various inflammatory diseases such as atherosclerosis, lupus erythematosis, and reperfusion injury following ischemia (Halliwell and Gutteridge, 1984; Repine et al, 1987; Halliwell, 1987; Werns and Lucchesi, 1988). Activated inflammatory cells are a major source of oxidative stress in inflammatory diseases and during secondary inflammation after an initial toxic insult. The endothelial cell, by virtue of its location, is especially susceptible to oxidantinduced injury from inflammatory cells. In the endothelial cell, lethal cell injury following oxidative damage is independent of lipid peroxidation and a critical target appears to be the D N A resulting in D N A strand scission (Thies and Autor, 1991; Kirkland, 1991). D N A strand breaks, resulting from the direct action of genotoxins, such as the hydroxyl free radical, activate the nuclear enzyme, poly(ADP-ribose) polymerase (PARP) (Benjamin and Gill, 1980) and contribute to oxidant-induced cell death.  2.2.  Metal  Toxicity  The ubiquitous nature of metals makes them a unique environmental and occupational hazard. Although metal exposure occurs naturally in the environment, it is the collecting, disposing, purifying, concentrating and using of metals that has increased the risk to humans of metal exposure. These activities pose a significant toxicological threat to all organisms. Although acute toxicities occur, often it is low-dose chronic exposure that results in an insidious onset of adverse biological consequences (i.e., toxicities such as abnormal development or carcinogenesis). Although metals may be detected in an organism during chronic low-dose exposure, it is difficult to correlate such tissue levels with the onset of these latent toxicities. This is partially due to a lack of understanding of the mechanism(s) of metal toxicity, and of knowing which biological markers are important and sensitive indicators of metal exposure and/or toxicity.  7  Metal toxicity is multifaceted. The genotoxic potential of some metals, such as nickel, arsenic and chromium, have been demonstrated in in vitro models, experimental animals and epidemiological studies. The inherent genotoxicity of metals may account, in part, for many of the gross toxicities observed such as carcinogenicity, embryotoxicity and teratogenicity. It is unlikely that a single mechanism of metal toxicity exists because metals are a diverse group and the potential for macromolecular interactions in the cell may be extensive. The chromosomal D N A or regulation of D N A function, however, stands out as a potential target for metals and could account for many of the toxicological effects. Genotoxic metals may react directly with D N A causing D N A damage (Swierenga et al, 1987; Cohen et al, 1990; Lin et al, 1992; Bridgewater et al, 1994). Redox active metal ions are able to serve as electron donors in Fenton-like reactions that produce the hydroxyl free radical, which damages D N A (Nackerdien et al, 1991; Ozawa et al, 1993). Indirect D N A damage, in the form of inhibition of D N A repair processes, can decrease the fidelity of D N A synthesis and alter gene expression. In general, metal interference in these metabolic processes involves some interaction between metals and the enzymes involved in D N A metabolism. This effect is not surprising because metals have a high affinity for proteins, especially residues containing sulfhydryl (cysteine) or nitrogen-containing groups (lysine and histidine). A basic premise of metal toxicity is the concept of metal-amino acid binding at a functionally critical site of an enzyme or displacement and substitution of essential metal ions in metalloproteins, which may disrupt enzyme activity and consequent cellular function.  Zinc-finger metalloproteins are one potential cellular target of metal replacement. Cadmium and mercury are metals of toxicological importance, and are also prime candidates for metal replacement in zinc-finger proteins because they are in the same grouping of the periodic table as zinc. The zinc-finger motif is widespread among eukaryotic DNA-binding proteins. It  8  has been estimated that between 300-700 human genes encode zinc-finger proteins (Klug and Schwabe,  1995).  PARP  is  an  abundant  zinc-finger  protein  (approximately  lxlO  6  copies/mammalian cell nucleus) that is expressed constitutively and therefore,represents a prime candidate to investigate metal replacement. In addition, the role of P A R P in regulating various processes involved in D N A metabolism provides a potential link between metal exposure and the toxicological effects of metals.  2.3.  Cell Death Cell death is the culmination of events beginning from an initial signaling episode  involving either an external environmental cue, such as a receptor-mediated signal or toxic injury, or a genetically programmed initiation that progresses through a complex biochemical pathway resulting ultimately in the demise and elimination of a cell. Under the umbrella of cell death are different processes that account for the way in which cells die. The two best known processes by which cells die have been referred to as cell death by suicide (apoptosis) and cell death by murder (accidental cell death) (Majno and Joris, 1995). Cell suicide is a purposeful cell death that proceeds under the direction of the cell. Alternatively, in response to a toxic stimulus, the cell may maintain homeostasis and repair toxicant-induced damage. However, in the case of extensive injury, where homeostatic and protective mechanisms are overwhelmed, an accidental type of cell death occurs. Two terms, necrosis and apoptosis, have been and continue to be used to differentiate the processes of cell death. However, the use of these terms has been vague and inconsistent. Apoptosis has been used to refer to individual cell suicide, while necrosis has become synonymous with accidental cell death occurring in large tracts of tissue. The concept of apoptosis, in contrast to necrosis, has been widely accepted, and great efforts have been extended  9  to describe the purported differences between these processes. Although the designation of apoptosis and necrosis are both based on morphological descriptors, many non-morphological markers have been used to distinguish between the two processes, especially in cells in culture. The use of experimental models employing cells in culture has, however, further complicated the field because cell death in vitro may not exactly represent the in vivo situation. The strict delineation between apoptosis and necrosis has caused much confusion among the scientific community. This is due, in part, to pathological findings of large clusters of cells undergoing apoptosis in animals exposed to toxicants (Weedon et al., 1979; Ray et al, 1996; Takeda et al, 1999) and the recognition of "single-cell necrosis" (Levin, 1998). Both ideas contradict previously held views. To clarify the terminology surrounding cell death, the Cell Death Nomenclature Committee of the Society of Toxicologic Pathologists put forward a set of terms to describe the processes involved (Trump et al, 1997). These terms are based on an original publication by Majno and Joris (1995). The Society of Toxicology subsequently adopted these guidelines (Levin, 1998). The term necrosis, derived from the Greek word neckrosis meaning deadness, is defined as the sum of the morphological changes occurring in cells after they have died, irrespective of the prelethal process (Levin, 1998). For example, unlike single apoptotic cells in situ, widespread and diffuse apoptosis, as in a toxin-induced apoptosis, does not necessarily resolve by engulfment of these apoptotic cells by neighboring cells and resident macrophages. Similarly, apoptotic cells in culture do not necessarily have healthy neighboring cells and macrophages to engulf them. As a result, apoptotic cells in culture may ultimately lose plasma membrane integrity. Therefore, cells of a tissue that are panapoptotic may eventually exhibit the classical morphological characteristics of necrosis (Weedon et al, 1979; Takeda et al, 1999). Typically, pathologists use the term necrosis to designate the presence of dead tissue (Trump et al, 1997). 10  Necrosis is then further combined with the descriptors, "coagulative" or "liquifactive", to more clearly describe the appearance of this dead tissue. The use of necrosis to describe the process leading to cell death is not adequate, especially when dealing with experimental cell culture models. Therefore, the recommendation has been made to use necrosis to recognize a cell after it has died (Trump et al, 1997). Majno and Joris (1995) propose that the term oncosis, with the root form onkos meaning swelling, be used to describe accidental cell death, and that oncosis, rather than necrosis, be used to contrast with apoptosis. Necrosis, then, is reserved as a general term describing the postmortem morphological changes that result after a cell has died from either apoptosis or oncosis (Majno and Joris, 1995; Trump et al, 1997; Levin, 1998). This would, therefore, not limit the use of necrosis to the in vivo situation but allow its use, where appropriate, in in vitro cell culture models. In accordance with these recommendations, this thesis will use the term necrosis to describe cells that are "dead" as defined by having lost plasma membrane integrity to large molecular weight compounds/components. Oncosis and apoptosis are used to describe the different processes of cell death that lead to necrosis. The entire process can, therefore, be more specifically called either oncotic or apoptotic necrosis. It is recognized that categorizing the process of cell death into two discrete processes, while providing some clarity, is perhaps somewhat restrictive, and that oncosis and apoptosis are only two forms of cell death among others that are yet to be described.  2.4.  The Processes of Cell Death  2.4.1.  Oncosis The morphology of oncosis is characterized by cell and organelle swelling, formation of  cytoplasmic blebs, karyolysis, and eventual loss of membrane integrity (Majno and Joris, 1995;  11  Trump et al, 1997). Typically, broad tracts of cells are affected by oncosis, which occurs in response to a toxic injury. The mechanism of oncosis is based on failure of the ionic pumps of the plasma membrane, resulting in increased plasma membrane permeability and cell lysis (Majno and Joris, 1995). As a consequence of cell lysis, the release of intracellular contents into the surrounding milieu usually results in a rapid and extensive recruitment of inflammatory cells, which may result in further tissue damage from reactive oxygen and enzymes released by these cells. 2.4.2.  Apoptosis "The term apoptosis is proposed for a hitherto little recognized mechanism of controlled  cell deletion, which appears to play a complementary regulation of animal cell populations"  but opposite role to mitosis in the  (Kerr et al, 1972). Apoptosis is derived from the Greek  terms apo, meaning from and ptosis, meaning fall, and therefore,describes a "falling o f f , similar to that of leaves from trees. The basis for the term apoptosis was the morphological observation of cell budding and the formation of smaller independent cell bodies (apoptotic bodies) that fall away from the original cell. The terms apoptosis and programmed cell death have been used interchangeably to describe a process of cell death that is under genetic control. However, these two terms describe distinct events. Programmed cell death refers to situations where cells are destined or programmed to die at a fixed time, as occurs during development. Majno and Joris (1995) state that "the genetic program of programmed cell death is a clock specifying the time for suicide, whereas the genetic program of apoptosis specifies the weapons (the means) to produce instant suicide".  Therefore, cells that are scheduled to die (i.e., programmed cell death) may carry out  that process by apoptosis.  12  Apoptosis is characterized by an orchestrated sequence of morphological and biochemical events that occurs in two phases. The initiation phase encompasses the pro-apoptotic signals that trigger apoptosis and include initiation through intracellular regulatory mechanisms, or by extracellular agents such as hormones, cytokines, killer cells and a variety of chemical, physical and viral agents. This is followed by an execution phase where the events leading to cell death are carried out. Apoptosis is considered an active process requiring energy for gene transcription and translation. These processes are thought to supply the biochemical machinery to carry out the apoptotic program. The morphological characteristics of apoptosis are cell shrinkage, chromatin condensation and margination, membrane budding, and formation of apoptotic bodies (Kroemer et al, 1995; Allen et al, 1997). In apoptosis, the cellular organelles (i.e., mitochondria, golgi) remain intact. The apoptotic bodies are extruded from the cell and may contain intact cellular organelles, as well as nuclear fragments. In situ, the apoptotic bodies are quickly engulfed by macrophages or neighboring cells. In vitro, the apoptotic bodies remain and proceed to lose membrane integrity, a process described as secondary or apoptotic necrosis. It is generally thought that apoptosis in vivo is not associated with inflammation, because apoptotic bodies are phagocytized prior to cell rupture and release of cellular contents. The loss of plasma membrane asymmetry with the translocation of phosphatidylserine from the inner surface to the outer surface of the plasma membrane is thought to play a role in identifying these cells for phagocytosis (Fadok et al, 1992; Savill et al, 1993). It would seem likely that when a death program is initiated following a toxic insult, the preferential process of death would be apoptosis as opposed to oncosis because there would be less damage as a consequence of secondary inflammation.  13  2.4.3.  The Role of Caspases in the Execution of Apoptosis The biochemistry of the intracellular signaling events and the operation of cellular  machinery leading to the execution of apoptosis is complex. A fundamental understanding of the apoptotic machinery arose from the study of the nematode, Caenorhabditis  elegans,  which  undergoes a loss of 131 cells from the total 1090 cells generated by apoptosis during development (Ellis et al, 1991). The study of both pro- (ced-3, ced-4) and anti-apoptotic (ced9)  genes and gene products in C. elegans has lead to the discovery of homologues in mammalian  cells, and suggests that the cellular machinery directing cell death has been evolutionarily conserved. The pro-apoptotic gene, ced-3, encodes a cysteine protease, CED-3, that was shown to be related to mammalian interleukin-l[3-converting enzyme (ICE) (Yuan et al, 1993). The importance of these proteases was shown by demonstrating that apoptosis could be induced in Rat-1 cells by overexpression of the murine ICE or CED-3 (Miura et al, 1993). Furthermore, the introduction of point mutations into a homologous region of either ICE or CED-3 disabled the ability of these proteins to cause cell death (Miura et al, 1993). Since the recognition of the relationship between CED-3 and ICE in cell death, a family of CED-3/ICE-like proteases has been identified and is referred to as the caspase family. The 'c' denotes a cysteine protease and 'aspase' refers to the ability of members of this family to cleave after an aspartic acid residue. There are now at least 13 members identified as caspase-1 to caspase-13 (Thornberry and Lazebnik, 1998). The caspase enzymes are thought to be the major participants in the initiation of cell disassembly in response to pro-apoptotic signals, as well as in the execution of this disassembly process. Caspases are expressed constitutively as inactive proenzymes. Following a pro-apoptotic signal, caspases are activated by cleavage at specific Asp sites to form the active protease. Different initiator caspases mediate distinct pro-apoptotic signals. Caspase-8 is associated with apoptosis involving death receptors, while caspase-9 is  14  involved in death induced by cytotoxins (Ashkenazi and Dixit, 1998). Caspase-8 and caspase-10 are able to cleave and activate all other caspase family members, and therefore,it has been suggested that these proteases are near the apex of an apoptotic cascade that functions to activate downstream members of caspases (Cohen, 1997). The cascade model for activation of effector caspases explains how distinct apoptotic signals induce the same cellular changes associated with apoptosis. Caspase enzymes are responsible for the proteolytic cleavage of a variety of intracellular substrates, and this activity is thought to ultimately result in the morphological and biochemical changes that are observed during apoptosis. The strict substrate specificity of caspase enzymes (i.e., an absolute requirement for Asp in the Pi position of the active site) suggests that the protein cleavage associated with apoptosis is not simply a random protein degradation event, but one that occurs in a coordinated manner to a select set of proteins. The caspase-mediated cleavage of proteins usually results in the loss of function of that protein, and appears to contribute to direct disassembly of cellular structures. For example, the cytoskeletal protein, fodrin, is cleaved by caspase-3 and this cleavage appears to be related to membrane budding and formation of apoptotic bodies (Cohen, 1997). Caspase-6 is involved in the disassembly of the nuclear lamina, a rigid structure underlying the nuclear membrane involved in chromatin organization, and this cleavage contributes to the chromatin condensation characteristic of apoptosis (Thornberry and Lazebnik, 1998). In contrast to the general finding of substrate inactivation by caspases, protein kinase C delta is proteolytically activated by caspase-3 at the onset of apoptosis induced by D N A damaging agents and tumor necrosis factor (Emoto et al, 1995). Protein kinase C delta activity contributes to the phenotypic changes associated with apoptosis (Ghayur et al, 1996).  15  Caspase-3, also referred to as CPP32, apopain or Yama, after the Hindu god of death (Fernandes-Alnemri et al, 1994), is a key member of the caspase family (Nicholson et al, 1995; Tewari et al, 1995). Caspase-3 is an effector caspase, and its activation has been reported to be a critical component of the apoptotic pathway, as inhibition of caspase-3 prevents downstream events associated with apoptosis in an in vitro system (Nicholson et al, 1995) and in whole cells (Schlegel et al, 1996). Measurement of caspase-3 activity has become a hallmark of apoptosis. One well-characterized proteolytic substrate for caspase-3 is the nuclear enzyme, poly(ADP-ribose) polymerase, the core subject of this thesis. 2.4.4.  Determination of the Mode of Cell Death in vitro In our laboratory, cell death has been defined as the loss of plasma membrane integrity.  The inability to exclude plasma membrane impermeant dyes, such as trypan blue, is a common measure of membrane integrity and cell viability. However, loss of plasma membrane integrity is not a specific measure of the process of cell death because cell lysis can occur by either oncotic or apoptotic necrosis. This is particularly true in vitro, where the phagocytic mediators are unavailable for "clean-up" of apoptotic bodies prior to apoptotic necrosis. Therefore, in order to distinguish the process of cell death, specific indicators of either apoptosis or oncosis are required. The morphological features of apoptosis, from cell shrinkage and chromatin condensation to apoptotic body formation, occur in concert with the underlying biochemistry of apoptosis, and it is from these events that specific indicators of apoptosis can be drawn. One of the earliest recognized events in apoptosis  was D N A fragmentation  at  internucleosomal sites to yield multiples of 180 bp, which form " D N A ladders" when subjected to electrophoresis on agarose gels (Walker and Sikorska, 1997). The presence of a D N A ladder has become a qualitative hallmark of apoptosis. Terminal deoxynucleotidyl transferase nick end-  16  labeling (TUNEL) was developed to detect apoptosis in situ in tissue sections. This method relies on detection of 3'-OH ends of D N A stand breaks generated during internucleosomal D N A fragmentation, and is more sensitive than agarose gel electrophoresis for fragmentation analysis (Allen et al, 1997). There are several caveats inherent in the assessment of internucleosomal D N A fragmentation as a marker of apoptosis. The first is that, although D N A cleavage events during oncosis are random and appear as a diffuse smear on D N A electrophoresis, 3'-OH ends of D N A may be generated and thus positively labeled by the T U N E L assay. This is of particular concern in studies where the initiating damage is an oxidant that generates random D N A strand scission. As a consequence, using the T U N E L assay as the sole marker of apoptosis may lead to erroneous conclusions. In addition, it is now recognized that during apoptosis D N A is initially fragmented into large 50 - 300 kbp fragments, which is followed in some models by internucleosomal D N A fragmentation (Cohen et al, 1992; Oberhammer et al, 1993; Walker et al, 1994; Allen et al, 1997). Cleavage of D N A into the larger fragments is sufficient to show chromatin condensation and subsequent sub-nuclear apoptotic character in the absence of oligonucleosome formation. Therefore, apoptosis can occur in the absence of D N A ladder formation. Once again, a negative result in either D N A ladder formation or the T U N E L assay may incorrectly suggest the lack of apoptosis in the model system tested.  Alternatively, a  positive T U N E L assay result in the case of oxidant injury may incorrectly suggest apoptosis. Phosphatidylserine is a normal phospholipid constituent located on the inner surface of the plasma membrane (Allen et al, 1997; Reutelingsperger and van Heerde, 1997). Externalization of phosphatidylserine from the inner surface to the outer surface is an early event in apoptosis, and is associated with the phagocytic removal of apoptotic bodies prior to lysis. The loss of plasma membrane asymmetry associated with phosphatidylserine translocation can be measured  17  as an indicator of apoptosis using annexin V (Reutelingsperger and van Heerde, 1997). Annexin V exhibits a Ca -dependent binding to negatively charged phospholipids, and has a high affinity 2+  for phosphatidylserine. In vivo, annexin V will bind to phosphatidylserine on an apoptotic cell and inhibit the procoagulant and pro-phagocytic activity associated with phosphatidylserine. Annexin V binding is usually assessed in combination with a marker of plasma membrane integrity, such as exclusion of propidium iodide, to rule out loss of membrane integrity concomitant with positive phosphatidylserine staining (Allen et al, 1997). Caspase-3 activity is currently the preeminent indicator of apoptosis, because its activity is thought to be required for apoptosis to occur. Thus, measuring increased caspase-3 activity in cytoplasmic extracts would-be highly confirmatory of an apoptotic mechanism. Caspase-3 activity can be measured in one of two ways. The first method is to examine the proteolytic cleavage of caspase-3 substrates and the most commonly measured substrate is P A R P (Duriez and Shah, 1997). P A R P (-113 kDa) is cleaved into a 24 kDa fragment and an 89 kDa fragment and either fragment, as well as the intact enzyme, may be identified by immunoblot analysis using the appropriate antibody. Most commonly, the intact enzyme and the 89 kDa fragment are analyzed in lysed cells to assess caspase-3 activity. Caspase-3 activity can also be measured directly in cell extracts using a tetrapeptide substrate, Asp-Glu-Val-Asp (DEVD), derived from the P A R P cleavage site, that is preferentially cleaved by caspase-3. The fluorescent tetrapeptide substrate conjugate, DEVD-7-amino-4-trifluoromethyl coumarin (DEVD-AFC), emits a blue light (^rnax = 400 niti), and upon cleavage by caspase-3, the free A F C undergoes a shift in fluorescence and emits a yellow-green light (^  max  = 505 nm) (Clontech Laboratories, 1997).  Caspase-3 is not the only caspase enzyme that cleaves PARP. Caspase-7 also participates in  18  P A R P cleavage, and therefore,the P A R P cleavage activity is usually referred to as caspase-3-like activity. The morphological appearance of a cell undergoing apoptosis is both distinct and specific, and therefore,all experiments assessing this process of death should include a microscopic analysis of cellular morphology (Allen et al, 1997). The ultrastructural changes occurring in an apoptotic cell, such as membrane budding and formation of apoptotic bodies, can best be assessed by scanning electron microscopy (Allen et al, 1997). Using the plasma membrane permeant nucleic acid stain, acridine orange, the characteristic apoptotic nuclear morphology can be visualized and distinguished from oncotic nuclear morphology. In apoptosis, the chromatin becomes pyknotic forming smooth globular masses, whereas karyolysis and non-specific nuclear degradation accompany oncosis (Majno and Joris, 1995).  When acridine orange is used in  combination with a plasma membrane impermeant nucleic acid stain, such as propidium iodide, not only is the nuclear morphology visualized, but the loss of membrane integrity associated with either oncotic or apoptotic necrosis can also be detected. Apoptosis is a form of cell death initially based only on the morphology of the cellular changes associated with it. Therefore, assessment of these features by microscopy should always accompany any biochemical method, regardless of its specificity (Allen et al, 1997). 2.4.5.  The Fate of a Cell: Apoptosis versus Oncosis An internal signal, involving cell cycle regulation factors, transcriptional regulation or  developmental cues, may initiate apoptosis in single cells within a population. However, an external stimulus, such as a toxic insult, may initiate more than one message in any given population of cells. Watson et al. (1995) report that hydrogen peroxide induces two distinct death pathways in an intestinal epithelial cell line: one that is immediate representing necrosis  19  (oncosis) and another that is delayed leading to apoptosis. The level or extent of initiating insult has also been reported to determine the pathway to cell death. For instance, higher concentrations of hydrogen peroxide or a longer incubation at higher temperature change the process of U937 human myeloid leukemic cell death from apoptosis to passive necrosis (oncosis) (Nosseri et al, 1994). Mild or intense insults with generators of nitric oxide or superoxide produce two distinct events, apoptosis or necrosis (oncosis), respectively, in cortical cell cultures (Bonfoco et al, 1995). Therefore, while it is convenient to describe cell death as a process occurring by either apoptosis or oncosis, it is likely that in any population of cells both processes are occurring simultaneously within the population following a toxic insult.  2.5.  Poly(ADP-ribose)  Polymerase  As discussed above, biochemical properties and cellular functions have implicated P A R P activity in both cell death and metal toxicity. P A R P is a metalloprotein that catalyzes the covalent attachment and elongation of ADP-ribose units onto a limited number of acceptor proteins (Benjamin and Gill, 1980; Ffayaishi and Ueda, 1982; Althaus and Richter, 1987). The substrate for this reaction is N A D , which, in contrast to its normal catalytic function as a +  cofactor in oxidation-reduction reactions, is catabolized in the ADP-ribosylation event releasing nicotinamide and liberating ADP-ribose monomers used for ADP-ribosylation of protein acceptors (Fig. 1). The protein acceptors are generally involved in maintenance of chromatin architecture and D N A metabolism, and therefore,poly-ADP-ribosylation is a post-translational modification of nuclear DNA-binding proteins induced by D N A strand scissions, as occurs with oxidant-induced D N A damage.  20  Poly(ADP-Ribose) Polymerase  W  N A D  '  Elongation  NH  \ JjHb  00 2  NHa  .Rib  Initiation Rib'  Rib 2"  Branching .Rib  ADP-Ribose  Rib  Rib  Wb  Fig. 1. Scheme showing the process of ADP-ribosylation of an acceptor protein. PARP catabolizes the substrate, N A D , releasing nicotinamide, and using the ADP-ribose monomers to ADP-ribosylate a protein acceptor. +  2.5.1.  PARP Functional Domains PARP is composed of 1,014 amino acid residues with a calculated molecular weight, based  on c D N A sequence, of 113,153 daltons (Da) (Kurosaki et al., 1987). The P A R P enzyme molecule can be proteolytically divided into three functional domains: a 46 kDa D N A binding domain in the N-terminal region, a central 22 kDa fragment containing the automodification domain and the 54 kDa fragment in the C-terminal region containing the N A D binding domain, +  as shown in Fig. 2 (Kameshita et ai, 1984).  21  DNA Binding Domain 46 kDa <  Automodification Domain 22 kDa  NAD Binding Domain 54 kDa +  •«<  Fl  *  Fll  Fig. 2. Scheme showing the functional domains of PARP. The three functional domains of PARP are: 1) the N-terminal D N A binding domain containing the zinc fingers (Fl and FII) and the nuclear location signal (NLS), which directs PARP to the nucleus, 2) the central automodification domain is where PARP ADP-ribosylates itself, and 3) the C-terminal catalytic domain, showing Lys-893, a residue thought to be critical for catalytic activity. Adapted from de Murcia etal. (1991).  The D N A binding domain is rich in Lys residues and has a basic character with a net charge of +15. Metal chelation experiments have demonstrated that both DNA-binding capacity (Mazen et al, 1989; Menissier-de Murcia et al, 1989) and PARP activity (Zahradka and Ebisuzaki, 1984) are dependent on the presence of zinc. Metal analysis of purified calf thymus P A R P has determined that 2 zinc ions are associated with each enzyme molecule and the zinc binding sites are localized within the D N A binding domain (Mazen et al, 1989). Based on analysis of the human PARP sequence deduced from cDNA, the existence of 2 putative zinc fingers in the DNA-binding region of PARP with the form: Cys - X  a a  2  - Cys - X  a a  28,30  - His - Xaa  2 - Cys was proposed (Mazen et al, 1989), as seen in Fig. 3. A n identity profile obtained after alignment of the human, murine, bovine and chicken PARP amino acid sequences revealed  22  greater than 70% conservation in the amino acids encoding the zinc-finger regions, suggesting that these regions are of functional importance (de Murcia et al, 1991).  ,MVQ A 1  M  R L S D K P I  C S  P  M 28aa  G K  K  w  _  K  Zn  ^  Q L G M  l  I  D R  W  °<L G  s  Zn  K  V 56  Finger 1  P E K  E  2,  F  v  D  P 30 aa  E  P H  V  K  v  % .if N-Terminal  V Q G  D  K  M  R  F  E  24V  K K S L  P g  125V 68 aa  Vl62  y  C-Terminal  v  Finger 2  Fig. 3. Amino acid sequence of the zinc fingers of PARP. The zinc-coordinating Cys and His residues are circled. Adapted from Gradwohl et al (1990).  The concept of repetitive zinc binding domains was first introduced by (Miller et al, 1985), who were investigating the Xenopus laevis transcription factor IIIA (TFIIIA). TFIIIA molecules contain 9 tandemly-repeated 30 amino acid units containing two invariant pairs of histidine and cysteine residues that tetrahedrally coordinate a Z n  2 +  ion. The classical zinc-finger  motif consists of a zinc atom tetrahedrally coordinated to a combination of cysteine and histidine residues with a loop of amino acids organized around the zinc atom at the base of the finger. Each independent zinc-binding domain of TFIIIA is capable of interacting with nucleic acids in a sequence-specific manner. In contrast, although the zinc fingers of P A R P are structurally similar, they do not bind to specific D N A sequences, but rather participate in the recognition and binding to D N A strand breaks (Menissier-de Murcia et al, 1989). The ability of D N A to support  23  poly(ADP-ribose) synthesis is dependent upon the number and type of strand breaks it contains (Benjamin and Gill, 1980). The D N A must be double stranded and contain either single or double strand breaks. Flush-ended double strand breaks are 10 times more effective at activating PARP than single strand breaks. Double strand breaks producing unpaired nucleotides extending from either the 3'- or 5'-strands are less effective than flush-ended D N A fragments. Finally, removal of the terminal 5'-phosphate groups markedly increases the ability of all types of D N A fragments to activate PARP (Benjamin and Gill, 1980). The ability to recognize and bind to D N A strand breaks is dependent on structurally-intact zinc coordinated in zinc-finger structures located in the D N A binding domain (Menissier-de Murcia et al, 1989; Ikejima et al, 1990). Binding of the PARP zinc fingers to a D N A strand break results in PARP activation and this dependence on D N A for stimulation of P A R P activity is almost absolute (Ueda et al, 1982). The binding of D N A is presumed to induce a conformational change in the enzyme that stabilizes its active form (Ueda et al, 1982; Ikejima et  al, 1990). However, this concept has not been directly tested. The automodification domain is also Lys rich and basic in nature, carrying a net charge of +7 (de Murcia et al, 1991). The automodification domain contains 15 glutamic acid residues that are considered to be potential sites for auto-poly(ADP-ribosyl)ation. The N A D binding domain, located in the C-terminal portion of the enzyme, is neutral in +  character (Kurosaki et al, 1987). It is the most highly conserved of the P A R P domains with greater than 83% sequence identity between human, murine, bovine and chicken PARP amino acid profiles (de Murcia et al, 1991). The N A D binding domain contains a block of 50 amino +  acids, residues 859 - 908, which are strictly conserved among vertebrates (100% sequence identity), and correspond to the catalytic site (de Murcia et al, 1994). Activity studies have been  24  performed using the bacterially-expressed carboxy-terminal 40 kDa domain of human PARP (Simonin et al, 1993), as well as the 89 kDa apoptotic fragment produced after in situ cleavage by caspase-3 (Simonin et al, 1993). These studies have shown that both the 40 and 89 kDa fragments retain a basal level of ADP-ribose polymer synthesis activity when immobilized on nitrocellulose, and that the extent of activity is similar to that of the unstimulated intact enzyme (Shah et al, 1995). However, the specific activity was approximately 500-fold lower than that of the whole enzyme when activated by D N A strand breaks (Simonin et al, 1993), illustrating, once again, the importance of D N A binding capacity on the ability of PARP to respond to D N A damage. 2.5.2.  Cellular Function of PARP ADP-ribosylation by P A R P is a common cellular event. Although the purpose of this post-  translational modification of nuclear proteins is not clear, two generalizations regarding the consequences of ADP-ribosylation of proteins were made by Althaus and Richter (1987): 1) the net result of ADP-ribosylation of acceptor proteins in vitro has been inhibition of enzyme activity, and 2) all physiologic protein acceptors identified so far act on deoxynucleotide or nucleotide polymers. ADP-ribosylations have, therefore, been suggested as a model of reversible DNA-protein interaction (Zahradka and Ebisuzaki, 1982). In this model, an ADP-ribosylated protein is electrostatically repulsed from D N A . Protein deactivation is then reversed by the action of poly(ADP-ribose) glycohydrolase, an ADP-ribose-catabolic enzyme that removes ADP-ribose polymers and reestablishes protein activity (Desnoyers et al, 1995). The activity of PARP itself is regulated in a similar manner. PARP binds to D N A strand breaks and its activity is stimulated. Once automodified, the affinity of PARP for D N A is decreased. The affinity of  25  PARP for binding D N A and its activity are reestablished through the action of poly(ADP-ribose) glycohydrolase. Nuclear targets of ADP-ribosylation by PARP that have been identified in vivo include the histones, as well as high mobility group proteins and topoisomerase I (Althaus and Richter, 1987). The localized synthesis of ADP-ribose onto histone acceptors at the site of a D N A strand break in vivo could increase the accessibility of D N A repair enzymes to the damaged site, thus facilitating D N A repair. Other potential targets that have been identified in vitro, include D N A polymerase a and p\ D N A ligase (Althaus and Richter, 1987), and Ca /Mg -dependent 2+  2+  endonuclease (Yoshihara et al, 1975). The major acceptor of ADP-ribosylation is, however, the enzyme itself (de Murcia et al, 1983). Although the biological function of PARP remains unclear, post-translational modification by PARP has been implicated in an array of cellular processes based on its ability to A D P ribosylate a variety of protein acceptors. As described above, PARP activity in vivo is thought to be involved in the modulation of histone and chromatin architecture (de Murcia et al, 1988; Realini and Althaus, 1992), genome surveillance and integrity (Satoh and Lindahl, 1992), and D N A repair (Molinete et al, 1993). In addition, PARP activity is implicated in the mechanism of cell death following D N A damage (Schraufstatter et al, 1986; Thies and Autor, 1991; Kirkland, 1991; Monti etal, 1994; Nosseri etal, 1994).  2.5.3.  Poly(ADP-ribose) Polymerase and Oxidant-induced Cell Death PARP activation mediates lethal cell injury ("cell death") following exposure to D N A  damaging agents, such as alkylating agents (Molinete et al, 1993) and reactive oxygen species (Schraufstatter et al, 1986; Thies and Autor, 1991; Kirkland, 1991). The D N A is a sensitive target for reactive oxygen-induced damage causing D N A strand scission, and PARP activation  26  increases in proportion to the extent of D N A damage. As a consequence of extensive PARP activation, there is a precipitous depletion in cellular N A D , followed by a decline in cellular +  ATP, other adenylates (ADP, AMP), as well as the calculated energy balance (ATP +  V2ADP /  ATP+ADP+AMP) of the cell (Thies and Autor, 1991). Previous studies in this laboratory have demonstrated that in oxidant-stressed endothelial cells, each of these metabolic changes precedes the loss of membrane integrity (Thies and Autor, 1991). A similar decline in energy status is observed in etoposide-treated HeLa cells as a consequence of P A R P activation (Bernardi et al, 1995) and in murine macrophages following H2O2 exposure (Schraufstatter et al, 1986). The PARP-dependent metabolic changes and their interrelationships are likely to have farreaching consequences on energy-requiring biosynthetic processes, such as D N A , R N A , protein and lipid synthesis, as well as on maintenance of cellular ion homeostasis (e.g., C a , M g , K , 2+  2 +  +  N a homeostasis). Following exposure to agents that damage D N A , the alteration in the cellular +  metabolic status, as a consequence of P A R P activation, is believed to contribute directly to an oncotic mode of cell death and subsequent loss of membrane integrity. 2.5.3.1.  PARP: A Caspase-3 Substrate in Apoptosis  P A R P is recognized as a substrate for caspase-3 (Nicholson et al, 1995) and caspase-7 (Cohen, 1997), proteases involved in the execution of apoptosis. It has been suggested that one function of caspase activation is to cleave proteins involved in D N A repair thereby incapacitating the cell's ability to repair itself and ensuring death by apoptosis (Casciola-Rosen et al, 1996). One result of caspase-3 and/or caspase-7 activity is the proteolytic cleavage of P A R P into a 24 kDa N-terminal fragment containing the DNA-binding domain, and an 89 kDa fragment containing both the automodification domain and the catalytic domain. This cleavage pattern, as detected by Western blot, is a common biochemical indicator of apoptosis (Duriez  27  and Shah, 1997). As discussed above, the ability to bind D N A through the function of the zinc fingers is essential for PARP activation in response to D N A damage. Therefore, caspase cleavage of PARP should effectively disable PARP activation in the event of D N A damage. The critical function of caspase-3 in the execution of apoptosis raises the question of the importance of PARP activity during this process. Caspase cleavage of P A R P implies that PARP activity is not necessary or desirable in an apoptotic process. This notion makes sense when the consequence of P A R P activity is considered. Following caspase activation, one of the later events in apoptosis is D N A fragmentation into 50 - 300 kbp fragments initially, followed by internucleosomal cleavage of D N A into multiples of 180 bp (Walker and Sikorska, 1997). It could be hypothesized that PARP activity at this stage, when D N A is being systematically degraded, is undesirable in apoptosis because a decline in energy status of the cell as a consequence of PARP activation, may compromise its ability to complete the active process of apoptosis. Similarly, Virag et al. (1998) have suggested that caspase cleavage of PARP acts as a preventative measure aimed at conservation of cellular energetics to ensure completion of the apoptotic program. Therefore, prior to extensive D N A degradation, P A R P activity is precluded by caspase cleavage. The time-course of PARP cleavage and D N A fragmentation has been investigated in a human osteosarcoma cell line that undergoes a slow spontaneous apoptosis (Rosenthal et al, 1997). PARP activity was evident before the commitment to apoptosis; however, shortly thereafter, the PARP cleavage product was observed concomitant with the appearance of the characteristic internucleosomal D N A fragmentation. Therefore, proteolytic cleavage and inactivation of PARP by caspase enzymes occurs prior to D N A fragmentation. However, in the case of oxidant injury, D N A is a prime target for oxidant damage and this damage is an early initiating event that results in the activation of PARP. How does PARP activation at an early stage after a genotoxic stimulus affect the ensuing survival/death signals?  