Computer-Assisted Drug Discovery Part I: Design, Development, Validation and Application of FRESH, a Novel In-Silico High-throughput Screening Program Part II: Monocarbonyl Curcumin Analogues: Heterocyclic Pleiotropic Kinase Inhibitors that Mediate Anticancer Properties Part III: Development of 2nd Generation NAMFIS Software Program by Java Open Access

Shi, Qi (2014)

Permanent URL: https://etd.library.emory.edu/concern/etds/4j03d035t?locale=en
Published

Abstract

There is an ever growing effort to apply computational power as a routine component of medicinal chemistry and drug discovery. In Part I of this dissertation, a novel in-silico high-throughput screening program was developed and applied to several drug discovery projects. The program, termed FRESH (FRagment-based Exploitation of modular Synthesis by virtual High Throughput Screening), combines virtual library enumeration, rapid vHTS (virtual High Throughput Screening), pharmacological property prioritizing and 2D/3D QSAR (Quantitative Structure-activity Relationship) construction. It is designed to address the issue of balancing multiple factors during drug lead-optimization of the drug discovery process. The workflow programming platform Pipeline Pilot and the corresponding programming language PilotScript were used to construct the program. The second part of the dissertation explores the mechanism behind the pleiotropic properties of mono-carbonyl curcumin analogues by molecular modeling calculations and protein sequence alignment. The last part of the dissertation reveals the mathematical principles and the Java programming approach behind the new generation of the NAMFIS (Nuclear magnetic resonance Analysis of Molecular Flexibility in Solution) software program, together with improvements on the old version.

Table of Contents

Part I: Design, Development, Validation and Application of FRESH, a Novel In-Silico High-throughput Screening Program
Chapter 1: Introduction...2

1.1 Drug molecules...2

1.1.1 Available chemical space...2
1.1.2 The concept of drug likeness...3
1.1.3 CNS drug likeness...5

1.2 Computer assisted drug discovery...7

1.2.1 Estimation of ligand-protein interactions in-silico...7
1.2.2 In-silico estimation of physical/ADMET properties...9

1.3 Lead optimization challenges...11

1.3.1 Multi-target/site therapies...11
1.3.2 Balancing multiple factors...13

Chapter 2: Design and Development of the FRESH Program...15

2.1 Overall design strategy...15
2.2 Screening software program selection...16
2.3 Construction of the FRESH program...19
2.4 Algorithm design and optimization...21
2.5 Advantages of FRESH...24

Chapter 3: FRESH Validation Case Studies...25

3.1 Introduction...25
3.2 Case I: Phosphoinositide 3-Kinase, α isoform. (PI3Kα) - Homology receptor model case...27

3.2.1 Case Background...27
3.2.2 Design of FRESH program, 1st round...28
3.2.3 Design of FRESH program, iterations...32

3.3 Case II: Carbonic Anhydrase 2 (CA II) - Crystal structure model case...34

3.3.1 Case Background...34
3.3.2 Design of FRESH program, 1st round...35
3.3.3 Design of FRESH program, iteration steps...38
3.3.4 Further experiments...41

3.4 Case III: Histone Deacetylase 1. (HDAC 1) - Ligand-only example...41

3.4.1 Case Background...41
3.4.3 Design of FRESH program, 1st round...42
3.4.4 Design of FRESH program, iterations...44

3.5 Conclusion and Future Work...45

Chapter 4: Application I: Designing novel SNRIs...47

4.1 Project background...47
4.2 Challenges in the lead optimization step...50
4.3 Receptor-based QSAR models of NET and SERT...51
4.4 Application of the FRESH program...55

4.4.1 The FRESH Program design...55
4.4.2 Resulting structures...57
4.4.3 Test result and comparison...58

4.5 Conclusion and future direction...59

Chapter 5: Application II: Identification of novel KCN1 analogs to block the p300/KCN1 interaction...61

5.1 Project background...61
5.2 Challenge in the lead optimization step...62
5.3 Receptor-based QSAR approach...63

5.3.1 Binding receptor selection...63
5.3.2 Binding site selection...66
5.3.3 Validate the scoring functions...69

5.4 Application of the FRESH program...73

5.4.1 The FRESH program design...73
5.4.2 Resulting structures...75

5.5 Application of FRESH to the new scaffold...79

5.5.1 New compound scaffold...79
5.5.2 Construction of QSAR...80
5.5.3 The FRESH program and the resulting structures...81
5.5.4 Future directions...83

Part I: Conclusions and Future Directions...85
Part II: Monocarbonyl Curcumin Analogues: Heterocyclic Pleiotropic Kinase Inhibitors That Mediate Anticancer Properties...86
Chapter 6: Curcumin analogs as Pleiotropic Kinase Blockers...87

6.1 Project background...87
6.2 Modeling of curcumin analogs...89

6.2.1 Comparison of AKT-1 and AKT-2...89
6.2.2 Docking pose analysis...90
6.2.3 Sequence Comparison of Kinase Binding Sites...96

6.3 Conclusion...97

Part III: Development of 2nd Generation NAMFIS Software Program by Java...99
Chapter 7: Design and Improvement of the NAMFIS Program...100

7.1 NAMFIS background...100

7.1.1 The Problem of Force Field Methods...100
7.1.2 Mathematical Background for NAMFIS...101
7.1.3 Problem with the Current Version of NAMFIS...102

7.2 New Generation of NAMFIS...104

7.2.1 Java VS Python...104
7.2.2 GUI and Backend Design...105
7.2.3 Minimization Step...107
7.2.4 Exception Handling...108

7.3 Conclusion and Future Work...110

Chapter 8: Experimental Details...112

8.1 Pipeline Pilot Tasks...112

8.1.1 Filter for desirable fragments/compounds by substructure matching...112
8.1.2 Select fragments/compounds according to physical/ADMET property values...114
8.1.3 Process the commercial library to synthetic fragments...116
8.1.4 Covalently attach fragments to the core structures...116
8.1.5 Remove duplicate structures...117
8.1.6 Merge data...118
8.1.7 Program debugging examples...118

8.2 Glide Task...120

8.2.1 Prepare protein receptor for docking and sequence comparisons...120
8.2.2 Generate receptor docking grids around ligand binding sites...121
8.2.3 Set ligand docking parameters to control output...121

8.3 MM-GBSA Task...121

8.3.1 Parameter settings for energy refinement...121

8.4 Miscellaneous Schrodinger Tasks...122

8.4.1 LigPrep - 2D to 3D structure conversion...122
8.4.2 QikProp - physical ADMET property estimates...122
8.4.3 Protein sequence alignment...122
8.4.4 Homology modeling...122
8.4.5 Induced fit docking - Flexible ligand and protein interaction...123

8.5 Additional Specific Details...123

8.5.1 PI3Kα case study...123
8.5.2 CA II case study...127

8.6 Java Programming Language Implementation for NAMFIS...129

8.6.1 Perform user input validation...138
8.6.2 Perform constrained geometry and population minimization...140

About this Dissertation

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
School
Department
Degree
Submission
Language
  • English
Research Field
Keyword
Committee Chair / Thesis Advisor
Committee Members
Last modified