Comparative Protein Modeling

Abstract

Chapter 2 provides a tutorial on the topic of comparative protein modeling, otherwise known as homology modeling. While many proteins from similar families have similar functions, it is common to find instances where proteins with similar structures have different functions. The authors describe in this chapter how to first construct a protein structure and then how to validate its quality as a model. The first step in homology modeling is to search for known, related sequences and structures by using, for example, the Protein Data Bank (PDB), or the Expert Protein Analysis System (ExPaSy) website which contains useful databases like SWISS-PROT, PROSITE, ENZYME, and SWISS-MODEL. Details about these databases along with pitfalls to avoid when using them are provided. The next step, which is most critical in a comparative modeling study, is sequence alignment. Both global, coarse-grained alignment strategies and local, fine-grained alignment strategies are described. The basics of alignment are given for the novice modeler, insights about sequence preparation are passed on to the reader, and commonly used alignment tools like BLAST, Clustal (and their progeny), T-Coffee, and Divide-and-Conquer are described. The differences between progressive and fragment-based methodologies are highlighted and a description about how one scores the final alignment so as to select the best model is given. The next two steps in homology modeling involve template selection and improving alignments. Methods like threading and uses of hydropathy plots are described before a tutorial is presented on how to actually construct a protein model. The difference between finding the best model versus a consensus model is highlighted, as is the need for satisfying spatial constraints. Segment match modeling, multiple template methods, hidden Markov modeling and other techniques are identified and explained for the novice. The penultimate step of refining the protein structures using, e.g., databases like Side-Chains with Rotamer Library (SCWRL) or by implementing atomistic simulation methods like simulated annealing is then described. Finally, the authors inform us about how to evaluate the validity of the derived protein structures using PROCHECK, Verify3D, ProSa, and PROVE in addition to existing tools from the realm of spectroscopy such as found in the OLERADO suite of programs. For each step of the homology modeling process the authors provide a worked example to illustrate some of the problems and pitfalls a novice could encounter, and, they provide tables of key websites containing databases and computational resources needed for homology modeling. The chapter consists of: Introduction Anatomy of a Comparative Model Searching for Related Sequences and Structures Expert Protein Analysis System (ExPASy) Protein Data Bank (PDB) Example: Finding Related Sequences and 3D Structures Sequence Alignment Preparing the Sequences Alignment Basics Similarity Matrices Basic Local Alignment Search Tool (BLAST) and PSI-BLAST Clustal Tree-based Consistency Objective Function for alignment Evaluation (T-Coffee) Divide-and-Conquer Alignment (DCA) Example: Sequence Alignment Selecting Templates and Improving Alignments Selecting Templates Threading Improving Sequence Alignments with Primary and Secondary Structure Analysis Example: Aligning the Target to the Selected Template Constructing Protein Models Overall Protein Model Construction Methods Satisfaction of Spatial Restraints Segment Match Modeling Multiple Template Method Example: Constructing a Protein Model Refinement of Protein Models Side-Chains with Rotamer Library (SCWRL) Energy Minimization Molecular Dynamics Molecular Dynamics with Simulated Annealing Evaluating Protein Models PROCHECK Verify3D ERRAT Protein Structure Analysis (ProSa) Protein Volume Evaluation (PROVE) Model Clustering Analysis Example: Evaluation of Protein Models Conclusions References