28  2.6.  Research  Hypothesis  It is hypothesized that PARP performs an important function in mediating the pattern by which cell death occurs following PARP  by DNA  strand scissions  exposure to agents that initiate DNA causes a decline  in the cellular  damage. Activation of energy balance  and  therefore,failure of energy-requiring synthetic and homeostatic processes. This would preclude apoptosis from occurring, and would direct the cell into an oncotic form of cell death. If, however, PARP activity is inhibited, either through pharmacological site or through disruption of DNA  intervention at the catalytic  binding, and its effect on cellular energy abrogated, an  alternate pathway of cell death would be allowed to proceed, that is an active energy-dependent form of cell death consistent with apoptosis.  2.6.1.  PARP Activation and the Turning-Point Between Oncosis and Apoptosis The dramatic effect that activated P A R P has on the cellular energy status produces a  paradoxical situation for the cell if it intends to operate an energy-dependent process. Simply stated, a cell can not operate an energy-dependent process, such as apoptosis, if it does not have the "currency" (i.e., ATP) to pay for such a program. Therefore, P A R P activity appears to be a pivotal determinant in the mechanism of oxidant-induced cell death. In response to D N A damage, the activation of P A R P circumvents any underlying default apoptotic signals and directs the cell into oncosis, as illustrated in Fig. 4. However, if P A R P activity is inhibited during oxidant-induced D N A damage, then cellular energy (the "currency") is maintained, and the underlying energy-dependent process of apoptosis is allowed to proceed. In addition, it would be considered more favorable to die by apoptosis than oncosis, because there is less inflammation and the potential for secondary inflammatory injury with apoptosis.  29  O X I D A T I V E  S T R E S S  • DNA, Protein, Lipid damage  M  poptosis initiation events  PARP  acf/Ve'  fNAD 1^ ^Energy Charge  (ONCOSIS  APOPTOSIS PARP INHIBITIO  4^  O X I D A T I V E  S T R E S S  DNA, Protein, Lipid damage  P A R P  a  C  # y g  Fig. 4. Scheme illustrating the proposed role of PARP in the modulation of oxidant-stressed endothelial cell death from oncosis to apoptosis. PARP activation, as a consequence of oxidantinduced DNA strand scission, depletes cellular N A D and energy charge ratio, forcing the cell into oncosis. However, alteration of PARP activity prevents the decline in cellular energy, maintains homeostasis, thus avoiding oncosis, and allows the active, energy-dependent process of apoptosis to occur. +  30  2.7.  Modulation of PARP  Activity  The scheme in Fig. 5 illustrates the possible role of PARP in mediating the process of cell death that may occur as a consequence of D N A damage following oxidant exposure. Following D N A strand scission, PARP activity may participate in D N A repair, resulting in survival of the cell. Under circumstances of extensive D N A damage, the ensuing increase in PARP activity may deplete cellular energy and result in oncosis. However, if PARP activity is modulated, either through pharmacological intervention at the catalytic site, through induction of automodification, or through altering D N A binding, execution of the apoptotic pathway would be allowed to proceed. Therefore, alteration of PARP activity following exposure to agents that damage D N A may provide a mechanism by which the process of cell death can be altered. This premise provided the basis for the research discussed in this thesis.  31  DNA Strand Scission  f  Modulate PARP Activity  P A R P Activity Alter DNA Binding  DNA Repair  Deplete Cellular Energy  I  AutoModification  Inhibit Enzymatic Activity  M e t a l  Replacement  4  P A R P Activity  I  Maintain Cellular Energy  Cell Survival  Oncosis  t Apoptosis  Fig. 5. Scheme illustrating the role of PARP mediating the process of cell death following exposure to a D N A damaging agent. Following D N A strand scission, PARP activation can result in cell survival or oncosis depending on the extent of the initial damage. However, i f P A R P activity is modulated, either through pharmacological inhibition, automodification or altering D N A binding, cellular energy is maintained and an apoptotic program is allowed to proceed.  32  2.7.1.  Inhibition of PARP Catalytic Activity PARP activity is commonly modulated with the use of competitive inhibitors of the  substrate, N A D . Inhibition of PARP activity, at the catalytic site, is concentration-dependent +  with some inhibitors showing higher affinity than others. In particular, 1,5-isoquinolinediol (DIQ), the PARP inhibitor used throughout this research, has an in vitro 50% inhibitory concentration (IC ) of 0.39 u M . Therefore, DIQ is 2 and 3 orders of magnitude more potent 50  than the other common PARP inhibitors, 3-aminobenzamide (3AJ3, IC50 = 33 fiM) and nicotinamide (IC50 = 210 (iM), respectively, (Banasik et al., 1992). P A R P inhibition is effective in vitro, ex vivo and in vivo (Rankin et al., 1989; Banasik et al, 1992; Lam, 1997; Cuzzocrea et  al, 1998). In the model of oxidant-stressed endothelial cells used in this research, it was previously shown that PARP activation caused rapid N A D  +  depletion, followed by a decline in cellular  energy stores (ATP, A D P , A M P ) (Thies and Autor, 1991). Each of these metabolic changes preceded the loss of cell membrane integrity, a process of death typical of oncosis. However, when oxidant-stressed endothelial cells were treated with a PARP inhibitor, 3AB, the metabolic alterations, as well as the loss of membrane integrity, were prevented without affecting the initial oxidative damage (i.e., D N A strand scission) (Thies and Autor, 1991). Therefore, inhibition of PARP appears to protect cells from the consequences of oxidant injury and prevent immediate oncosis. P A R P inhibition concomitant with oxidant stress has also been demonstrated to prevent the decline in N A D and A T P in HT-29-18-C1 intestinal epithelial cells (Watson et al, 1995), +  murine thymocytes (Virag et al, 1998) and U937 cells (Nosseri et al, 1994; Coppola et al, 1995). Therefore, the rationale behind the competitive inhibition of PARP activity is to preserve  33  the metabolic status of the oxidant-stressed cell and allow operation of the apoptotic machinery for self-destruction. 2.7.2.  Modulation of PARP DNA-Binding Capacity Inhibition of the catalytic site is not the only way to modulate P A R P activity. The D N A -  binding domain provides another site where P A R P activity could be manipulated through interference of zinc-finger-mediated DNA-binding capacity. To illustrate the effect of interfering with P A R P DNA-binding capacity, Schreiber et al (1995) established stable HeLa cell lines that constitutively produced the 46 kDa D N A binding domain of human PARP. Constitutive expression of the P A R P D N A binding domain in this model strongly inhibited endogenous P A R P activity following treatment of cells with hydrogen peroxide (Masson et al, 1995). In addition, cell survival and chromosomal stability were greatly decreased following treatment with A^-methyl-Af'-nitro-A^nitrosoguanidine. These experiments indicate that inhibition of P A R P activity, by preventing its ability to bind to D N A strand breaks, sensitizes these cells to alkylating agent- and oxidant-induced D N A damage, likely as a consequence of inhibition of resident P A R P activity. A number of strategies have been developed to alter P A R P activity and dissect the physiological role of P A R P in cellular response to D N A damaging agents. These strategies include the use of competitive inhibitors, generation of cell lines deficient in P A R P activity (Chatterjee et al, 1989), depletion of endogenous P A R P activity by antisense R N A expression (Ding et al, 1992; Ding and Smulson, 1994), as well as the generation of a dominant-negative mutant of P A R P as described above. One mechanism for modulating P A R P activity that has undergone only a cursory investigation is the potential for disrupting D N A binding capacity by replacing the zinc in the zinc-finger motif with an alternate metal. It is clear that structurally  34  intact zinc fingers are critical for DNA-binding capacity and P A R P activation. Therefore, it would seem plausible that substitution of the zinc in the zinc-finger motif with a different metal may alter the D N A binding capacity and consequently affect overall P A R P activity. 2.7.2.1.  Evidence for Metal Replacement in Zinc-finger Motifs  Reports of metal ion substitution in zinc fingers are present for several metalloproteins. The bovine estrogen receptor, a D N A binding protein containing 2 zinc fingers, has been studied in terms of metal replacement of zinc in the finger motif. Competition experiments  in vitro  performed by dialyzing the bovine estrogen DNA-binding apopolypeptide against buffer containing Z n , C d 2+  2 +  or C o , but not C u 2+  2 +  or N i , restored specific D N A binding properties as 2 +  measured by the mobility shift assay (Predki and Sarkar, 1992). However, the Co -reconstituted 2+  protein had a decreased affinity for the estrogen response element compared with the native protein. Thus, not only could the zinc in the zinc-finger be replaced, but substitution of an alternate metal appeared to restore, at least in part, the original function of the protein. Alternate metal incorporation into bacterially-synthesized zinc-finger proteins has been demonstrated by X u et al. (1993). A plasmid-encoded fragment of methionyl-tRNA synthetase, containing the single putative zinc-finger motif and the ATP binding site, was overexpressed in Plasmid containing cells, grown in a zinc-depleted media supplemented with C0CI2 or  E. coli. 113  C d C l , incorporated C o 2  2+  or  1 1 3  Cd  2 +  into the protein. The enzyme activity of methionyl-tRNA  synthetase containing an alternate metal ion in the zinc-finger motif was compared to the native 2+ or Zn -substituted protein. The kinetic parameters of the metal-substituted proteins were found not to differ from the native enzyme. This study is of particular importance because it demonstrates that alternate metals can be incorporated into newly synthesized protein in an intact, although artificial, cell system.  35  2.7.2.2.  Evidence for Metal Replacement in the Zinc Fingers of PARP  Classical zinc blot competition experiments have been performed to assess the ability of various divalent metal ions to compete with Z n 6 5  65  nitrocellulose paper. Using  2 +  for binding sites on PARP immobilized on  2+  Zn -loaded PARP, Mazen et al. (1989) have shown that the bound  zinc could be displaced by various other bivalent metal ions, including Z n  2 +  or C u , and 2+  partially exchanged by Cd . Buki et al. (1991) demonstrated that alternate divalent metal ions can "rescue" P A R P from loss of activity following zinc removal. A PARP inhibitor, 6-nitroso1,2-benzopyrone (6-NOBP), which acts via destabilization of zinc in one finger motif of PARP, was used to effect zinc removal. Exogenously Zn -loaded PARP activity was shown to 65  decrease in a concentration-dependent  2+  manner with increasing 6-NOBP  concentrations.  Coincidental with an almost total loss of polymerase activity, nearly half of the Z n  2+  content of  P A R P was ejected by incubation with 6-NOBP. However, protection against 6-NOBP-induced inactivation was afforded by either 0.5 m M ZnCl2 or CdCl2. These investigations provide some indirect evidence that alternate metals can replace the zinc in the zinc-finger site in PARP, and that in vitro enzyme activity can be maintained with an alternate metal. However, direct measurements of metal-binding stoichiometry of PARP with alternate metals have not been performed. In addition, the potential for and consequences of in vivo  metal incorporation, in terms of PARP function and repercussions to the cell, are not clear  and may be metal dependent. From a toxicological perspective, the chronic sub-lethal exposure of cells to metals may result in incorporation of contaminant metals into zinc-finger proteins, which, consequently, may have far-reaching effects on the function of the protein within the cell. The apparent biological function of PARP in regulating critical cellular processes, particularly its role in cell death, and preliminary experiments demonstrating the potential for metal replacement  36  in zinc fingers, provide ample reason to further investigate this as a mechanism to modulate PARP activity. 2.7.3.  Mass Spectrometric Analysis of Metal-Protein Complexes As alternatives to the use of radioactive metal isotopes for the study of metal replacement  in the PARP zinc fingers, other methods were sought to assess the metal-protein complexes. Several techniques have been used to study the interaction between metal ions and proteins, and include  absorption  spectrometry,  circular dichroism, electron  paramagnetic  resonance  spectrometry and nuclear magnetic resonance spectrometry (Loo, 1997). Recently, various mass spectrometric techniques have gained prominence for the analysis of noncovalently bound protein complexes, including matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) (Hutchens et al, 1992b; Scobie et al, 1993) and electrospray ionization mass spectrometry (ESI-MS) (Loo, 1997). 2.7.3.1.  Matrix-Assisted Laser Desorption Ionization - Time of Flight Mass Spectrometry  M A L D I - T O F MS is an extremely accurate and sensitive technique for the determination of molecular masses extending to greater than 200,000 Da (Scobie et al, 1993). This technique involves mixing the protein with a solution of a chromaphoric matrix. The analyte/matrix mixture is deposited and dried on a sample stage, and then irradiated with a pulsed laser at an absorption maximum of the matrix. The interaction of photons with the matrix and sample results in the formation of intact ions related to the molecular mass of the protein. In M A L D I TOF M S , the mass of an ion is calculated based on the time required to travel from the point of ion formation to the point of detection. This method has a typical mass accuracy of 0.1% (Scobie etal, 1993).  37  Hutchens et al. (1992b) investigated the use of M A L D I - T O F M S for examination of peptide-metal interactions. For this study, a 26-residue sequence, (GHHPH) G, was synthesized, 5  mixed in equal proportion with a saturated solution of the matrix, 2,5-dihydroxybenzoic acid, and analyzed in the presence or absence of metal ions. In the absence of metal ions, the apopeptide was detected. When the peptide-matrix mixture was analyzed in the presence of 1.7 m M copper sulfate, five additional peaks were observed on the mass spectrum corresponding to the binding of one copper ion per metal binding unit (GHHPH). Therefore, M A L D I - T O F M S has been used successfully to study the stoichiometric binding of copper ions to the 26-residue peptide. This study suggests that the detection of protein-metal complexes using M A L D I - T O F M S is possible, at least when metal ions are added exogenously. 2.7.3.2.  Electrospray Ionization Mass Spectrometry  ESI is a gentle ionization method yielding no molecular fragmentation and allows weakly bound complexes to be detected. In ESI-MS, the sample is introduced as a liquid and ions are formed as a result of spraying the liquid into a region of high electric potential at atmospheric pressure (Scobie et al, 1993). The droplets formed are desolvated and the resulting gas-phase ions are directed towards the mass analyzer. A unique aspect of electrospray ionization is the formation of a number of signals in the mass spectrum, which result from the multiple charge states of the protein. These multiply charged ions can be correlated to the molecular mass of the sample by determining the observed mass and charge state of each ion. A method based on maximum entropy (MaxEnt) has been developed to automatically disentangle the m/z spectrum and produce a deconvoluted spectrum presenting the data from each protein as a single peak on a true molecular weight scale (Green et ai, 1996). Generally, the molecular weights determined by ESI-MS combined with MaxEnt are within + 0.01% of the values calculated from the sequences.  38  In ESI-MS, the most common system for analyzing peptides and proteins is a volatile buffer consisting of water and an organic solvent (i.e., acetonitrile) containing a small amount of acid (e.g., formic acid or acetic acid). The addition of a small amount of acid promotes protonation and detection of molecules as positive ions. However, acidic conditions and organic solvent environments can lead to protein denaturation and destabilization of the metal-protein complex. Therefore, the problem of analyzing metal-protein complexes is not trivial, as it is necessary to maintain the specific noncovalent interactions between the protein and the metal during the generation and the transfer of ions from liquid to gas phase. For this purpose, the protein must be injected in a solution that preserves the form with the metal bound, and therefore,the composition and pH of the solution are of prime importance (Lafitte et al, 1995). Metal-protein complexes are generally found in an aqueous environment in situ. Preserving solution associations before a phase change is a prerequisite for successful detection of the complex by ESI-MS and aqueous solutions are preferable to those containing excessive amounts of organic solvent (Hu et al, 1994). However, it is more difficult to vaporize an aqueous solvent than an organic one, and therefore,from a technical perspective, it is desirable to have the sample in a volatile organic solvent. The balance between aqueous and organic composition must be determined empirically. Finally, for the analysis of zinc-finger proteins, such as PARP, that contain a histidine ligand in the zinc coordination site, a pH environment above 5 is required because histidine becomes protonated at a pH below 5 and the zinc is ejected (Surovoy et al, 1992). Taking into account the specific requirements of the metal-protein complex to be analyzed, ESI-MS appears to provide the mass accuracy and resolution to unambiguously determine the metal-binding stoichiometries with a protein. In a review describing the study of noncovalent protein complexes, Loo (1997) has illustrated the potential of ESI-MS to determine metal-protein stoichiometry. ESI-MS has been  39  used to determine the calcium-binding stoichiometry of Ca -binding proteins, such as +  parvalbumin, lactalbumin, calmodulin and calbindin (11,976.7, 14,182.2, 16,791.5 and 29,866 Da, respectively) using both positive and negative electrospray conditions and non-acidic solvents (Hu et al, 1994; Lafitte et al, 1995; Veenstra et al, 1997). These proteins are able to bind multiple calcium ions through a helix-loop-helix structural motif where each C a  2+  ion is  coordinated to 6 amino acid residues (Lafitte et al, 1995) Metallothioneins (MW 6,000 - 7,000 Da) are unique metalloproteins that contain 20 cysteine residues in their 61-amino acid structure. Metallothioneins are capable of binding up to 7 divalent transition metals, and all cysteine residues are thought to participate in coordinating the metal ions through metal-thiolate bonds. The metal-thiolate bond is pH sensitive due to protonation of the cysteine ligands and subsequent metal ejection below pH 3. However, the detection of 7 metal ions, either zinc or cadmium, has been accomplished by ESI-MS at pH 10 (Yu etal, 1993).  The steroid family of nuclear receptors are transcription factors that contain zinc-finger regions in their D N A binding domains and are involved in the sequence-specific binding to D N A . Typically, these transcription factors have the zinc ion coordinated to four Cys residues, the classical type of zinc-finger motif. The zinc-binding stoichiometry of the estrogen receptor (MW = 8,248 Da), the vitamin D receptor (MW = 12,819 Da) and the glucocorticoid receptor (MW = 9,474 Da) have been analyzed by ESI-MS under non-acidic conditions (Hutchens et al, 1992a; Witkowska et al, 1995; Craig et al, 1997). These studies also demonstrated that the zinc fingers of both the glucocorticoid and estrogen receptor D N A binding domains were also able to bind cadmium and copper, respectively. The human immunodeficiency virus nucleocapsid protein, Ncp7 ( M W = 6,444 Da), contains two zinc fingers of the form Cys-X2-Cys-X4-His-X4-Cys. This motif is similar to that  40  of the P A R P zinc fingers, except the "loop" region is shorter (i.e., 4 residues in each Ncp7 motif compared to 28 and 30 in the P A R P zinc fingers). The zinc fingers of Ncp7 are involved in the encapsulation of genomic R N A during HIV viral assembly. Surovoy et al. (1992) and Loo et al. (1996) have demonstrated that ESI-MS can be used to determine the zinc stoichiometry for Ncp7. These studies illustrate that metal-binding stoichiometry of noncovalent metal-protein complexes, including zinc-finger motifs similar to the ones in PARP, can be analyzed by ESIM S . However, many of the zinc-finger proteins that have been analyzed to date are fragments of the intact protein. These fragments are composed of either the entire D N A binding domain or only a peptide fragment encompassing the zinc-finger region, and the molecular weight range of these fragments is between 6,400 and 13,000 Da. In addition, most studies report in vitro metal binding stoichiometries and have not isolated the metalloprotein from an intact cell system exposed to alternate metals. Therefore, the concept of alternate metal incorporation into native zinc-finger proteins in growing cells has yet to be fully investigated. Finally, the metal-binding property of a zinc-finger protein in the molecular weight range of P A R P (-113 kDa) has not been analyzed when isolated from intact cells.  41  2.8.  Thesis Objectives This thesis will examine examined how alteration of PARP activity influences the process  of cell death following an oxidant exposure. Specific objectives of this project were:  1.) To investigate metal replacement in the zinc fingers of PARP as a mechanism to modulate DNA-binding capacity, and therefore,PARP activity, and further, to determine the effect of metal replacement in the PARP zinc fingers on cellular response to oxidative stress. Specific objectives were developed to investigate metal replacement in the zinc fingers of PARP and the potential effect that metal replacement may have on cellular response to oxidative stress were: a. ) To produce and purify recombinant human PARP using the baculovirus expression system as a source of the purified enzyme for subsequent studies. b. ) To develop and optimize a PARP enzyme assay that will be subsequently used to determine the effect of metal replacement on PARP activity. c. ) To develop an analytical method to measure the extent and type of metal replacement in the zinc fingers of PARP. The ability to verify metal replacement is a critical aspect of the proposed research, and therefore, the development of this technique is essential for progression to the next steps.  d. ) To assess the effect of metal replacement in the zinc fingers of PARP in terms of in vitro enzyme activity and D N A binding capacity. e. ) To grow endothelial cells in the presence of an alternate metal and assess metal incorporation into the zinc fingers of PARP using the analytical technique developed.  42  f.) To determine the effect of metal replacement in the zinc fingers of P A R P in an ex vivo model of oxidant injury. Specifically, endothelial cells grown in the presence of an alternate metal would be subjected to an oxidant stress, and the time-course and extent of cell death, as well as caspase-3 activity and nuclear morphology, would be assessed and compared to native cells.  2.) To determine the effect of pharmacological inhibition of PARP activity following oxidative injury in an ex vivo endothelial cell culture system as a way of modulating cell death. Specific objectives were developed to guide the investigation of oxidant-induced endothelial cell injury and cell death and to evaluate the mechanism of cell death following oxidant stress were: a. ) To assess the time-course and extent of oxidant-induced loss of endothelial cell membrane integrity as a marker of cell death, using hydrogen peroxide as the initiator of oxidant stress. b. ) To determine the effect of the presence of P A R P inhibitors and/or inhibitors of protein synthesis or endonuclease activity, on the time-course and extent of oxidant-induced endothelial cell death. c. ) To assess caspase-3-like enzyme activity and substrate cleavage activity over the timecourse of oxidant-induced endothelial cell death in PARP-inhibited and non-inhibited cells as an indicator of apoptosis. d. ) To assess morphological changes in the nucleus of PARP-inhibited and non-inhibited cells, using fluorescence microscopy as a way to differentiate between oncosis and apoptosis.  43  3.  3.1.  EXPERIMENTAL  Materials Unless otherwise noted, all chemicals were obtained from Sigma Chemical Co. (St. Louis,  MO),  and  all cell  culture  plastic-ware  and  chemicals  were  obtained  from  Life  Technologies/Gibco B R L (Burlington, O N , Canada). The recombinant baculovirus containing the full-length gene for human P A R P (AcPARP) was obtained from Dr. Gilbert de Murcia (Centre National de la Recherche Scientifique, Strasbourg, France). Recombinant caspase-3 was obtained from Dr. D.W. Nicholson  (Merck Frosst Centre for Therapeutic Research, Pointe  Clare-Dorval, QC, Canada). The defined bovine calf serum (BCS) and characterized fetal calf serum (FCS) used for the culture of the bovine pulmonary artery endothelial cells were purchased from Hyclone Laboratories (Logan, UT, USA). Phenylmethylsulfonylfluoride (PMSF), leupeptin, pepstatin A , trypsin, chymotrypsin and catalase were obtained from Boerhinger Mannheim (Laval, QC, Canada). The P A R P inhibitors, 1,5-dihydroxyisoquinoline (DIQ) and 3-aminobenzamide (3AB), as well as l,l,l,3,3,3-hexafluoro-2-propanol (HFTPA) were obtained from Aldrich Chemical Co. (Milwaukee, WI, USA). Acrylamide, bisacrylamide, Coomassie Brilliant  Blue  G-250, N-A-A'-N'-tetramethylethylenediamine (TEMED)  and  ammonium persulfate, (NH^SaOs, were obtained from BioRad Laboratories (Hercules, C A , USA). HPLC-grade methanol, isopropanol and acetonitrile (ACN) were purchased from Fisher Scientific (Fair Lawn, NJ, USA), as was trifluoroacetic acid (TFA) and acetic acid. CytoScint " 1  scintillation fluid was obtained from ICN (Costa Mesa, C A , USA). Nicotinamide [U- C] 14  adenine dinucleotide [ C - N A D ] was obtained from Amersham Canada Ltd. (Oakville, O N , 14  Canada).  +  Histone H I was obtained from Life Technologies/Gibco B R L (Burlington, O N ,  44  Canada). Propidium iodide (PI), YO-PRO-1  iodide (YO-PRO), fura-2 and acridine orange  (AO) were obtained from Molecular Probes Inc. (Eugene, OR, USA). CPP-32 Fluorescent Assay Kit was obtained from Clontech Laboratories, Inc. (Palo Alto, C A , USA). A Milli-Q water purification unit (Millipore Canada Ltd., Mississauga, O N , Canada) system was the source of purified water. Milli-Q water is distilled water that was passed through a series of cartridges to remove organics and inorganics. The purified water underwent a final filtration through a 0.22 (im filter to remove microorganisms and particles.  3.2.  Cell Culture  Techniques  3.2.1.  Cell Culture of Bovine Pulmonary Artery Endothelial Cells Primary cultures of bovine pulmonary artery endothelial cells (EC) were routinely  harvested by scraping the lumenal surface of the artery with a scalpel blade. The E C obtained from several arteries were pooled, and subsequently maintained in continuous culture at 3 7 C under humidified 95% air and 5% CO2. The culture medium was M199 supplemented with 15% BCS, 5% FCS, M E M vitamins, B M E amino acids and L-glutamine, as described previously (Thies and Autor, 1991). Cells were grown as adherent monolayers in Nunc T25 tissue culture flasks and routinely sub-cultured when they reached confluency. One cell passage was defined as the period of time from initial plating at 30-50% cell density to a 100% confluent monolayer. This usually required 72 h under standard culture conditions. A l l experiments utilized cells passaged from 3-20 times. 3.2.2.  Cell Culture of Spodoptera frugiperda (S/9) Cells o  S/9 cells were continuously maintained at 27 C in either suspension or monolayer culture in Grace's Insect Cell Culture Media supplemented with 3.3 g/L yeastolate, 3.3 g/L lactalbumin hydrolysate and 10% qualified FBS, as described by Summers and Smith (1988). The complete 45  medium is referred to as TC-100. Antibiotics were not used for routine culturing of S/9 cells. S/9 cells in suspension culture were maintained in 50-mL spinner flasks (Bellco Glass Inc., Vineland, NJ, USA) with constant stirring at 50 - 60 rpm. Every 5 days, the density of suspension cultures was determined by counting with a hemocytometer. Spinner flasks were reestablished at a density of 5 x 10 cells/mL and the volume brought up to 50 mL with fresh 5  TC-100. Each 50-mL spinner flask was used for 3 consecutive passages before being cleaned and sterilized. For monolayer culture, 1 x 10 cells were seeded into Nunc T25 cell culture flasks 6  and 5 mL of fresh medium was added. When monolayer cultures reached confluency, the cells were dislodged by rapidly pipetting the media across the monolayer with a sterilized Pasteur pipet. For passage of these cells, a new T25 flask containing 4.5 mL fresh medium was seeded with 0.5 mL of the cell suspension.  3.3.  Recombinant  PARP  Production  Using  the  Baculovirus  Expression  System 3.3.1.  AcPARP Virus Propagation The baculovirus vector was Autographa californica multicapsid nucleopolyhedrosis virus  (AcMNPV), and preparation of the recombinant virus was described by Giner et al. (1992). Initially, the recombinant virus (AcPARP) stock, which had a titre of 3 x 10 plaque forming 8  units per mL (pfu/mL), was propagated to produce a stock of virus for protein expression. Virus propagation involved infecting S/9 cells at a low multiplicity of infection (MOI) and, after a period of infection, harvesting the virus-containing medium. M O I is defined as the number of infectious units added per cell. For virus propagation, a MOI of 0.5 - 1.0 plaque forming units (pfu) per cell was used, with 1 x 10 cells being infected in a total volume of 50 mL. 8  Approximately 6 days post-infection, the medium was harvested by centrifugation of the cells at  46  500g for 5 min. The medium was transferred to sterile culture tubes, and the viral stocks were stored at 4 C protected from the light (Jarvis and Garcia, Jr., 1994). 3.3.2.  Assessment of Viral Titre The titre of the virus stock was determined by assessment of the tissue culture infection  dose (TCID ) according to the method of Summers and Smith (1988). Log dilutions of the virus 50  stock were prepared (10"' to 10" dilutions), and each dilution was used to infect 1 row of a 968  well micro-titre plate containing S/9 cells (1.5 x 10 cells/well). The infection process was 4  monitored for 7-10 days. The lack of cell growth and the inability of the cells to reach confluence identified wells of infected cells. The TCID50 is defined as the dilution that would give rise to 50% positive wells and 50% negative wells. The titre of the virus in infectious doses per unit of inoculum was obtained from calculating the reciprocal of the TCID50 value and converting it to pfu/mL (King and Possee, 1992). 3.3.3.  Effect of MOI on Extent and Time-Course of Recombinant PARP Expression In order to characterize the AcPARP/S/9 expression system for recombinant P A R P  (rePARP), the effect of M O I on the time-course and the extent of protein expression was examined. S/9 cells (1 x 10 cells/50 mL) were infected with AcPARP at a MOI of 0.1, 1.0 or 10 8  pfu/mL. For the initial infection period, S/9 cells were placed in separate 50-mL Falcon tubes (VWR Canlab, Mississauga, ON, Canada) and the appropriate amount of virus stock was added. To facilitate virus contact with the cells and synchronization of infection, the total volume in each tube was kept between 5 - 1 0 mL. The cells were placed at 27 C for 2 h with periodic gentle mixing. Following the 2 h incubation, the infected cells were transferred to 50-mL spinner flasks containing fresh TC-100 and returned to the 27°C incubator.  47  At 2, 12, 24, 36, 48, 72, 96 and 120 h post-infection, a 5-mL aliquot of the cell suspension was removed and placed in a 15-mL Falcon tube. The cells were pelleted by centrifugation at 500g for 5 min, resuspended in 5 m L I X Dulbecco's phosphate-buffered  saline (PBS,  components in g/L: KC1, 0.2; K H P 0 , 0.2; M g C l anhydrous, 0.047; NaCl, 8.0; N a H P 0 , 1.15) 2  4  2  2  4  (Life Technologies/Gibco B R L , Burlington, O N , Canada), spun again, and finally resuspended in 1 mL ice-cold lysis buffer (25 m M Tris-HCl pH 8.0, 50 m M glucose, 10 m M ethylenediamine tetraacetic acid (EDTA), 1 m M |3-mercaptoethanol ((3-ME), 1 m M phenylmethylsulfonylfluoride (PMSF) freshly added). The cell suspension was transferred to a microcentrifuge tube and frozen at -80 C until analysis by SDS-polyacrylamide gel electrophoresis (SDS-PAGE). 3.3.4.  Routine Expression of Recombinant PARP For routine expression of recombinant PARP, a maximum of 5 x 10 cells in a total of five  spinner flasks (1 x 10 cells / 50-mL spinner flask) were infected at a M O I of 5. The initial low 8  volume, 2 h incubation period, described in section 3.3.3, was used. At 72 h post-infection, the cells from each spinner flask were transferred to 50-mL Falcon tubes and spun in a centrifuge at 500g for 5 min. The cell pellets were washed with 5 mL ice-cold PBS and pelleted again by centrifugation. The pellets were then resuspended in 5 mL ice-cold lysis buffer and pooled. The protease inhibitors, leupeptin, pepstatin A and P M S F were added to the cell lysate at final concentrations of 0.5, 0.7 and 174.2 U.g/mL, respectively. Purification of recombinant P A R P was carried out immediately after harvesting cells.  3.4.  Purification of Recombinant  PARP  3.4.1.  Preparation of a Crude Cell Extract The purification procedure for recombinant P A R P produced in the baculovirus expression  system was performed according to the method of Giner et al. (1992). This procedure involved  48  an initial D N A precipitation step followed by selective protein precipitation using 30%-70% ammonium sulfate. The pooled cell lysates of AcPARP-infected S/9 cells were homogenized on ice using 2 x 10 strokes on a Potter-Elvenhjem tissue grinder mounted on a T-Line Laboratory Stirrer (Montrose, P A , USA) with a setting of 40. Tween-20 (0.2%), Nonidet P-40 (0.2%) and NaCl (0.5 M ) were added and the cell lysate was incubated at 4°C for 60 min with agitation. The solution was cleared by centrifugation (20 min at 20,000g) at 4°C. Protamine sulfate (1 mg/mL) was added to the supernatant to precipitate D N A . The lysate was incubated at 4°C for 60 min with agitation followed by centrifugation (10 min at 20,000g) at 4°C. Solid ammonium sulfate was slowly mixed into the supernatant to reach 30% saturation and then incubated for 60 min at o  4 C with continuous agitation. The lysate was spun in a centrifuge at 20,000g for 15 min. The supernatant was collected and solid ammonium sulfate was added to reach 70% saturation. The lysate was incubated for a further 60 min at 4°C with agitation. After a 15 min centrifugation at 20,000g, the supernatant was discarded and the precipitated material was resuspended in 2.5 mL / 10 cells in the purification buffer (100 m M Tris-HCl pH 7.4, 10% glycerol, 0.5 m M E D T A , 1 8  m M PMSF, 12 m M p-ME). The protease inhibitors, leupeptin, pepstatin A and PMSF were added to the cell lysate at final concentrations of 0.5, 0.7 and 174.2 (ig/mL, respectively. At each step of the purification procedure, fresh PMSF was added to the supernatant because it has a short half-life in aqueous solutions. To monitor the progress of the purification, 20-|iL aliquots were taken at each stage in the purification procedure for analysis by SDS-PAGE. The partially purified sample was then subjected to a final purification step by affinity chromatography.  49  3.4.2.  Preparation of the 3AB-Affigel Affinity Chromatography Resin A specific affinity chromatography resin was prepared by coupling the P A R P inhibitor, 3-  aminobenzamide (3AB), to an activated agarose gel, Affigel  10™ (BioRad Laboratories,  Hercules, C A , USA). Affigel 10 is an N-hydroxysuccinimide ester of a derivatized cross-linked agarose gel bead support and contains a neutral 10-atom spacer arm (BioRad Bulletin 1085). Ligands with free primary amino groups will couple spontaneously with Affigel 10 through a stable amine bond, releasing N-hydroxysuccinimide. A n anhydrous coupling method in the presence of isopropanol was performed as outlined by Ushiro et al. (1987) and in BioRad Bulletin 1085. The coupled resin was washed thoroughly with isopropanol in a sintered-glass funnel to remove unbound 3AB, and then gradually equilibrated into water and finally into the affinity chromatography loading buffer (100 m M Tris-HCl p H 8.0, 17% glycerol, 0.5 m M E D T A , 0.3 m M KC1, 12 m M (3-ME). After degassing, the coupled resin was poured into a 10 x 1.0 cm I D . glass chromatography column to achieve a bed height of approximately 7 cm (bed volume of approximately 7 mL). 3.4.3.  Final Purification of PARP by Affinity Chromatography The final purification step involved 3AB-affinity chromatography  as described by  Burtscher et al. (1986). Prior to application of the partially purified cell extract to the column, the column was equilibrated in loading buffer at a flow rate of 15 mL/h. The partially purified cell extract was applied to the column, followed by loading buffer. The U V absorbance of the column effluent was monitored at 280 nm at all times. When the unretained proteins had eluted from the column, as evidenced by a large initial peak, the elution buffer (loading buffer containing 5 u M DIQ) was applied to the column. Elution of P A R P from the affinity column occurred shortly after application of the elution buffer and was seen as a large absorbance peak  50  (Fig. 10, page 88). The baseline of the elution buffer was slightly higher than that of the loading buffer due to the inherent absorbance of the P A R P inhibitor, DIQ. The effluent from the affinity column was collected in approximately 7.5-mL fractions and the entire affinity chromatography procedure was carried out at 4 C. A n aliquot of 20 | i L was removed for analysis by SDS-PAGE to assess the protein profile of the various fractions. Fractions containing the purified P A R P protein were pooled and concentrated to <10 mL using a 50-mL Amicon stirred cell apparatus equipped with a YM-30 membrane (Amicon, Beverly, M A ) . The concentrated protein was placed in 0.75" diameter dialysis tubing with a molecular weight exclusion limit of 12,000 - 14,000 Da (Life Technologies/Gibco B R L ) . Dialysis of the concentrated protein was carried out against 2 L of dialysis buffer (50mM TrisHC1 pH 8.0, 1 m M DTT, 4 m M MgCl ) for 6 h at 4°C. The dialysis buffer was replaced, and the 2  protein concentrate dialyzed overnight. A final dialysis in 1 L of fresh dialysis buffer was performed for 2 h. The dialyzed protein solution was divided into 1-mL aliquots and stored at -80°C. To achieve a uniform supply of purified P A R P for subsequent experimental procedures, the expression and purification of P A R P was repeated until approximately 25 mg of P A R P had been purified. The various batches of purified P A R P were allowed to thaw on ice and then pooled. The pooled P A R P was divided into single-use aliquots of 50 - 100 uL and stored at 80°C. A protein assay was performed to determine the concentration of the pooled PARP.  3.5.  Protein  Quantitation  The protein concentration of the purified recombinant P A R P extract was determined according to the dye-binding method of Bradford (1976), using the Protein Assay Quantitation Reagent obtained from BioRad (Hercules, C A , USA). Standard curves were prepared over a range of 0 - 600 (Xg/mL bovine serum albumin using the BioRad Protein Assay Standard II. To  51  account for the volume of dialysis buffer in the protein samples measured, an equivalent volume of dialysis buffer, usually 75 uL, was added to the standard samples. The absorbances from triplicate PARP samples and duplicate standard samples were measured at 595 nm using a Hewlett Packard 8452A Diode Array Spectrophotometer.  3.6.  