A Personal Foreword 7
Contents 9
Preface 16
List of Contributors 18
I Introduction 21
1 The Basic Principles of GRID 23
1.1 Introduction 23
1.2 Philosophy and Objectives 23
1.3 Priorities 24
1.4 The GRID Method 25
1.4.1 GRID Probes Are Anisometric 26
1.4.2 The Target \u201cResponds\u201d to the Probe 28
1.4.3 The Target is Immersed in Water 30
1.5 The GRID Force Field 30
1.5.1 The Lennard-Jones Term 31
1.5.2 The Electrostatic Term 31
1.5.3 The Hydrogen Bond Term 32
1.5.4 The Other Terms 32
1.6 Nomenclature 34
1.6.1 \u201cATOM\u201d Records 34
1.6.2 \u201cHETATM\u201d Records 35
1.7 Calibrating the GRID Force Field 36
1.7.1 Checking the Calibration 37
1.7.2 Checking Datafile GRUB 37
1.8 The Output from GRID 38
1.8.1 GRID Maps from Macromolecules 39
1.8.2 GRID Maps from a Small Molecule 44
1.9 Conclusions 45
2 Calculation and Application of Molecular Interaction Fields 47
2.1 Introduction 47
2.2 Calculation of MIFs 47
2.2.1 The Target 47
2.2.2 The Probe 48
2.2.3 The Interaction Function 49
2.2.3.1 Van der Waals Interactions 49
2.2.3.2 Electrostatic Interactions 50
2.2.3.3 Hydrogen Bonds 51
2.2.3.4 Entropy 52
2.3 Selected Applications of MIFs 53
2.3.1 Mapping a Ligand Binding Site in a Protein 53
2.3.2 Deriving 3D-QSARs 54
2.3.3 Similarity Analysis of a Set of Related Molecules 56
2.4 Concluding Remarks and Outlook 58
II Pharmacodynamics 63
3 Protein Selectivity Studies Using GRID-MIFs 65
3.1 Introduction 65
3.2 GRID Calculations and Chemometric Analysis 66
3.2.1 Source and Selection of Target Structures 66
3.2.2 Selection and Superimposition of Binding Sites 67
3.2.3 Calculation of the Molecular Interaction Field 67
3.2.4 Matrix Generation and Pretreatments 70
3.2.4.1 Region Cut-outs 71
3.2.5 GRID/PCA 71
3.2.5.1 Score Plots 72
3.2.5.2 Two-Dimensional Loading Plots 73
3.2.5.3 Loading Contour Maps 74
3.2.5.4 Problems of GRID/PCA 74
3.2.6 GRID/CPCA 75
3.2.6.1 Block Unscaled Weights 76
3.2.6.2 CPCA 78
3.2.6.3 Identification of Important Variable Blocks for Selectivity 79
3.2.6.4 Contour Plots 79
3.3 Applications 80
3.3.1 DNA Minor Groove Binding \u2013 Compare AAA and GGG Double Helix 80
3.3.2 Dihydrofolate Reductase 81
3.3.3 Cyclooxygenase 81
3.3.4 Penicillin Acylase 82
3.3.5 Serine Proteases 83
3.3.5.1 S1 Pocket 84
3.3.5.2 P Pocket 84
3.3.5.3 D Pocket 86
3.3.6 CYP450 87
3.3.7 Target Family Landscapes of Protein Kinases 89
3.3.8 Matrix Metalloproteinases (MMPs) 90
3.3.9 Nitric Oxide Synthases 94
3.3.10 PPARs 95
3.3.11 Bile Acid Transportation System 95
3.3.12 Ephrin Ligands and Eph Kinases 96
3.4 Discussion and Conclusion 97
4 FLAP: 4-Point Pharmacophore Fingerprints from GRID 103
4.1 Introduction 104
4.1.1 Pharmacophores and Pharmacophore Fingerprints 104
4.1.2 FLAP 106
4.2 FLAP Theory 106
4.3 Docking 108
4.3.1 GLUE: A New Docking Program Based on Pharmacophores 109
4.3.2 Case Study 111
4.4 Structure Based Virtual Screening (SBVS) 112
4.5 Ligand Based Virtual Screening (LBVS) 114
4.6 Protein Similarity 115
4.7 TOPP (Triplets of Pharmacophoric Points) 117
4.8 Conclusions 121
5 The Complexity of Molecular Interaction: Molecular Shape Fingerprints by the PathFinder Approach 123
5.1 Introduction 123
5.2 Background 124
5.3 The PathFinder Approach 125
5.3.1 Paths from Positive MIF 125
5.3.2 Paths from Negative MIF 127
5.4 Examples 129
5.4.1 3D-QSAR 129
5.4.2 CYP Comparison 132
5.4.3 Target\u2013Ligand Complexes 132
5.5 Conclusions 135
6 Alignment-independent Descriptors from Molecular Interaction Fields 137
6.1 Introduction 137
6.1.1 The Need for MIF-derived Alignment-independent Descriptors 137
6.1.2 GRIND Applications 139
6.2 GRIND 140
6.2.1 The Basic Idea 140
6.2.1.1 Computation of MIF 141
6.2.1.2 Extraction of Highly Relevant Regions 142
6.2.1.3 MACC2 Encoding 144
6.2.2 The Analysis of GRIND Variables 148
6.3 How to Interpret a GRIND-based 3D QSAR Model 150