Protein Analysis by SDS-Polyacrylamide  Gel  Electrophoresis  SDS-PAGE was carried out using the buffer system of Laemmli (1970). The discontinuous gel system, composed of the stacking and resolving gel, as well as all buffers, was prepared according to the protocols in Molecular Cloning: A Laboratory Manual (Sambrook et al, 1989). The resolving gel was prepared by diluting an acrylamide:bisacrylamide mixture (29:1) to 812% (g/mL) in a buffer containing 5% glycerol, 1% SDS and 375 m M Tris-HCl pH 8.8. The stacking gel was prepared by diluting an acrylamide:bisacrylamide mixture (29:1) to 5% (g/mL) in a buffer containing 1% SDS and 125 m M Tris-HCl pH 6.8. Polymerization of the gels was initiated by adding (NH ) S20 at 0.1% and a trace amount of T E M E D (4 uL to both 10 mL 4  2  8  resolving gel mixture and 4 mL stacking gel mixture). The protein samples were diluted with an equal volume of the SDS-PAGE sample dilution buffer containing 4% SDS, 200 m M dithiothreitol (DTT), 24% glycerol, 0.2% bromophenol blue and 100 m M Tris-HCl pH 6.8. Immediately before loading the gel, the protein samples were boiled for 2 min and then placed on ice. Twenty microlitres of each protein sample were loaded on the polymerized SDSpolyacrylamide gel (7x8 cm ). The electrophoresis buffer (25 m M Tris base, 250 m M glycine, 2  1% SDS, pH 8.3) was used to run the gel in a BioRad Mini-PROTEIN® II electrophoresis system under constant voltage of 60 V for the stacking gel and 100 V for the resolving gel. The gel was stained for 1 h with 0.25% Coomassie Brilliant Blue G-250 in methanol:water:acetic acid (45:45:10 v/v), and destained for 1 h in methanol:water:acetic  acid (45:45:10 v/v).  52  Consecutive washes in 7% acetic acid and then in 3% glycerol were performed prior to gel drying. The stained gel was dried in a model SE 540 gel drier (Hoefer Scientific Instruments, C A , USA).  3.7.  PARP Enzyme Activity Assay PARP activity is measured as incorporation of ADP-ribose, derived from N A D , into acid+  insoluble precipitates. Initial experiments, which assessed the effect of D N A and histones on P A R P activity, were performed under a variety of conditions (e.g., incubation time, [NAD ]), +  and are specified in the figure legends. To prepare sonicated D N A , 1-mL aliquots of calf thymus D N A (1000 (Xg/mL), dissolved in Milli-Q-grade water, were sonicated on ice (microtips at limit, 40% duty cycle) (Vibra Cell™, Sonics & Materials, Danbury, C T , USA). Different sonication times were evaluated (0 - 90 s), and for the standard P A R P assay, D N A sonicated for 30 s was routinely used. The sonicated D N A was pooled, re-aliquoted and stored at -80 C until use. Stock solutions of P A R P (50-100 [ig/mL) and histone HI (1000 |ig/mL) (Life Technologies/Gibco B R L ) were freshly prepared in Milli-Q-grade water for each experiment. When lysine-rich histones (Worthington Biochemical Corp., Freehold, NJ, USA) were used, a 1000 |ig/mL stock was prepared in Milli-Q-grade water. In preparation for each reaction, the following sequence of addition of reaction components was performed: the reaction buffer was added first, followed by PARP, D N A and, finally, histone H I . To develop a standardized assay, the effect of the amount of PARP, length of incubation and concentration of N A D were determined, and combined with the results from preliminary +  experiments on the effect of D N A and histone HI on P A R P activity. Therefore, in the standardized P A R P activity assay, the final reaction mixture contained 0.5 |ig P A R P , 10 (ig D N A , 10 \ig histone H I , 100 m M Tris-HCl pH 8.0, 10 m M M g C l , 5 m M DTT and 500 u M 2  53  N A D containing 1.0 uCi/umol [ C - N A D ] in a total volume of 100 uL. The tubes were kept on +  14  +  ice until the reaction was initiated by addition of [ C- N A D ] . Under standard conditions, the 14  +  samples were incubated at 37 C for 2 min. To ensure the contents were at optimal temperature at the start of the reaction, each tube was incubated at 37 C for 1 min immediately prior to addition of [ C- N A D ] . The reaction was stopped adding of 700 uL ice-cold 10% trichloroacetic acid 14  +  (TCA) followed by placing the samples on ice. Protein precipitates were collected on Whatman GFC filter discs (Whatman International, Maidstone, England) and the  unincorporated  radioactivity was removed by 2 x 5 mL washes of 5% T C A . The radioactivity was counted on a Beckman LS 6000TA scintillation counter. PARP activity was expressed as nmol ADP-ribose incorporated per min per fig PARP (nmol/min/fig protein). PARP activity was assessed in the presence of known inhibitors, as well as a novel PARP inhibitor. Nicotinamide, a known PARP inhibitor (Banasik et al, 1992), is released as a product of the ADP-ribosylation reaction. Therefore, a concentration-response curve was constructed comparing the effect of increasing concentrations of nicotinamide on P A R P activity. Each sample contained 1.0 ug PARP, 5 ug D N A , 5 fig histone H I , 100 m M Tris-HCl pH 8.0, 10 m M M g C l , 5 m M DTT, 100 u M N A D containing 1.25 uCi/umol [ C - N A D ] and either 0.1, 1, 10, +  14  +  2  100, 1000 and 10,000 u M nicotinamide in a total volume of 100 uL. The samples were o  incubated for 2 min at 37 C. The reaction was stopped by addition of 700 uL 10% T C A and placing the samples on ice. Protein precipitates were collected, the unincorporated radioactivity was removed by 2 x 5 mL washes of 5% T C A , and the radioactivity counted. Concentrationresponse curves were also constructed for two other known inhibitors of PARP, 3AB and DIQ (Banasik et al, 1992), as well as a suspected PARP inhibitor, allopurinol. Each sample contained 1.0 fig PARP, 10 |ig D N A , 10 ug histone H I , 100 m M Tris-HCl pH 8.0, 10 m M M g C l , 5 m M 2  54  DTT, 100 u M N A D containing 1.25 uCi/umol [ C - N A D ] . The samples were incubated for 2 +  14  +  min at 37 C and the reaction stopped by precipitation of protein with the addition of 10% T C A . Protein precipitates were collected, the unincorporated radioactivity was removed and the radioactivity counted. The effect of DIQ and 3AB on PARP activity was assessed over the concentration range of 0.001 - 100 | i M and 0.01 - 1000 | i M , respectively. Allopurinol was prepared in dimethylsulfoxide (DMSO) at a concentration of 33.90 m M and dilutions of this stock solution were prepared in water. The effect of allopurinol on P A R P activity was assessed over the concentration range of 1 - 3390 u M . Control samples were assessed to determine the effect of D M S O alone on PARP activity.  3.8.  Proteolytic Digestion of PARP  3.8.1.  Trypsin Digestion A stock solution of trypsin (33.5 (ig/mL) was prepared in Milli-Q-grade water and added  to 98 |j,g of PARP in a solution containing 50mM Tris-HCl pH 8.0, 1 m M DTT, 4 m M M g C l , to 2  achieve a trypsin:PARP ratio of 1:500 (g/g). The digest was incubated at room temperature (20°C) for 0.5, 15, 30, 60, 90, 120 and 150 min. At each time-point, 20 [iL of the digest were removed, diluted with an equal volume of SDS-PAGE sample dilution buffer and placed on ice until analysis by SDS-PAGE. 3.8.2.  Caspase-3 Digestion Recombinant caspase-3 was obtained as a 40 n M solution in 0.1 M H E P E S - K O H pH 7.5,  10% sucrose, 0.1% CHAPS and 10 m M DTT. The caspase-3 was divided into small aliquots (10 - 20 |iL) and stored at -80 C. The time-course of PARP digestion by caspase-3 was determined over a 120 min period. Caspase-3 was added to 50 u L of purified P A R P (940 jJ.g/mL) in a  55  solution containing 50mM Tris-HCl pH 8.0, 1 m M DTT, 4 m M MgCl2, to achieve a caspase3:PARP ratio of 1:1000 (mol/mol). The digest was performed at 37 C, and 10 fiL aliquots were removed at 5, 15, 30, 60, 90 and 120 min after the digestion was started. The samples of the digest were diluted with 10 uL Milli-Q-grade water and an equal volume of SDS-PAGE sample dilution buffer and placed on ice until analyzed on a 12% gel by SDS-PAGE. For routine digestion of PARP with caspase-3, up to 2 mg of PARP were placed in a 5-mL capped polypropylene tube and digested at a PARP:caspase-3 ratio of 2000:1 for 4 h at 37 C. The digest was monitored by SDS-PAGE at 1, 2, 3 and 4 h by removing a 10-uL aliquot, and diluting it with 10 | i L Milli-Q-grade water and an equal volume of SDS-PAGE sample dilution buffer. The samples were analyzed on a 12% gel by SDS-PAGE to verify complete digestion of PARP. The remaining digest was placed on ice to be used immediately, or divided into aliquots and stored at -80°C.  3.9.  Separation of Caspase-3-Digested PARP  Fragments  3.9.1.  Centrifugal Ultrafiltration Under Non-Denaturing Conditions It was postulated that the caspase-3-digested PARP fragments could be separated by the  selective retention of the large 89 kDa PARP fragment by centrifugal ultrafiltration. The small 24 kDa fragment could then be concentrated from the filtrate. Caspase-3-digested PARP (1.2 mg) was placed in an Ultrafree-4 centrifugal filter unit (Millipore Canada Ltd., Mississauga, ON, Canada) containing a Biomax filter with a nominal molecular weight limit (NMWL) of 50,000 Da. The filter unit was pre-rinsed prior to use by spinning in a centrifuge with 2 mL Milli-Qgrade water at 3,500g for 15 min at room temperature in a J6B centrifuge (Beckman, Palo Alto, C A , USA) equipped with a JS-4.2 swinging-bucket rotor. The water retained by the filter, as well as the filtrate, was discarded. The PARP digest (-1.5 mL) was placed in the device and 1.5  56  mL of ice-cold Milli-Q-grade water was added. The digest was spun in a centrifuge at 4 C for 15 min at 3,500g. The filtrate was removed and placed in a separate 50-mL Falcon tube on ice. The retained digest was washed with 3 mL of Milli-Q-grade water and spun in a centrifuge twice at o  4 C for 15 min at 3,500g. The filtrate was collected after each spin and pooled. The final retentate (total volume of 250 (iL) was divided into aliquots and stored at -80C. The pooled filtrate, approximately 7.5 mL, was concentrated using an Ultrafree-4 centrifugal filter device containing a Biomax filter with a 10,000 Da N M W L to approximately 500 | i L by spinning in a centrifuge at 4 C for 3 x 10 min at 3,500g, adding 3 mL of retentate at each step. The concentrated retentate was then desalted by 3 consecutive spins in 3 mL of ice-cold M i l l i - Q grade water. The concentrated retentate was divided into aliquots and stored at -80C. A 20 | i L sample from each fragment concentrate was removed and diluted with an equal volume of SDSP A G E sample dilution buffer. The samples were analyzed on a 10% gel by SDS-PAGE to assess the effectiveness of the separation of PARP fragments by centrifugal ultrafiltration. 3.9.2.  Centrifugal Ultrafiltration Under Denaturing Conditions The inability to separate the PARP fragments by centrifugal ultrafiltration under non-  denaturing conditions prompted an attempt to separate the fragments under conditions where potential non-specific association of the fragments would be restricted. Therefore, caspase-3  r  digested PARP was subjected to centrifugal ultrafiltration in the presence of 8 M urea, to completely denature the fragments and limit any non-specific binding that may have prevented their separation under non-denaturing conditions. Centrifugal ultrafiltration under denaturing conditions was carried out in filter units containing Biomax filters with either a 50,000 Da N M W L or a 100,000 Da N M W L . Aliquots containing 497 u.g of caspase-3-digested PARP were placed in each Ultrafree-4 centrifugal filter  57  unit. Both filter units were pre-rinsed prior to use by spinning in a centrifuge with 2 mL M i l l i - Q grade water at 3,500g for 15 min at room temperature. Three milliliters of ice-cold 8 M urea were added, and each device was spun in a centrifuge at 4°C for 15 min at 3,500g. There was approximately 800 u L of retentate remaining following the initial centrifugation. The filtrate was removed from the collection tube and placed in a 50-mL Falcon tube on ice. Two more successive washes were performed by centrifugation in 2 mL 8 M urea at 4°C for 20 min at 3,500g. The retentate was then desalted by 3 successive washes in 3 mL ice-cold Milli-Q-grade water to yield approximately 700 uL from the 50,000 Da N M W L filter and 500 uL from the 100,000 Da N M W L filter. The filtrate from each wash was collected, pooled, and, finally, concentrated and desalted in an Ultrafree-4 filter containing a Biomax filter with 10,000 N M W L by 3 successive washes in 3 mL ice-cold Milli-Q-grade water. The filtrates from the 50,000 Da N M W L and 100,000 Da N M W L filters were concentrated to approximately 300 uL and 400 | i L , respectively. A 10-uL sample from each fragment concentrate was removed and diluted with an equal volume of SDS-PAGE sample dilution buffer. The samples were analyzed on a 10% gel by SDS-PAGE to assess separation of the PARP fragments by centrifugal ultrafiltration under denaturing conditions. 3.9.3.  Affinity Chromatography An affinity matrix that specifically binds PARP through interaction with the catalytic site  had been prepared in our lab. Therefore, it was postulated that separation of the DNA-binding domain containing the zinc-finger region (24 kDa fragment) from the substrate binding region, within the 89 kDa fragment, was possible by the differential interaction with the affinity matrix. The 3AB-Affigel affinity resin was prepared as a slurry containing 20 mL of the gel bed resuspended in a total volume of 50 mL of affinity chromatography loading buffer (100 m M  58  Tris-HCl pH 8.0, 17% glycerol, 0.5 m M E D T A , 12 m M (3-ME) that did not contain KC1. This slurry is referred to as the gel suspension. Separation of the PARP fragments was performed using a batch method in 1.5-mL microcentrifuge tubes. Initial experiments examined the effect of increasing KC1 concentration and quantity of affinity matrix on the ability to separate the PARP fragments. The KC1 concentration was varied to determine the concentration that allowed binding of the large fragment to the 3AB ligand and prevented aggregation of the small and large fragment. KC1 concentrations evaluated were 0, 100, 200, 300, 400 and 600 m M . Two different amounts of freshly coupled 3AB-Affigel (200 and 500 (iL of the gel suspension) were used at all KC1 concentrations to determine the binding capacity of the resin for the 89 kDa PARP fragment. To each tube 100 uL purified PARP (860 ug/mL), 800 uL loading buffer without KC1, and either 200 or 500 uJL of the gel suspension were added. The KC1 concentration was adjusted by addition of appropriate volumes of a 4.5 M stock KC1 solution. A tube containing no affinity gel ("no-gel control") acted as a control and all samples were adjusted to the same volume with loading buffer. The tubes were tumbled end-over-end at 4°C overnight. The following morning, the tubes were spun in a microcentrifuge (Eppendorf Scientific, Inc., Westbury, N Y , USA) at ll,750g for 1 min. Twenty microlitres of the supernatant were removed from each tube and diluted with an equal volume of SDS-PAGE sample dilution buffer. The samples were analyzed on a 10% gel by SDS-PAGE to detect changes in the protein profile of the PARP fragments following differential binding to the affinity resin (i.e., if the 3AB-Affigel selectively bound the 89 kDa PARP fragment, then the supernatant would reveal only the unbound 24 kDa fragment). The lack of selective binding of the PARP fragments with this 3AB-Affigel preparation suggested that the affinity ligand, 3AB, was not efficiently coupled to the resin, and therefore,the 89 kDa fragment was unable to bind. A second experiment was performed to examine the ability  59  of different batches of 3AB-Affigel to bind the PARP fragments selectively. The 3AB-Affigel gel suspension (lot #1) was used in the experiments described above. A second preparation of 3AB-Affigel (lot #2) was also evaluated for its ability to bind the PARP fragments differentially. Lot #2 came from a 3AB-Affigel preparation that had been used previously in the purification of baculovirus-expressed recombinant PARP, and therefore,had a demonstrated ability to bind fulllength PARP. The two preparations of 3AB-Affigel  (lot #1 and lot #2) were re-examined for  their ability to bind the PARP fragments differentially. A 0.5-mL aliquot of both lots of 3ABAffigel was transferred into two separate 5-mL polypropylene tubes in the presence of loading buffer containing 300 m M KC1 and 100 uL of P A R P (860 Ug/mL). The KC1 concentration was the same as that used in the purification of baculovirus-expressed PARP, and therefore,under these conditions, the affinity resin has previously been demonstrated to bind full-length PARP. A tube containing no affinity gel ("no-gel control") acted as a control and all samples were adjusted to the same volume (1.6 mL) with loading buffer. The tubes were tumbled end-over-end at 4 C overnight. The following morning, the tubes were spun in an Eppendorf microcentrifuge at ll,750g for 1 min. Twenty microlitres of the supernatant from each tube were removed and diluted with an equal volume of SDS-PAGE sample dilution buffer. The samples were analyzed on a 10% gel by SDS-PAGE to detect changes in the protein profile with differential binding of the PARP fragments to the affinity resin. As there was no selective binding of the PARP fragments with either batch of 3ABAffigel, an experiment was performed to verify the binding capacity of the 3AB-Affigel preparations for full-length PARP. For each lot of affinity resin, 0.5, 1.0 and 1.5 mL of 3ABAffigel were transferred into separate 5-mL polypropylene tubes in the presence of loading buffer containing 300 m M KC1 and 100 uL of PARP (940 Ug/mL). A no-gel control was  60  included and all samples were adjusted to the same volume (1.6 mL) with loading buffer. The tubes were tumbled end-over-end at 4 C overnight. The following morning, the tubes were spun in an Eppendorf microcentrifuge at ll,750g for 1 min. Twenty microlitres of the supernatant from each tube were removed and diluted with an equal volume of SDS-PAGE sample dilution buffer. The samples were analyzed on a 10% gel by SDS-PAGE to detect binding of PARP to the affinity resin and subsequent disappearance of PARP from the supernatant. 3.9.4.  HPLC Separation of PARP Fragments Under Acidic Conditions In the absence of a method to separate the caspase-3-digested P A R P fragments under non-  denaturing conditions, it was decided that reverse phase H P L C , under acidic conditions, would be used to separate the fragments and thus obtain a purified sample of the 24 kDa fragment for subsequent analysis by mass spectrometry. The H P L C consisted of an HP 1050 Liquid Chromatograph (Hewlett Packard, Avondale, P A , USA) equipped with a variable wavelength U V detector. The PRP-1 polymer analytical column (250 x 4.1 mm I.D., 5 urn), obtained from Hamilton Co. (Reno, N A , USA), was used in combination with the following mobile phases: (A) 20% A C N - 0.1% T F A (v/v), and (B) 90% A C N - 0.1% T F A (v/v), delivered at a flow rate of 1 mL/min. The protein fragments were eluted under a gradient system programmed as follows: 100% A at 1 min, going to 60% A/40% B at 10 min then back to 100% A at 11 min. Repeated 50-uE injections of caspase-3-digested PARP were performed and detection of eluted proteins was by U V absorbance at 280 nm. Three peaks were apparent in the chromatogram and the eluant from each peak was collected separately and pooled. After the collection was complete, the pooled fractions were lyophilized using the Freezone 4.5 Freeze Dry System (Labconco Corp. Kansas City, MI, USA). The lyophilized samples were resuspended in a small volume of  61  Milli-Q-grade water (i.e., < 1 mL) for analysis. The identity of the 24 and 89 kDa fragments was' confirmed by the location of the protein bands on a gel analyzed by SDS-PAGE. Although an H P L C method was available to separate the caspase-3-digested PARP fragments, the acidic conditions were not conducive to retention of metal in the zinc fingers located within the DNA-binding domain. Therefore, attempts were made to develop a method for the separation of proteins by H P L C under neutral conditions. Development of a non-acidic separation technique was desirable because the liquid chromatograph could be interfaced to the mass spectrometer for direct analysis of proteins following separation. For developmental work, an HP 1050 Liquid Chromatograph (Hewlett Packard, Avondale, P A , USA) equipped with a variable wavelength U V detector was used. Two different columns were evaluated: the PRP-1 polymer column (250 x 4.1 mm I D . , 5 (xm) obtained from Hamilton Co. (Reno, N V , USA) and the Jupiter C 4 column (50 x 2.0 mm I.D., 5 |xm) obtained from Phenomenex (Torrance, C A , USA). Detection of eluting proteins was by U V absorbance at 280 nm. Table 1 shows the various mobile phases and gradient programs evaluated. Model proteins were used to assess the various separation conditions, including trypsinogen, cytochrome c, as well as intact PARP and the caspase-3-digested P A R P fragments. Stock solutions of trypsinogen and cytochrome c were prepared in water and diluted in the appropriate mobile phase to achieve a concentration of 500 |Xg/mL. Purified PARP and the PARP digest were also diluted in mobile phase to approximately 500 |Xg/mL. Injections of 20 (xL were performed.  62  Table 1. Summary of H P L C conditions used in the development of a method to separate PARP fragments under acidic and non-acidic conditions.  HPLC  Mobile Phase Composition  Column  A  B  Jupiter C 4  20% A C N 0.1% T F A  90% A C N 0.1% T F A  PRP-1  20% A C N 0.1% T F A  90% A C N 0.1% T F A  Jupiter C 4  20% A C N  90% A C N  Jupiter C 4  Jupiter C 4 PRP-1 / Jupiter C 4 PRP-1  PRP-1  20% methanol 200 m M HFIPA  Gradient Parameters  Flow  Time Gradient (min) Type  Rate  0  90% methanol 200 m M HFIPA  20% methanol 800 m M HFIPA  90% methanol 800 m M HFIPA  20% A C N 4 m M NH4HCO3 pH 8.0 20% A C N 0.3% HCOONH4 pH4.3  90% A C N 4 m M NH4HCO3 pH 8.0 90% A C N 0.3% H C O O N H pH4.3  20% A C N 0.3%HCOONH4 pH 3.5  90% A C N 0.3%HCOONH pH 3.5  4  4  100 0 100 100 60 100 100 0 100 100 0 100 100 0 100 100 0 100 100 0 100 100 0 100  1 10 11 1 10 11 1 10 11 1 10 11 1 10 11 1 10 11 1 10 11 1 10 11  (ml/min)  Linear  1  Linear  1  Linear  1  Linear  1  Linear  1  Linear  1  Linear  1  Linear  1  63  3.10.  Mass Spectrometric Analysis of Native PARP and PARP  Fragments  3.10.1. Laser Ablation-Inductively Coupled Plasma Mass Spectrometry (LA-ICPMS) of PARP Bands on Acrylamide Gels Elemental Research Inc. (North Vancouver, B C , Canada) performed all ICPMS analyses. Laser ablation-ICPMS was performed on PARP protein bands isolated on acrylamide gels. Increasing concentrations of purified PARP (0.4 - 3.9 |0,g) were loaded onto a 12% gel and analyzed by SDS-PAGE as described in section 3.6. The destained gel was dried onto a clear cellophane membrane (BioRad, Hercules, C A , USA). The dried gel was submitted to Elemental Research Inc. for LA-ICPMS analysis of the protein bands. Fig. 6 shows a scheme depicting the process of L A - I C P M S of PARP protein bands localized on a polyacrylamide gel. Analysis of the zinc content of each protein band was performed by ablating a 5-jim section of the leading edge of the protein band. The ablated material was then carried by a flow of argon gas into the ICPMS for analysis. The abundance of the four zinc isotopes, Z n , Z n , Z n 64  66  67  and Z n  68  was determined  by performing, at least, duplicate measurements of each protein band. The abundance of the zinc isotopes was also assessed in different areas of the blank gel (n = 9) to provide a background value for zinc in the gel. The data presented for abundance of zinc isotopes in PARP bands represents the background-subtracted values. In order to determine the zinc abundance per amount of protein, an isotopic marker was sought that would be representative of the quantity of protein ablated. The abundance of sulfur isotopes was evaluated as one potential marker. Therefore, in each protein sample, where the leading edge of the protein band was ablated, the abundance of S  33  and S  34  was also determined.  The abundance of the sulfur isotopes was also evaluated in the blank gel. A total of 9 measurements of the blank gel were performed, and at least duplicate measurements of each  64  protein band were made. The data presented for abundance of sulfur isotopes in PARP bands represents the background-subtracted values. 3.10.2. Inductively Coupled Plasma Mass Spectrometry (ICPMS) of PARP in Solution Zinc content of PARP was also estimated directly by ICPMS analysis of a solution of PARP. Dilutions of purified PARP (0, 1, 5, 10, 50, 100 and 200 Ug/mL) were prepared in dialysis buffer (50 m M Tris-HCl pH 8.0, 1 m M DTT, 4 m M MgCl ) and submitted to Elemental 2  Research Inc. for analysis by ICPMS. The most abundant zinc isotope, Z n , was analyzed in 64  each PARP solution. The values for Z n  64  abundance (ppb) do not represent background  subtracted values.  65  Plasma Torch  To ICPMS Laser ontrollerj  TolCP-MS Computer  Computer i  irror  Nd-YAG Laser Lens  Stage Controller  microscope  Udser B e a m P A R P  protein band  Electrophoresis Gel  Fig. 6. Scheme depicting the process of laser ablation - inductively coupled plasma mass spectrometry (LA-ICPMS) of PARP protein bands localized on a polyacrylamide gel by SDSPAGE.  66  3.10.3. Protein Analysis by Matrix Assisted Laser Desorption Ionization Time-of-Flight Mass Spectrometry (MALDI-TOF MS) M A L D I - T O F mass spectrometry of PARP and the caspase-3-digested P A R P fragments was performed on a Bruker Biflex Mass Spectrometer by the Mass Spectrometry Facility in the Department of Chemistry, University of British Columbia (Vancouver, B C , Canada). Samples of purified PARP and caspase-3-digested PARP were desalted by 3 consecutive spins in 3 mL of ice-cold Milli-Q-grade water at 4 C for 15 min at 3,500g, using an Ultrafree-4 centrifugal filter containing a Biomax filter with a 10,000 Da N M W L . Each sample was spun in a centrifuge until the volume of retentate was approximately 100 uL. The desalted P A R P and the caspase-3digested PARP fragments were stored in small aliquots at -80 C until analysis by M A L D I - T O F M S . The protein concentration of each sample was calculated to be 5 - 7 j i M , based on the assumption that 100% of the protein was recovered from the filtration devices. While the membranes have low protein-binding characteristics, it is unlikely that protein recovery was 100%. However, due to the limited amount of both purified P A R P and the caspase-3-digested PARP fragments, a protein assay was not performed. Therefore, the protein concentration was an estimate only. M A L D I - T O F M S analysis of PARP and the PARP fragments was performed in the presence of a matrix mixture composed of a saturated solution of sinapinic acid (approximately 22 mg/mL) prepared in a water/ACN mixture (2:1) containing 0.1% T F A . Bovine serum albumin was added to the PARP solutions as an internal calibration standard. A second analysis of the PARP fragments was carried out in the presence of the same matrix mixture, except that trifluoroacetic acid was removed. For this analysis, bovine insulin and cytochrome c were added  67  as internal calibration standards and the mass range was focused around the expected molecular weight of the small P A R P fragment (i.e. 24 kDa). Although the analysis of the 24 kDa PARP fragment without trifluoroacetic acid present in the matrix mixture was successful, the pH of the matrix solution (pH 3.4) was too low to assess the molecular weight of the PARP fragment with the zinc ions intact. Therefore, an attempt was made to analyze the PARP fragments with a sinapinic acid matrix mixture at a pH above 5. The pH of a saturated solution of sinapinic acid prepared in water/ACN (2:1) was adjusted to 5.8 with 1 M N H O H . This matrix was then used to analyze the PARP fragments by M A L D I - T O F 4  MS. 3.10.4. Protein Analysis by Electrospray Ionization Mass Spectrometry (ESI-MS) Protein molecular weight was determined by electrospray ionization mass spectrometry (ESI-MS) in positive ion mode on a V G Q U A T T R O triple quadrupole mass spectrometer (MicroMass Canada Inc., Pte-Claire, QC, Canada) equipped with an electrospray interface. The instrument was interfaced to an HP Series II 1080 Liquid Chromatograph (Hewlett Packard, Avondale, PA, USA), which delivered samples to the mass spectrometer for flow injection. The H P L C was operated without an in-line chromatography column. Samples were introduced into the ESI source via a 75 urn stainless steel probe. Instrument control and data acquisition were under the operation of the MassLynx Version 2.1 software package (MicroMass Canada Inc.). Multiply charged spectra were deconvoluted to yield a molecular weight estimation using the maximum-entropy (MaxEnt) algorithms supplied with the instrument data system. For the determination of protein molecular weight, calibration of the m/z axis was performed with horse heart myoglobin ( M W = 16951.499, Sigma Chemical Co., Cat. No. M-1882), according to the instrument manufacturer's  directions. The instrument parameters were optimized for the  68  ionization of myoglobin using a solvent system consisting of 15% A C N / 4 m M NH4HCO3 pH 8.0. The general conditions used for positive ion ESI-MS of the model proteins, trypsinogen, cytochrome c, chymotrypsin, as well as the 24 kDa PARP fragment, are listed in Table 2. Most parameters are listed as a range, and therefore, varied depending on the specific conditions of analysis (i.e., organic/aqueous composition of the solvent system). For the analysis, concentrated stock solutions of proteins were prepared in Milli-Q-grade water at a concentration of 1 mg/mL and working stock solutions were prepared by diluting the concentrated stock solution (1:6) with the solvent used for analysis. The solvent systems evaluated for ESI-MS analysis of proteins under acidic and neutral conditions are shown in Table 3. ESI-MS was performed in positive ion mode on a V G Q U A T T R O I mass spectrometer (MicroMass Inc.) by the Mass Spectrometry Facility in the Department of Pharmacy, University of Washington (Seattle, W A , USA), The instrument was interfaced to a Shimadzu H P L C for delivery of samples to the mass spectrometer.  Table 2. The operating conditions for the V G QUATTRO mass spectrometer used for positive ion ESI-MS of proteins.  Parameter  Setting  Flow Rate  30 -50 uL/min  Injection Volume  5-10uL  Cone Voltage  30- 55 V  Source Temperature  100- 130°C  Bath Gas  250 - 450 L/h  Scanning Range  6 0 0 - 1500 tn/z  Electrospray Ionization  Positive Ion Mode  69  Table 3. The solvent systems evaluated for the ESI-MS analysis of proteins under acidic and neutral conditions.  Solvent System Components  Composition  A C N / formic acid  15 - 2 0 % / 0 . 1 %  A C N / acetic acid  20% / 26.5 m M  A C N / propionic acid  20% / 26.5 m M  A C N / octanoic acid  20%/26.5 m M 20% / 8 m M pH 5.0  A C N / CH3COONH4  20% / 8 m M pH 6.7 20% / 1 m M pH 5.0 20% / 1 m M pH 6.7  A C N / triethylamine acetate  15%/4mMpH7.0  A C N / pyridine acetate  15%/50mMpH5.9  A C N / NH4HCO3  1 5 % / 4 m M p H 8.0  A C N / HFIPA A C N / NH4HCO3 / deferoxamine  20% / 26.5 m M 20%/53 m M p H 5 . 1 15% / 4 m M pH 8.0 / 100 u M 1 5 % / 4 m M p H 8.0/ 1 m M  A C N / NH4HCO3 / E D T A  15% / 4 m M pH 8.0 /100 u M 15%/4mMpH8.0/ 1 mM  70  3.11.  Endothelial Cell Injury  Experiments  3.11.1. Preparation of E C for Cell Injury Experimentation To prepare multiwell plates and culture dishes for experimentation, cells were removed from a culture flask containing a confluent monolayer of E C (approximately 2 x 10 cells) using 6  0.05% trypsin-EDTA. The cells were collected by centrifugation (200g for 5 min), washed with PBS, counted, spun again, and resuspended in fresh M199 culture media. Cells were placed into culture at a density of 12,000 cells/well in 96-well plates, and 400,000 cells/well in 6-well plates and 35 mm culture dishes. At these densities, confluence was achieved within 48 h. In preparation for an experiment, growth medium was removed to avoid potential confounding serum antioxidants. Cells were washed with, and experimental treatments conducted in, Hanks balanced salt solution (HBSS, composition in mM: CaCl , 1.3; KC1, 5.0; K H P 0 , 0.3; M g S 0 , 2  2  4  4  0.8; NaCl, 138.0; N a H P 0 , 0.3; D-glucose, 5.6, pH 7.4) without phenol red. The HBSS was 2  4  buffered for ambient (0.5%) C 0 conditions with the addition of 4 m M N a H C 0 and 10 m M 2  3  HEPES (Freshney, 1987). The final volume for experiments performed in the 96-well plate was 300 fiL, and 4 mL for the 35-mm culture dishes or 6-well plates. Periodic assessment of E C viability by microscopic visualization using trypan blue exclusion demonstrated 100% viable E C were routinely being prepared and used for each experiment. For each of the experimental treatments, working stock solutions of H 0 , DIQ, cycloheximide, aurintricarboxylic acid and 2  2  staurosporine were prepared in Milli-Q-grade water and added in small volumes (< 10% of the total experimental volume) to achieve the desired final concentrations. Non-treated control cells received an equivalent volume of vehicle (i.e., Milli-Q-grade water). In cell injury experiments using the 96-well plate format, oxidant concentrations greater than 50 |J,M were pre-determined  71  to be necessary to reproducibly overcome antioxidant defenses and cause significant cell death within a 12-hour period. 3.11.2. Cell Injury Assessment Irreversible E C injury (oncotic and/or apoptotic necrosis) was assessed by monitoring the loss of plasma membrane integrity with one of two plasma membrane-impermeant fluorescent nucleic acid dyes, PI or YO-PRO (Nieminen et al, 1992). The fluorescence was detected and quantified using a Millipore CytoFluor 2350 fluorescence plate scanner adapted to operate with an internal environment of 37°C. Working stock solutions of PI (120 |iM) or YO-PRO (60 uM) were prepared in Milli-Q-grade water and added to each well of the 96-well plate to achieve a final concentration of 4 | i M or 2 | i M , respectively. The fluorescence from YO-PRO or PI was detected using an excitation/emission filter set at 485 ± 22 nm / 530 ± 30 nm or 530 ± 30 nm / 645 ± 50 nm, respectively. Cells were maintained at 37 C and the fluorescence was continuously monitored at 20-min intervals. At 2 h prior to the final scan, digitonin (118 u M final concentration) was added to each well to permeabilize any remaining intact cells, which provided an end-point fluorescence reading of 100% dead cells in each specific well. 3.11.3. Cell Injury Data Analysis The fluorescence data were expressed as percent dead cells at each time-point. The last fluorescence value collected, after addition of digitonin, was recorded as 100% dead E C in each experiment. The initial scan served as the background fluorescence, representing 0% dead cells for each individual well. Therefore, calculations for percent dead cells were made for each well, at every data collection period, according to the formula: [(F - B)/(Fi o - B)] * 100 = % dead t  0  cells, where F is the fluorescence at an individual time-point for a particular well, B is the initial t  background fluorescence of live cells, or time = 0 fluorescence for that well, and Fioo is the  72  maximum fluorescence obtained after addition of digitonin to the well. Each time-course experiment for cell injury was repeated at least three times. Within an experiment, each experimental treatment was conducted in triplicate or quadruplicate, and the values for each experimental group were expressed as averages.  3.12.  Assessment of Caspase Activity in Oxidant-Stressed Endothelial  Cells  E C were prepared in 96-well plates as described above using HBSS buffered for 5.0% C 0  2  conditions with the addition of 24 m M HEPES and 50 m M N a H C 0 (Freshney, 1987). The cells 3  were then treated with the following agents: H 0 (200 uM), H 0 + DIQ (50 uM), and H 0 + 2  2  2  2  2  2  3AB (1 mM). Each treatment was performed, at least, in triplicate and the experiment repeated on 3 separate occasions. The plate was incubated at 37 C under 5% C 0 for the duration of the 2  treatment period. The protocol for measuring caspase-3 activity was performed essentially as described by the manufacturer using the ApoAlert™ CPP-32 (caspase-3) Fluorescent Assay Kit, except that it was modified for use in a 96-well plate with a fluorescence plate scanner. Following the treatment period, the plate was removed from the incubator and the HBSS was aspirated. To each well, 50 uL of cold lysis buffer (supplied by the manufacturer) was added. The plate was placed on ice for 10 min, followed by addition of 50 uE of 2X reaction buffer (supplied by the manufacturer), containing 10 m M DTT, to each well. The assay was started by addition of 5 uL of the fluorogenic caspase-3  substrate,  DEVD-AFC  (50 u M final  concentration). The plate was incubated at 37 C for 4 h to permit generation of the fluorescent product at a signal-to-noise ratio that allowed for adequate comparison between treatment groups. The fluorescence of each well was measured using an excitation/emission filter set of 360 ± 40 nm / 530 ± 30 nm, respectively. Although the D E V D substrate is primarily cleaved by caspase-3, it may be cleaved by other members of the caspase family, and therefore,this assay is  73  a measure of caspase-3-like activity. Caspase-3-like activity is expressed as the percent increase in fluorescence over the non-treated control group. To derive a total caspase-3-like activity for each of the different treatment groups, the area under the caspase-3-like activity versus time curve was calculated by the linear trapezoid method (Gibaldi and Perrier, 1975) over the entire course of an experiment.  3.13.  Immunoblot Analysis of PARP Cleavage by Caspase Confluent E C were prepared in 6-well plates and exposed to the various treatments at 37°C.  After exposure, the supernatant was removed and spun at 200g for 5 min at 4°C to collect any floating cells. Ice cold PBS (4 mL) was added to each well, the cells removed by scraping with a rubber policeman, and collected in a tube following centrifugation. For each treatment group and time-point, the floating and adherent cell pellets from two identical wells were combined. Cells were resuspended at 10 x 10 cells/mL of sample dilution buffer (62.5 m M Tris-HCl pH 6.8, 6 M 6  urea, 10% glycerol, 2% SDS, 0.003% bromophenol blue, 5% (3-ME freshly added). Cells were lysed by sonication on ice for 2 x 10 s (microtips at limit, 40% duty cycle) (Vibra Cell™, Sonics & Materials, Danbury, CT) and incubated for 15 min at 6 5 C before loading (equivalent of 200,000 cells/well) onto a SDS-polyacrylamide minigel (8%). Purified P A R P and/or the purified 89 kDa P A R P cleavage product were loaded onto the gel as standards. The 89 kDa apoptotic P A R P fragment was obtained from the caspase-3 digestion of purified P A R P and the fragment was subsequently purified by HPLC. Electrotransfer of protein onto nitrocellulose membranes (Trans-Blot® Transfer Medium, o  BioRad) was performed at 4 C using a mini-transblot cell apparatus (BioRad) in a buffer containing 25 m M Tris, 192 m M glycine and 20% v/v methanol at 100 V for 1 h. A l l subsequent steps were carried out at room temperature. The blots were blocked in a solution  74  containing I X PBS, pH 7.4, 5% non-fat powdered milk and 0.1% Tween-20 (PBSMT) for 2 h with gentle shaking. Incubation with the primary antibody, anti-PARP C2-10 (PharMingen, Mississauga, ON, Canada), was carried out overnight at a dilution of 1:10,000 in fresh P B S M T containing 1 m M sodium azide. The membrane was washed 1 x 15 min and 2 x 5 min in 0.1% Tween-20 in PBS, pH 7.4. The blots were then incubated with the secondary antibody, antimouse IgG labeled with horseradish peroxidase (Amersham, Buckinghamshire, England), at a dilution of 1:2500 in fresh PBSMT for 2 h, followed by washing in 0.1% Tween-20 in PBS. Chemiluminescent detection of immunoreactivity was performed using the ECL™ Western Blotting Analysis System (Amersham).  3.14.  Fluorescence Microscopy and Morphologic  Assessment  Confluent E C prepared in 35 mm culture dishes were exposed to various treatments and incubated at 37 C. At various times during the experiment, both A O (15 u,M) and PI (5 uM) were added to the cells and further incubated for 20 min prior to visualization. The cells were viewed with a Zeiss Universal R microscope equipped with a standard fluorescein filter set (450 - 490 nm excitation and 510 nm dicroic) and a long band pass emission filter at 520 nm. The P A R P inhibitor, DIQ, did not exhibit auto-fluorescence with these excitation and emission settings. The plasma membrane permeant A O is used as a nuclear stain for live cells, and yields a green fluorescence when it is bound to double-stranded nucleic acids (Zelenin, 1993). In addition, A O in live cells is actively taken up and trapped by acidic organelles, such as lysosomes. Uptake and trapping of A O into the lysosome is dependent upon these organelles maintaining an ATP-dependent proton gradient. When inside the lysosome, A O turns from an initial green to an orange-red fluorescence as it forms dimers when it increases in concentration (Zelenin, 1993). Therefore, the red cytoplasmic fluorescence in live cells is due to lysosomal  75  uptake of A O . PI is a plasma membrane impermeant dye used to demonstrate loss of plasma membrane integrity, a marker of either oncotic or apoptotic necrosis. Following the loss of plasma membrane integrity, PI binds to nucleic acids and yields a red fluorescence. Representative micrographs that depict the typical, specific morphological changes and features from each treatment group, are presented.  3.15.  Assessment  Endothelial  of  Intracellular  Free  Calcium  in  Oxidant-Stressed  Cells  Intracellular free calcium, [Ca ]i, was assessed by the dual-excitation fluorescence ratio 2+  method (Grynkiewicz et al, 1985), using the Ca -sensitive fluorophore, fura-2, with an 2+  Attofluor digital fluorescence microscopy system (Atto Instruments Inc., Carl Zeiss Canada Ltd.). A ratio of fluorescence emissions from fura-2 at excitation wavelengths of 334 nm and 380 nm is related to [Ca ]i and is independent of variation caused by differences in cell 2+  thickness, intracellular dye quantity, or photobleaching (Mattson et al, 1995). Therefore, an increase in the ratio of 334/380 nm is indicative of increased [Ca ]j. Confluent E C on a glass 2+  coverslip were incubated with the cell-permeant fura-2/acetoxymethyl ester (7.5 jiM) in HBSS containing 5% bovine serum albumin for 45 min at room temperature. The acetoxymethyl ester linkage is cleaved by ubiquitous cellular esterases, effectively trapping the charged fura-2 in the cytoplasm.  Free fura-2/acetoxymethyl ester was removed by washing with HBSS and the  coverslip placed in a temperature-controlled imaging chamber. Oxidant stress was initiated by replacement of the HBSS in the chamber with fresh HBSS solution (1 mL) containing 200 | i M H2O2. Using the excitation wavelengths of 334 and 380 nm, fluorescence intensities (at 510 nm) were simultaneously obtained from multiple E C (15-20 cells) in HBSS at 37C.  The raw  intensity data at each excitation wavelength were corrected for background levels before  76  calculation of the ratio. A ratio was obtained every 2 min. The data were expressed as an average ratio of the viable cells at each time point. In this group of calcium studies, E C were considered non-viable when there was a concurrent and continuous decline in the fluorescence emissions at both the 334 and 380 nm excitation wavelengths in conjunction with a rise and subsequent fall in the 334/380 nm ratio. This was indicative of increased plasma membrane permeability resulting in a loss of intracellular fura-2 concurrent with a rise of [Ca ]i. The PARP 2+  inhibitor, DIQ, and the antioxidant, catalase, were added in small volumes to the chamber to achieve desired concentrations of 50 flM and 1000 U/mL, respectively. When DIQ or catalase were added following initiation of oxidant stress, the condition for addition was preset as the time when 5-10% of the cells were observed as non-viable.  3.16.  Data Analysis Where appropriate, the linearity of standard curves was calculated by the method of least  squares and expressed as the coefficient of determination, r . 2  77  4.  4.1.  Recombinant  PARP  RESULTS  Production  Using  the  Baculovirus  Expression  System 4.1.1.  AcPARP Virus Propagation and Viral Titre Determination Each round of AcPARP propagation produced a virus stock that underwent viral titre  determination. The first round of virus propagation produced a virus stock with a titre of 3.2xl0  8  pfu/mL, which was used for subsequent viral propagations. A l l virus stocks subsequently produced had a titre in the range of 10 - 10 pfu/mL. 7  4.1.2.  8  Effect of MOI on Time-Course of PARP Expression The effect of M O I on the time-course of PARP expression in S/9 cells is shown in Fig. 7.  The protein profile from S/9 cells infected with wtAcMNPV at a M O I of 2 is shown in lane 2 of each gel. The absence of a protein band at 116 kDa demonstrated that wtAcMNPV-infected S/9 cells do not produce a protein in this molecular weight range. However, when Sf9 cells were infected with the recombinant virus containing the full-length gene for human PARP, the protein profile was altered and a protein band at 116 kDa became visible approximately 36 h after infection (Fig. 7, lanes 3 - 10). The M W of this protein band corresponds to the apparent M W of PARP. The effect of a M O I of 0.1, 1 and 10 on the time-course of protein expression in AcPARPinfected S/9 cells is shown in Fig. 7A, B , and C, respectively. 5/9 cells infected at a M O I of 0.1 or 1 (A and B) showed a 116 kDa band that increased in intensity up until 72 h post-infection, after which time the band disappeared. The decline in intensity of the protein bands at later timepoints is likely the result of the lytic nature of the infection process. Cells infected at a M O I of  78  10 (Fig. 7C) showed a different time-course of protein expression with the 116 kDa band increasing in intensity until 96 h post-infection. Maximum expression of P A R P occured between 72 - 96 h.  Infection at a high M O I ensures simultaneous infection of all cells and  therefore,synchronizes protein production. Therefore, all subsequent infections were carried out at an M O I of 5 and cells were harvested at 72 h post-infection. A M O I of 5 was selected as a compromise between synchronicity of infection and amount of virus stock required for each infection.  79  T i m e P o s t - I n f e c t i o n (h)  A .  kDa 200 116 97.4 66 45  MW  V  12  2 4 3 6 4 8 7 2 96 120  £2 7  — — —  -  mmm>  MOI=0.1  B.  200 116 97.4 66  __  _  <  .• •  ^ ^ ^ ^  ^U^^UMMMAB^^^^  45  M0I= 1  200 116 97.4 66 45  M 0 I = 10 Lane  1  2  3  4  5  6  7  8  9  10  Fig. 7. Effect of different M O F s on the time-course of protein expression in ^cPARP-infected SJ9 cells. Flasks containing 1x10 cells were infected at a MOI of 0.1 (A.), 1 (B.), or 10 ( C ) . Crude cell extracts were prepared in a lysis buffer (25 m M Tris-HCl pH 8.0, 50 m M glucose, 10 m M EDTA, 1 m M B-ME, 1 m M PMSF) and analyzed by 0.1% SDS -10% PAGE. Protein bands were stained with 0.25% Coomassie Brilliant Blue G250. Lane identification: M W markers (lane 1); crude cell extract from wt4cMNPV-infected S/9 cells harvested 72 h after infection (lane 2); crude cell extracts from ^cPARP-infected Sf9 cells harvested 2, 12, 24, 36,48, 72, 96 and 120 h after infection (lanes 3 - 1 0 ) . s  80  4.2.  Purification of Recombinant  PARP  The purification procedure for recombinant P A R P produced in the baculovirus expression system was a 3 step procedure that involved solubilization of cellular membranes and proteins, D N A precipitation followed by selective protein precipitation using 30%-70% ammonium sulfate, as described by Giner et al. (1992). The partially purified cell extract then underwent final purification by affinity chromatography, which involved binding of P A R P from the crude extract to a specific affinity ligand coupled to a supporting resin. 4.2.1.  Preparation of a Crude Cell Extract In order to monitor the purification procedure, samples at each stage of the purification  were taken for analysis by SDS-PAGE (Fig. 8). Crude cell lysates from individual spinner flasks of AcPARP-infected S/9 cells showed intense protein staining and a strong band at 116 kDa, which corresponded to the apparent M W of P A R P (Fig. 8A., lanes 3 - 6). Following homogenization, the addition of detergents to solubilize membranes and proteins (Fig. 8A., lane 8) did not appear to alter the protein profile. A band at 116 kDa was apparent in the cell pellet from this initial step in the purification (Fig. 8A., lane 9), indicating incomplete solubilization of proteins and loss of P A R P in the first centrifugation. D N A precipitation did not appear to alter the protein profile of the recovered supernatant (Fig. 8B., lane 3). However, the resuspended pellet from this stage in the purification contained many proteins, including a strong band at 116 kDa, suggesting that some proteins were co-precipitating with the D N A (Fig. 8B., lane 4). The reduced protein staining of the supernatant following treatment with 30% ammonium sulfate (Fig. 8B., lane 5) indicated removal of proteins. The final step, protein precipitation with 70% ammonium sulfate, resulted in complete removal of proteins from the supernatant (Fig. 8B., lane 7). The resuspended pellet from this step (Fig. 8B., lane 8) showed a very strong band at 116  81  kDa, with many other proteins present in the extract. This extract was applied to the affinity chromatography column for the final purification step.  82  Fig. 8. Purification of rePARP from AcPARP-infected Sf9 cells. A total of 5.64 x 10 cells were infected with AcPARP at an M O I of 5 and harvested 72 h post-infection. Samples at various stages in the purification were analyzed by 0.1% SDS - 10% P A G E and the gel was stained with 0.25% Coomassie Brilliant Blue G250. Gel A lane identification: blank wells (lanes 1, 10); M W markers (lane 2); crude cell lysate from individual flasks (lanes 3 - 6 ) ; crude cell lysate after homogenization (lane 7); supernatant from cell lysate (lane 8); cell pellet (lane 9). Gel B lane identification: blank wells (lanes 1, 10); M W markers (lane 2); supernatant following D N A precipitation (DNA ppt'n) (lane 3); cell pellet following D N A precipitation (lane 4); supernatant following 30% ( N H ) S 0 precipitation (lane 5); cell pellet following 30% ( N H ) S 0 precipitation (lane 6); supernatant following 70% ( N H ) S 0 precipitation (lane 7); cell pellet following 70% ( N H ) S 0 precipitation (lane 8); purified rePARP 2.5 pig (lane 9). s  4  2  4  4  4  4  2  2  2  4  4  4  83  kDa  Lane  1  2  3  4  5  6  7  8  9  10  4.2.2.  Affinity Chromatography for PARP Purification  4.2.2.1.  Modification of the Elution Buffer Composition  Elution of PARP from the 3AB-Affigel affinity column relies on displacement of the bound PARP by an inhibitor of higher affinity for PARP than 3AB. Burtscher et al. (1986) described the elution of PARP from the 3 AB-affinity column with an elution buffer containing 1 m M 3-methoxybenzamide. Although 3AB and 3-methoxybenzamide are competitive inhibitors, with similar 50% inhibitory concentrations (IC50) of 33 | i M and 17 fiM, respectively (Banasik  et  al, 1992), the elution profile was difficult to monitor with these compounds in the elution buffer because of their inherent absorbance at 280 nm. To reduce the amount of inhibitor required for PARP elution from the affinity column, and therefore,reduce background U V absorbance, an alternate inhibitor was sought. An elution buffer containing DIQ, which has an IC50 of 0.39 | l M (Banasik et al, 1992), was evaluated. Elution of P A R P from the 3AB-affinity column with a buffer containing increasing concentrations of DIQ is shown in Fig. 9. The lowest concentration of DIQ found to elute PARP effectively from the column was 1 (iM. In all subsequent purifications, the elution buffer contained 5 u M DIQ to ensure efficient elution of P A R P from the column. Switching to an elution buffer containing low concentrations of a high affinity inhibitor, such as DIQ, efficiently displaced PARP from column binding sites with less inherent absorbance. In addition, a high affinity inhibitor may also improve the elution profile by eluting P A R P in a narrower volume. PARP binding to the affinity resin appeared to be a dynamic process because continued flow of the loading buffer through the column resulted in a slow continuous elution of PARP even in the absence of DIQ (data not shown). A comparison between the U V absorbance tracing and the protein profile of fractions eluted from the affinity column is presented in Fig. 10. The initial large peak eluting from the  85  affinity column corresponded to contaminant proteins in the extract that did not bind to the affinity column (Fig. 10B., lanes 5 - 8 ) . Shortly after switching to the elution buffer, a large peak was observed that contained P A R P eluted from the column (Fig. 10B., lanes 13 - 15). The increase in baseline following the elution of the P A R P peak indicated the increased background absorbance due to DIQ in the elution buffer. 4.2.2.2.  Direct Application of the Cell Extract to 3AB-Affinity Column  Potential methods for shortening the purification procedure were investigated. Initially, the crude cell extract was applied directly to the column prior to D N A precipitation with protamine sulfate. However, column flow was severely reduced and recovery of P A R P was low. The extract was then prepared to the point of D N A precipitation and then applied to the affinity column. While the shortened procedure appeared to purify P A R P effectively from the cell extract (data not shown), problems were still encountered with reduced column flow rates, likely the result of the heavy protein load applied to the column. It was necessary to regenerate the column by digestion of contaminating D N A with DNase I (0.82 units/|iL), and removal of proteins by continuous washing with a high salt buffer (1.0 M KC1). This procedure adequately regenerated the column. However, as a result of column obstruction with these shortened procedures, it was decided that the full purification procedure would be used.  86  Fig. 9. Concentration of DIQ required to elute PARP from the 3AB-affinity column. A total of 4x10 cells were infected with AcPARP at a MOI of 5 and harvested 72 hours post-infection. The crude cell extract was partially purified and applied to a 3AB affinity chromatography column equilibrated with loading buffer (100 m M Tris-HCl pH 7.4, 17% glycerol, 12 m M BM E , 0.5 m M E D T A , 0.3 M KC1). Following application of sample, the column was washed with loading buffer and fractions corresponding to initial contaminant protein peak were collected. After washing the column with loading buffer, elution of rePARP was carried out by applying the loading buffer containing progressively increasing concentrations of DIQ (0.1, 1.0, 10, 25, 82 uM). Elution fractions from each DIQ concentration were pooled and concentrated. Samples were analyzed by 0.1% SDS - 7.5% P A G E and the gel stained with 0.25% Coomassie Brilliant Blue G250. Lane identification: M W markers (lane 1); pooled fractions from initial contaminant protein peak (lane 2); pooled and concentrated fractions from affinity column eluted with 0.1, 1.0, 10, 25 and 81 u M DIQ in elution buffer (lanes 3 - 7 , respectively). 8  87  1  1  r  F10  F15  F20  T F5  B. 3AB-Affigel Elution Fractions kDa  Lane  MW  1  F5  2  F6  3  F7  F8  F9  F10  4  5  6  7  3AB-Affigel Elution Fractions  F11  F12  F13  8  9  10  MW F14  F15  F16  F17  F18  F19  F20  F21  F22  11  13  14  15  16  17  18  19  20  12  Fig. 10. Purification of rePARP from a partially purified AcPARP-infected SjV cell extract by 3AB-Affigel affinity chromatography. The protein pellet obtained following 70% (NH )2S0 precipitation of the partially purified S/9 cell extract was resuspended in loading buffer (100 m M Tris-HCl pH 7.4, 17% glycerol, 12 m M p-ME, 0.5 m M EDTA, 0.3 M KC1) and applied to the 3AB-Affigel affinity column. Protein elution from the affinity column was measured by U V absorbance at 280 nm (A.). Fractions (5 mL) eluted from the column were collected and analyzed by 0.1 % SDS - 10% P A G E and the gels stained with 0.25% Coomassie Brilliant Blue G25 (B.). Lane identification: M W markers (lane 1 and 11); fractions collected during application of protein extract to column (lanes 2 - 10); fractions collected following elution of rePARP with the elution buffer (loading buffer containing 5 u M DIQ) (lanes 12 - 20). 4  4  88  4.2.3.  Yield of PARP from the Purification Procedure Individual batches of PARP were prepared using the complete purification procedure. The  identity of purified PARP was based on apparent molecular weight on SDS-PAGE, selective binding to the 3AB-Affigel affinity column, and recognition by an anti-PARP antibody in a dotblot analysis (data not shown). The individual batches were combined into a uniform preparation of PARP to be used in all subsequent assays. The protein yield from individual batches ranged from 276 - 679 (ig protein/mL. Following concentration of the combined batches, a total of 28 mL of protein solution was obtained with a concentration of 939 u,g/mL. Therefore, 26.3 mg of PARP were purified from a total of 48.8 x 10 AcPARP-infected SJ9 cells. 8  The yield of P A R P using this purification procedure (approximately 26 mg of PARP from 48.8xl0 5/9 cells) was considerably less than the 20 mg of PARP from 5 x l 0 5/9 cells reported 8  8  by Giner et al. (1992). Potential reasons for this discrepancy are numerous and include subtle differences in the host cells and conditions of culture, as well as differences in the purification procedure. It is evident from Fig. 8 that the PARP protein is present in many of the cell pellets that were discarded. This was a significant site of loss of the protein and suggested that the purification procedure may not have been fully optimized. However, the complete purification procedure, as described, resulted in adequate yield of purified PARP, and therefore,the decision was made not to invest more time in obtaining a greater yield from the expression system.  89  4.3.  PARP Enzyme Assay Development Standard parameters that were optimized during the development of the in vitro enzyme  assay to measure PARP activity included enzyme and substrate amount, as well as incubation time and temperature. However, to assess PARP activity in vitro, the enzyme assay was further characterized with respect to the co-substrates in the assay mixture. In the enzyme assay for PARP, D N A is required for activation of the enzyme, and histones are present as the protein acceptor for ADP-ribose. Limitations in either D N A or histones could result in decreased activity and non-linear response of PARP, and it was therefore, important to optimize the amount of each of these components. 4.3.1.  Effect of Temperature on PARP Activity PARP activity in vitro has been reported to be higher at 25 C than at 37 C (Niedergang et  al, 1979). To establish the optimum incubation temperature for the P A R P assay in our laboratory, the assay was performed under identical reaction conditions, except that the o  o  incubation temperature was either 25 C or 37 C. Fig. 11 shows the effect of incubation o  o  temperature on PARP activity. PARP activity was higher at 37 C than at 25 C over the 2 min incubation period. Therefore, 37 C was selected as the incubation temperature for the standard assay procedure. 4.3.2.  Effect of DNA Sonication Time on PARP Activity Because D N A strand breaks stimulate PARP activity, the extent of D N A strand breakage  required to maximally activate P A R P was determined. Calf thymus D N A was dissolved in water and strand breaks were introduced by sonication. The effect of changes in D N A sonication time on PARP activity is shown in Fig. 12. PARP activity increased sharply upon sonication of D N A and reached a plateau between 10 and 30 s of sonication. PARP activity was observed when non90  sonicated D N A was present indicating that either the non-sonicated D N A may contain some double or single strand breaks, or that PARP has basal activity in the absence of activation by D N A . Maximal stimulation of PARP was achieved with a sonication time of 30 s. In all further P A R P assay studies, D N A sonicated for 30 s was used.  91  Fig. 11. Effect of incubation temperature on rePARP enzyme activity. Purified rePARP (1.0 pg) activity was assayed for 2 min at 25 C or 37 C in a reaction mixture containing 100 m M TrisH C l pH 8.0, 10 m M M g C l , 5 m M DTT, 10 pg sonicated D N A , 10 pg histone HI and 100 p M N A D (1.25 uCi C - N A D / p m o l N A D ) in a final volume of 100 pL. The reaction was terminated by addition of 10% T C A and the acid-insoluble radioactivity determined. Values for PARP activity represent mean ± SD of triplicate determinations in a single experiment. 2  +  14  +  +  92  1.25  100  DNA Sonication Time (s) Fig. 12. Effect of duration of D N A sonication on rePARP activity. Aliquots of calf thymus D N A (1000 u.g/mL), prepared in water, were sonicated on ice. Purified rePARP (1.0 |ig) activity was assayed for 4 min at 37 C in a reaction mixture containing 100 m M Tris-HCl pH 8.0, 10 m M M g C l , 5 m M DTT, 10 jxg sonicated D N A , 10 |ig histone H I and 200 u M N A D (1.25 uCi C NAD /jxmol N A D ) in a final volume of 100 | l L . The reaction was terminated by addition of 20% T C A and the acid-insoluble radioactivity determined. Values for P A R P activity represent mean ± SD of triplicate determinations in a single experiment. +  1 4  2  +  +  93  4.3.3.  Effect of the Presence or Absence of Histones on PARP Activity To determine the effect of histones on PARP activity, the enzyme assay was performed  under identical reaction conditions in the presence or absence of histones. Elimination of histones from the reaction mixture resulted in a 75% decrease in P A R P activity, as seen in Fig. 13. In the absence of an alternate protein acceptor, PARP ADP-ribosylates itself. Therefore, this finding illustrates the ability of PARP to ADP-ribosylate itself, thereby decreasing its own activity.  0.6  c 'E  0.5H  o  0.4H  > CoL o c  O < Q.  _c <D  OC -Q < Q_  n  o  0.3 H  Q.  *i_ i  0.2 H  Q_ Q < "o E c  o.H 0.0-  +HIS  -HIS  Fig. 13. Effect of the presence or absence of histones on PARP activity. Samples containing 0.5 pg P A R P were incubated for 4 min at 37°C in a reaction buffer containing 100 m M Tris-HCl pH 8, 10 m M M g C l , 5 m M DTT, 5 pg sonicated D N A , 5 pg histone H I and 90 p M N A D (1.26 pCi C - N A D / p m o l N A D ) in a final volume of 100 pL. The reaction was terminated by addition of 10% T C A and the acid-insoluble radioactivity was determined. Values for PARP activity represent mean ± SD of triplicate determinations in a single representative experiment. +  2  14  +  +  94  4.3.4.  Effect of Different Preparations of Histones on PARP Activity PARP activity was assessed in side-by-side experiments using either Gibco histone H I or  Worthington lysine-rich histone subgroup (Fig. 14). P A R P activity was linear over the 2 min incubation when Gibco histone H I was used. However, when the Worthington lysine-rich histone subgroup was used, PARP activity was non-linear over the 2 min incubation period. Therefore, Gibco histone H I was used in all subsequent assays. 4.3.5.  Effect of Histone Concentration on PARP Activity To determine the effect of varying the histone concentration on PARP activity, the enzyme  assay was performed under identical conditions except that the amount of histones was varied in relation to the D N A (Fig. 15). Maximal PARP activation was obtained when the amount of histones was equivalent to the amount of D N A in the reaction mixture (i.e., 10 |ig each of D N A and histones). When the amount of histones was either greater (50 pig) or less (1 p,g) than the amount of D N A (10 u.g), PARP activity was diminished at all time-points measured. This pattern of enzyme activity in relation to the amount of D N A and histones was the same regardless of the amount of PARP present (i.e., 0.7 u.g (Fig. 15), and 1.36 or 2.1 (ig of PARP (data not shown)). Therefore, an equal quantity of both D N A and histones provided the conditions for optimal PARP activity.  95  A.  5n  Incubation Time (min) Fig. 14. Comparison of Gibco histone H I (A.) and Worthington lysine-rich histones (B.) on PARP activity. Samples containing 0.5 pg PARP were incubated at 37°C in a reaction buffer containing 100 m M Tris-HCl pH 8, 10 m M M g C l , 5 m M DTT, 10 pg sonicated D N A , 10 pg histones and 500 p M N A D (1.00 p C i C-NAD /pmol N A D ) in a final volume of 100 pL. The reaction was terminated by addition of 10% T C A and the acid-insoluble radioactivity was determined. Values for PARP activity represent mean ± S E M of duplicate determinations in 3 independent experiments. 2  +  14  +  +  96  2.0i T3 CD ft  1.5-  -•s o 4->  —  " <  CD CO  Q..8 DC  1.0-  v  < o L  CL Q  <oE 0.5c  0.00.0  2.5  5.0 7.5 10.0 12.5 Incubation Time (min.)  15.0  17.5  Fig. 15. Effect of histone amount on PARP activity. Purified P A R P (0.7 |lg) activity was assayed at 37 C in the presence of 1, 10 or 50 fig of histone H I . The reaction mixture consisted of 100 m M Tris-HCl pH 8.0, 10 m M M g C l , 5 m M DTT, 10 u.g sonicated D N A and 200 u M N A D (1.25 u £ i C-NAD /(imol N A D ) in a final volume of 100 uL, The reaction was terminated by addition of 20% T C A and the acid-insoluble radioactivity determined. Values for P A R P activity represent the mean of duplicate determinations in a single experiment. +  2  14  +  +  97  4.3.6.  Effect of the Presence or Absence of DNA on PARP Activity To determine if D N A was required for P A R P activity, the enzyme assay was performed  under identical reaction conditions in the presence or absence of sonicated D N A . Nearly complete loss of P A R P activity was found when D N A was eliminated from the reaction mixture (Fig. 16).  1.00-1  & CO  0.75H  i_  o CL o > oc c 1_  o <  CD  o. 8 f  < CL  ol  O CL  0.50H  'E  Q <  0.25H  o E c 0.00-^  ^  -  -DNA  +DNA  Fig. 16. Effect of the presence or absence of D N A on P A R P activity. Samples containing 1.0 pg P A R P were incubated for 4 min at 37°C in a reaction buffer containing 100 m M Tris-HCl pH 8, 10 m M M g C l , 5 m M DTT, 10 pg sonicated D N A , 10 pg histone H I and 200 p M N A D (1.25 pCi C - N A D / p m o l N A D ) in a final volume of 100 pL. The reaction was terminated by addition of 10% T C A and the acid-insoluble radioactivity was determined. Values for P A R P activity represent mean ± SD of triplicate determinations in a single representative experiment. +  2  14  +  +  98  4.3.7.  Effect of the Amount of Histones and DNA on PARP Activity The effect of increasing the amount of both D N A and histones on P A R P activity is seen in  Fig. 17. PARP activity in the first 2 min of incubation was the same for tubes containing 5, 10 or 43 |ig of both D N A and histone H I . After longer incubations, P A R P activity in the samples containing 43 fig of both D N A and histones declined compared to samples containing less D N A and histones. When the levels of D N A and histones were both increased to 86 u,g, the activity of PARP was lower at all time-points tested compared to the other groups. The lower activity with 43 and 86 fig of D N A and histones was most likely due to the formation of a cloudy precipitate, which occurred when these amounts of D N A and histones were mixed. PARP activity was highest when either 5 or 10 (ig of both D N A and histones were present in the reaction mixture. Therefore, either 5 or 10 |ig of both D N A and histones were used in all subsequent assays.  99  DNA / Histones  •  c  5ug/5ug  * 10 ug/10 ug  CD  o  v 43|ig/43ug =5 >. £ .*=: eg > o  a  86 ug/ 86 ug  < 8 CL . £ DC CD  i CL Q <  o E  6  9  Incubation Time (min) Fig. 17. Effect of various amounts of histone H I and D N A on P A R P activity. Purified PARP (1.0 |lg) activity was assayed at 37 C in the presence of 5, 10, 43 or 86 pg of both histone H I and sonicated D N A . The reaction mixture consisted of 100 m M Tris-HCl pH 8.0, 10 m M MgCl2, 5 m M DTT, and 100 p M N A D (1.25 p C i C - N A D / p m o l N A D ) in a final volume of 100 pL. The reaction was terminated by addition of 10% T C A and the acid-insoluble radioactivity determined. Values for P A R P activity represent the mean ± S E M of duplicate determinations in 3 independent experiments. +  14  +  +  100  4.3.8.  Effect of Increasing Amounts of PARP on PARP Activity The graph of PARP activity versus increasing amounts of P A R P protein is shown in Fig.  18. PARP activity was linear over the range of 0.1 - 0.8 |lg of P A R P with a r value of 0.997. 2  Based on these results, 0.5 (0,g of PARP was selected as the amount of PARP to be used in the standard enzyme assay.  1.6-1  ) W  0.0  (  0.2  (  !  1  0.4  0.6  0.8  Amount of PARP (pig) Fig. 18. Effect of increasing PARP amount on PARP activity. Samples were incubated for 2 min at 37°C in a reaction buffer containing 100 m M Tris-HCl pH 8.0, 10 m M M g C l , 5 m M DTT, 10 (Xg sonicated D N A , 10 \ig histone HI and 500 u M N A D (1.00 jxCi C - N A D / u m o l N A D ) in a final volume of 100 |0,L. The reaction was terminated by addition of 10% T C A and the acidinsoluble radioactivity determined. Values for PARP activity represent the mean ± S E M of duplicate determinations in 3 independent experiments. 2  +  14  +  +  101  4.3.9.  The Kinetics of PARP Activity The kinetics of PARP activity over the period of 0 - 3 min is shown in Fig. 19. PARP  activity was linear over this time period with a r value of 0.941. A n incubation time of 2 min 2  was selected for all subsequent assays.  4n  0.0  0.5  1.0  1.5  2.0  2.5  3.0  3.5  Incubation Time (min) Fig. 19. Kinetics of P A R P activity. Samples containing 0.5 pg PARP were incubated at 37°C in a reaction buffer containing 100 m M Tris-HCl pH 8.0, 10 m M M g C l , 5 m M DTT, 10 pg sonicated D N A , 10 pg histone H I and 500 p M N A D (1.00 p C i C - N A D / p m o l N A D ) in a final volume of 100 pL. The reaction was terminated by addition of 10% T C A and the acidinsoluble radioactivity determined. Values for PARP activity represent the mean ± S E M of duplicate determinations in three independent experiments. 2  +  14  +  +  102  4.3.10. Substrate Saturation Curve for PARP Activity A saturation curve for the PARP substrateO, N A D , was constructed over the concentration +  range of 0 - 1000 JJM (Fig. 20). The saturation curve appeared to plateau between 500 and 1000 (IM indicating that saturation of the enzyme was achieved over this range of N A D  +  concentrations. Therefore, the 500 | i M N A D concentration was considered a saturating amount +  of substrate and was used for the standard PARP enzyme assay. 4.3.11. Standardized Assay for PARP Activity The standardized assay conditions for the PARP enzyme activity assay are shown in Table 4.  Table 4. Standardized assay components and incubation time for the PARP enzyme assay. Assay Components  Concentration  PARP  0.5 |ig  NAD  500 u M  +  [ C-NAD ]  1.0u€i/umol N A D  Sonicated D N A  10 Jig  Histone HI  10|ig  Tris-HCl pH 8.0  100 m M  MgCl  10 m M  14  +  2  DTT Incubation Time  +  5 mM 2  min  103  Fig. 20. Substrate saturation curve for N A D in the PARP enzyme assay. Samples containing 0.5 pg PARP were incubated for 2 min at 37 °C in a reaction buffer containing 100 m M Tris-HCl pH 8, 10 m M M g C l , 5 m M DTT, 10 pg sonicated D N A , 10 pg histone H I in a final volume of 100 pL. N A D concentration range was from 50 - 1000 m M , with the specific activity of each concentration held constant at 1.25 p C i C-NAD /mmol N A D . The reaction was terminated by addition of 10% T C A and the acid-insoluble radioactivity was determined. Values for PARP activity represent mean ± S E M of duplicate determinations in at least 3 independent experiments. +  2  +  14  +  +  104  4.4.  Inhibition of PARP Activity  4.4.1.  Reaction Product Inhibition by Nicotinamide Inhibition of PARP activity may occur if the nicotinamide released upon cleavage of the  substrate, N A D , reaches inhibitory concentrations. Inhibition of PARP activity by nicotinamide +  was examined and the results are presented in Fig. 21. The I C  50  for nicotinamide, calculated  under these reaction conditions, is 174 | i M . Uninhibited P A R P released 1.50 nmol nicotinamide during the 2 min incubation period in a parallel experiment, which corresponds to 15 |i,M nicotinamide in the reaction mixture. This is well below any concentration where significant inhibition would occur. Therefore, product inhibition of P A R P is unlikely to occur over the 2 min incubation period and under the.reaction conditions described. 4.4.2.  Inhibition of PARP Activity with Known and Novel Inhibitors PARP activity over a range of inhibitor concentrations was examined for the known PARP  inhibitors, DIQ and 3AB, as well as a novel PARP inhibitor, allopurinol (Fig. 22). The I C  50  values for DIQ and 3AB were estimated to be 0.51 and 23.2 | i M , respectively, under the reaction conditions used. Therefore, in agreement with published reports (Banasik et al, 1992), DIQ was shown to have a higher potency than 3 A B as a PARP inhibitor. Allopurinol was tested for its ability to inhibit PARP. Although allopurinol is a less potent inhibitor of PARP activity compared to DIQ and 3AB, an inhibition curve was constructed over the concentration range of 1 - 3500 fiM, and the IC50 for allopurinol was estimated to be 713 | 1 M . Because allopurinol was initially dissolved in a small amount of DMSO, a control sample was run in the presence of an equivalent volume of D M S O to determine the effect on PARP activity. PARP activity in the presence of 2 |0,L of D M S O was no different than PARP activity in the absence of D M S O (data not shown). 105  Fig. 21. Inhibition of PARP activity by nicotinamide. Samples containing 1.0 pg P A R P were incubated for 2 min at 37 C in a reaction buffer containing 100 m M Tris-HCl pH 8.0, 10 m M M g C l , 5 m M DTT, 5 pg sonicated D N A , 5 pg histone H I and 100 p M N A D (1.25 p C i C NAD /pmol N A D ) in a final volume of 100 pL. The reaction was terminated by addition of 10% T C A and the acid-insoluble radioactivity determined. Values for P A R P activity represent the mean of duplicate determinations in a single experiment. +  1 4  2  +  +  106  1.00n  0.75H  To c  > O < OL  0.50H  'CD CD CO O  £  •s CL CD  .E ri  0.25-  0.00H  -0.25  -  J  3  -  2  -  1  0  1  2  Log Inhibitor Concentration (uM) Fig. 22. Inhibition of PARP activity by the known PARP inhibitors, DIQ and 3 A B , and allopurinol (AP). Samples containing 1.0 pg PARP were incubated in the presence of various concentrations of DIQ, 3AB or A P for 2 min at 37°C in a reaction buffer containing 100 m M Tris-HCl pH 8.0, 10 m M M g C l , 5 m M DTT, 10 pg sonicated D N A , 10 pg histone H I and 100 p M N A D (1.25 pCi C - N A D / p m o l N A D ) in a final volume of 100 pL. PARP inhibitors were prepared in H 0 . A stock solution of A P was prepared in D M S O and diluted in H 0 . The maximum concentration of D M S O was 2%. The reaction was terminated by addition of 10% T C A and the acid-insoluble radioactivity determined. Values for PARP activity represent the mean ± SD of triplicate determinations in a single experiment. 2  +  14  2  +  +  2  107  4.5.  Proteolytic Digestion of PARP Due to the large size of intact PARP and the size limitations for analysis of proteins by  electrospray ionization mass spectrometry, a specific and reproducible fragmentation method for digestion of PARP was required to produce a lower molecular weight fragment that contained the DNA-binding domain. 4.5.1.  PARP Digestion by Trypsin Fragments expected following digestion of PARP with trypsin are 66 and 54 kDa initially,  then 41, 36, 29, and 17 kDa after prolonged digestion, as shown in Fig. 23. The 66 and 29 kDa fragments contain the D N A binding domain. Digestion of purified PARP was monitored over time by SDS-PAGE using different ratios of trypsin:PARP (1:10, 1:20, 1:25, 1:50, 1:100, 1:200, 1:250, 1:500, 1:1000). Digestion at a trypsin:PARP ratio of 1:500 gave the best fragmentation profile in a relatively short period of time. The time-course of PARP digestion with trypsin is shown in Fig. 23B. Within 30 s of digestion, there were many protein bands apparent in the digest (Fig. 23, lane 2), including the full-length PARP at 116 kDa. By 30 - 60 min of digestion, there were 4 main protein fragments at 41, 36, 29 and 17 kDa and several minor protein bands present (Fig. 23, lane 5). Samples from later time-points showed that some of the bands lose intensity following prolonged digestion, particularly the 36 and 29 kDa bands (Fig. 23, lane 7 and 8). The protein digest obtained by trypsin digestion of purified PARP produced a complex mixture of fragments that would likely be difficult to separate for further analysis by mass spectrometry. Therefore, to reduce the complexity of the PARP digestion mixture, a protease was sought that had fewer potential cleavage sites on PARP.  108  4.5.2.  PARP Digestion by Caspase-3 Caspase-3 cleaves PARP at a single unique site,  DEVD214-G215  (human numbering), to  produce a 24 kDa fragment containing the DNA-binding domain and an 89 kDa fragment containing the automodification and catalytic domains (Duriez and Shah, 1997). The time-course of PARP digestion with caspase-3 at a PARP:caspase-3 ratio of 1000:1 is shown in Fig. 24. Within 5 min of digestion, 3 protein bands of approximately 116, 89 and 24 kDa were evident in the digest. The 24 and 89 kDa bands increased in size over time, concomitant with a decrease in the size of the 116 kDa band. By 120 min, the digest contained only 2 bands of 24 and 89 kDa, indicating complete digestion of the full-length PARP. Caspase-3 digestion of purified PARP was adopted as the standard method to produce smaller fragments that were more amenable to mass spectrometry.  109  A.  P A R P f r a g m e n t a t i o n pattern f o l l o w i n g trypsin d i g e s t i o n . *  17 k D a  B_  Trypsin-Digested PARP (min) PARP  kDa  MW  0.5  1  2  15  30  60  90  120  150  T R P 2.6 ug  7  8  9  200  Lane  3  4  5  6  10  Fig. 23. Time-course of PARP digestion with trypsin. The expected fragmentation pattern for PARP following digestion with trypsin is shown in (A.). The starred fragments contain the DNA-binding domain. Purified PARP (98 pg) was digested for 150 min at 22°C with trypsin (TRP) at a TRP:PARP ratio of 1:500. Samples of the digest were removed at various times (B., lanes 2 - 8 ) , placed on ice and an equal volume of sample dilution buffer was added for analysis by 10% SDS-PAGE. The protein bands, equivalent to 6.3 pg of digested PARP loaded into each well, were stained with Coomassie Brilliant Blue G250. Trypsin (lane 9), and purified PARP (2.6 pg, lane 10) were run as standards.  110  Purified PARP kDa  Lane  MW  2.35pg  1  2  Caspase-3 Digested PARP 5  3  15  30  60  90  120  4  5  6  7  8  (min)  Fig. 24. Time-course of P A R P digestion by caspase-3. Purified PARP (47 pg) in dialysis buffer (50 m M Tris-HCl pH 8.0, 1 m M DTT, 4 m M MgCl ) was digested for 2 h at 37°C with caspase-3 at a PARP:caspase-3 ratio of 1000:1. Samples were removed after 5, 15, 30, 60, 90 and 120 min of incubation (lanes 3 - 8 ) for analysis by SDS-PAGE and protein bands (equivalent to 2.9 pg PARP starting material) were stained with Coomassie Blue. Lane 1, molecular weight markers; lane 2, purified PARP, 2.35 pg. 2  Ill  4.6.  Separation of Caspase-3-Digested PARP  Fragments  As a component of developing a mass spectrometric technique to examine metal content in the zinc fingers of PARP, a method was required to obtain the purified 24 kDa PARP fragment containing the zinc fingers. A number of different approaches were taken to separate the caspase-3-digested  PARP  fragments,  including  centrifugal  ultrafiltration,  affinity  chromatography, and HPLC. 4.6.1.  Centrifugal Ultrafiltration of PARP Fragments  4.6.1.1.  Centrifugal Ultrafiltration Under Non-Denaturing Conditions  A centrifugal filter device with a 50,000 nominal molecular weight limit (NMWL) was used to separate fragments based on differences in molecular weight. Ultrafiltration was carried out under non-denaturing conditions in the presence of water. Samples of purified PARP, caspase-3, the PARP digest before centrifugal ultrafiltration, as well as the retentate and filtrate obtained following centrifugal ultrafiltration were analyzed by SDS-PAGE, and the protein profile of each sample is shown in Fig. 25. Purified PARP, a single protein band at 116 kDa, is seen in lane 2. Caspase-3 is a 29 kDa protein composed of 12 kDa and 17 kDa subunits. Under the denaturing conditions of the gel, the subunits should be detected individually (Fig. 25, lane 3). However, the lack of protein bands in this lane suggested that the amount of protein loaded on the gel is below the limit of detection for Coomassie Blue staining. The PARP digest after a 3-h incubation with caspase-3 is shown in lane 4. There are two strong bands present at approximately 89 and 24 kDa, which indicated PARP digestion. The faint band present at 116 kDa, suggested that a small amount of PARP remained undigested, even after 3 h of incubation. A sample of retentate from the centrifugal filtration device is shown in lane 5. Two large protein bands are seen at approximately 89 and 24 kDa, as well as a number of fainter protein bands in  112  the range of 40 - 116 kDa. The undigested PARP at 116 kDa, as well as the other contaminating proteins, are more visible in this protein preparation due to the 5-fold concentration of the sample during ultrafiltration. No protein bands are visible in the concentrated sample of filtrate from the centrifugal filtration device (lane 6). Therefore, separation of the 24 and 89 kDa digestion fragments of PARP was not accomplished by centrifugal ultrafiltration in the presence of water. 4.6.1.2. The  Centrifugal Ultrafiltration Under Denaturing Conditions inability to separate the caspase-3-digested  PARP  fragments  by centrifugal  ultrafiltration under non-denaturing conditions suggested that the fragments may be aggregating or undergoing a non-specific association that prevented their separation. Therefore, ultrafiltration under denaturing conditions, using 8 M urea, was attempted with membranes of 2 different nominal molecular weight limits, 50,000 and 100,000 N M W L . Samples of the retentate and filtrate from each ultrafiltration device were analyzed by SDS-PAGE, and the results are shown in Fig. 26. The retentate from both the 50,000 N M W L and 100,000 N M W L filters (lanes 3 and 5) contained protein bands at 24 and 89 kDa, whereas no protein bands were present in the filtrate from each filter. Therefore, separation of the PARP digestion fragments was not achieved by centrifugal ultrafiltration under denaturing conditions using either a 50,000 or 100,000 NMWL.  113  50 000 NMWL  Fig. 25. Separation of PARP fragments by centrifugal ultrafiltration. Caspase-3-digested PARP (1.2 mg) was placed in a Millipore Ultrafree-4 centrifugal filter unit with a 50,000 N M W L . Centrifugal ultrafiltration was carried out in the presence of Milli-Q-grade water at 4 C. The filtrate from 3 successive spins in Milli-Q-grade water was collected. The retentate and filtrate were desalted and concentrated using Ultrafree-4 centrifugal filter devices (10,000 N M W L ) . For SDS-PAGE analysis, a 20 | l L sample of caspase-3-digested PARP, the retentate and filtrate from each filter device was removed and run on a 10% gel, followed by Coomassie Blue staining of protein bands. Lane 1; M W markers; lane 2, purified PARP (4.7 (Xg); lane 3, caspase-3; lane 4, caspase-3-digested PARP; lane 5, retentate from 50,000 N M W L filtration device; lane 6, filtrate from 50,000 N M W L filtration device.  