6.3.1 Overview 150
6.3.2 Interpreting Correlograms 151
6.3.3 Interpreting Single Variables 153
6.3.4 GRIND-based 3D QSAR Models are not Pharmacophores 154
6.4 GRIND Limitations and Problems 155
6.4.1 GRIND and the Ligand Conformations 155
6.4.2 The Ambiguities 157
6.4.3 Chirality 159
6.5 Recent and Future Developments 159
6.5.1 Latest Developments 159
6.5.1.1 Shape Description 159
6.5.1.2 Anchor GRIND 160
6.5.2 The Future 160
6.6 Conclusions 161
7 3D-QSAR Using the GRID/GOLPE Approach 165
7.1 Introduction 165
7.2 3D-QSAR Using the GRID/GOLPE Approach 167
7.3 GRID/GOLPE Application Examples 169
7.3.1 Estrogen Receptor Ligands 169
7.3.2 Acetylcholinesterase Inhibitors 178
7.4 Conclusion 185
III Pharmacokinetics 191
8 Use of MIF-based VolSurf Descriptors in Physicochemical and Pharmacokinetic Studies 193
8.1 ADME Properties and Their Prediction 193
8.2 VolSurf Descriptors 194
8.3 Application Examples 199
8.3.1 Aqueous Solubility 200
8.3.2 Octanol/Water Partition Coefficients 204
8.3.3 Volume of Distribution (VD) 210
8.3.4 Metabolic Stability 212
8.4 Conclusion 213
9 Molecular Interaction Fields in ADME and Safety 217
9.1 Introduction 217
9.2 GRID and MIFs 218
9.3 Role of Pgp Efflux in the Absorption 219
9.3.1 Materials and Methods 219
9.3.1.1 Dataset 219
9.3.1.2 Computational Methods 219
9.3.1.3 ALMOND Descriptors 220
9.3.2 Results 220
9.3.3 Pharmacophoric Model Interpretation 222
9.3.4 Discussion 223
9.4 HERG Inhibition 224
9.4.1 Materials and Methods 224
9.4.1.1 Dataset 224
9.4.1.2 Computational Methods 224
9.4.2 Results 225
9.4.2.1 Nonbasic Nitrogen Subset 225
9.4.2.2 Ionizable Nitrogen Subset 226
9.4.2.3 Interpretation of Pharmacophoric Models 228
9.5 CYP 3A4 Inhibition 229
9.5.1 Materials and Methods 229
9.5.1.1 Dataset 229
9.5.1.2 Computational Methods 230
9.5.1.3 Ligand GRIND Descriptors 230
9.5.1.4 Protein GRIND Descriptors 230
9.5.1.5 Overlap of Structures 231
9.5.2 Results 231
9.5.2.1 Distances in the Protein Pocket 234
9.5.3 Discussion 235
9.6 Conclusions 236
10 Progress in ADME Prediction Using GRID-Molecular Interaction Fields 239
10.1 Introduction: ADME Field in the Drug Discovery Process 239
10.2 Absorption 243
10.2.1 Passive Transport, Trans-cellular Pathway 245
10.2.2 Active Transport 247
10.3 Distribution 248
10.3.1 Solubility 248
10.3.2 Unspecific Protein Binding 250
10.3.3 Volume of Distribution 250
10.4 Metabolism 252
10.4.1 Cytochrome P450 Inhibition 253
10.4.2 Site of Metabolism Prediction 253
10.4.3 Metabolic Stability 254
10.4.4 Selectivity Analysis 255
10.5 Conclusions 262
11 Rapid ADME Filters for Lead Discovery 269
11.1 Introduction 269
11.2 The Rule of Five (Ro5) as ADME Filter 270
11.3 Molecular Interaction Fields (MIFs): VolSurf 271
11.4 MIF-based ADME Models 273
11.5 Clinical Pharmacokinetics (PK) and Toxicological (Tox) Datasets 274
11.6 VolSurf in Clinical PK Data Modeling 276
11.7 ChemGPS-VolSurf (GPSVS) in Clinical PK Property Modeling 277
11.8 ADME Filters: GPSVS vs. Ro5 281
11.9 PENGUINS: Ultrafast ADME Filter 284
11.10 Integrated ADME and Binding Affinity Predictions 287
11.11 Conclusions 288
12 GRID-Derived Molecular Interaction Fields for Predicting the Site of Metabolism in Human Cytochromes 293
12.1 Introduction 293
11.2 The Human Cytochromes P450 294
12.3 CYPs Characterization using GRID Molecular Interaction Fields 295
12.4 Description of the Method 299
12.4.1 P450 Molecular Interaction Fields Transformation 300
12.4.2 3D Structure of Substrates and Fingerprint Generation 301
12.4.3 Substrate\u2013CYP Enzyme Comparison: the Recognition Component 302
12.4.4 The Reactivity Component 303
12.4.5 Computation of the Probability of a Site being the Metabolic Site 304
12.5 An Overview of the Most Significant Results 305
12.5.1 Importing Different P450 Cytochromes 307
12.6 Conclusions 309
12.7 Software Package 309
Index 311