114  kDa  Purified PARP MW 4.7 pg 9*  100 000 NMWL A0  50 000 NMWL A©  200 89 kDa  24kDa  Lane  1  Fig. 26. Separation of PARP fragments by centrifugal ultrafiltration in the presence of 8 M urea. Aliquots of caspase-3-digested PARP (497 pg) were placed in Millipore Ultrafree-4 centrifugal filter units with either a 50,000 or 100,000 N M W L . Centrifugal ultrafiltration was carried out in the presence of 8 M urea at 4 C. The filtrate from 3 successive spins in 8 M urea was collected. The retentate and filtrate were desalted and concentrated using Ultrafree-4 centrifugal filter devices (10,000 N M W L ) . For SDS-PAGE analysis, a 10 pL sample of retentate and filtrate from each filter device was removed and run on a 10% gel, followed by Coomassie Blue staining of protein bands. Lane 1, M W markers; lane 2, purified PARP 4.7 pg; lane 3, retentate from 100,000 N M W L filtration device; lane 4, filtrate from 100,000 N M W L filtration device; lane 5, retentate from 50,000 N M W L filtration device; lane 6, filtrate from 50,000 N M W L filtration device.  115  4.6.2.  Affinity Chromatography for Separation of PARP Fragments It was hypothesized that separation of the 24 kDa PARP fragment, containing the D N A -  binding domain, from the 89 kDa PARP fragment, containing the substrate binding site, could be accomplished by the differential binding of the PARP fragments to 3AB-Affigel. SDS-PAGE analysis was used to detect changes in the protein profile following interaction of the PARP fragments with the affinity resin. Initial separations were performed in batch method examining the effect of increasing KC1 concentration and quantity of affinity matrix used (Fig. 27). The KC1 concentration was varied from 0 - 600 m M to determine the concentration that allowed binding of the large fragment to the 3AB ligand and prevented aggregation of the small and large fragments. Fig. 27 shows protein bands recovered in the supernatant from separations using either 200 (iL (A.) or 500 |J.L (B.) of 3AB-Affigel under increasing KC1 concentrations. When the 89 and 24 kDa bands from each sample (lanes 4 - 9 ) were compared with the no-gel control (lane 3), it was apparent that the intensity of the bands did not change either with increasing KC1 concentration or quantity of Affigel used. This unexpected result was verified in a second experiment where the freshly coupled 3AB-Affigel  (lot #1) was compared to an old batch of  3AB-Affigel (lot #2) that had previously been demonstrated to bind PARP. Both lots of 3ABAffigel were reassessed for their ability to bind the 89 kDa fragment (Fig. 28). The intensity of both P A R P fragments was similar to the no-gel control, which indicated the inability of either batch of Affigel to selectively bind the large fragment. However, both batches of 3AB-Affigel were able to bind full-length PARP (Fig. 29). This result suggested that the inability to bind the 89 kDa fragment was not due to poor coupling efficiency of the 3AB ligand to the affinity matrix.  116  A. 2 0 0 ul 3 A B Affigel KCI Concentration (mM) kDa  No MW P A R P Gel  0  100 200 300 400 600 MW  200 116 97  89 k D a  66  45  31  24 k D a  21  B. 500 u13AB Affigel 200 89 k D a  mm  Lane  1  2  3  24 k D a  10  Fig. 27. Differential binding affinity of 3AB-Affigel for the 89 kDa PARP fragment under different KCI concentrations. Aliquots of either 200 \iL (A.) or 500 \iL (B.) 3AB-Affigel gel suspension were incubated in loading buffer (100 m M Tris-HCl pH 8.0, 10% glycerol, 0.5 m M E D T A and 12 m M (J-ME) containing 100 uL of PARP digest (equivalent to 86 (ig purified PARP starting material) in a total volume of 1.6 mL. The KCI concentration ranged from 0 - 600 mM. Each sample was rotated overnight at 4 C. For SDS-PAGE analysis, a 20 [iL sample of supernatant from each tube was analyzed on a 10% gel and protein bands stained with Coomassie Blue. Lane 1 and 10, M W markers; lane 2, purified PARP 2.4 fig; lane 3, no-gel control containing the PARP digest in 0 m M KCI; lanes 4 - 9 , supernatant from 3AB-Affigel incubations with the PARP digest under various concentrations of KCI.  117  kDa MW  PARP 2.4 p g  0 ^ °  89 kDa  24 kDa  Lane  1  Fig. 28. Binding affinity of two different preparations of 3AB-Affigel for digested PARP. 3 A B Affigel (0.5 mL of gel suspension) was incubated in loading buffer (100 m M Tris-HCl pH 8.0, 10% glycerol, 0.5 m M EDTA, 300 m M KC1 and 12 m M B-ME) containing 100 p L of PARP digest (equivalent to 86 pg PARP starting material) in a total volume of 1.6 mL. Each sample was rotated overnight at 4°C. For SDS-PAGE analysis, a 20 pL sample of supernatant from each tube was removed and analyzed on a 10% gel, protein bands stained with Coomassie Blue. Lane 1, M W markers; lane 2, purified PARP 2.4 pg; lane 3, supernatant from no-gel control sample containing the PARP digest in 0 m M KC1; lane 4, supernatant from Lot #1 3AB-Affigel; lane 5, supernatant from Lot #2 3AB-Affigel.  118  PARP No kDa MW 2.4 pg Gel 200  3AB Affigel Lot#1 (ml) 0.5 1.0 1.5  3AB Affigel Lot#2(ml) 0.5 1.0 1.5  Lane  Fig. 29. Binding affinity of two different preparations of 3AB-Affigel for native PARP. 3 A B Affigel (0.5, 1.0 or 1.5 mL of gel suspension) was incubated in a loading buffer (100 m M TrisH C l pH 8.0, 10% glycerol, 0.5 m M EDTA, 300 m M KC1 and 12 m M B-ME) containing 94 pg purified PARP in a total volume of 1.6 mL. Each sample was rotated overnight at 4°C. For SDSP A G E analysis, a 20 p L sample of supernatant from each tube was removed and analyzed on a 10% gel, followed by Coomassie Blue staining of protein bands. Lane 1, M W markers; lane 2, purified PARP 2.4 pg; lane 3, no-gel control sample containing PARP in 0 m M KC1; lanes 4-6, supernatant from samples containing Lot #1 3AB-Affigel ; lanes 7-9, supernatant from samples containing Lot #2 3AB-Affigel.  119  4.6.3.  Separation of Caspase-3-Digested PARP Fragments by HPLC  4.6.3.1.  HPLC of Basic Proteins Under Acidic Conditions  Because neither centrifugal ultrafiltration nor affinity chromatography was successful as a separation method for the caspase-3-digested P A R P fragments, the decision was made to separate the fragments by H P L C under acidic conditions. Although this meant that the apo- form of the 24 kDa P A R P fragment would be produced, it provided a method to obtain the fragment for evaluation of the ESI-MS method that was being developed. Protein elution from either a silica-based reverse phase column (Jupiter C 4 ) or a polymeric reverse phase column (PRP-1) was readily achieved under acidic conditions using a mobile phase consisting of A C N and 0.1% T F A (Table 5). Trypsinogen and intact P A R P eluted as single peaks (chromatograms  not shown), whereas  cytochrome c eluted as 2 peaks  (chromatogram not shown), suggesting potential degradation or contamination of the sample. The retention times of cytochrome c, trypsinogen and P A R P seem to reflect the differences in their molecular weights (cytochrome c, = 12 kDa; trypsinogen, = 24 kDa; P A R P =113 kDa), with the smaller proteins eluting sooner than the larger proteins. The chromatogram showing the elution profile of the caspase-3-digested P A R P fragments is shown in Fig. 30A. The two peaks that eluted at 8.3 and 11 min were collected separately. The initial peak at 4.5 min was also present in the blank injection and was determined to be a component of the protein digestion buffer, likely the DTT and/or Tris. The identity of peaks 2 and 3 was confirmed by SDS-PAGE as the 24 and 89 kDa fragments, respectively (Fig. 30B, lanes 3 and 4). While peak 2 appeared as a single band at 24 kDa on SDS-PAGE, peak 3 appeared as a large band at 89 kDa, as well as many other less intense bands, including one at 116 kDa, which represents intact PARP.  120  Table 5. Summary of results from the development of an H P L C assay to separate caspase-3digested P A R P fragments under acidic and neutral conditions. Specific information regarding mobile phase composition and gradient elution profiles is found in Materials and Methods section 3.9.4  Analysis Condition  Mobile Phase Components  HPLC Column  Jupiter C 4  ACN/ 0.1% T F A Acidic  PRP-1  ACN/ HCOONH4 pH3.5  ACN / H 0 2  PRP-1  Protein Analyzed  Y  5.3/5.8  trypsinogen  Y  6.2  intact P A R P  Y  7.3  PARP Fragments  Y  8.3/11.0  trypsinogen  Y  4.0  intact PARP  Y  4.5  PARP Fragments  Y  3.6/4.3  trypsinogen  N  Jupiter C 4  N  trypsinogen  N  cytochrome c  N  trypsinogen  N  pH8.0  intact P A R P  N  ACN/ HCOONH4 pH4.3  trypsinogen  Y  intact PARP  N  HFIPA / Methanol/ H 0  Jupiter C 4  2  ACN/ NH4HCO3  PRP-1  PRP-1  Retention Time (min)  cytochrome c  cytochrome c  Neutral  Elution Achieved (Yes/No)  4.3  121  Fig. 30. H P L C separation (A.) and SDS-PAGE identification (B.) of caspase-3-digested PARP fragments. A n injection (50 uL) of dialysis buffer (50 m M Tris-HCl pH 8.0, 1 m M DTT, 4 m M MgCl ) (lower chromatogram) was compared to an injection of the PARP digest reconstituted in dialysis buffer (upper chromatogram). Separation was performed on a Hamilton PRP- 1 column (25 cm x 4.6 mm ID.) at a flow rate of 1 ml/min with mobile phases consisting of (a.) 20% A C N / 0.1% T F A , and (b.) 90% A C N / 0.1% TFA. Gradient elution was performed as follows: 100% mobile phase (a.) at 1 min, changing to 60% mobile phase (a.) by 10 min, and back to 100% mobile phase (a.) by 11 min. Detection of proteins was by U V absorbance at 280 nm. For SDS-PAGE analysis, the PARP digest and a sample of each of the purified P A R P fragments were analyzed on a 10% gel, followed by Coomassie Blue staining of protein bands. Lane 1, M W markers; lane 2, purified PARP 2.1 |ig; lane 3 - 4, peaks #2 and #3, respectively, eluting from H P L C column. 2  122  PARP Digest Dialysis Buffer i  0  B.  1  r  5 10 Time (min)  p HPLC Eluant MW 2.1 pg Peak 2 Peak 3 A R P  kDa 200 116 97  89 kDa  66 45  31  24 kDa  123  4.6.3.2.  Development of an HPLC Method for Separation of PARP Fragments Under  Neutral Conditions H P L C is considered the best method for separation of the PARP fragments because less manipulation of the protein is required and the L C can be directly interfaced to the mass spectrometer for immediate analysis of the protein. However, one limitation of this technique is that the H P L C separation must be performed at a pH > 5 to prevent metal ejection from the zincfinger motifs of the PARP fragment. Chromatography of proteins is commonly performed under acidic conditions. When protein separation is achieved at neutral conditions, a common component of the mobile phase is phosphate salts, which act as an ion-pairing reagent thus limiting solute interaction with the stationary phase. However, buffer salts are generally not compatible with mass spectrometric analysis. Therefore, the choice of mobile phase components for H P L C is limited when the ultimate goal is to perform a mass spectrometric analysis of the protein. A summary of the various conditions evaluated during development of a method for separating proteins by H P L C is presented in Table 5 (page 121). Model basic proteins,.such as trypsinogen and cytochrome c, were used to assess various mobile phases. Specific details about mobile phase composition and gradient profiles can be found in Table 1 (page 63). A number of approaches to designing a mobile phase system have been attempted that would satisfy both H P L C separation and mass spectrometry compatibility demands. H P L C of the model basic proteins was attempted initially using A C N without other mobile phase modifiers. However, gradients from 20% A C N to 90% A C N did not result in elution of either trypsinogen or cytochrome c. Hexafluoroisopropanol (HFIPA) is a weak organic acid with a pH in solution > 5. The favorable ionization characteristics suggested that HFIPA may be a good mobile phase modifier (Apffel et al, 1997). Gradients from 20% to 90% methanol, containing  124  26.5 m M HFIPA, did not result in elution of either trypsinogen or cytochrome c. Increasing the HFIPA concentration to 200 or 800 m M did not improve the chromatographic characteristics. Several buffers, which were compatible with mass spectrometry, were evaluated as mobile phase modifiers. Chromatography was attempted using mobile phases consisting of A C N containing 4 m M NH4HCO3, pH 8.0. However, elution of either trypsinogen or intact PARP was not achieved. Gradient elution using a mobile phase consisting of A C N containing 0.3% HCOONH4  pH 4.3 was evaluated. Elution of trypsinogen was achieved in this system (Table 5,  page 121). However, when intact PARP was injected, no peak elution was observed. To determine if the lack of PARP elution was related to the pH of the mobile phase, chromatography was reevaluated using A C N containing 0.3% H C O O N H 4 , pH 3.5. Under these conditions, elution of both trypsinogen at 4 min and intact PARP at 4.9 min was achieved (chromatograms not shown). Separation and elution of caspase-3-digested PARP fragments was also achieved under these mobile phase conditions (chromatogram not shown). These results were sufficient to suggest that the elution of PARP and its digestion fragments was, in part, pH dependent. Although elution of intact PARP and the PARP fragments was achieved using A C N / 0.3% H C O O N H 4 , pH 3.5, the pH of this mobile phase was too low to permit analysis of metalintact PARP by ESI-MS. Mobile phase modifiers that appeared suitable for mass spectrometric analysis (HFIPA, NH4HCO3, H C O O N H 4 ) either did not result in elution of PARP or did not provide an adequate chemical environment for mass spectrometric analysis of PARP.  125  4.6.3.3.  Choice of HPLC Column for Protein Separation  Initially, a silica-based reverse phase column, Jupiter C 4 (50 x 2.0 mm I.D., 5 (im, Phenomenex) was evaluated. This column was recommended for protein separations due to the low degree of exposed silanol groups on the stationary phase. A polymeric reverse phase H P L C column, PRP-1 (250 x 4.1 mm I D . , 5 |0,m, Hamilton), composed of a poly(styrene-divinyl) benzene support phase, was also investigated for protein separations. Although the potential for hydrogen bonding with residual silanol groups on a silica-based reverse phase column has been eliminated with the use of the polymeric column, there are other hydrophobic interactions between the polymeric stationary phase and protein that prevented elution at pH above 3.5 (Table 5). The lack of success in achieving elution of basic proteins from either a reverse phase C 4 column or a reverse phase polymeric column under non-acidic conditions prevented development of an H P L C method to separate the P A R P fragments prior to mass spectrometric analysis.  126  4.7.  Mass Spectrometric Analysis of Native PARP and PARP  4.7.1.  Development of a LA-ICPMS Method for Analysis of PARP Metal Content  4.7.1.1.  Fragments  LA-ICPMS Analysis of Zinc Isotopes in PARP Bands on a Polyacrylamide Gel  Analysis of the zinc isotopes, Z n , Z n , Z n 64  66  67  and Z n , in PARP bands isolated on SDS68  P A G E gels, was performed by laser ablation-inductively coupled plasma mass spectrometry (LA-ICPMS). Isotopic abundance of the blank gel was determined and used for background subtraction. Fig. 31A shows the abundance of Z n , Z n , Z n 64  66  67  and Z n  68  in the protein bands as  the amount of P A R P loaded onto the gel was increased. There did not appear to be a clear relationship between the abundance of zinc isotopes and the amount of P A R P protein loaded onto the gel. The abundance of zinc isotopes in the blank gel (Fig. 3 IB) was high, especially for Z n . The high background levels of zinc isotopes resulted in many negative values for zinc 68  abundance in the protein bands, particularly at the lower concentrations of protein. Therefore, the high background values of the zinc isotopes did not allow for the detection of small amounts of zinc in the protein bands. 4.7.1.2.  LA-ICPMS Analysis of Sulfur Isotopes in PARP Bands on a Polyacrylamide Gel  The abundance of sulfur isotopes was evaluated as a potential marker of the quantity of protein so that the amount of metal detected could be normalized to the amount of protein in the ablated sample. Fig. 32 shows the abundance of S  33  and S in the protein bands as the amount of 3 4  PARP loaded onto the gel increases. There did not appear to be a relationship between sulfur isotope abundance and amount of P A R P protein loaded onto the gel. The abundance of sulfur isotopes in the blank gel (Fig. 32B) was high in comparison to the amount of sulfur being measured in the PARP protein bands. Therefore, sulfur did not appear to be a good marker for protein quantity.  127  20001500•= 5?  £ °  1000-  s  500-  3  o ro  <m o *-  0J  •  *; OO  (0 Q .  Zn<34 Zn66  -500-1000-  1  2  3  V  Zn67  •  Znea  i 5  4  Amount of PARP in Band (ug)  B. 10000-1 0)  C5  5000-12000  c w  m o  1500-^  <D ^ O w C c «J 3 "D o ° 1000H 3  4 Qi  Q. O O  500 H  (/) Zn64  Znee  Zn67  Znes  Fig. 31. Abundance of zinc isotopes in PARP protein bands isolated by SDS-PAGE (A.) and in the blank gel ( B.) measured by L A - I C P M S . The gel was dried onto a cellophane support prior to L A - I C P M S analysis. A section (5 (im in diameter) of the leading edge of each protein band was ablated and isotopic abundance (counts/s) of zinc isotopes was measured. L A - I C P M S was performed by Elemental Research Inc. (North Vancouver, BC). Each value represents the background-subtracted average of at least duplicate measurements of the protein band. Areas of blank gel were analyzed (B.) for abundance of the zinc isotopes and the mean value for each isotope was used as the background value. Each blank value represents the mean ± S E M of 9 measurements.  128  A. 100000l c CD +-*  80000H  O 0) O C  S34 S33  "E El o  60000H  co a. CA T3 T3 C CO  40000-  ffl  a)  20000H  CL  o CO  1  2  3  -1  5  Amount of PARP in Band (ug)  B. 1000000)  o  80000-  c  « m a) v>  60000-  § § c  n <  a> Q.  o +*  o  40000  20000-  o  S33  S34  Fig. 32. Abundance of sulfur isotopes in PARP protein bands isolated by SDS-PAGE (A.) and in the blank gel (B.) measured by LA-ICPMS. The gel was dried onto a cellophane support prior to LA-ICPMS analysis. A section (5 pm) of the leading edge of each protein band was ablated and isotopic abundance (counts/s) of sulfur isotopes was measured (A.). LA-ICPMS was performed by Elemental Research Inc. (North Vancouver, BC). Each value represents the backgroundsubtracted average of at least duplicate measurements of the protein band. Areas of blank gel were analyzed for abundance of the sulfur isotopes (B.) and the mean value for each isotope was used as the background value. Each blank value represents the mean ± S E M of 9 measurements.  129  4.7.1.3.  Analysis of Zinc in a PARP Solution by ICPMS  An analysis of the Z n  64  isotope was performed on a solution of purified P A R P to determine  the detection limit for zinc in a protein in solution. Dilutions of purified PARP were analyzed by ICPMS, and the abundance of the Z n observed between Z n  64  64  isotope is shown in Fig. 33. A linear relationship was  abundance and the concentration of PARP in solution. The abundance of  zinc, measured in the dilute PARP samples (1,5 and 10 pg/mL), was not above that measured in the buffer itself. Therefore, the results indicated the limit of detection for Z n  64  x  in a PARP  solution is between 10 and 50 pg/mL of PARP protein.  0.30-.  0.251  0.0040  •  1 50  .  1 100  .  1 150  .  1 200  PARP Concentration (pg protein/mL) Fig. 33. Analysis of purified P A R P in solution by ICPMS. Purified PARP was diluted with dialysis buffer (50 m M Tris-HCl pH 8.0, 1 m M DTT, 4 m M MgCl ) to achieve concentrations of 1, 5, 10, 50, 100 and 200 pg PARP/mL. ICPMS analysis of P A R P solutions was performed by Elemental Research Inc. (North Vancouver, BC). 2  130  4.7.2.  MALDI-TOF Mass Spectrometric Analysis of Intact PARP and PARP Fragments Matrix assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF  MS) is a technique that may be used to measure molecular weight differences between native and metal-replaced PARP. Preliminary investigations were carried out with native PARP and caspase-3-digested PARP fragments. The molecular weight of full-length P A R P determined by M A L D I - T O F M S was between  112,496 -  112,574 Da (Fig. 34), which was within  approximately 0.5% of the calculated molecular weight based on amino acid sequence (113,143 Da) (Cherney et al, 1987). It is likely that improved accuracy for the molecular weight measurement would be obtained with a more concentrated protein sample. The molecular weight estimations for the caspase-3-digested PARP fragments (23,998 and 89,101 Da) (Fig. 35) were within 0.2% of the expected mass based on amino acid sequence (24,041 and 89,120 Da, respectively). M A L D I - T O F M S is performed in the presence of a matrix compound, and for this research, a saturated solution of sinapinic acid prepared in a mixture of 2:1 water:ACN containing 0.1% T F A was used. Therefore, protein analysis occurred in an acidic environment. Due to the presence of T F A in the matrix mixture, the molecular weight of the PARP fragment containing the zinc fingers likely represents the apo- or metal-free protein. Analysis of metalintact PARP requires a pH above 5.0 because the histidine ligand of the zinc-finger coordination site becomes protonated below pH 5.0, which results in metal ejection (Surovoy et al, 1992) M A L D I - T O F MS analysis of the 24 kDa PARP fragment was performed with the sinapinic acid matrix mixture in the absence of 0.1% T F A (Fig. 36). The analysis was performed in the presence of bovine insulin (MW=5,742 Da) and cytochrome c (MW=12,357 Da) as internal calibration standards. Two separate measurements were performed and the resulting mass of the  131  small PARP protein fragment was 24,565 and 24,572 Da (average = 24,568.5 Da). This molecular weight estimation is approximately 2% greater than the expected molecular weight of 24,042 Da. However, the pH of the matrix mixture, composed of sinapinic acid in H 0 : A C N 2  (2:1), was 3.4. Therefore, the pH of the matrix mixture was still too low to observe the 24 kDa PARP fragment in its native, metal-intact form. A sample of sinapinic acid was prepared with the pH buffered to 6.0 with ammonium hydroxide. Using this preparation in the matrix mixture, no signal was obtained when the PARP digest was reanalyzed.  132  (BSA) 13752  5  0.0120  0.0100 H  Fig. 34. M A L D I - T O F M S analysis of native PARP (expected molecular weight 113,135 Da). Matrix is a saturated solution of 22 mg/mL sinapinic acid dissolved in 0.1% T F A in H i O : A C N (2:1). B S A was added to the sample as an internal calibration standard.  133  (BSA) 13618  5+  0.0120  Fig. 35. M A L D I - T O F M S analysis of caspase-3-digested PARP. The calculated molecular weights for the 2 fragments based on c D N A sequence are 24,040.84 and 89,120.37 Da. The matrix is a saturated solution of 22 mg/mL sinapinic acid dissolved in 0.1% T F A in H 0 : A C N (2:1). B S A was added to the sample as an internal calibration standard. 2  134  Bovine Insulin  0.0160-  5742  0.0140-1  0.0120^  g c  0.0100«  CO T3 C  -9 <  Cytochrome c  0.0080  12357 PARP  Fragment  24565  0.0060 •  0.0040  0.0020 10000  20000  30000  40000  50000  Fig. 36. M A L D I - T O F M S analysis of caspase-3-digested PARP without T F A present in the matrix mixture. The calculated molecular weight of small P A R P fragment based on amino acid sequence is 24,040.84 Da. Matrix is a saturated solution of 22 mg/mL sinapinic acid dissolved in H.20:ACN (2:1) pH 3.4. Bovine insulin and cytochrome c were added to the sample as internal calibration standards.  135  4.7.3.  ESI-MS Analysis of Native PARP and PARP Fragments ESI-MS is an accurate and sensitive technique for determining the molecular mass of a  protein (Scobie et al, 1993). The ability to distinguish small changes in molecular weight should provide a means to assess metal replacement in the zinc fingers of P A R P based on differences in molecular weight of native (zinc-intact), apo- and metal-replaced PARP. ESI-MS has a size limitation for protein ionization and, as a consequence, fragmentation of P A R P and purification of the 24 kDa PARP fragment containing the zinc fingers was required prior to analysis. Analysis of proteins by ESI-MS is usually accomplished in the presence of an organic solvent (i.e., acetonitrile) under acidic conditions (0.1% formic acid) to facilitate protein ionization. A low pH environment is not compatible with maintaining the metal ion coordination in the zinc-finger site. Therefore, development of a ESI-MS method for basic proteins under neutral conditions (pH > 5) was necessary. The basic proteins, trypsinogen, chymotrypsin, and cytochrome c, as well as myoglobin, were used for analytical method development. These proteins were basic in nature and were in the molecular weight range of interest (Table 6) and therefore,made appropriate models for the 24 kDa PARP fragment. 4.7.3.1.  ESI-MS Analysis of Basic Proteins Under Acidic Conditions  ESI-MS performed under acidic conditions (i.e., 20% A C N / 0.1% formic acid) produced good protein spectra for trypsinogen, chymotrypsin, cytochrome c and myoglobin. Fig. 37 and Fig. 38 show the raw spectrum (A), MaxEnt transformed spectrum (B) and the mock spectrum (C), which was used for molecular weight calculations, for trypsinogen and cytochrome c, respectively. The raw spectrum for each protein showed a typical protein charge envelope with major peaks evenly spaced across the spectrum. A protein spectrum was also obtained for the 24 kDa PARP fragment under acidic conditions (Fig. 39).  136  Table 6. Comparison of molecular weights and PI of model proteins used for ESI-MS method development.  Protein  MW (Da)  pi  Trypsinogen  23,981  9.3  Chymotrypsin  25,686  9.1  Cytochrome c  12,384  10.5  Myoglobin  16,951.50  PARP (Cherney et al, 1987)  113,135  P A R P (In-house Calculation)  113,143  P A R P Small Fragment  24,041  P A R P Large Fragment  89,120  9.8  137  A.) Raw spectrum  100-1  B.) MaxEnt Transformed Spectrum  171 4.0 1600.4 ,,1714.2 1500.3. /1714.5 1499.9, 1499.6, 1499.4,  23980.0  100  1599.81714.0  100  / 1500.0  \  23992.0  1845.5  23964.0  Ii 846.6  1845.6 24012.0  %  .1847.7  /  .1848.3  1999.5  ,2001.5  23945.0  |^2002.9 2180.4  24928.0  23095.0  1500  C.) Mock Spectrum  2000  m/z  23000  24000  2181.2  1500  2000  Fig. 37. ESI-MS analysis of trypsinogen (expected molecular weight 23,980.89 Da) under acidic conditions (20% A C N / 0.1% formic acid). Raw mass spectrum (A.), MaxEnt transformed spectrum (B.) showing the calculated molecular weight of 23,980.0 Da, and the mock spectrum (C). The solvent system consisted of 20% A C N . Trypsinogen was prepared in 20% A C N / 0.1% formic acid.  138  B.) MaxEnt Transformed Spectrum  A.) Raw spectrum J03JY001 78(13.155) 1373.0  S c a n E S + J 0 3 J Y 0 0 1 78(1 3.155) 1.47e6 12348.0 1 UU-i  C.) Mock Spectrum  Scan ES + J03JY001 78(13.155) 8.21e6 ^1235.3 1 0 0  Scan ES + 1.19e6  12343.2,  1373.3  1545.0 1544.7  1 2408.8  ,1 545.3  12413.6 12453.6 12336.8. 1500  2000  m/z  1 ..f  0 10000  12000  1552.3 1765.2  ,1 2510.4  1768.42059.2  14155.2  14000  mass  T  1500  S>  k ii  2000  m/z  Fig. 38. ESI-MS analysis of cytochrome c (expected molecular weight 12,384 Da) under acidic conditions (20% A C N / 0.1% formic acid). Raw mass spectrum (A.), MaxEnt transformed spectrum (B.) showing a major peak with the calculated molecular weight of 12,348.0 Da, and the mock spectrum (C). The solvent system consisted of 20% A C N . Cytochrome c was prepared in 20% A C N / 0.1% formic acid.  139  B.) MaxEnt Transformed Spectrum  A.) Raw spectrum J M Y 2 0 0 0 3 9 0 (7.675)  Scan ES+  888.8  100  8.49e5 959.9  J M Y 2 0 0 0 3 9 0 (7.675) ,„„  S c a n E S + J M Y 2 0 0 0 3 9 0 (7.675)  23968.8  100  C.) Mock Spectrum  3.09e6  100  959.8  Scan ES+ 3.83e5  888.6  1142.3 11143.3  23989.6 / ^23995.2 22891.2  o-ippiH  750  vt».. Mfwjnfm^mw\ 1000  1250  1500  m/z  0 22000  25682.4  1498.8  25047.2  Mill  24000  mass 26000  m/z 1000  1500  Fig. 39. ESI-MS analysis of the 24 kDa P A R P fragment (expected molecular weight 24,040.84 Da) under acidic conditions (15% A C N / 0.1% formic acid). Raw mass spectrum (A.), MaxEnt transformed spectrum (B.) showing the calculated molecular weight of 23,968.8 Da, and the mock spectrum ( C ) . The solvent system consisted of 15% A C N / 4 m M ammonium bicarbonate (NH4HCO3), pH 8.0. The PARP fragment was prepared in 15% A C N / 0.1% formic acid.  140  4.7.3.2.  Comparison Between Positive and Negative Ion Electrospray MS  The metal binding stoichiometry of two metalloproteins, the vitamin D receptor DNAbinding domain, M W = 12,819 Da (Craig et al, 1997), and calbindin, a calcium binding protein of M W = 29,866 Da (Veenstra et al, 1997), has been determined using ESI-MS under nonacidic conditions in a solvent system containing 4 m M ammonium bicarbonate (NH4HCO3), pH 8.0/15% methanol. The vitamin D receptor DNA-binding domain was analyzed under positive electrospray conditions, whereas calbindin was analyzed under negative electrospray conditions. The success of both these investigations using the ammonium bicarbonate buffer system, and the molecular weight of the proteins that were analyzed, particularly calbindin, suggested that this solvent system may work with the model proteins and the P A R P fragment used in this study. Myoglobin was used to investigate positive and negative ESI-MS using either an acidic solvent system (ACN/TFA) or an ammonium bicarbonate solvent system (4 m M NH4HCO3, pH 8.0 / 15% ACN). A comparison between the raw spectra obtained for myoglobin under positive and negative electrospray conditions is shown in Fig. 40. Under acidic conditions using positive electrospray (Fig. 40A.), a typical protein spectrum with a well defined charge envelope and little noise was obtained. Under positive electrospray conditions using the ammonium bicarbonate buffer system, a protein charge envelope was still apparent but more noise was evident in the spectrum (Fig. 40B.). Using the same ammonium bicarbonate buffer system, myoglobin was analyzed under negative electrospray conditions (Fig. 40C). A more concentrated sample of myoglobin was required to obtain an adequate signal, indicating that sensitivity was lost for this protein under negative conditions. The 10-fold loss of signal intensity also indicated the loss of sensitivity in negative mode compared to positive ES mode. Again, a charge envelope was observed, but the signal to noise ratio was even less than that obtained in  141  positive ion M S . Therefore, the raw spectra obtained for myoglobin became progressively worse as the electrospray conditions were changed from acidic positive ES, to neutral positive ES, and finally to neutral negative ES. The transformed spectra for myoglobin under these three conditions (i.e., acidic positive ES; neutral positive ES; neutral negative ES) are shown in Fig. 41. The expected molecular weight for myoglobin is 16,951.499 Da. Under acidic positive ES conditions (Fig. 41A.), the MaxEnt transformed spectrum showed a clean spectrum (i.e., no noise) with a major protein peak at 16,949.6 Da and a much smaller adduct protein at 17,041.4 Da. Under neutral positive ES conditions (Fig. 41B.), the transformed spectrum had a major peak at 16,950.8 Da; however, the adduct protein had greater intensity than under acidic positive ES conditions. The transformed spectrum obtained under neutral negative ES conditions (Fig. 41C), showed a major protein at M W = 16,951.3 Da and an adduct protein at 16,973.5 Da. The baseline from this spectrum showed much more noise than the spectra obtained using positive ES. It is interesting to note that even when the raw mass spectra can be very noisy, the MaxEnt program can still calculate a relatively accurate molecular weight for myoglobin. As expected, acidic conditions provided a transformed spectrum with the least noise in the baseline. Under neutral positive ES conditions, a reasonably clean transformed spectrum and correct M W determination were obtained for myoglobin. The spectra obtained with negative ES under neutral conditions were not as good as those for positive ES under neutral conditions. Therefore, all further analyses were performed under positive ion mode.  142  Fig. 40. Comparison between the raw spectra obtained following ESI-MS analysis of myoglobin (MG) under acidic positive electrospray conditions (A.), neutral positive electrospray conditions (B.), and neutral negative electrospray conditions ( C ) . The solvent systems used for analysis under acidic and neutral conditions were 15% A C N / 0.1% formic acid or 15% A C N / 4 m M ammonium bicarbonate (NH4HCO3), pH 8.0, respectively. For acidic or neutral analysis, myoglobin was prepared in either 15% A C N / 0.1% formic acid or 15% A C N / 4 m M NH4HCO3, pH 8.0, respectively.  143  A.) ESI-MS Analysis of MG Under Acidic, Positive Electrospray Conditions  B.) ESI-MS Analysis of MG Under Neutral, Positive Electrospray Conditions  1150  C.) ESI-MS Analysis of MG Under Neutral, Negative Electrospray Conditions  1590.7.  1728.1.  1200  1250  1300  1350  1«  1621.8.  1604.3 IBM.  144  Fig. 41. Comparison between the MaxEnt transformed spectra obtained following ESI-MS analysis of myoglobin (MG) under acidic positive electrospray conditions (A.), neutral positive electrospray conditions (B.), and neutral negative electrospray conditions ( C ) . The solvent systems used for analysis under acidic and neutral conditions were 15% A C N / 0.1% formic acid or 15% A C N / 4 m M ammonium bicarbonate (NH4HCO3), pH 8.0, respectively. For acidic or neutral analysis, myoglobin was prepared in either 15% A C N / 0.1% formic acid or 15% A C N / 4 mM NH4HCO3, pH 8.0, respectively.  145  A.) MaxEnt Transformed Spectrum of MG Analyzed Under Acidic, Positive Electrospray Conditions  ±o.: 16000  16200  15400  M00  ' 17000  17200  17400  17600  17800  16000  B.) MaxEnt Transformed Spectrum of MG Analyzed Under Neutral, Positive Electrospray Conditions  18109.0 1W71.C  „HV,, 17000  U.-ytLl lL 16000  f  ) l„ti  10000  „  20000  i.n, 21000  i LI  II, 22000  L  23000  i,,,.L>  IN.,,,,,, 24000 25000  JL., i ,•„• >. It,,,. I.-, N mt 26000 27000 28000  C.) MaxEnt Transformed Spectrum of MG Analyzed Under Neutral, Negative Electrospray Conditions  7660  '  i7too  irtoo  146  4.7.3.3.  Development of an ESI-MS Method for Analysis of Basic Proteins Under Neutral  Conditions The choice of a solvent for the ESI-MS analysis of zinc-finger proteins in their native form required not only consideration of the pH of the solvent, but also compatibility with the mass spectrometer and the stringent requirements for ionization. Therefore, the literature describing similar analyses with zinc-finger peptides provided a good starting point for method development. Table 3 (page 70) shows a summary of the various solvent systems that were evaluated. Ammonium acetate buffers have previously been used to analyze zinc-finger peptides by ESI-MS (Surovoy et al, 1992; Loo et al, 1996). The responses in the total ion chromatogram of trypsinogen, prepared in either an acidic solvent containing formic acid or a solvent containing ammonium acetate pH 5 - 6.68, were compared (Fig. 42). A large protein peak in the total ion chromatogram was obtained when the acidic solvent was used. However, this signal was completely lost when ammonium acetate was included in the sample buffer (Fig. 42A.). Decreasing the concentration of ammonium acetate from 8 m M to 1 m M seemed to improve the signal in the total ion chromatogram somewhat (Fig. 42B.). However, the raw spectra obtained when 1 m M ammonium acetate was included as part of the buffer were not typical protein spectra and did not provide an accurate molecular weight measurement for trypsinogen (data not shown). Therefore, the switch from a solvent system containing formic acid, in which a good protein signal and spectrum was obtained, to a solvent system containing ammonium acetate resulted in an almost complete loss of the protein signal. The lack of success with ammonium acetate led to the investigation of other volatile, weak organic acids that might have a pH above 5. Formic acid (26.5 m M solution, pH 2.88), propionic  147  acid (26.5 m M solution, pH 3.30), acetic acid (26.5 m M solution, pH 3.33) and octanoic acid (26.5 m M solution in 40% A C N , pH 3.85) were examined. Although these acids appeared to provide adequate protein spectra of trypsinogen (data not shown), the pH of these solutions was still too low to examine native metal content of a zinc-finger protein. The glucocorticoid receptor DNA-binding domain (MW < 10,000 Da), a zinc-finger protein, has been analyzed under neutral electrospray conditions using pyridine acetate, pH 5.9 as the electrospray solvent (Witkowska et al, 1995). Myoglobin was analyzed under these conditions; however, no protein signal was obtained (data not shown). Samples of both myoglobin and cytochrome c were prepared in 4 m M triethylamine acetate, pH 7.0 / 15% A C N . The samples were analyzed using SIR monitoring in positive electrospray mode to increase the sensitivity of detection. No peak was detected for either myoglobin or cytochrome c (data not shown). The lack of response with either triethylamine acetate or pyridine acetate may be the result of ion competition. It is much easier to protonate triethylamine than it is to protonate a protein. Furthermore, there is much more triethylamine available, so charge preferentially goes there. As evidence of this, the most abundant peak in the spectrum was ion 101, triethylamine (data not shown). It is likely that the same process of ion competition was occurring when pyridine acetate was used as in the electrospray solvent and therefore,protein ionization in these solvents was hindered. Using the ammonium bicarbonate buffer system (4 m M NH4HCO3, pH 8.0 / 15% ACN), protein spectra were obtained for myoglobin (data not shown). Analysis of trypsinogen in this solvent system failed to produce a signal in the raw spectrum. The MaxEnt transformed spectra did not show a protein peak indicative of trypsinogen (data not shown). The lack of a trypsinogen protein peak reflected the lack of ionization apparent in the raw spectrum. Due to the positive results with myoglobin, ESI-MS analysis of the 24 kDa P A R P fragment was performed  148  using the ammonium bicarbonate buffer system. Direct injection of the 24 kDa P A R P fragment dissolved in water and diluted with the electrospray solvent resulted in a poor response with only a small peak apparent on the total ion chromatogram. The raw spectrum of this peak revealed not a typical protein spectrum as expected, but a dominant ion species at approximately 615 - 616 m/z (data not shown). It was determined that much of the intensity of the peak in the total ion chromatogram upon injection of the P A R P fragment was due to ion 616 (data not shown). Therefore, the P A R P fragment did not appear to become ionized under these neutral conditions. Hexafluoroisopropanol (HFIPA), a weak organic acid with a pH in solution > 5, has previously been used as a solvent for ESI-MS (Apffel et al, 1997). To evaluate HFIPA as a potential solvent for analysis of native metalloproteins, a solvent system containing 20% A C N / 53 m M HFIPA was developed. The pH of 53 m M HFIPA was 5.1. Protein spectra were obtained for cytochrome c (Fig. 43) and myoglobin (Fig. 44). These spectra provided sufficient data to calculate the molecular weight of each protein to within < 1% of the expected molecular weight. The spectrum obtained for cytochrome c has less noise than that obtained for myoglobin, indicating that, under these conditions, ionization of a lower molecular weight protein may be easier. The 24 kDa fragment of P A R P was also analyzed in the solvent system containing 20% A C N / 53 m M HFIPA. Fig. 45 shows the raw spectrum and the MaxEnt transformed spectrum for the 24 kDa P A R P fragment. The raw spectrum showed a charge envelope typical of a protein spectrum that contained evenly spaced peaks, and a noisy baseline. The MaxEnt transformed spectrum shows a major peak with a molecular weight of 23,969.8 Da, which is within 0.3% of the expected molecular weight for the fragment (24,041 Da). The minor peaks seen in the baseline of the MaxEnt spectrum likely reflect the noise apparent in the raw spectrum.  149  IJ26JU005 100-  Scan ES+ TIC | 3.10e9  5.92. 6 76 7.43 .27 20% ACN 8 m M CH3COONH4 pH 5.01  20% ACN 0 . 1 % Formic Acid  20% ACN 8 mM CH3COONH4 pH 6.68  • IJ27JU004 100-  0/ /o -  0.00  20% ACN 1 mM C H C O O N H pH 6.68 3  4  •  3  Scan ES+ TIC 8.64e8  26.79 4  20% ACN 0 . 1 % Formic Acid  16.6S  6.42  5.00  20% ACN 1 mM C H C O O N H pH 5.01  10.00  15.00  20.00  i  <  1 1 1  1  25.00  30.00  1 Time 35.00  Fig. 42. Relative abundance of trypsinogen in the total ion chromatogram following ESI-MS analysis under acidic and neutral conditions. Trypsinogen was dissolved in either 20% A C N / 0.1% formic acid, 20% A C N / 8 m M ammonium acetate (CH3COONH4), pH 5.01, or 20% A C N / 8 m M CH3COONH4, pH 6.68 (A.). For comparison, trypsinogen was dissolved in 20% A C N /1 m M C H 3 C O O N H 4 , pH 6.68, 20% A C N / 1 m M C H 3 C O O N H 4 , pH 5.01, or 20% A C N / 0.1%  formic acid (B.). The electrospray solvent system was 20% A C N .  150  C.) Mock Spectrum  B.) MaxEnt Transformed Spectrum 3.96e  100  B  100-,  A.) Raw spectrum Scan ES+ 4.10e6  ,1235.5  1373.2  12343.9  13737  1544.9  1544.9 1545.3  1544.3 12352.0  1765.4  1379.2  1379.1  1551.6  JV-.J  0 1200  1400  1600  1800  2000  Da/e 2200  1542.7| ,1551.71 1575.2  13723.5  2059.4  14810.0 10000  12000  ML  14000  ^ mass  1765.4 1764.9^  0 1200  m  MiffliUirWimlliiiilt ' mir  1400 ' 1600 ' 1800  Dale 2000 ' 2200  Fig. 43. ESI-MS analysis of cytochrome c (expected molecular weight 12,384 Da) under neutral conditions using hexafluoroisopropanol (HFIPA). Raw mass spectrum (A.), MaxEnt transformed spectrum (B.) showing a major peak with the calculated molecular weight of 12,350.0 Da, and the mock spectrum ( C ) . The solvent system consisted of 20% A C N / 53 m M HFIPA. Cytochrome c was prepared in 20% A C N / 53 m M HFIPA.  151  C.) Mock Spectrum 100  B.) MaxEnt Transformed Spectrum 100  "j  A.) Raw spectrum  100  1303.4  [1303.9  1412.5  1412.4  ,1412.6  ,1412.7 1464.2  1417.4 1541.0  1540.6  15719.0  12350.0  16993.5 .17552.5  1694.6 1756.2 14331.0 1765.6  0 1200  1400  1600  1800 ' 2000  17575.5  Dale 2200  iWo'^'i'z'Ioo^  Dale  Fig. 44. ESI-MS analysis of myoglobin (expected molecular weight 16,951.50 Da) under neutral conditions using hexafluoroisopropanol (HFIPA). Raw mass spectrum (A.), MaxEnt transformed spectrum (B.) showing a major peak with the calculated molecular weight of 16,937.0 Da, and the mock spectrum ( C ) . The solvent system consisted of 20% A C N / 53 m M HFIPA. Myoglobin was prepared in 20% A C N / 53 m M HFIPA.  152  A.) Raw spectrum  100  827.6 1048.2  B.) MaxEnt Transformed Spectrum  1 152.0  100  23969.6  C.) Mock Spectrum  1043.2  100  653.3  9  2  2  9  /  1443.1  800.3  1090.5  /o  23992.8  1262.6 ,1332.5  24780.0 25476.0  1 000  1500  m/z  24000  mass  26000  1499.8  m/z  1000  Fig. 45. ESI-MS analysis of the purified 24 kDa PARP fragment (expected molecular weight 24,040.84 Da) under neutral conditions using hexafluoroisopropanol. Raw mass spectrum (A.), MaxEnt transformed spectrum (B.) showing a major peak with the calculated molecular weight of 23,969.6 Da, and the mock spectrum (C). Solvent system consisted of 20% A C N / 53 m M HFIPA. The purified PARP fragment was prepared in 20% A C N / 53 m M HFIPA.  153  4.7.3.4.  Protein Adsorbance to the Instrument Components  Over the course of the analysis of the model proteins in various non-acidic solvents systems, it became apparent that these proteins were "sticking" to components of the H P L C and/or mass spectrometer. Protein adsorption to instrument components was detected when an injection of 0.1% formic acid was made following the injection of a protein in a non-acidic solvent. For example, when an injection of cytochrome c was performed in a non-acidic solvent, and this injection was followed by a series of formic acid injections (i.e., formic acid "chasers"), a peak was detected following each acid injection (Fig. 46). The raw spectra of each of these peaks could be deconvoluted to produce accurate molecular weight estimations for cytochrome c (data not shown). Therefore, cytochrome c was eluting from unknown adsorption sites within the instrument with each injection of acid. This adsorption phenomenon was observed for all of the model proteins tested (i.e., cytochrome c, myoglobin, trypsinogen), although the extent of adsorption appeared to vary somewhat, with cytochrome c being particularly affected. Because protein adsorption appeared to be a general phenomenon, it was hypothesized that part of the inability to analyze PARP under non-acidic conditions was due to its tendency to stick to the instrument. To test this hypothesis, an injection of the 24 kDa PARP fragment was performed using the 4 m M ammonium bicarbonate, pH 8.0 / 15% A C N solvent system. The injection of PARP produced a small response at 3.94 min in the total ion chromatogram, which corresponded primarily to ion 615 (Fig. 47A. and B.). A typical protein spectrum was not obtained from this peak (data not shown) and the transformed spectrum (Fig. 47C.) did not show a single prominent protein species. A subsequent injection of 15% A C N / 0.1% formic acid, the formic acid "chaser" (Fig. 47A.), produced a large peak at 7.68 min. The raw spectrum of this peak following acid injection did show a typical protein response (data not shown) that was  154  transformed into a protein species with molecular weight of 23,970.4 Da (Fig. 47D.). The expected molecular weight for the 24 kDa PARP fragment was 24,040.84 Da. It is likely that the peak eluting, following the formic acid injection, was in fact, the 24 kDa PARP fragment. These results confirm the belief that injection of PARP under neutral conditions results in its adherence to some component of the L C or M S , and that subsequent injections of formic acid are able to elute the protein from its binding sites within the system. Using cytochrome c as the model "sticky" protein, ways to reduce protein adsorption to the instrument with the addition of modifiers to the 4 m M ammonium bicarbonate, pH 8.0 / 15% A C N solvent system were investigated. Initially, E D T A , a metal chelator, and deferoxamine, an iron chelator, were added to the solvent system with the hypothesis that the protein was adhering to the iron of the stainless steel components of the instrument. Addition of the metal chelators did not alter the adsorption characteristics of cytochrome c. Individual amino acids (25 mM) were added to the solvent with the hypothesis that a particular amino acid may be responsible for the absorption phenomenon and that exogenous addition of the amino acid would saturate the binding sites on the instrument and allow the protein to pass through the system. Several amino acids were evaluated, such as aspartic acid, arginine, glutamine, glycine, lysine, methionine and phenylalanine, which covered the range of amino acid groups (basic, acidic, aromatic and S-. containing amino acids). However, none of these modifiers improved the adsorption characteristics of cytochrome c. Further investigation of the protein "sticking" phenomenon was performed in the Mass Spectrometry Facility at the School of Pharmacy, University of Washington (Seattle, WA). The organic content of the non-denaturing mobile phase (ACN/ammonium bicarbonate, pH 8 or ACN/ammonium acetate, pH > 6) was increased with the addition of isopropanol as an organic modifier. When the model "sticky" protein, cytochrome c, was analyzed using a solvent system  155  containing 50% organic content, the sticking phenomenon was not reproduced. However, ionization of cytochrome c was not accomplished in this solvent system. Reducing the organic content to 25% resulted in cytochrome c sticking to the instrument. It was then released by a subsequent injection of formic acid. Therefore, the inability to reproduce the adsorption phenomenon with a solvent system containing 50% organic content suggested that cytochrome c had increased solubility in this solvent and did not adsorb to the instrument components under these conditions. However, neither did it ionize. Attempts were made to adjust the organic content of the solvent to achieve increased solubility in the solvent and protein ionization. However, it appeared that in order to prevent protein adsorption, the organic content had to be so high that ionization was also inhibited. Injection of the 24 kDa PARP fragment in a solvent system containing 25% organic modifier resulted in protein adsorption and poor response in the total ion chromatogram. A subsequent injection of formic acid released the P A R P fragment from its binding sites, and a reasonable protein spectrum was obtained. The observations with cytochrome c suggested that an increase in the organic content of the solvent would not help in the electrospray analysis of PARP.  156  Fig. 46. Adsorption of cytochrome c (CC) to components of the HPLC and/or mass spectrometer and its release following injection of formic acid. The total ion chromatogram following analysis of cytochrome c is shown. An injection of cytochrome c dissolved in 20% ACN / 53 mM hexafluoroisopropanol (HFIPA) was performed, followed by sequential injections of 0.1% formic acid. Raw spectra of the initial cytochrome c injection and each subsequent peak following the formic acid injections revealed protein spectra characteristic of cytochrome c. The solvent system consisted of 20% ACN / 53 mM HFIPA.  157  Fig. 47: Adsorption of the 24 kDa PARP fragment to components of the H P L C and/or mass spectrometer and its release following injection of formic acid. A n injection of the 24 kDa PARP fragment dissolved in 15% A C N / 4 m M ammonium bicarbonate pH 8.0 was performed, followed by an injection of 15% A C N / 0.1% formic acid. The total ion chromatogram is shown in (A.). Single ion monitoring of ion 615 is shown in (B.). The MaxEnt spectrum calculated by transforming the combined spectra from retention times (RT) 3.0 - 4.3 min is shown in ( C ) . The MaxEnt spectrum calculated by transforming the combined spectra from retention times (R ) 7.5 - 8.4 min is shown in (D.). T  158  JMY20003 100-,  Injection of 15% ACN / 0.1% Formic  PARP in N H 4 H C O 3 Buffer  Scan ES+ TIC I 1 29e9  7 6 8  3.94  1 11 1 1 1 1 1 1 1 1 • 1 3.00 4.00 5.00  2.00  0.00 " ' " 1.00  JMY20003  326  100-  3 9 4  11  1  1111  0.00  1  1111  1 "  1.00  i  1 1 1 1  1  3.00  4.00 2  0 0 l  1 • 1 • 6.00 1 1  1  111  1 1 7.00 1111  11  • 1 • 1 8.00 1  1  11  11  • • 1 • • 1 ' '1 • 1 9.00 10.00 1 1  11  3  i i | i I i i | IT i i | T i i i | i  5.00  8.00  9  6  0  1  i Time 9.00 10.00  i i | I I i i | i i l l | l i i i | i i i l | l i l l | (l  6.00  1 1  Scan ES+ 615 7.00e7  Single Ion Monitoring of Ion 615  i  1 1 | T  2.00  JMY20003 46 (3.935) 1  T l  1111  7.00  Scan ES+ 8.24e5 I  0  MaxEnt Analysis of Combined Spectra from R 3.0 - 4.3 min  22165.0  T  %  23878.0 \ 24498.0  23358.0 22915.0 22000  22500  23000  25760.0  24562.0  22734.0 22761.0  4Si  ll 23539.0 23500  24569.0 V  \  24170.0  24000  24500  25172.0  25072.0  25000  25718.0 \  25436.0 25802.0 mass 26000  25500  Scan ES+ 3.08e6 I  JMY20003 90(7.675) 23970.4  100-  MaxEnt Analysis of Combined Spectra from R 7.5 - 8.4 min T  %  23988.8  c 22414.4 22261.6/ 22000  22500  23108.823188.823275.2 , , 1 I f ^23385.6 23000  23500  24054.4  24000  24931.2. 25047.2 ^24436.8 24500  25000  25682.4  25916.0 ,  25500  i ^25946.41 mass 26000  159  1 1 1  '1  1  4.8.  Endothelial  4.8.1.  Effect of H 0 on the Time-Course and Extent of E C Death 2  Cell Injury 2  The time of onset, time-course and extent of H 0 -induced E C death were concentration 2  2  dependent, as depicted in Fig. 48. Loss of plasma membrane integrity of E C did not begin until approximately 2-4 h following initial oxidant exposure. In this set of experiments, a maximum loss of plasma membrane integrity was achieved with 200 p M H 0 2  2  over the duration of the  experiment. The onset of cell death varied from 2-7 h, with the shortest delay (at approximately 2-3 h) being achieved with 200 p M F£ 0 . Higher concentrations of H 0 (300 - 700 pM) did 2  2  2  2  not shorten the onset of cell death (data not shown). 4.8.2.  Effect of the Combination of PARP and Protein Synthesis Inhibition on H2O2-  Induced E C Death Apoptosis is thought of as an active process, often requiring synthesis of new proteins. Fig. 49 shows the effect of cycloheximide, a protein synthesis inhibitor, in conjunction with P A R P inhibition on the time-course and extent of E C injury following H 0 2  2  exposure. In  oxidant-injured E C treated with the P A R P inhibitor, DIQ, the onset and time-course of cell lysis was delayed by approximately 2 h compared to cells treated with H 0 2  2  alone. Inhibition of  protein synthesis with cycloheximide did not have any effect on the onset of cell lysis or extent of cell death in H 0 -treated cells. However, the combined treatment of DIQ + cycloheximide 2  2  further delayed and attenuated the loss of membrane integrity compared to treatment with either DIQ or cycloheximide alone. The ability to further delay the onset of cell lysis with cycloheximide suggested that an active mode of cell death, dependent upon protein synthesis, was occurring in oxidant-stressed E C when P A R P was inhibited.  160  120-1  Time Post-H 0 Exposure (h) 2  2  Fig. 48. Time-course of irreversible E C injury after exposure to increasing concentrations of H2O2. Confluent E C were exposed to increasing concentrations of H2O2 (50 - 200 pM) and loss of plasma membrane integrity was monitored with the use of the plasma membrane-impermeant fluorescent probe, YO-PRO. Each data point represents the mean ± SD of 3-4 wells in a single representative experiment. NTC, non-treated control cells.  161  Time Post-100 jiM H 0 Exposure (h) 2  2  Fig. 49. Effect of PARP and protein synthesis inhibition on the time-course of H 0 -induced E C death. Confluent E C were treated with DIQ (10 pM), cycloheximide ( C H X , 50 uM), or DIQ+CHX, followed by the addition of 100 p M H 0 . Non-treated controls (NTC) received vehicle (i.e., water). Loss of plasma membrane integrity was monitored continuously over the course of the experiment using the plasma membrane-impermeant fluorescent probe, YO-PRO. Each data point represents the mean ± SD of 3 wells from a single representative experiment. 2  2  2  2  162  4.8.3.  Effect of PARP and Endonuclease Inhibition on H 0 -Induced E C Death 2  2  Ca / M g -dependent endonucleases are thought to be involved in the internucleosomal cleavage of D N A during apoptosis (Wyllie, 1980). Aurintricarboxylic acid, an inhibitor of C a 2+  stimulated endonuclease activity (Jones et al, 989; Ray et al, 1992), was examined to determine if it would diminish oxidant-induced E C death. The effect of the combination of endonuclease inhibition and P A R P inhibition on the time-course and extent of E C injury following H 0 2  2  exposure is shown in Fig. 50. Inhibition of P A R P with DIQ, or treatment with aurintricarboxylic acid, delayed the onset and diminished the extent of oxidant-induced cell lysis. However, in the presence of H 0 , the combined treatment of DIQ + aurintricarboxylic acid further attenuated the 2  2  loss of membrane integrity compared to treatment with either DIQ or aurintricarboxylic acid alone. The time-course of loss of membrane integrity in control E C treated with DIQ or aurintricarboxylic acid alone, in the absence of oxidant (data not shown), was no different than non-treated control cells. The ability to further protect against cell lysis with aurintricarboxylic acid was suggestive of an active mode of cell death in oxidant-injured E C following P A R P inhibition.  163  100n  Time (h) Fig. 50. Effect of P A R P and endonuclease inhibition on the time-course of H 02-induced E C death. Confluent cells were treated with DIQ (10 pM), aurintricarboxylic acid (ATA, 100 pM), or DIQ+ATA, followed by the addition of 100 p M H 0 . Non-treated controls (NTC) received vehicle (i.e., water). Loss of plasma membrane integrity was monitored continuously over the course of the experiment with the use of the plasma membrane-impermeant fluorescent probe, YO-PRO. Each data point represents the mean ± SD of 4 wells from a single representative experiment. 2  2  2  164  4.8.4.  Effect of Staurosporine on the Time-Course and Extent of E C Death Staurosporine is a protein kinase C inhibitor commonly used to initiate apoptosis. It was  used in these studies as a positive control for apoptosis. Therefore, as a comparison to E C treated with H2O2, the time-course and extent of death in E C treated with increasing concentrations of staurosporine were determined (Fig. 51). Upon treatment with 5 p M staurosporine, E C began to lose plasma membrane integrity by 1 h. The onset of cell lysis was progressively delayed with exposure to lower concentrations of staurosporine. The extent of cell death was concentration dependent and varied with the time of assessment. However, after 24 h, 100% cell death was achieved with all concentrations tested. 120-.  .20-1 0  ,  5  ,  1  10  15  ,  20  ,  25  Time (h) Fig. 51. Time-course of irreversible E C injury after exposure to increasing concentrations of staurosporine. Confluent E C were exposed to increasing concentrations of staurosporine and loss of plasma membrane integrity was monitored using the plasma membrane-impermeant fluorescent probe, PI. Non-treated controls (NTC) received vehicle (i.e., water). Each data point represents the mean ± SD of 3 wells in a single representative experiment.  165  4.9.  Caspase Activation in Endothelial  Cells  4.9.1.  Caspase-3-like Activity in Oxidant-Stressed E C To test whether the cell death occurring in oxidant-stressed cells following P A R P  inhibition was apoptosis, the activity of caspase-3 was examined in intact EC. The time-course of caspase-3-like activation along with the concurrent cell membrane integrity assessment of E C treated with 200 p M H2O2 in the absence or presence of a P A R P inhibitor are shown in Fig. 52. Following P A R P inhibition with DIQ, increased caspase-3-like activity occurred prior to the loss of plasma membrane integrity, with a peak activity at 4 h. The subsequent decline in caspase-3like activity over the period of 5-8 h coincided temporally with the loss of plasma membrane integrity. In contrast, caspase-3-like activity in cells treated with H2O2 alone was negligible. Similar results for the time-course of caspase-3-like activity and loss of membrane integrity were obtained in cells treated with 200 p M H2O2 in the absence or presence of 1 m M 3AB (data not shown). The inset in Fig. 52A shows the combined total caspase-3-like activity from 3 independent experiments. Oxidant-stressed E C treated with either of the P A R P inhibitors, DIQ or 3AB, had dramatically increased caspase-3-like activity compared to cells treated with H2O2 alone. Treatment with DIQ or 3AB alone (data not shown) had no effect on increasing caspase3-like activity compared to the non-treated control cells. The intensity of the damaging insult may determine whether a cell undergoes apoptosis or oncosis. Therefore, the relationship between total caspase-3-like activity and H2O2 concentration in the absence or presence of P A R P inhibition was examined (Fig. 53). Although an inverse relationship between caspase-3-like activity and H2O2 concentration was observed, the extent of caspase-3-like activation in cells treated with H2O2 alone was much less compared to PARPinhibited cells at every H2O2 concentration tested. This increase in caspase-3-like activity  166  suggests a dramatic change in the process by which oxidant-stressed cells die, and is consistent with an active mechanism of apoptosis.  167  Fig. 52. Comparison between the time-course of caspase-3-like activation (A.) and loss of cell membrane integrity (B.) in PARP-inhibited E C treated with H 0 . A parallel assessment of caspase-3-like activity and cell injury was performed in E C treated with 200 uM H2O2 in the absence or presence of 50 uM DIQ. Non-treated controls (NTC) received vehicle (i.e., water). Caspase-3-like activity was measured using the ApoAlert CPP32 Assay Kit every hour between 1 and 12 h. Membrane integrity was monitored continuously over the course of the experiment using the plasma membrane-impermeant fluorescent probe, PI. Each data point represents the average + SD of 3 wells from a single representative experiment. Inset graph: total caspase-3like activity in E C treated with 200 uM H2O2 in the absence or presence of DIQ (50 (iM) or 3AB (1 mM). Caspase-3-like activity was also assessed in control E C treated with DIQ or 3AB in the absence of H 0 (data not shown) and was no different than that in N T C . Data are expressed as total caspase-3-like activity by determining the area under the caspase-3-like activity versus time curve for each experimental group from 0 - 6 h after oxidant exposure. Values for caspase-3-like activity are derived from triplicate determinations in three independent experiments. 2  2  2  2  168  A.  1400-  • H0 A H 0 + DIQ A DIQ 2  1200-  2  2  2  > 2^ 1000._ •*-•  c  CD  o  o o < o 800-  *6  600-  CO © • w <D CO  w 2 400(0 u Q. _E  TO o ~  2000-  -200-  0  i  6  8  10  12  B.  T i m e (h)  169  6000-1  H 0 + DIQ  m^mm  >  Z£  2  2  H0  2  2  J  <D Z 4000H  5 (!) CD § s 2000H  co  (0 CD (/) c (C — o  g_  o 0  J  50  100  200  300  400  [ H 0 ] (uM) 2  2  Fig. 53. Relationship between H2O2 concentration and total caspase-3-like activity in PARPinhibited and non-PARP-inhibited EC. Cells were exposed to 50 - 400 uJVI H2O2 in the absence or presence of 50 u M DIQ. Caspase-3-like activity was measured using the ApoAlert CPP32 Assay Kit over the course of the experiment. Data are expressed as total caspase-3-like activity by determining the area under the caspase-3-like activity versus time curve for each experimental group from 0 - 24 h after oxidant exposure. Values for caspase-3-like activity are derived from triplicate determinations in a single representative experiment.  170  4.9.2.  Caspase-3-like Activity in E C Treated with Staurosporine In parallel experiments, the time-course of caspase-3-like activation along with the  concurrent plasma membrane integrity were assessed in EC treated with either 1.25 or 5.0 (iM staurosporine (Fig. 54). EC treated with 5.0 \iM staurosporine began to lose plasma membrane integrity by 1 h. In the corresponding experiment, caspase-3-like activity reached a maximum at 3 h when there was approximately 50% loss of plasma membrane integrity. In contrast, EC treated with a lower concentration of staurosporine (1.25 uM) showed a delayed onset of cell lysis beginning at 6 h. The peak caspase-3-like activity in EC treated with 1.25 uM staurosporine occurred prior to loss of membrane integrity in 50% of cells. Although the pattern of peak caspase activity prior to loss of membrane integrity in 50% of cells was observed for both concentrations of staurosporine used the peak caspase-3-like activity in EC treated with 1.25 fiM staurosporine was approximately 2.5-fold greater compared to 5.0 | i M staurosporine-treated cells. Fig. 55 shows the relationship between staurosporine concentration and total caspase-3like activity. A concentration-dependent increase in total caspase-3-like activity was observed with decreasing concentrations of staurosporine. By comparison, in a side-by-side experiment, the total caspase-3-like activity in EC treated with 200 (iM H2O2 in the presence of DIQ is also shown. Total caspase-3-like activity in EC treated with H2O2 + DIQ was of approximately the same magnitude as EC treated with 0.3125 ixM staurosporine, and was greater than that observed for all other concentrations of staurosporine.  171  A.  600n 500H  >o  • SS1.25uM A SS5.0 uM  400H  < >-  _g <§ "T CD  300-  «? $ 200Qi £ W £ o 100(0 c Q. (0 2 (0 ^  O  o-  -1000  B.  100  —r-  2  4  6 Time (h)  8  10  12  8  10  12  * SS5.0 U.M  • SS 1.25 uM  1"  • NTC  o O a> Q  -200  —r  -  4  6 Time (h)  Fig. 54. Comparison between the time-course of caspase-3-like activation (A.) and loss of cell membrane integrity (B.) in E C treated with staurosporine. A parallel assessment of caspase-3like activity and cell injury was performed in E C treated with 5.0 or 1.25 uJVl staurosporine (SS). Non-treated controls (NTC) received vehicle (i.e., water). Caspase-3-like activity was measured using the ApoAlert CPP32 Assay Kit every hour between 1 and 12 h. Plasma membrane integrity was monitored continuously over the course of the experiment using the plasma membrane-impermeant fluorescent probe, PI. Each data point represents the mean ± SD of 3 wells from a single representative experiment.  172  3000 >  n  2500-  a> z 2000 CO  C)  d) a)  w  1500-  CO  (0 CD  1000-^  CO c  ro — iS 500o  05.0  2.5  1.25  0.625  [Staurosporin]  0.3125 (uivr)  H 0 + DIQ 2  2  Fig. 55. Concentration dependence of total caspase-3-like activity in E C treated with staurosporine. E C were treated with 0.3125 to 5.0 p M staurosporine. For comparative purposes, caspase-3-like activity was also assessed in E C treated with 200 p M H2O2 in the presence of 50 p M DIQ. Data are expressed as total caspase-3-like activity by determining the area under the caspase-3-like activity versus time curve for each experimental group from 0 - 24 h after exposure. Values for caspase-3-like activity are derived from triplicate determinations in a single representative experiment.  173  4.10.  Caspase-Induced Cleavage of PARP in Oxidant-Injured EC  4.10.1. Determination of Cell Number Required for Western Blot Analysis of P A R P An immunoblot analysis of PARP in E C is shown in Fig. 56. The PARP protein band was detected with as little as 25,000 EC. To obtain a strong signal of PARP, and its 89 kDa cleavage product, 200,000 E C were analyzed routinely.  P A R P (ng) 67  34  # E n d o t h e l i a l C e l l s (x 1 0 E 3 ) 10  25  50  1 00  1 35  1 75  225  116 k D a Lane  1  8  Fig. 56. Determination of the number of E C required for immunoblot analysis of PARP. A Western blot for native PARP was performed with increasing numbers of E C (lanes 3 - 9) to determine the cell number required to detect PARP. Purified PARP (lane 1, 67 ng; lane 2, 34 ng).  4.10.2. Western Blot of P A R P Cleavage in Oxidant-Injured E C As a confirmatory measure of caspase-3-like activity, the time-course of PARP cleavage in oxidant-stressed E C was assessed (Fig. 57). In E C treated with oxidant alone, only intact PARP was evident over the time-course of experimental exposure (1 to 8 h). No PARP cleavage was observed up to 4 h in PARP-inhibited cells treated with oxidant. However, the 89 kDa PARP cleavage product was evident from 5 to 8 h after oxidant exposure in E C treated with the PARP inhibitor. The intensity of the 89 kDa PARP fragment increased from 5-8 h concomitant with a decrease in the intact PARP band. Cleavage of PARP by caspase-3-like activity occurred before  174  the majority of cells had lost membrane integrity (Fig. 52 B). Appearance of the apoptotic PARP fragment confirmed caspase-3-like activity in PARP-inhibited E C undergoing an oxidant stress.  C P3 PARP Cleavage 45 ng P r o d u c t 1  Time P o s t - E x p o s u r e (h)  116 k D a  200 u.M H 0 .  116 k D a 89 kDa  200 U.M H 0 , + 50 uM DIQ  89 kDa  B.  Lane  2  1  10  Fig. 57. PARP cleavage in oxidant-stressed E C treated with a PARP inhibitor. E C were treated with 200 u M H 2 O 2 in the absence (A.) or presence (B.) of 50 | l M DIQ. PARP cleavage was assessed hourly by immunoblot analysis of cell extracts (lanes 3 - 10). Caspase-3-like activity cleaves intact PARP ( M R =116 kDa) into two fragments of 89 and 24 kDa. The anti-PARP C210 clone used in the immunoblot analysis recognizes the 116 and 89 kDa proteins. Purified PARP (45 ng, lane 1) and the purified 89 kDa apoptotic P A R P fragment (lane 2) were included as standards.  175  4.11.  Morphological Assessment of Apoptosis in EC To further examine the mechanism of cell death following oxidant exposure with and  without P A R P inhibition, morphological indicators of apoptosis were examined over an 8-h experimental period. Fluorescent micrographs of E C 4 to 5 h following treatment with H2O2 alone are shown at high and low magnification in Fig. 58A and B. Consistent with the data in Fig. 52B, approximately half of the oxidant-stressed E C at this time-point showed an orange-red fluorescence of the nucleus due to PI staining, which indicated loss of membrane integrity. The green fluorescence of the nuclei in the remaining oxidant-stressed cells was due to A O staining and indicated intact, viable cells. The chromosomal material of E C treated with oxidant alone appeared rough, wispy and clumpy, with non-descript, raggedy borders, and a thin margination around the edge of the nuclear membrane. In contrast, after 4 to 5 h of incubation in H2O2 and the P A R P inhibitor, DIQ, the E C had not lost membrane integrity as evidenced by the lack of red nuclear staining by PI and green fluorescent staining of the nucleus with A O (Fig. 58C and D). Furthermore, the nuclear material exhibited a dense, globular appearance with sharp, welldefined edges existing as multiple and single globules, consistent with other reports of the appearance of nuclei in E C apoptosis (Thomas et al, 1998). In addition, some of the nuclei showed the characteristic half moon-shaped pyknotic chromatin of apoptotic cells (Majno and Joris, 1995). At later time points (data not shown), when these cells underwent apoptotic necrosis and.lost plasma membrane integrity, the nuclear material showed the same dense globular appearance but was stained red by PI. E C incubated for 11 to 13 h with staurosporine, a recognized initiator of apoptosis, showed similar nuclear morphology to the oxidant-stressed E C treated with a P A R P inhibitor (Fig. 58E and F).  176  A striking feature of oxidant-stressed cells treated with DIQ was the intense red fluorescence in the cytoplasm (Fig. 58C and D), which was not present in cells exposed to H2O2 alone (Fig. 58A and B). This red fluorescence is due to active (ATP-dependent) uptake and concentration of A O in acidic organelles such as lysosomes and golgi (Zelenin, 1993). E C incubated with staurosporine (Fig. 58E and F), the positive control for apoptosis, showed a similar red cytoplasmic fluorescence to E C treated with H2O2 and the P A R P inhibitor (Fig. 58C and D). Control E C treated with the P A R P inhibitor alone (Fig. 58G and H) also exhibited the orange-red cytoplasmic fluorescence of lysosomal uptake of A O , as well as a diffuse, dullstained, green nuclear material that was indistinguishable from non-treated control cells (data not shown). These cells did not exhibit any fluorescence in the absence of A O and/or PI indicating that the fluorescence was not due to DIQ. In addition, control E C treated with the PARP inhibitor alone maintained the same morphologic characteristics throughout the duration of the experiment. Interestingly, the intense red lysosomal fluorescence in the cytoplasm of staurosporine-treated and PARP-inhibited oxidant-stressed E C is more globular and less punctate than that of the control cells, which is suggestive of lysosomal fusion in cells undergoing apoptosis. Finally, the nuclear and cytoplasmic morphology observed in individual cells under high magnification (Fig. 58A, C, E and G) was representative of the population of cells as viewed under lower magnification (Fig. 58B, D, F and H).  177  Fig. 58. Morphological assessment for apoptosis and oncosis in oxidant-stressed E C following P A R P inhibition. E C were exposed to H2O2 alone (A and B), H2O2 in the presence of the PARP inhibitor, DIQ (C and D), or 0.612 p M staurosporine (E and F), which acted as a positive control for initiation of apoptosis and its morphological characteristics. Control E C were treated for 2 h with DIQ alone (G and H). Prior to visualization under the fluorescence microscope, A O and PI were added to the cells. The morphology of DIQ-treated controls was maintained throughout the course of the experiment (up to 10 h) and was not distinguishable from non-treated controls (data not shown). Panels (A, C, E and G) and (B, D, F and H) represent cells magnified 320X and 200X, respectively. The bar in each panel represents 10 pm.  178  179  4.12.  The Effect of PARP Inhibition on [Ca 7, in Oxidant-Stressed +  EC  Calcium is an important intracellular messenger involved in a wide array of cellular processes, including signal transduction, muscle contraction, cell proliferation and gene transcription (Berridge et al, 1998). Calcium is also involved in cell death, both oncosis (Farber, 1990; Kristian and Siesjo, 1998) and apoptosis (McConkey and Orrenius, 1997; Nicotera and Orrenius, 1998). Intracellular calcium signaling is a highly regulated process, involving both internal (endoplasmic reticulum, mitochondria) and external stores that are regulated through voltage-operated, receptor-operated and store-operated channels (Berridge et al, 1998). The intracellular and extracellular free calcium concentrations are normally approximately 0.1 and 1000 pmol/L, respectively, due, in part, to the sequestration by intracellular stores of calcium, and, in part, to the energy-dependent extrusion of the C a - H exchanger (Kristian and Siesjo, 2+  +  1998). Because P A R P inhibition during an oxidant stress appeared to maintain energy-dependent lysosomal uptake of A O , the effect of P A R P inhibition on [Ca ]i, another energy-dependent 2+  process, was assessed. The effect of P A R P inhibition on [Ca ]i during an oxidant stress was 2+  measured as the 334/380 nm fura-2 fluorescence ratio (Fig. 59). Following treatment of E C with H2O2 alone, there was a continuous, sharp increase in [Ca ]j that both preceded the onset of, and 2+  occurred in conjunction with, the decline in viable cells. When oxidant-stressed E C were treated with DIQ, [Ca ]i remained at normal cellular levels, rising only slightly when the cells began to 2+  lose viability. There was also a delay in the onset of loss of viability in PARP-inhibited cells, consistent with Fig. 52B. Therefore, in E C undergoing an oxidant stress, concomitant treatment with DIQ appeared to prevent the rapid rise in [Ca ]i and delayed the loss of viability. 2+  180  Time (min) Fig. 59. Intracellular free calcium concentration ([Ca ]j) in oxidant-stressed EC following PARP inhibition. EC were treated with 200 u M H2O2 in the absence (black lines) or presence (red lines) of 50 u M DIQ. The [Ca ]; was assessed by the dual-excitation fluorescence ratio method, using the Ca -sensitive fluorophore, fura-2. Cell viability was determined as described in section 3.15. A rise in [Ca ]i is represented by an increase in the ratio of the excitation wavelengths at 334 and 380 nm. The 334/380 nm ratio is noted by the narrow lines and the % viable cells noted by the thick lines. Non-treated controls and EC treated with DIQ alone did not show an increase in [Ca ]j over the course of the experiment. 2+  2+  2+  2+  2+  181  4.13.  The Effect of DIQ is Not a Result of Antioxidant  Activity  It is possible that the shift from oncosis to apoptosis, when DIQ was added concurrently with H2O2, was a result of DIQ acting as an antioxidant, thus decreasing the extent of oxidative damage and oncosis. It is well established that antioxidants, such as catalase, added at the onset of oxidant stress, prevent oxidative damage and cell death. To determine whether or not DIQ was acting as an antioxidant, it was added to E C long after oxidant exposure, when an antioxidant would no longer be effective at limiting cell death.  If DIQ showed an effect in  altering cell death, it would suggest that DIQ was acting "down-stream" of the initiating oxidant  2+ damage. Changes in [Ca ]j and caspase-3-like activity were examined when the DIQ was added long after the oxidant damage had occurred. Fig. 60 shows the time-course of changes in [Ca ]j 2+  and cell viability for E C that had been treated with either the P A R P inhibitor, or catalase, long after the initiation of oxidative stress. Prior to addition of DIQ, the [Ca ]j was sharply increased 2+  in oxidant-injured E C (Fig. 60A). However, with the addition of DIQ at approximately 2 h, the elevated [Ca ]i was reversed to normal values, and remained low until much later when the 2+  remaining cells began to lose viability (Fig. 60B.). By comparison, the late addition of catalase failed to alter the rising [Ca ]j (Fig. 60A.) or delay the onset of loss of viability (Fig. 60B.). The 2+  time-course of cell viability and the rise in [Ca ]; mirrored that of E C treated with H2O2 alone 2+  (Fig. 59), suggesting that the antioxidant, catalase, had no effect when added long after the oxidant injury was initiated. Fig. 61 shows that the late addition of DIQ (i.e., 2 h after oxidant exposure) resulted in a dramatic increase in total caspase-3-like activity over the course of the experiment compared to E C treated with oxidant alone. Further, the increase in total caspase-3like activity was comparable to that observed when DIQ was added at the start of oxidant  2+ exposure. The ability of DIQ to increase caspase-3-like activity and reverse the rising [Ca ]i,  182  when added long after the initiation of oxidant exposure, is strong evidence that. DIQ is acting "down stream" of the initiating damage of cell death and is not acting as an antioxidant.  183  A.  o  00  to  CO CO  500  Time (min) Catalase added X DIQ added  200  250  300  350  400  450  500  Time (min) Fig. 60. Effect of P A R P inhibition and antioxidant treatment on reversing the oxidant-induced rise in [Ca ]; and loss of cell viability. E C were treated with 200 u M H2O2, and the cells monitored for changes in [Ca ]i (A.) and cell viability (B.). When 5-10% of cells were observed to become non-viable, either DIQ (50 pM) (red lines) or catalase (100 U/mL) (blue lines) were added to the cells. [Ca ]j was assessed by the dual-excitation fluorescence ratio method using the Ca -sensitive fluorophore, fura-2, and the cell viability determined as described in section 3.15. A rise in [Ca ]i is represented by an increase in the ratio of the excitation wavelengths at 334 nm and 380 nm. E C treated with either DIQ or catalase alone did not show an increase in [Ca ]i over the course of the experiment. 2+  2+  2+  2+  2+  2+  184  6000n  4000H  2000H  H2^2 0 2  H0 + DIQ 0h 2  2  H0 + DIQ 2h 2  2  Fig. 61. Effect of addition of a P A R P inhibitor at 0 h or 2 h after the initial oxidant exposure on caspase-3-like activity. Caspase-3-like activity was measured using the ApoAlert CPP32 Assay Kit. Data are expressed as total caspase-3-like activity by determining the area under the caspase-3-like activity versus time curve for each experimental group from 0 - 12 h after oxidant exposure. Values for caspase-3-like activity are derived from triplicate determinations in a single representative experiment.  185  5.  DISCUSSION  The overall goal of this research was to investigate potential mechanisms for alteration of PARP activity and to examine the effect of modulating PARP activity on the cellular response to oxidative stress. Two experimental approaches were taken to investigate this research question: 1) competitive inhibitors of the substrate binding site were used to inhibit P A R P activity, and 2) the potential for alteration of PARP activity, through disruption of zinc fingers within the D N A binding domain, was investigated. Specifically, the question central to this thesis asked if alteration of PARP activity could influence the process by which cells died.  5. / .  Zinc Finger Metal Replacement and Modulation of PARP  Activity  One major goal of this thesis was to examine the potential for metal replacement in the zinc-fingers of P A R P as a way to modulate PARP activity and as a potential mechanism of metal toxicity. It is well established that structurally intact zinc fingers are required for P A R P D N A binding capacity and PARP activity. The hypothesis was that replacement of the zinc-finger zinc with an alternate metal ion will alter the function of the zinc fingers and have repercussions on DNA-binding capacity and therefore,on PARP activity and function within the cell. In order to address this question, considerable preliminary work was necessary, which included: 1) expression and purification of recombinant PARP as a source of purified enzyme for subsequent experiments, 2) optimization of an in vitro enzyme assay to assess P A R P activity, 3) development of a method to produce and purify a fragment of the intact enzyme that contained the zinc-finger region which was amenable to mass spectrometry, and 4) development of an analytical technique to measure metal-replacement in the zinc fingers of purified PARP and PARP in vivo. This series of developmental and in vitro investigations were to be followed up  186  with ex vivo experiments to examine the effect of metal replacement in the zinc fingers of PARP on cellular response to oxidative stress. Mass spectrometric analysis of metal replacement in the zinc fingers of PARP required that the intact enzyme be cleaved into smaller fragments more amenable to ionization. A method was developed to digest P A R P into smaller fragments and subsequently separate the fragments prior to analysis by mass spectrometry. The results showed that caspase-3 readily produced two fragments of PARP, a 24 kDa fragment containing the D N A binding domain, and an 89 kDa fragment containing both the automodification and the catalytic domains. Prior to mass spectrometric analysis, the PARP fragments had to be separated to obtain the purified 24 kDa fragment containing the D N A binding domain. Several approaches were taken to separate and purify the PARP fragments, with the stipulation that any separation procedure had to maintain the native metal configuration of the zinc-finger motifs. Initial attempts at separation of the PARP fragments were based on the size difference of the fragments. Ultrafiltration using a centrifugal membrane device was one technique that was evaluated for separation of caspase-3-digested PARP fragments based on differences in molecular weight of the fragments. It was reasoned that selection of the membrane with an appropriate molecular weight cut-off would allow passage of the 24 kDa PARP fragment while retaining the 89 kDa fragment, thus effectively and rapidly purifying the fragments. Initial attempts at separation of the fragments under non-denaturing conditions using a membrane with 50 kDa N M W L were unsuccessful. The lack of success with this technique was thought to be due to aggregation or non-specific association of the fragments that prevented their separation. Therefore, a separation under denaturing conditions using 8 M urea was undertaken. It was recognized that these conditions would likely disrupt metal binding to the zinc finger. However, the denaturing conditions would restrict non-specific association of the fragments, therefore,  187  indicating the feasibility of the technique, as well as providing a sample of the purified 24 kDa fragment that could be used for development of a mass spectrometric analytic procedure. However, the separation of P A R P fragments under denaturing conditions using centrifugal ultrafiltration was also unsuccessful. There are several possible reasons why centrifugal ultrafiltration did not work as expected. These filtration devices have been designed primarily for desalting and buffer exchange processes, and it is generally suggested that the N M W L should be 3 times larger than the size of the molecule that will be passing through the membrane. The N M W L or molecular weight cutoff refers to the molecular weight of a globular solute at which the solute is 90% rejected (Amicon, 1995). However, separation of the 24 kDa fragment from the 89 kDa fragment was not achieved even when a membrane with 100,000 N M W L was used. Factors other than size, such as degree of hydration, aggregation, conformation and electrical characteristics of the solute, will affect retention by the membrane. Therefore, the N M W L rating is not an absolute indication of the fractionation characteristics of the membrane. Furthermore, an effect called concentration polarization can add an additional layer to the filtration process (Amicon, 1995). During centrifugation, the solvent and the solute are forced onto the membrane surface resulting in an accumulation of retained solute, which eventually leads to the formation of a gel layer, or secondary membrane. At moderate to high concentrations of solute, the resistance of the gel layer is significantly greater than that of the membrane, and flux (ultrafiltration flow rate) becomes independent of membrane permeability. Therefore, due to concentration polarization, the effective permeability of the membrane may be much lower than indicated by the N M W L rating. After consideration of these various factors, it was concluded that centrifugal ultrafiltration was not an appropriate method for separation and purification of the P A R P fragments.  188  Affinity chromatography was investigated as a method to separate the P A R P fragments as it was already in use for purification of P A R P from cell extracts. The affinity matrix contained a PARP-specific ligand, 3 A B , that binds PARP through interaction with the catalytic site. Therefore, it was reasoned that separation of the 24 kDa fragment, containing the zinc-finger region, from the substrate binding region could be accomplished by the differential interaction of the fragments with the affinity matrix. SDS-PAGE was used to detect changes in the protein profile with differential binding of the PARP fragments to the affinity resin. It was hypothesized that the affinity matrix would selectively bind the 89 kDa fragment, and the 24 kDa fragment would remain unbound in solution. If this were true, SDS-PAGE analysis would reveal the disappearance of the 89 kDa fragment from the solution, concomitant with no change in intensity of the 24 kDa fragment. The experiments performed revealed that the differential binding of the caspase-3 cleavage products of PARP to the 3AB-Affigel affinity matrix was not achieved under conditions of varying salt concentration or amount of affinity resin used (Fig. 27, page 117). Using two different preparations of 3AB-Affigel, and optimal buffer conditions for binding full-length PARP (i.e., 300 m M KCI in the buffer) (Fig. 29, page 119), the differential binding of the 89 kDa fragment was not achieved (Fig. 28, page 118). Therefore, the binding characteristics of the different preparations of the Affigel were ruled out as possible reasons for the lack of binding of the 89 kDa fragment. These results were unexpected, but nonetheless demonstrated that the separation of the P A R P fragments could not be accomplished using affinity chromatography. The ability of both batches of 3AB-Affigel affinity matrix to bind full-length PARP, but not the 89 kDa fragment,  suggested that cleavage of PARP by caspase-3 confers a  conformational change in the large fragment rendering it unable to bind to the substrate analogue, 3AB, and possibly N A D . Therefore, these findings initially suggested that the 89 kDa +  189  fragment, which contains the catalytic domain, is unable to bind to the substrate when it has been cleaved by caspase-3. However, this conclusion is in disagreement with reports demonstrating that while the P A R P catalytic domain is unable to be stimulated in the presence of D N A , it retains basal ADP-ribosylation activity. A deletion analysis of the carboxy-terminal domain of human PARP determined that the 40 kDa C-terminal region of the enzyme was an autonomous catalytic domain (Simonin et al, 1990). This domain exhibited the same affinity for N A D as the +  full-length enzyme, catalyzed the initiation, elongation and branching of ADP-ribose polymers, but exhibited no D N A dependence. The 40 kDa fragment was found to have a 500-fold lower specific activity than that of the whole enzyme activated by D N A strand breaks.  Furthermore,  the 40 kDa P A R P fragment was shown to retain basal catalytic activity when immobilized on nitrocellulose and when in solution. Shah et al. (1995) developed an activity-Western blot technique to detect ADP-ribose polymers on automodified P A R P in cell extracts following electrophoresis and transfer to a nitrocellulose membrane. Renatured immobilized PARP was incubated with N A D to permit polymer synthesis, and detection of the polymer was achieved +  using anti-poly(ADP-ribose) antibodies. Using this method, it was demonstrated that the 89 kDa PARP fragment generated in apoptotic cells retained the ability to generate poly(ADP-ribose). The addition of nicked D N A during the activity blotting procedure resulted in an increased modification of the full-length enzyme, but not of the 89 kDa fragment. These investigations suggested that the 89 kDa apoptotic PARP fragment retained some ADP-ribose polymer synthesis capability, at least when immobilized on nitrocellulose, and that the 40 kDa carboxyterminal portion of the enzyme was the minimum unit required for catalytic activity. The possible reasons for the discrepancy between published studies and the results presented in this thesis warrant some discussion. In the present work, the inability to bind the 89 kDa fragment differentially to the 3AB-Affigel may reflect a conformational change in the  190  binding site that prevents competitive inhibitors, such as 3 A B , from binding. However, conclusions regarding the ability of the 89 kDa fragment to bind to the substrate, N A D , cannot +  be made based on this data. The ability of the 89 kDa fragment to catalyze ADP-ribosylation reactions when immobilized on nitrocellulose (Shah et al, 1995) suggests that this fragment is able to bind to the substrate at some basal level. However, the catalytic activity of the apoptotic 89 kDa P A R P fragment has yet to be demonstrated in solution. The ability of the 40 kDa fragment to bind substrate and be catalytically active in solution (Simonin et al, 1993), and the inability of the 89 kDa fragment to bind the 3AB-Affigel in the present work, may suggest that the longer fragment folds over in solution blocking the substrate binding site. When the 89 kDa fragment is immobilized on nitrocellulose, this folding over might be prevented, and therefore,following renaturation, the fragment still exhibits ADP-ribosylation activity (Shah et al, 1995). It remains to be seen if the 89 kDa PARP fragment demonstrates ADP-ribosylation activity in solution following caspase cleavage in response to an apoptotic stimulus in cells. It could be postulated that the cleavage of P A R P at the caspase-3 cleavage site, although far removed from the catalytic site, could affect conformation throughout the molecule and prevent substrate binding. Alternately, cleavage of PARP separating the D N A binding domain from the rest of the molecule may prevent any conformational change that PARP undergoes following binding to a D N A strand break, resulting in enzyme activation. In the present work, the finding that the 89 kDa PARP fragment was unable to bind to a substrate analogue in solution agrees with the possibility that PARP is inactivated by caspase-3 during the execution phase of apoptosis. This finding needs to be explored further in apoptotic cells where PARP is cleaved. The purpose of developing a separation method for the PARP fragments was to facilitate the mass spectrometric analysis of the 24 kDa fragment. Because neither centrifugal  191  ultrafiltration nor affinity chromatography were successful as separation methods, the decision was made to separate the P A R P fragments by H P L C under acidic conditions and collect fractions corresponding to the eluting fragments. This provided a way of obtaining a pure 24 kDa fragment that could then be used to evaluate the electrospray mass spectrometric method. Separation of the PARP fragments was readily achieved by H P L C under acidic conditions (Fig. 30, page 122). Although separation of the PARP fragments was accomplished using H P L C under acidic conditions, the purified 24 kDa fragment was likely present as the apo-protein without intact zinc ions. Therefore, it was still desirable to achieve separation of the fragments under non-acidic conditions to maintain native conformation of the protein. H P L C is considered the best method for separation of the PARP fragments because less manipulation of the protein is required and the liquid chromatograph can be interfaced directly to the mass spectrometer for immediate analysis of the protein. A number of approaches toward designing a mobile phase system were attempted that would satisfy both H P L C separation and mass spectrometry compatibility demands. Solvent systems that seemed promising for protein ionization under neutral conditions in the mass spectrometer (HFIPA, ammonium bicarbonate) did not provide adequate elution characteristics for the model proteins or the P A R P fragments. In the absence of buffer salts, the pH of the mobile phase appears to be the key determinant for protein elution from a reverse phase column. Elution of the PARP fragments was achieved in a mobile phase consisting of acetonitrile and ammonium formate, pH 3.5, but not when the pH was increased to 4.3. In conclusion, reverse phase chromatography of proteins under neutral conditions proved difficult, especially when the separation conditions were limited by the solvent requirements of the mass spectrometer. Based on the solvent systems and chromatography columns evaluated, it seems unlikely that a mobile phase will be found which satisfies: 1) the pH dependency for  192  elution from the chromatography column; 2) the pH requirement for maintenance of the zincfinger motif; and 3) the requirements for chemical compatibility for ionization in the mass spectrometer. It was essential to develop an analytical method to assess the metal content in the zinc fingers of PARP in the planned studies. Mass spectrometry appeared to provide the sensitivity required to address this type of analytical question. Several approaches were taken in the development of a technique to assess metal content in PARP, including direct measurement of the metal within the protein (LA-ICPMS), as well as assessment of molecular weight changes between the zinc-intact and metal-replaced PARP (MALDI-TOF M S , and ESI-MS). Initially, the metal content in a PARP protein band immobilized on an acrylamide gel was analyzed by L A - I C P M S . It was projected that this method could be adapted for use in examining metal incorporation into PARP in growing cells. Cell extracts from metal-exposed cells would be subjected to electrophoresis, and the metal content of the PARP protein band assessed by L A ICPMS (Fig. 6, page 66). Analysis of zinc isotopes in purified P A R P protein bands immobilized on a SDS-polyacrylamide gel revealed that the amount of zinc measured was independent of the amount of protein loaded. In addition, the background levels of zinc isotopes were, in most cases, as high or higher than those detected in the protein band (Fig. 31, page 128). It is likely that the denaturing condition of the gel system caused ejection of the zinc ions from the finger motif. Therefore, the zinc levels measured in the protein bands were essentially no different than the background levels. ICPMS analysis of zinc was performed on a solution of P A R P to determine the detection limit for zinc in a protein in solution. The most abundant zinc isotope, Z n , was used for this 64  analysis because it gave the most linear response between the samples. A linear relationship between Z n  64  abundance and PARP concentration was observed (Fig. 33, page 130). However,  193  due to the high background levels of zinc in the buffer, the detection limit for Zn  was  determined to be between 10-50 pg/mL of P A R P protein. The slope of the line from the standard curve indicated that the quantity of zinc detected per pg of protein was 0.0011 pg Zn / pg PARP protein. Relating this detection limit back to the analysis of zinc in a PARP band immobilized on a gel, it would appear that if 1 pg of PARP was loaded on a gel, and the entire sample was ablated, then it would be expected to measure 1.1 ng of Zn. However, usually only a 5 pm section of the sample is ablated, and therefore,the amount of zinc in the ablated material would be well below the detection limit for zinc by ICPMS. Development of a LA-ICPMS method to analyze zinc content of a PARP protein band isolated by SDS-PAGE did not appear feasible due to the denaturing conditions of the electrophoretic system and subsequent ejection of the zinc from the finger motif. It is possible that a non-denaturing gel system would solve the problem of metal ejection. However, there is some concern as to whether analysis of PARP bands by LA-ICPMS would provide the sensitivity required to measure zinc in the ablated sample. An alternate approach to examining metal content of PARP is to compare the molecular weights of the apo-, zinc-intact and metal-replaced enzyme. The changes in molecular weight due to the presence of an alternate metal in the zinc fingers would allow identification of that metal and determination of the extent of replacement. Two mass spectrometric methods were evaluated for this type of analysis: M A L D I - T O F MS and ESI-MS. Both methods are recognized for their ability to generate accurate molecular mass information on intact proteins (Scobie et al, 1993). Due to the high mass limit of M A L D I - T O F M S , both full-length PARP and the caspase-3digested P A R P fragments were analyzed. The matrix used for analysis was sinapinic acid, which  194  was prepared as a saturated solution in E ^ O i A C N (2:1) containing 0.1% T F A . The molecular weight determination for full-length PARP by M A L D I - M S was within approximately 0.5% of the calculated molecular weight based on amino acid sequence. It is likely that improved accuracy for the molecular weight measurement would be obtained with a more concentrated protein sample. The molecular weights determined for the caspase-3-digested P A R P fragments (23,998 and 89,101 Da) were well within the expected mass based on amino acid sequence (24,041 and 89,120 Da). However, due to the presence of T F A in the matrix mixture, the molecular weight of the PARP fragment containing the zinc fingers likely represents the apo-, or metal-free protein. In the absence of T F A in the matrix mixture, the molecular weight of the small PARP fragment was 24,565-24,572 Da. Without T F A present, the mass accuracy appeared to be within 2% of the expected value for this fragment. The increase in molecular weight (-528 Da) did not appear to represent the increased molecular weight due to stoichiometric binding of two zinc ions to the zinc-finger motifs (-130 Da). In the absence of T F A , the pH of the matrix mixture was still acidic (pH 3.4), and therefore,below the range for analysis of the metal-intact PARP fragment. Ionization of the small PARP fragment was not accomplished when the matrix mixture was buffered to pH 6.0, suggesting that the increase in pH suppressed protein ionization, or that the protein became adsorbed to instrument components and was not available for ionization. Hutchens et al. (1992b) used M A L D I - T O F M S to study the stoichiometric binding of copper ions to a 26 residue metal-binding motif, (GHHPH) G, using 2,5-dihydroxybenzoic acid 5  as the matrix. Although 2,5-dihydroxybenzoic acid has a pKa of 2.93, stoichiometric binding of exogenously-added copper ions to the 26 residue peptide was observed by M A L D I - T O F M S . This study suggests that the detection of protein-metal complexes is possible, at least when metal  195  ions are added exogenously. When present at a relatively high concentration in the sample matrix, a metal ion would compete effectively with hydrogen ions for protein-binding sites, even at low pH. However, the ability to detect metalloproteins in their native conformation, without addition of exogenous metals, has yet to be demonstrated by M A L D I - T O F M S . Hutchens et al. (1992b) also indicated that when the analysis was repeated under the identical experimental conditions (except that sinapinic acid was used as the matrix) there was little or no evidence of peptide-bound copper ions. Therefore, the type of matrix used for M A L D I - T O F M S analysis may influence the detection of metal-protein complexes. The results presented in this thesis suggest that although M A L D I - T O F M S provided adequate mass accuracy for the molecular weight determination of full-length PARP and the PARP fragments, the environment under which it was performed (acidic conditions, sinapinic acid matrix) did not permit analysis of metal-intact PARP. A more rigorous assessment of the native conformation of the PARP zinc fingers could be conducted with different matrices (i.e., 2,5-dihydroxybenzoic acid and nicotinic acid) to determine the potential of M A L D I - T O F M S in addressing this type of analytical problem. Another accurate and sensitive technique that provides molecular mass determination for intact proteins is ESI-MS (Scobie et al, 1993). The model proteins, trypsinogen, chymotrypsin, cytochrome c and myoglobin, were used for method development. Three of these proteins, trypsinogen, chymotrypsin and cytochrome c, were selected because of their size and basic characteristics and therefore,were thought to be good models for the DNA-binding domain of PARP. Myoglobin was used as a model protein because it is the "gold standard" for protein mass spectrometry and therefore.readily permitted evaluation of various electrospray solvent systems and instrument conditions. In addition, the molecular weights of these model proteins were in the range of the small P A R P fragment.  196  ESI-MS analysis of trypsinogen, chymotrypsin, cytochrome c, myoglobin and the 24 kDa PARP fragment was readily accomplished under acidic conditions. The mass of the PARP fragment determined by ESI-MS was 23,968.8 Da, which was within 0.3% of the mass calculated from the amino acid sequence (24,041 Da). This margin of error is greater than the 0.01% mass accuracy reported for ESI-MS using a double-focusing mass spectrometer (Scobie et al, 1993), but still within acceptable limits. The analysis of intact metal-binding proteins by ESI-MS poses a more difficult analytical problem due to the pH requirements for maintenance of the metal-protein complex, as well as the ionization requirements of the mass spectrometer. The metal-binding properties of various metalloproteins, including calcium-binding proteins (Hu et al, 1994; Lafitte et al, 1995; Veenstra et al, 1997), metallothioneins (Yu et al, 1993), and various zinc-finger proteins (Surovoy et al, 1992; Hutchens et al, 1992a; Witkowska et al, 1995; Craig et al, 1997), have been accomplished by ESI-MS. Therefore, ESI-MS appeared to be a promising technique for analysis of metal-binding in the zinc fingers of PARP. A number of buffer systems that were compatible with the mass spectrometer and permitted pH regulation were evaluated. It immediately became apparent that the signal obtained for a basic protein (trypsinogen) in a neutral buffer system (ammonium acetate) was dramatically reduced compared to the signal obtained in an acidic solvent (Fig. 37, page 138). The loss of sensitivity when protein ESI-MS analysis was performed at neutral pH became a consistent finding during the development of a method to analyze P A R P zinc fingers. The decrease in sensitivity may be related to two factors: 1) the decreased ionization of the analyte at a nonacidic pH; and 2) ionization of abundant buffer components in preference to the less abundant protein species, which is referred to as ion competition. Triethylamine acetate and pyridine acetate, two solvents that had previously been used for analysis of zinc-finger peptides, also  197  exhibited this ion competition phenomenon, resulting in a complete absence of a protein signal from either myoglobin and cytochrome c (see section 4.7.3.3, page 147). One solvent system that showed promise for the analysis of the PARP zinc fingers was an ammonium bicarbonate (pH 8.0) / acetonitrile mixture. This solvent system had been used for the ESI-MS analysis of the native metal-binding properties of a zinc-finger protein, the vitamin D receptor DNA-binding domain (Craig et al, 1997; Veenstra et al, 1998) and the calciumbinding protein, calbindin (Veenstra et al, 1997). The report of calbindin ( M W = 29,866 Da) analysis by ESI-MS  was particularly interesting because it described the metal-binding  properties of a protein in the molecular weight range of PARP. Up to that point, much of the ESI-MS analysis had been performed on relatively low molecular weight peptides and proteins (< 12,000 Da). In the present work, the analysis of myoglobin, but not trypsinogen or the 24 kDa PARP fragment, was accomplished with the ammonium bicarbonate solvent system (see section 4.7.3.3, page 147). The lack of ionization of the 24 kDa PARP fragment under these conditions, and previous evidence suggesting that protein adsorbance to instrument components occurred under neutral conditions, prompted an investigation of whether the lack of P A R P signal was due, in part, to this protein "sticking" phenomenon. Injection of the 24 kDa P A R P fragment in the ammonium bicarbonate buffer, followed by a formic acid chaser, illustrates the adsorbance phenomenon (see section 4.7.3.4, page 154). The adsorbance of injected proteins to instrument components was observed to some extent for all protein species analyzed (i.e., myoglobin, trypsinogen, cytochrome c and the 24 kDa PARP fragment), and appeared to be particularly severe for some proteins, namely cytochrome c and the 24 kDa PARP fragment. Attempts to reduce the sticking by addition of mobile phase modifiers (i.e., metal chelators, individual amino acids) (section 4.7.3.4, page 154) were unsuccessful. 198  Further investigation of the relationship between solvent composition and protein adsorbance was performed, and suggested that, at least for cytochrome c, a critical determinant for adsorbance was the organic content of the solvent system. At relatively high organic content (i.e., 50%), cytochrome c did not become adsorbed to the instrument components, as any bound protein could not be displaced with a formic acid chaser. These results suggested an increased solubility of cytochrome c in this solvent. However, at this level of organic content, the ability to ionize cytochrome c appeared to be suppressed. It was concluded that while an increase in the organic modifier increases protein solubility in the solvent and reduces adsorption to the instrument components, the increased organic content seemed to suppress protein ionization and impede detection. Therefore, it is likely that in order to analyze a protein under non-acidic conditions, some protein adsorption to the instrument will occur. Furthermore, the phenomenon of protein adsorption at neutral pH is probably not unique to ESI-MS, and may have contributed to the decreased sensitivity observed with ICPMS as well as M A L D I - T O F M S . Therefore, assay development using any technique will have to take into account this potential loss of protein. The use of a solvent system containing a mixture of acetonitrile and HFIPA allowed for accurate molecular mass determinations for myoglobin (Fig. 44, page 152) and cytochrome c (Fig. 43, page 151), as well as the 24 kDa PARP fragment (Fig. 45, page 153). The raw spectrum of the 24 kDa PARP fragment contained more noise compared to the spectrum for cytochrome c. However, this did not impede accurate molecular mass determination for the PARP fragment because the molecular weight calculated for the fragment in a HFIPA solvent (23,969.6 Da) was virtually identical to that determined under acidic conditions (23,968.8 Da). This is the first time a spectrum and molecular weight estimation for the 24 kDa PARP fragment has been obtained in a non-acidic solvent system, and represents a major step forward in the development of a technique for analyzing the native conformation of the zinc fingers of PARP. However, the  199  solution of HFIPA (53 mM) was determined to have a pH of 5.1. Ideally, for the analysis of the native metal conformation of the 24 kDa PARP fragment, the pH should be at least 6 to preserve the majority of the histidine-metal binding in the zinc coordination site. Therefore, further developmental work is required to achieve optimum conditions for the analysis of the PARP zinc fingers in their native state. One caveat in the literature describing metal-protein binding properties is worth consideration: metal-binding stoichiometry was commonly evaluated by ESI-MS in the presence of added metal ion (i.e., the metal ion was added to the electrospray solvent and protein sample during analysis under neutral conditions) (Hutchens et al, 1992a; Witkowska et al, 1995; Veenstra et al, 1997; Craig et al, 1997; Veenstra et al, 1998). This practice enhanced metal binding to the protein and allowed detection of the protein with different amounts of metal bound. As discussed previously for M A L D I - T O F M S , the addition of exogenous metal ions to the sample probably encourages the binding of the metal to the protein. The ability to monitor the metal content of a metalloprotein extracted from cells or tissues from organisms exposed to a metal presents a more complex scenario than simply being able to measure metal-binding stoichiometry in vitro. To date, there is little evidence of the use of ESI-MS for this type of analysis.  200  5.2.  Biological  Consequences  of  Pharmacological  Inhibition  of  PARP  Activity The question central to this thesis asked if alteration of P A R P activity could influence the process of cell death following an oxidant stress. The first indication that the process of oxidantinduced cell death may be altered by inhibition of P A R P was revealed by the phenomenon of enhanced protection against cell death when PARP-inhibited E C exposed to H2O2 were treated with either an inhibitor of protein synthesis (Fig. 49, page 162), or an inhibitor of endonuclease activity (Fig. 50, page 164). These preliminary findings suggested that the process of cell death in oxidant-stressed E C treated with the P A R P inhibitor was different than in cells treated with oxidant alone. To further address this question, E C were subjected to oxidant injury from H2O2 in the presence or absence of a P A R P inhibitor, and cell death was characterized in terms of the process by which the cells died, oncosis or apoptosis. The studies presented here demonstrate that P A R P inhibition in oxidant-stressed E C can alter the process by which these irreversibly-injured cells are destined to die. Oxidant-injured E C undergo an oncotic mode of cell death, whereas inhibition of P A R P in these cells prevents oncosis and permits apoptosis. The shift of oxidant-stressed E C into apoptosis following PARP inhibition was demonstrated by increased caspase-3-like activity (Fig. 52, page 168 and Fig. 57, page 175), as well as by morphological assessment using fluorescent microscopy (Fig. 58, page 178). In order to interpret cellular morphology and caspase-3-like activity as indicating apoptosis, E C were also treated with a known initiator of apoptosis, staurosporine. In staurosporine-treated EC, caspase-3-like activity (Fig. 54, page 172 and Fig. 55, page 173) and cellular morphology (Fig. 58, page 178) were comparable to oxidant-stressed E C treated with a P A R P inhibitor.  201  Following exposure to D N A damaging agents, the predominant mode of cell death has been, found to depend upon the extent of damage, with oncotic death occurring after extensive damage and apoptotic mechanisms operating after moderate damage (Nosseri et al, 1994; Bonfoco et al., 1995; Coppola et al, 1995; Watson et al, 1995). The inverse relationship of toxin amount and extent of apoptosis was true in the E C model presented in this study (Fig. 53, page 170). However, inhibition of P A R P substantially increased the caspase-3-like activity regardless of the amount of oxidant stress placed on the cells. Interestingly, when E C were treated with staurosporine (Fig. 55, page 173), which is considered an initiator of apoptosis, the extent of apoptosis, as measured by caspase-3-like activity, was also found to be inversely related to staurosporine concentration. Therefore, high concentrations  of an  apparent  apoptogenic stimulus may actually induce oncosis. It could be argued that the increase in apoptosis caused by P A R P inhibition was a result of the inhibitor acting as an antioxidant, which would mimic a lower concentration of oxidant and theoretically result in less oxidative damage. Other investigators (Schraufstatter et al, 1986) have shown little or no antioxidant effect of the P A R P inhibitor, 3AB, and in the present study, a similar increase in caspase-3-like activity was observed with either DIQ or 3AB treatment (Fig. 52A, inset, page 168). Furthermore, the ability to increase capase-3-like activity (Fig. 61, page 185) and reverse the elevated [Ca ]i (Fig. 60, page 184) by inhibiting P A R P with DIQ long after 2+  the initial oxidant damage had occurred, is strong evidence that DIQ is acting "down-stream" of the initiating damage of cell death and not as an antioxidant. The availability of cellular energy, as A T P (Leist et al, 1997; Tsujimoto, 1997; Eguchi et al, 1997; Lelli, Jr. et al, 1998) or an elevated energy charge ratio (ATP + Vi A D P / ATP+ADP+AMP) (Thies and Autor, 1991), may be the critical factor determining the fate of the cell and the mode of cell death. A T P depletion by >50% in Jurkat cells, prior to exposure to  202  apoptosis-inducing agents, was sufficient to change the process of death from apoptosis to oncosis (Leist et al, 1997). Apoptosis proceeded in U937 human myeloid leukemia cells only after recovery of N A D levels following an initial drop after mild oxidant exposure (Nosseri et +  al, 1994). Recently, Lelli et al. (1998) reported that in oxidant-stressed EC, a threshold level of ATP 25% of the basal levels is required for apoptosis to proceed. Therefore, it appears that maintenance of energy stores is essential for an apoptotic mode of death. Following exposure to a variety of D N A damaging agents, the extent of P A R P activation and its effect on available cellular energy may dictate the process of cell death. In this research, it was hypothesized that oncosis predominates as a result of P A R P activation after a lethal oxidant injury (Fig. 4, page 30). However, if, P A R P activity is inhibited, cellular energy is maintained and an apoptotic mode of cell death (initiated at the onset of oxidant-induced damage) is allowed to proceed. P A R P inhibition concomitant with oxidant stress has been shown to prevent the decline in N A D and ATP in HT-29-18-C1 intestinal epithelial cells (Watson et +  al, 1995), murine thymocytes (Virag et al, 1998) and U937 cells (Nosseri et al, 1994; Coppola et al, 1995). It has been demonstrated previously that P A R P inhibition prevents the decline in N A D , ATP and the energy charge ratio in oxidant-stressed EC. Furthermore, when P A R P was +  inhibited long after oxidant exposure, the energy charge ratio returns to normal control levels (Thies and Autor, 1991). The results of the present study also indicate indirectly that cellular energy is being maintained in PARP-inhibited cells undergoing an oxidant-stress compared to cells treated with H2O2 alone. During oxidant injury and consequent depletion of cellular energy, the characteristic swelling of oncosis occurs due to inhibition of the energy-dependent Na /K -ATPase and C a +  +  2+  ATPase on the plasmalemmal surface, and the subsequent influx of water (Trump and Berezesky, 1996). In the present study, the maintenance of low [Ca ]; in oxidant-stressed E C 2+  203  treated concurrently with DIQ, but not in E C treated with oxidant alone (Fig. 59, page 181), suggests that the energy-dependent ion homeostatic processes remained operational during PARP inhibition. Perhaps even more convincing is the reversal of the elevated [Ca ]j in oxidant2+  stressed E C by inhibiting PARP as much as 2 h following the initial oxidant exposure (Fig. 60, page 184). Furthermore, the ability of lysosomes to concentrate A O (i.e., red fluorescence in the cytoplasm) requires an energy-dependent membrane-bound proton pump that maintains an acidic pH inside the lysosome (Zelenin, 1993). In E C treated with oxidant alone, the cytoplasm lacked the red lysosomal fluorescence evident in PARP-inhibited E C treated with an oxidant, apoptotic SS-treated cells, control cells treated with DIQ alone (Fig. 58, page 178) and in N T C cells (not shown). The normal functioning of the lysosomal proton pump in PARP-inhibited E C compared to E C exposed to H2O2 alone suggests that PARP inhibition protects cellular energy stores. Therefore, PARP inhibition either during or after an oxidant stress appears to preserve or restore, respectively, energy dependent processes. PARP degradation by caspase enzymes during the execution phase of apoptosis is one of the earliest protein cleavage events that occurs after high molecular weight fragmentation of chromatin, but before internucleosomal D N A fragmentation (Kaufmann, 1996; Greidinger et al, 1996). It could be hypothesized that PARP activity at this stage in apoptosis when D N A is systematically being degraded is undesirable because a decline in energy status of the cell, as a consequence of PARP activation, may compromise the cell's ability to complete the active process of apoptosis. Virag et al (1998) have recently suggested that caspase cleavage of PARP acts as a preventative measure aimed at conservation of cellular energetics to ensure completion of the apoptotic program. Therefore, PARP activity is precluded through its cleavage by caspase prior to the extensive D N A degradation that occurs in apoptosis. Alternatively, caspase-induced cleavage of P A R P may interfere with its ADP-ribosylation capabilities. As a consequence, the  204  function of PARP in the post-translation modification, and thus inhibition, of nuclear proteins involved in apoptosis (i.e., endonucleases) (Tanaka et al, 1984; Nelipovich et al, 1988; Rice et  al, 1992), or in D N A repair (Casciola-Rosen et al, 1996) could be compromised. Caspase cleavage of PARP may remove a potential point of control involved in the regulation of these apoptotic processes. The question is: what is the role of PARP in cell death? The results from this study suggest that active PARP does not have a role in apoptosis following an oxidative insult. In fact, PARP activation following oxidant injury results in oncosis and cell rupture. However, if PARP is inhibited, either through pharmacological or physical intervention (i.e., PARP cleavage by caspase), an apoptotic program is allowed to proceed. In fact, under the circumstances of oxidant-induced D N A damage, it could be suggested that P A R P inactivation must be ensured in order for apoptosis to proceed. Conversion of oxidant-induced oncosis to apoptosis following PARP inhibition has also been observed in thymocytes (Virag et al, 1998) and in human myeloid leukemia U937 cells (Palomba et al, 1996). Therefore, PARP activity appears to be a pivotal determinant in the mechanism of oxidant-induced cell death. In response to D N A damage, the activation of P A R P circumvents any underlying apoptotic signals, blocking apoptosis and directing the cell into oncosis (Fig. 4, page 30). Caspase-induced cleavage and resulting inactivation of PARP, suggest that PARP activity may be unnecessary or undesirable during apoptosis. Potentially, some of the strongest arguments for the lack of involvement of PARP in apoptosis come from studies of PARP knockout animals and cells. Nicotera's group (Leist et al, 1997) have compared the sensitivity of thymocytes, hepatocytes and neurons obtained from wild-type and PARP-null mice to undergo apoptosis. They demonstrated that there was no difference between wild-type and PARP-null cells in any of the apoptotic criteria measured. Therefore, the absence of PARP did not appear to  205  alter the ability of these cells to undergo apoptosis. Wang et al. (1997) have also suggested that PARP is dispensible for apoptosis because PARP -/- cells underwent apoptosis normally in response to a variety of apoptosis inducing agents including anti-Fas antibody, tumor necrosis factor a, y-irradiation, and dexamethasone. The rate and extent of apoptosis was examined in wild-type and PARP deficient (-/-) primary bone marrow cells that were exposed to the alkylating agent, methylnitrosourea (Oliver et al, 1998). Methylnitrosourea is a DNA-damaging agent that activates PARP catalytic function and initiates the base-excision repair pathway. PARP-deficient cells were drastically more sensitive to apoptosis induced by methylnitrosourea as compared to their wild-type counterparts. Taken together, these studies suggest that PARP activity is not required for apoptosis. To further clarify the role of PARP in mediating the process of oxidant-induced cell death, caspase activitation and nuclear morphology should be examined in PARP-null cells exposed to an oxidant stress. The results in this thesis suggest that PARP-deficient cells would behave similarly to wild-type P A R P cells exposed to a pharmacological inhibitor of PARP and preferentially undergo apoptosis following oxidantinduced D N A damage. The ability to modulate the process of cell death may have clinical relevance for diseases mediated by oxidative stress, such as inflammatory injury. Because inflammation is associated with oncosis and not apoptosis, switching from oncosis to apoptosis may limit cell and tissue loss that is a consequence of secondary inflammation. PARP activation has been implicated in oxidant injury of the heart (Gilad et al, 1997; Zingarelli et al, 1998) and in other disease states with an inflammatory component (Szabo, 1998). Pharmacological inhibition of P A R P has been shown to reduce local inflammation in rats (Cuzzocrea et al, 1998) and mice (Szabo et al, 1997), and limit tissue damage following ischemia/reperfusion injury (Lam, 1997; Thiemermann et al, 1997) and during splanchnic artery occlusion (Cuzzocrea et al, 1997). PARP inhibition 206  has also been shown to protect against drug-induced hepatic toxicity (Kroger et al, 1996), which has a secondary inflammatory component. The results presented in this paper are consistent with these studies and provide a potential explanation for the effects of P A R P inhibition in these various disease states and toxic injuries. The results demonstrating that allopurinol has inhibitory activity towards P A R P may explain, in part, why allopurinol affords protection in models of ischemia-reperfusion that do not contain xanthine oxidase, the putative source of oxidative stress. The association of caspase-3 activation with apoptosis has spurred research to find and develop therapeutically useful caspase inhibitors to prevent apoptosis. However, in oxidantinjury with extensive D N A damage, attempts at caspase inhibition without concurrent inhibition of PARP is unlikely to have any effect, as oncosis will predominate due to PARP activation. The present study provides evidence and rationale to inhibit PARP (i.e., to recruit more damaged cells from oncosis into apoptosis) as an adjunct to caspase inhibition. Finally, alteration of PARP activity following exposure to oxidant stress would appear to provide a mechanism to alter the process of cell death from oncosis to apoptosis, which ultimately has the potential to be regulated and controlled.  207  6. SUMMARY & CONCLUSIONS  This thesis has focused on the toxicological implications of PARP in terms of its potential to act as a site for metal replacement, as well as its role in oxidant-induced cell death. The hypothesis of this research was that alteration of P A R P activity, either through pharmacological inhibition of catalytic activity or disruption of DNA-binding capacity, would alter the process of oxidant-induced cell death from oncosis to apoptosis. Therefore, the overall goal of this research was to investigate the mechanisms for alteration of PARP activity and to examine the effect that alteration of PARP activity had on cellular response to oxidative stress. As such, this research advanced in two directions: 1.) Investigation of the potential for metal replacement in PARP zinc fingers as a mechanism of altering enzyme activity Of primary importance in this avenue of research was the ability to detect and validate changes in metal composition of the zinc fingers in both purified PARP and in P A R P isolated from intact cells. Development of an analytical method for detection of metal replacement in PARP focused on mass spectrometry, as this technique provided the sensitivity required to measure metal replacement in endogenous PARP. Several techniques were investigated to address this problem, including L A - I C P M S , M A L D I - T O F M S , as well as ESI-MS. However, the development of a MS technique proved to be a difficult technical problem due to the size of PARP in combination with its chemical properties and stability requirements for maintenance of the metal-intact zinc-finger motif. ESI-MS appeared to offer the most promising results for analysis of PARP under neutral pH conditions. A variety of neutral solvent systems were assessed; however only one solvent, 20% ACN/53 m M HFIPA pH 5.1, permitted detection and molecular weight estimation for PARP in the absence of acid. This was the first time a spectrum  208  and molecular weight estimation for the 24 kDa PARP fragment had been obtained in a nonacidic solvent system and represented a major step forward in the development of a technique for analyzing the native conformation of the zinc fingers of PARP. However, due to the pH of this solvent, it was unclear whether analysis of metal-intact PARP was achieved. Therefore, further developmental work will be required to determine whether these conditions permit analysis of the PARP zinc fingers in their native state. In addition, an understanding of the problems associated with this type of analytical development was attained, which may contribute to the scientific knowledge base in this area and may assist future attempts at assay development. The technical problems encountered during the development of a technique to detect metal replacement in PARP stalled research progress in this area and precluded any conclusion being reached regarding the effect of metal replacement on PARP activity and D N A binding capacity. In anticipation of this research going forward, a number of techniques were established to address the central question of the effect of metal replacement in the zinc fingers of PARP. First, a baculovirus expression system for production of recombinant PARP was established. Recombinant P A R P was produced and purified for subsequent in vitro experiments. In order to assess the effect of metal replacement on PARP activity, an enzyme assay was characterized and optimized in terms of substrates and co-substrates, incubation time, temperature and amount of PARP. In addition, allopurinol was identified as a novel inhibitor of PARP activity in vitro. Finally, in order to facilitate analysis of metal replacement in the zinc-finger region, a method for digestion of PARP into smaller fragments using caspase-3 and separation of the digestion fragments by H P L C was developed. This technique may have other experimental applications, such as in the investigation of the isolated PARP D N A binding domain or catalytic domain.  209  2.) Examination of the role of PARP inhibition on the mode of oxidant-induced cell death One of the central questions in this thesis asked if alteration of P A R P activity could influence the process by which oxidant-stressed endothelial cells died. The time-course and extent of cell death, defined as loss of membrane integrity, were characterized in a fluorescence assay of cell injury following treatment of endothelial cells with various concentrations of H2O2 in the absence or presence of a P A R P inhibitor. The mode of cell death, oncosis or apoptosis, in oxidant-injured endothelial cells was assessed in terms of nuclear morphology and caspase-3 activity, a hallmark of apoptosis. The studies described in this thesis demonstrated that PARP inhibition in oxidant-stressed endothelial cells altered the process by which these irreversiblyinjured cells were destined to die. Oxidant-injured endothelial cells underwent an oncotic mode of cell death, whereas inhibition of PARP in these cells prevented oncosis and permitted apoptosis. The shift of oxidant-stressed endothelial cells from oncosis into apoptosis following P A R P inhibition was demonstrated by increased caspase-3-like activity, as well as by morphological assessment using fluorescent microscopy. In addition, preservation of energydependent processes in oxidant-stressed endothelial cells treated with a P A R P inhibitor was demonstrated by the maintenance of low [Ca ]i and evidence of a normal functioning of the lysosomal proton pump, which actively concentrates A O in lysosomes. Therefore, PARP inhibition during an oxidant stress appeared to preserve energy dependent processes. In conclusion, alteration of P A R P activity following exposure to oxidant stress would appear to provide a mechanism to alter the process of oxidant-induced cell death from oncosis to apoptosis, which ultimately has the potential to be regulated and controlled.  210  7.  FUTURE DIRECTIONS  Previous studies have clearly shown that, following oxidant stress, there is a rapid depletion of N A D and a decline in cellular energy as a result of P A R P activation. Based on this +  fundamental knowledge, it was predicted that oxidant-stressed E C would undergo oncosis due to the consequences of P A R P activation. However, when P A R P was inhibited, the active process of apoptosis would then be able to proceed. The research in this thesis has demonstrated that this is the case. The ability to switch oxidant-induced cell death from what is, in essence, virtually complete oncosis to apoptosis, affords an interesting model for follow-up studies comparing and contrasting various biochemical and molecular aspects of cell death between: 1) oxidant-induced oncosis; 2) oxidant-induced apoptosis; and 3) apoptosis induced by non-genotoxic agents (e.g., staurosporine). Future research should be carried out in the following areas: I) In vitro studies on:  •  Confirmation of the altered energy status following oxidant stress and energy recovery when P A R P is inhibited after oxidant damage, and correlation with: a) energy-dependent processes such as maintenance of [Ca ]i and lysosomal proton gradient; and b) the timecourse and extent of mitochondrial membrane potential and release of cytochrome c.  •  Confirmation of the switch from oncosis to apoptosis using a PARP-deficient cell.  •  The time-course of gene expression preceding necrosis, focusing on genes associated with apoptosis, (e.g., bcl-2 and bax).  •  The importance of caspases in the execution of oxidant-induced apoptosis. Can the process of cell death, and the fate of the cell, be altered when both P A R P and caspase-3 are inhibited? Can oxidant-induced cell death be converted to a form of cell death that can be controlled?  211  2) New avenues of research in vivo: •  Does PARP inhibition in disease states associated with oxidant injury (e.g., pulmonary oxygen toxicity, ischemia-reperfusion injury) increase the extent of apoptosis and decrease inflammation and tissue loss that is associated with necrosis?  The potential for replacement of the zinc in zinc-finger proteins by an alternate metal is well established, at least in vitro. PARP, being a zinc-finger enzyme that is constitutively expressed and in abundant supply in the cell nucleus, would appear to be a prime candidate as a prototype to investigate metal replacement in vivo. In addition, the role of P A R P in cell death, D N A repair and genomic surveillance suggest that metal replacement in the PARP zinc fingers may contribute to the mechanism of metal toxicity. A key requirement of these studies is having a technique to assess whether an alternate metal has replaced zinc in vivo. This research has provided the basis for a mass spectrometric technique that would measure metal replacement in the PARP zinc fingers. Future research should be carried out in the following areas: •  Determine if the ESI-MS method and the ACN/HFIPA solvent system allow for analysis of metal-intact PARP. Additional developmental work may be required to adjust this solvent system for this purpose. Further enzymatic digestion of the 24 kDa PARP fragment to produce a smaller fragment may aid in the ionization by ESI-MS.  •  M A L D I - T O F MS could also be further evaluated with different matrices to determine the potential for analysis of metal-intact PARP.  •  The series of in vitro studies originally outlined in this thesis, which included investigation of metal-replaced PARP activity and DNA-binding capacity, would elucidate the effect of zinc-finger metal replacement on PARP activity. A metal-replaced  212  24 kDa P A R P fragment could be prepared in vitro and used to examine the effect on in vitro  PARP activity in a dominant-negative fashion. Confirmation of metal replacement  would, however, be required in order to interpret a negative effect. Ultimately, metal incorporation into zinc-finger proteins, in particular PARP, in a living system, should be the direction of future research, both as a potential mechanism of metal toxicity and as a biomarker of metal exposure. Alternate metal incorporation into zincfinger proteins in cells grown in the presence of an alternate metal and subsequent consequences for the function of the protein and the cell, would indicate the toxicological potential of this mechanism of metal toxicity.  213  8.  REFERENCES  Allen, R.T., Hunter, W.J., and Agrawal, D . K . (1997). Morphological and biochemical characterization and analysis of apoptosis. J.Pharmcol.Toxicol.Methods 37, 215-228. Althaus, F.R. and Richter, C. (1987). ADP-Ribosylation of proteins: Enzymology and biological significance. Mol.Biol.Biochem.Biophys. 37, 1-126. Amicon, Inc. Membrane Filtration / Chromatography Catalog. (1995). Apffel, A . , Chakel, J.A., Fischer, S., Lichtenwalter, K . , and Hancock, W.S. (1997). New procedure for the use of high-performance liquid chromatography-electrospray ionization mass spectrometry for the analysis of nucleotides and oligonucleotides. J.Chromatogr.A 777,  3-21.  Ashkenazi, A . and Dixit, V . M . (1998). Death receptors: signaling and modulation. Science 281, 1305-1308. Banasik, M . , Komura, H . , Shimoyama, M . , and Ueda, K . (1992). Specific inhibitors of poly(ADP-ribose) synthetase and mono(ADP- ribosyl)transferase. J.Biol.Chem. 267, 1569-1575. Basaga, H.S. (1990). Biochemical aspects of free radicals. Biochem.Cell Biol. 68, 989-998. Benjamin, R.C. and Gill, D . M . (1980). ADP-ribosylation in mammalian cell ghosts. Dependence of poly(ADP-ribose) synthesis on strand breakage in D N A . J.Biol.Chem. 255, 1049310501. Benjamin, R.C. and Gill, D . M . (1980). Poly(ADP-ribose) synthesis in vitro programmed by damaged D N A . A comparison of D N A molecules containing different types of strand breaks. J.Biol.Chem. 255, 10502-10508. Bernardi, R., Negri, C , Donzelli, M . , Guano, F., Torti, M . , Prosperi, E., and Scovassi, A.I. (1995). Activation of poly(ADP-ribose)polymerase in apoptotic human cells. Biochimie 77, 378-384. Berridge, M.J., Bootman, M.D., and Lipp, P. (1998). Calcium - a life and death signal. Nature 395, 645-648. Bonfoco, E., Krainc, D., Ankarcrona, M . , Nicotera, P., and Lipton, S.A. (1995). Apoptosis and necrosis: Two distinct events induced, respectively, by mild and intense insults with N methyl-D-aspartate or nitric oxide/superoxide in cortical cell cultures. Proc.Natl.Acad.Sci.U.S.A. 92, 7162-7166. Bradford, M . M . (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal.Biochem. 72, 248-254.  214  Bridgewater, L.C., Manning, F.C., Woo, E.S., and Patierno, S.R. (1994). D N A polymerase arrest by adducted trivalent chromium. Mol.Carcinog. 9, 122-133. Buki, K.G., Bauer, P.I., Mendeleyev, J., Hakam, A . , and Kun, E. (1991). Destabilization of Z n coordination in ADP-ribose transferase (polymerizing) by 6-nitroso-l,2-benzopyrone coincidental with inactivation of the polymerase but not the D N A binding function.  2+  FEBSLett. 290, 181-185.  Burtscher, H.J., Auer, B., Klocker, H . , Schweiger, M . , and Hirsch-Kauffmann, M . (1986). Isolation of ADP-ribosyltransferase by affinity chromatography. Anal.Biochem. 152, 285-290. Cadenas, E. (1989). Biochemistry of oxygen toxicity. Annu.Rev.Biochem. 58, 79-110. Casciola-Rosen, L., Nicholson, D.W., Chong, T., Rowan, K.R., Thornberry, N.A., Miller, D.K., and Rosen, A . (1996). Apopain/CPP32 cleaves proteins that are essential for cellular repair: a fundamental principle of apoptotic death. J.Exp.Med. 183, 1957-1964. Chatterjee, S., Hirschler, N.V., Petzold, S.J., Berger, S.J., and Berger, N . A . (1989). Mutant cells defective in poly(ADP-ribose) synthesis due to stable alterations in enzyme activity or substrate availability. Exp.Cell Res. 184, 1-15. Cherney, B.W., McBride, O.W., Chen, D.F., Alkhatib, H . , Bhatia, K., Hensley, P. and Smulson, M.E. (1987). cDNA sequence, protein structure, and chromosomal location of the human gene for poly(ADP-ribose) polymerase. Proc.Natl.Acad.Sci.U.S.A. 84, 83708374. Clontech Laboratories, Inc. ApoAlert™ CPP32 Assay Kits Product Protocol. (1997). Cohen, G . M . (1997). Caspases: the executioners of apoptosis. Biochem.J. 326, 1-16. Cohen, G . M . , Sun, X . M . , Snowden, R.T., Dinsdale, D., and Skilleter, D . N . (1992). Key morphological features of apoptosis may occur in the absence of internucleosomal D N A fragmentation. Biochem.J. 286, 331-334. Cohen, M . , Latta, D., Coogan, T., and Costa, M . (1990). Biological Effects of Heavy Metals (Foulkes, E.C., Ed.) C R C Press, Boca Raton, Florida. 19-75. Coppola, S., Nosseri, C , Maresca, V . , and Ghibelli, L . (1995). Different basal N A D levels determine opposite effects of poly(ADP-ribosyl)polymerase inhibitors on H 02-induced apoptosis. Exp.CellRes. 221,462-469. 2  Craig, T.A., Veenstra, T.D., Naylor, S., Tomlinson, A.J., Johnson, K L , Macura, S., Juranic, N . , and Kumar, R. (1997). Zinc binding properties of the D N A binding domain of the 1,25dihydroxyvitamin D3 receptor. Biochemistry 36, 10482-10491.  215  Cuzzocrea, S., Zingarelli, B., Costantino, G., Szabo, A., Salzman, A . L . , Caputi, A.P., and Szabo, C. (1997). Beneficial effects of 3-aminobenzamide, an inhibitor of poly (ADP-ribose) synthetase in a rat model of splanchnic artery occlusion and reperfusion. Br. J.Pharmacol. 121, 1065-1074. Cuzzocrea, S., Zingarelli, B., Gilad, E., Hake, P., Salzman, A . L . , and Szabo, C. (1998). Protective effects of 3-aminobenzamide, an inhibitor of poly (ADP-ribose) synthase in a carrageenan-induced model of local inflammation. Eur. J.Pharmacol. 342, 67-76. de Murcia, G., Huletsky, A . , and Poirier, G.G. (1988). Modulation of chromatin structure by poly(ADP-ribosyl)ation. Biochem.CellBiol.66, 626-635. de Murcia, G., Jongstra-Bilen, J., Ittel, M.E., Mandel, P., and Delain, E. (1983). Poly(ADPribose) polymerase auto-modification and interaction with D N A : electron microscopic visualization. EMBO J. 2, 543-548. de Murcia, G., Menissier-de Murcia, J., and Schreiber, V. (1991). Poly(ADP-ribose) polymerase: molecular biological aspects. Bioessays 13, 455-462. de Murcia, G., Schreiber, V., Molinete, M . , Saulier, B . , Poch, O., Masson, M . , Niedergang, C , and Menissier de Murcia, J. (1994). Structure and function of poly(ADP-ribose) polymerase. Mol.Cell.Biochem.13S, 15-24. Desnoyers, S., Shah, G.M., Brochu, G., Hoflack, J.C., Verreault, A . , and Poirier, G.G. (1995). Biochemical properties and function of poly(ADP-ribose) glycohydrolase. Biochimie 11, 433-438. Ding, R., Pommier, Y . , Kang, V . H . , and Smulson, M . (1992). Depletion of poly(ADP-ribose) polymerase by antisense R N A expression results in a delay in D N A strand break rejoining. J.Biol.Chem. 261, 12804-12812. Ding, R. and Smulson, M . (1994). Depletion of nuclear poly(ADP-ribose) polymerase by antisense R N A expression: influences on genomic stability, chromatin organization, and carcinogen cytotoxicity. Cancer Res. 54, 4627-4634. Duriez, P.J. and Shah, G . M . (1997). Cleavage of poly(ADP-ribose) polymerase: a sensitive parameter to study cell death. Biochem.Cell Biol.lS, 337-349. Eguchi, Y., Shimizu, S., and Tsujimoto, Y . (1997). Intracellular ATP levels determine cell death fate by apoptosis or necrosis. Cancer Res. 51, 1835-1840. Ellis, R.E., Yuan, J.Y., and Horvitz, H.R. (1991). Mechanisms and functions of cell death. Annu.Rev.Cell.Biol. 1, 663-698.  Emoto, Y., Manome, Y., Meinhardt, G., Kisaki, H., Kharbanda, S., Robertson, M . , Ghayur, T., Wong, W.W., Kamen, R., and Weichselbaum, R. (1995). Proteolytic activation of protein kinase C delta by an ICE-like protease in apoptotic cells. EMBO J. 14, 61486156.  216  Fadok, V . A . , Voelker, D.R., Campbell, P.A., Cohen, J.J., Bratton, D.L., and Henson, P . M . (1992). Exposure of phosphatidylserine on the surface of apoptotic lymphocytes triggers specific recognition and removal by macrophages. J.Immunol. 148, 2207-2216. Farber, J.L. (1990). The role of calcium in lethal cell injury. Chem.Res. Toxicol. 3, 503-508. Farber, J.L., Kyle, M.E., and Coleman, J.B. (1990). Mechanisms of cell injury by activated oxygen species. Lab.Invest. 62, 670-679. Fernandes-Alnemri, T., Litwack, G., and Alnemri, E.S. (1994). CPP32, a novel human apoptotic protein with homology to Caenorhabditis elegans cell death protein Ced-3 and mammalian interleukin-1 beta-converting enzyme. J.Biol.Chem. 269, 30761-30764. Freeman, B . A . and Crapo, J.D. (1982). Biology of disease: free radicals and tissue injury. Lab.Invest. 47, 412-426.  Freshney, R.I. (1987). Culture of Animal Cells. A Manual of Basic Technique. 2nd edition. Alan R. Liss, Inc., New York, N Y . 57-84. Ghayur, T., Hugunin, M . , Talanian, R.V., Ratnofsky, S., Quinlan, Emoto, Y., Pandey, P., Datta, R., Huang, Y . , Kharbanda, S., Allen, H., Kamen, R., Wong, W., and Kufe, D. (1996). Proteolytic activation of protein kinase C delta by an ICE/CED 3-like protease induces characteristics of apoptosis. J.Exp.Med. 184, 2399-2404. Gibaldi, M . and Perrier, D. (1975). Pharmacokinetics Marcel Dekker, Inc., New York. 293-296. Gilad, E., Zingarelli, B., Salzman, A . L . , and Szabo, C. (1997). Protection by inhibition of poly (ADP-ribose) synthetase against oxidant injury in cardiac myoblasts in vitro. J.Mol. Cell. Cardiol. 29, 2585-2597. Giner, H., Simonin, F., de Murcia, G., and Menissier-de Murcia, J. (1992). Overproduction and large-scale purification of the human poly(ADP-ribose) polymerase using a baculovirus expression system. Gene 114, 279-283. Gradwohl, G., Menissier de Murcia, J.M., Molinete, M . , Simonin, F., Koken, M . , Hoeijmakers J.H. and de Murcia, G. (1990). The second zinc-finger domain of poly(ADP-ribose) polymerase determines specificity for single-stranded breaks in D N A . Proc.Natl.Acad.Sci.U.S.A. 87, 2990-2994. Green, B.N., Hutton, T., and Vinogradov, S.N. (1996). Methods in Molecular Biology: Protein and Peptide Analysis by Mass Spectrometry (Chapman, J.R., Ed.) Humana Press Inc., Totowa, NJ. 279-294. Greidinger, E.L., Miller, D.K., Yamin, T.T., Casciola-Rosen, L . , and Rosen, A . (1996). Sequential activation of three distinct ICE-like activities in Fas-ligated Jurkat cells. FEBSLett. 390, 299-303.  217  Grynkiewicz, G., Poenie, M . , and Tsien, R.Y. (1985). A new generation of Ca indicators with greatly improved fluorescence properties. J.Biol.Chem. 260, 3440-3450. +  Halliwell, B . (1987). Free radicals and metal ions in health and disease. Proc.Nutr.Soc. 46, 1326. Halliwell, B . and Gutteridge, J.M. (1984). Oxygen toxicity, oxygen radicals, transition metals and disease. Biochem.J. 219,  1-14.  Hayaishi, O. and Ueda, K . (1982). ADP-Ribosylation Reactions - Biology and (Hayaishi, O. and Ueda, K., Eds.) Academic Press, New York. 3-16.  Medicine  Hille, R. and Nishino, T. (1995). Flavoprotein structure and mechanism. Xanthine oxidase and xanthine dehydrogenase. FASEB J. 9, 995-1003. Hu, P., Ye, Q.Z., and Loo, J.A. (1994). Calcium stoichiometry determination for calcium binding proteins by electrospray ionization mass spectrometry. Anal.Chem. 66, 4190-4194. Hutchens, T.W., Allen, M . H . , L i , C M . , and Yip, T.T. (1992a). Occupancy of a C2-C2 type 'zinc-finger' protein domain by copper. Direct observation by electrospray ionization mass spectrometry. FEBS Lett. 309, 170-174. Hutchens, T.W., Nelson, R.W., Allen, M . H . , L i , C M . , and Yip, T.T. (1992b). Peptide-metal ion interactions in solution: detection by laser desorption time-of-flight mass spectrometry and electrospray ionization mass spectrometry. Biol.Mass Spectrom. 21, 151-159. Ikejima, M . , Noguchi, S., Yamashita, R., Ogura, T., Sugimura, T., Gill, D . M . , and Miwa, M . (1990). The zinc fingers of human poly(ADP-ribose) polymerase are differentially required for the recognition of D N A breaks and nicks and the consequent enzyme activation. J.Biol.Chem. 265,21907-21913. Imlay, J.A. and Linn, S. (1988). D N A damage and oxygen radical toxicity. Science 240, 13021309. Jarvis, D.L. and Garcia, A., Jr. (1994). Long-term stability of baculoviruses stored under various conditions. Biotechniques 16,508-513. Jones, D.P., McConkey, D.J., Nicotera, P., and Orrenius, S. (1989). Calcium-activated D N A fragmentation in rat liver nuclei. J.Biol.Chem. 264, 6398-6403. Kalyanaraman, B . , Perez-Reyes, E., and Mason, R.P. (1980). Spin-trapping and direct electron spin resonance investigations of the redox metabolism of quinone anticancer drugs. Biochim.Biophys.Acta 630, 119-130.  Kameshita, I., Matsuda, Z., Taniguchi, T., and Shizuta, Y . (1984). Poly (ADP-Ribose) synthetase. Separation and identification of three proteolytic fragments as the substratebinding domain, the DNA-binding domain, and the automodification domain. J.Biol.Chem. 259, 4770-4776.  218  Kaufmann,  S.H. (1996). Proteolytic cleavage  MoLMedToday  during chemotherapy-induced  apoptosis.  2 ,298-303.  Kerr, J.F., Wyllie, A . H . , and Currie, A.R. (1972). Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br. J.Cancer 26, 239-257. King, L . A . and Possee, R.D. (1992). The Baculovirus Expression System - A Laboratory Guide., Chapman & Hall, London, England. Kirkland, J.B. (1991). Lipid peroxidation, protein thiol oxidation and D N A damage in hydrogen peroxide-induced injury to endothelial cells: role of activation of poly(ADPribose)polymerase. Biochim.Biophys.Acta 1092, 319-325. Klug, A. and Schwabe, J.W. (1995). Protein motifs 5. Zinc fingers. FASEB J. 9, 597-604. Kristian, T. and Siesjo, B.K. (1998). Calcium in ischemic cell death. Stroke 29, 705-718. Kroemer, G., Petit, P., Zamzami, N . , Vayssiere, J.L., and Mignotte, B.(1995). The biochemistry of programmed cell death. FASEB J. 9, 1277-1287. Kroger, H . , Ehrlich, W., Klewer, M . , Gratz, R., Dietrich, A . , and Miesel, R. (1996). The influence of antagonists of poly(ADP-ribose) metabolism on acetaminophen hepatotoxicity. Gen. Pharmacol. 27, 167-170. Kurosaki, T., Ushiro, H., Mitsuuchi, Y . , Suzuki, S., Matsuda, M . , Matsuda, Y . K . N . , Kangawa, K., Matsuo, H . , Hirose, T., (1987). Primary structure of human poly(ADP-ribose) synthetase as deduced from c D N A sequence. J.Biol.Chem. 262, 15990-15997. Laemmli, U . K . (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685. Lafitte, D., Capony, J.P., Grassy, G., Haiech, J., and Calas, B. (1995). Analysis of the ion binding sites of calmodulin by electrospray ionization mass spectrometry. Biochemistry 34, 13825-13832. Lam,  T.T. (1997). The effect of 3-aminobenzamide, an inhibitor of poly-ADP-ribose polymerase, on ischemia/reperfusion damage in rat retina. Res.Commun.Mol.Path.Pharmacol. 95, 241-252.  Leist, M . , Single, B., Castoldi, A.F., Kuhnle, S., and Nicotera, P. (1997). Intracellular adenosine triphosphate (ATP) concentration: a switch in the decision between apoptosis and necrosis. J.Exp.Med. 185, 1481-1486. Leist, M . , Single, B., Kunstle, G., Volbracht, C , Hentze, H., and Nicotera, P. (1997). Apoptosis in the absence of poly-(ADP-ribose) polymerase. Biochem.Biophys.Res.Commun. 233 , 518-522.  219  Lelli, J.L., Jr., Becks, L . L . , Dabrowska, M.I., and Hinshaw, D.B. (1998). A T P converts necrosis to apoptosis in oxidant-injured endothelial cells. Free Radic.Biol.Med. 25, 694-702. Levin, S. (1998). Apoptosis, necrosis, or oncosis: what is your diagnosis? A report from the Cell Death Nomenclature Committee of the Society of Toxicologic Pathologists. Toxicological Sciences 41, 155-156. Lin, X . , Zhuang, Z., and Costa, M . (1992). Analysis of residual amino acid-DNA crosslinks induced in intact cells by nickel and chromium compounds. Carcinogenesis 13, 17631768. Loo, J.A. (1997). Studying noncovalent protein complexes by electrospray ionization mass spectrometry. Mass Spectrom Rev 16, 1-23. Loo, J.A., Holler, T.P., Sanchez, J., Gogliotti, R., Maloney, L . , and Reily, M . D . (1996). Biophysical characterization of zinc ejection from HrV nucleocapsid protein by antiHIV 2,2'-dithiobis[benzamides] and benzisothiazolones. J.Med.Chem. 39, 4313-4320. Majno, G. and Joris, I. (1995). Apoptosis, oncosis, and necrosis. A n overview of cell death. Am.J.Pathol 146,  3-15.  Mason, R.P. and Chignell, C.F. (1981). Free radicals in pharmacology and toxicology-selected topics. Pharmacol.Rev. 33, 189-211. Masson, M . , Rolli, V., Dantzer, F., Trucco, C , Schreiber, V., Fribourg, S.M.M., Ruf, A . , Miranda, E.A., Niedergang, C , Hunting, D., Gowans, B . , Schulz, G.E., Menissier de Murcia, J. and de Murcia, G. (1995). Poly(ADP-ribose) polymerase: structure-function relationship. Biochimie 77, 456-461. Mattson, M.P., Barger, S.W., Begley, J.G., and Mark, R.J. (1995). Cell Death (Schwartz, L . M . and Osbornw, B.A., Eds.) Academic Press, Inc., San Diego, C A . 187-216. Mazen, A . , Menissier-de Murcia, J., Molinete, M . , Simonin, F., Gradwohl, G.P.G., and de Murcia, G. (1989). Poly(ADP-ribose)polymerase: a novel finger protein. Nucl.Acids Res. 17, 4689-4698. McConkey, D.J. and Orrenius, S. (1997). The role of calcium in the regulation of apoptosis. Biochem.Biophys.Res.Commun. 239, 357-366.  McCord, J . M . and Fridovich, I. (1978). The biology and pathology of oxygen radicals. Ann.Intern.Med. 89, 122-127.  Menissier-de Murcia, J., Molinete, M . , Gradwohl, G., Simonin, F., and de Murcia, G. (1989). Zinc-binding domain of poly(ADP-ribose)polymerase participates in the recognition of single strand breaks on D N A . J.Mol.Biol. 210, 229-233. Miller, J., McLachlan, A.D., and Klug, A . (1985). Repetitive zinc-binding domains in the protein transcription factor IIIA from Xenopus oocytes. EMBO J. 4, 1609-1614.  220  Miura, M . , Zhu, H., Rotello, R., Hartwieg, E.A., and Yuan, J. (1993). Induction of apoptosis in fibroblasts by IL-1 beta-converting enzyme, a mammalian homolog of the C. elegans cell death gene ced-3. Cell 75, 653-660. Molinete, M . , Vermeulen, W., Burkle, A . , Menissier-de Murcia, J., Kupper, J.H.H.J., and de Murcia, G. (1993). Overproduction of the poly(ADP-ribose) polymerase DNA-binding domain blocks alkylation-induced D N A repair synthesis in mammalian cells. EMBO J. 12,2109-2117. Monti, D., Cossarizza, A., Salvioli, S., Franceschi, C., Rainaldi, G., Straface, E., Rivabene, R., and Malorni, W. (1994). Cell death protection by 3-aminobenzamide and other poly(ADP- ribose)polymerase inhibitors: different effects on human natural killer and lymphokine activated killer cell activities. Biochem.Biophys.Res.Commun.199, 525-530. Nackerdien, Z., Kasprzak, K.S., Rao, G., Halliwell, B., and Dizdaroglu, M . (1991). Nickel(II)and cobalt(II)-dependent damage by hydrogen peroxide to the D N A bases in isolated human chromatin. Cancer Res. 51, 5837-5842. Nelipovich, P.A., Nikonova, L.V., and Umansky, S.R. (1988). Inhibition of poly(ADP-ribose) polymerase as a possible reason for activation of Ca /Mg -dependent endonuclease in thymocytes of irradiated rats. International Journal of Radiation Biology & Related Studies in Physics, Chemistry & Medicine 53, 749-765. 2+  2+  Nicholson, D.W., A l i , A., Thornberry, N.A., Vaillancourt, J.P., Ding, C.K., Gallant, M . , Gareau, Y . , Griffin, P.R., Labelle, M . , Lazebnik, Y . A . , Munday, N.A., Raju, S.M., Smulson, M.E., Yamin, T., Yu, V . L . , and Miller, D.L. (1995). Identification and inhibition of the ICE/CED-3 protease necessary for mammalian apoptosis. Nature 376, 37-43. Nicotera, P. and Orrenius, S. (1998). The role of calcium in apoptosis. Cell Calcium 23, 173180. Niedergang, C , Okazaki, H . , and Mandel, P. (1979). Properties of purified calf thymus poly(adenosine diphosphate ribose) polymerase. Comparison of the DNA-independent and the D N A - dependent enzyme. Eur.J.Biochem. 102, 43-57. Nieminen, A . L . , Gores, G.J., Bond, J.M., Imberti, R., Herman, B., and Lemasters, J.J. (1992). A novel cytotoxicity screening assay using a multiwell fluorescence scanner. Toxicol. Appl. Pharmacol. 115, 147-155. Nosseri, C , Coppola, S., and Ghibelli, L . (1994). Possible involvement of poly(ADP-ribosyl) polymerase in triggering stress-induced apoptosis. Exp.Cell Res. 212, 367-373. Oberhammer, F., Wilson, J.W., Dive, C , Morris, I.D., Hickman, J.A., Wakeling, A.E.Walker, P.R., and Sikorska, M . (1993). Apoptotic death in epithelial cells: cleavage of D N A to 300 and/or 50 kb fragments prior to or in the absence of internucleosomal fragmentation. EMBO J. 12, 3679-3684.  221  Oliver, F.J., de la Rubia, G., Rolli, V . , Ruiz-Ruiz-, M.C., de Murcia, G., and Menissier-de Murcia, J. (1998). Importance of poly(ADP-ribose) polymerase and its cleavage in apoptosis. J.Biol.Chem. 273(50), 33533-33539. Ozawa, T., Ueda, J., and Shimazu, Y . (1993). D N A single strand breakage by copper(II) complexes and hydrogen peroxide at physiological conditions. Biochemistry & Molecular Biology International 31, 455-461. Palomba, L . , Sestili, P., Cattabeni, F., Azzi, A., and Cantoni, O. (1996). Prevention of necrosis and activation of apoptosis in oxidatively injured human myeloid leukemia U937 cells. FEBSLett. 390,91-94. Predki, P.F. and Sarkar, B . (1992). Effect of replacement of "zinc finger" zinc on estrogen receptor D N A interactions. J.Biol.Chem. 267, 5842-5846. Pryor, W.A. (1986). Oxy-radicals and related species: their formation, lifetimes, and reactions. Annu.Rev.Physiol. 48, 657-667.  Rankin, P.W., Jacobson, E.L., Benjamin, R.C., Moss, J., and Jacobson, M . K . (1989). Quantitative studies of inhibitors of ADP-ribosylation in vitro and in vivo. J.Biol.Chem. 264,4312-4317. Ray, S.D., Mumaw, V.R., Raje, R.R., Fariss, M.W. (1996). Protection of acetaminophen-induced hepatocellular apoptosis and necrosis by cholesteryl hemisuccinate pretreatment. J.Pharmacol.Exp.Ther. 279, 1470-1483. Ray, S.D., Sorge, C.L., Kamendulis, L . M . , and Corcoran, G.B. (1992). Ca -activated D N A fragmentation and dimethylnitrosamine- induced hepatic necrosis: effects of C a endonuclease and poly(ADP-ribose) polymerase inhibitors in mice. (++)  (++)  J.Pharmacol.Exp.Ther. 263, 387-394.  Realini, C . A . and Althaus, F.R. (1992). Histone shuttling by poly(ADP-ribosylation). J.Biol.Chem. 267, 18858-18865. Repine, J.E., Cheronis, J.C., Rodell, T.C., Linas, S.L., and Patt, A . (1987). Pulmonary oxygen toxicity and ischemia-reperfusion injury. A mechanism in common involving xanthine oxidase and neutrophils. Am.Rev.Respir.Dis. 136, 483-485. Reutelingsperger, C P . and van Heerde, W . L . (1997). Annexin V , the regulator of phosphatidylserine-catalyzed inflammation and coagulation during apoptosis. Cellular & Molecular Life Sciences 53, 527-532. Rice, W.G., Hillyer, C D . , Harten, B., Schaeffer, C A . , Dorminy, M . , Lackey III, D.A., Kirsten, E., Mendeleyev, J., Buki, K . G . , Hakam, A . , and Kun, E. (1992). Induction of endonuclease-mediated apoptosis in tumor cells by C-nitroso-substituted ligands of poly(ADP-ribose) polymerase. Proc.Natl.Acad.Sci.U.S.A. 89, 7703-7707.  222  Rosenthal, D.S., Ding, R., Simbulan-Rosenthal, C . M . , Vaillancourt, J.P., Nicholson, D.W., and Smulson, M . (1997). Intact cell evidence for the early synthesis, and subsequent late apopain-mediated suppression, of poly(ADP-ribose) during apoptosis. Exp. Cell Res. 232, 313-321. Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual, 2nd edition. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Satoh, M.S. and Lindahl, T. (1992). Role of poly(ADP-ribose) formation in D N A repair. Nature 356, 356-358. Savill, J., Fadok, V . , Henson, P., and Haslett, C. (1993). Phagocyte recognition of cells undergoing apoptosis. Immunol.Today 14, 131-136. Schlegel, J., Peters, I., Orrenius, S., Miller, D.K., Thornberry, N.A., Yamin, T.T., and Nicholson, D.W. (1996). CPP32/apopain is a key interleukin 1 beta converting enzyme-like protease involved in Fas-mediated apoptosis. J.Biol.Chem. 271, 1841-1844. Schraufstatter, I.U., Hinshaw, D.B., Hyslop, P.A., Spragg, R.G., and Cochrane, C.G. (1986). Oxidant injury of cells. D N A strand-breaks activate polyadenosine diphosphate-ribose polymerase and lead to depletion of nicotinamide adenine dinucleotide. J.Clin.Invest. 77, 1312-1320. Schreiber, V . , Hunting, D., Trucco, C , Gowans, B . , Grunwald, D., de Murcia, G., De, and Murcia, J . M . (1995). A dominant-negative mutant of human poly(ADP-ribose) polymerase affects cell recovery, apoptosis, and sister chromatid exchange following D N A damage. Proc.Natl.Acad.Sci.U.S.A. 92,4753-4757. Scobie, H.A., Vath, J.E., Y u , W., and Martin, S.A. (1993). A Practical Guide to Protein and Peptide Purification for Microsequencing. 2nd edition. Academic Press, Inc., 125-153. Shah, G . M . , Kaufmann, S.H., and Poirier, G.G. (1995). Detection of poly(ADP-ribose) polymerase and its apoptosis-specific fragment by a nonisotopic activity-western blot technique. Anal.Biochem. 232,251-254. Simonin, F., Hofferer, L . , Panzeter, P.L., Muller, S., de Murcia, G., and Althaus, F.R. (1993). The carboxyl-terminal domain of human poly(ADP-ribose) polymerase. Overproduction in Escherichia coli, large scale purification, and characterization. J.Biol.Chem. 268, 13454-13461. Simonin, F., Menissier-de, M.J., Poch, O., Muller, S., Gradwohl, G., Molinete, Penning, C , Keith, G., and de, M . G . (1990). Expression and site-directed mutagenesis of the catalytic domain of human poly(ADP-ribose)polymerase in Escherichia coli. Lysine 893 is critical for activity. J.Biol.Chem. 265, 19249-19256. Sugden, K.D., Geer, R.D., and Rogers, S.J. (1992). Oxygen radical-mediated D N A damage by redox-active Cr(III) complexes. Biochemistry 31, 11626-11631.  223  Summers, M.D. and Smith, G.E. (1988). A Manual of Methods for Baculovirus Vectors and Insect Cell Culture Procedures. Texas Agricultural Experiment Station, College Station. Surovoy, A., Waidelich, D., and Jung, G. (1992). Nucleocapsid protein of HTV-1 and its Zn2+ complex formation analysis with electrospray mass spectrometry. FEBS Lett. 311, 259262. Swierenga, S.H., Gilman, J.P., and McLean, J.R. (1987). Cancer risk from inorganics.  Cancer  Metastasis Rev.6, 113-154.  Szabo, C. (1998). The role of poly(ADP-ribose) synthetase in inflammation. Eur. J.Pharmacol. 350, 1-19. Szabo, C , Lim, L.H.K., Cuzzocrea, S., Getting, S.J., Zingarelli, J3., Flower, R.J., Salzman, A . L . , and Perretti, M . (1997). Inhibition of poly (ADP-ribose) synthetase attenuates neutrophil recruitment and exerts antiinflammatory effects. J.Exp.Med. 186, 10411049. Takeda, M . , Shirato, I., Kobayashi, M . , and Endou, H . (1999). Hydrogen peroxide induces necrosis, apoptosis, oncosis and apoptotic oncosis of mouse terminal proximal straight tubule cells. Nephron 81:234-238. Tanaka, Y . , Yoshihara, K., Itaya, A . , Kamiya, T., and Koide, S.S. (1984). Mechanism of the inhibition of C a , Mg -dependent endonuclease of bull seminal plasma induced by ADP-ribosylation. J.Biol.Chem. 259, 6579-6585. 2+  2+  Tewari, M . , Quan, L.T., O'Rourke, K., Desnoyers, S., Zeng, Z., Beidler, D.R., Poirier, G G , Salvesen, G.S., and Dixit, V . M . (1995). Yama/CPP32 beta, a mammalian homolog of CED-3, is a CrmA- inhibitable protease that cleaves the death substrate poly(ADPribose) polymerase. Cell 81, 801-809. Thiemermann, C , Bowes, J., Myint, F.P., and Vane, J.R. (1997). Inhibition of the activity of poly(ADP ribose) synthetase reduces ischemia-reperfusion injury in the heart and skeletal muscle. Proc.Natl.Acad.Sci.U.S.A. 94, 679-683. Thies, R.L. and Autor, A.P. (1991). Reactive oxygen injury to cultured pulmonary artery endothelial cells: mediation by poly(ADP-ribose) polymerase activation causing N A D depletion and altered energy balance. Arch.Biochem.Biophys. 286, 353-363. Thomas, H.C., Lame, M . W . , Dunston, S.K., Segall, H.J., and Wilson, D.W. (1998). Monocrotaline pyrrole induces apoptosis in pulmonary artery endothelial cells. Toxicol.Appl.Pharmacol. 151, 236-244.  Thornberry, N . A . and Lazebnik, Y . (1998). Caspases: enemies within. Science 281, 1312-1316. Trump, B.F. and Berezesky, I.K. (1996). The role of altered [Ca ]; regulation in apoptosis, oncosis, and necrosis. Biochim.Biophys.Acta 1313, 173-178. 2+  224  Trump, B.F., Berezesky, I.K., Chang, S.H., and Phelps, P.C. (1997). The pathways of cell death: oncosis, apoptosis, and necrosis. Toxicol.Pathol. 25, 82-88. Tsujimoto, Y . (1997). Apoptosis and necrosis: Intracellular ATP levels as a determinant for cell death modes. Cell Death and Differentiation 4, 429-434. Ueda, K . , Kawaichi, M . , and Hayaishi, O. (1982). ADP-Ribosylation Reactions. Biology and Medicine. (Ueda, K. and Hayaishi, O., Eds.) Academic Press, Inc., 117-155. Ushiro, H . , Yokoyama, Y . , and Shizuta, Y . (1987). Purification and characterization of poly (ADP-ribose) synthetase from human placenta. J.Biol.Chem. 262, 2352-2357. Veenstra, T.D., Johnson, K . L . , Tomlinson, A.J., Craig, T.A., Kumar, R., and Naylor, S. (1998). Zinc-induced conformational changes in the DNA-binding domain of the vitamin D receptor determined by electrospray ionization mass spectrometry. J.Am.Soc.Mass Spectrom. 9,  8-14.  Veenstra, T.D., Johnson, K . L . , Tomlinson, A . J . , Naylor, S., and Kumar, R. (1997). Determination of calcium-binding sites in rat brain calbindin D28K by electrospray ionization mass spectrometry. Biochemistry 36, 3535-3542. Virag, L . , Scott, G.S., Cuzzocrea, S., Marmer, D., Salzman, A . L . , and Szabo, C. (1998). Peroxynitrite-induced thymocyte apoptosis: the role of caspases and poly (ADP-ribose) synthetase (PARS) activation. Immunology 94, 345-355. Walker, P.R. and Sikorska, M . (1997). New aspects of the mechanism of D N A fragmentation in apoptosis. Biochem.CellBiol. 75,287-299. Walker, P.R., Weaver, V . M . , Lach, B . , LeBlanc, J., and Sikorska, M . (1994). Endonuclease activities associated with high molecular weight and internucleosomal D N A fragmentation in apoptosis. Exp.Cell Res. 213,100-106. Wang, Z., Stingl, L., Morrison, C , Jantsch, M . , Los, M . , Schulze-Osthoff, K., and Wagner, E.F. (1997). P A R P is important for genomic stability but dispensible in apoptosis. Genes Dev. 11, 2347-2358. Watson, A . J . , Askew, J.N., and Benson, R.S. (1995). Poly(adenosine diphosphate ribose) polymerase inhibition prevents necrosis induced by H2O2 but not apoptosis. Gastroenterology 109, 472-482.  Weedon, D., Searle, J., and Kerr, J.F. (1979). Apoptosis. Its nature and implications for dermatopathology. Am.J.Dermatopathol. 1, 133-144. Weiss, S.J. and LoBuglio, A.F. (1982). Phagocyte-generated oxygen metabolites and cellular injury. Lab.Invest. 47, 5-18.  Werns, S.W. and Lucchesi, B.R. (1988). Leukocytes, oxygen radicals, and myocardial injury due to ischemia and reperfusion. Free Radic.Biol.Med. 4, 31-37.  225  Witkowska, H.E., Shackleton, H.L., Dahlman-Wright, K . , K i m , J.Y., and Gustafsson, J.A. (1995). Mass spectrometric analysis of a native zinc-finger structure: the glucocorticoid receptor D N A binding domain. J.Am.Chem.Soc. Ill, 3319-3324. Wyllie, A . H . (1980). Glucocorticoid-induced thymocyte apoptosis is associated with endogenous endonuclease activation. Nature 284, 555-556. Xu, B., Krudy, G.A., and Rosevear, P.R. (1993). Identification of the metal ligands and characterization of a putative zinc finger in methionyl-tRNA synthetase. J.Biol.Chem. 268, 16259-16264. Yoshihara, K . , Tanigawa, Y . , Burzio, L . , and Koide, S.S. (1975). Evidence for adenosine diphosphate ribosylation of C a , Mg -dependent endonuclease. Proc.Natl.Acad. Sci. 2+  2+  U.S.A. 72, 289-293.  Yu, X . , Wojciechowski, M . , and Fenselau, C. (1993). Assessment of metals in reconstituted metallothioneins by electrospray mass spectrometry. Anal.Chem. 65, 1355-1359. Yuan, J., Shaham, S., Ledoux, S., Ellis, H . M . , and Horvitz, H.R. (1993). The C. elegans cell death gene ced-3 encodes a protein similar to mammalian interleukin-1 beta-converting enzyme. Cell 75, 641-652. Zahradka, P. and Ebisuzaki, K . (1982). A shuttle mechanism for DNA-protein interactions. The regulation of poly(ADP-ribose) polymerase. Eur.J.Biochem. 127, 579-585. Zahradka, P. and Ebisuzaki, K . (1984). Poly(ADP-ribose) polymerase is a zinc metalloenzyme. Eur.J.Biochem. 142, 503-509.  Zelenin, A . V . (1993). Fluorescent and Luminescent Probes for Biological Activity. A Practical Guide to Technology for Quantitative Real-Time Analysis. (Mason, W.T., Ed.) Academic Press Ltd., San Diego. 83-99. Zingarelli, B . , Salzman, A . L . , and Szabo, C. (1998). Genetic disruption of poly (ADP-ribose) synthetase inhibits the expression of p-selectin and intercellular adhesion molecule-1 in myocardial ischemia/reperfusion injury. Circ.Res. 83, 85-94.  226  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0089331/manifest

Comment

Related Items