Targeting metastable coiled-coil domains by computational design

Approximately 30% of eukaryotic genomes are predicted to encode partially unfolded proteins. Many of these unstructured domains contact multiple partners in short-lived interactions critical for cellular homeostasis. Understanding the functional implications of these transient binding events is a current challenge that could be addressed with designed peptide inhibitors. Most current protein design methodologies, however, target only structurally well-defined, stable structures. To address this limitation, we implemented a computational design strategy that alternates between a fixed backbone sequence search for binding specificity and structural optimization of the designed interfaces. We applied this method to create specific peptide inhibitors of the C-terminal metastable coiled-coil domain of the essential yeast septin Cdc12p. Specific binding of the designed sequences was demonstrated by circular dichroism and equilibrium ultracentrifugation. Our results validate computational methods to design specific peptide ligands to protein domains lacking intrinsic structural stability and set the stage for functional analysis of Cdc12p coiled coil function in vivo.     

Pairwise decomposition of an MMGBSA energy function for computational protein design

Computational protein design (CPD) aims at predicting new proteins or modifying existing ones. The computational challenge is huge as it requires exploring an enormous sequence and conformation space. The difficulty can be reduced by considering a fixed backbone and a discrete set of sidechain conformations. Another common strategy consists in precalculating a pairwise energy matrix, from which the energy of any sequence/conformation can be quickly obtained. In this work, we examine the pairwise decomposition of protein MMGBSA energy functions from a general theoretical perspective, and an implementation proposed earlier for CPD. It includes a Generalized Born term, whose many‐body character is overcome using an effective dielectric environment, and a Surface Area term, for which we present an improved pairwise decomposition. A detailed evaluation of the error introduced by the decomposition on the different energy components is performed. We show that the error remains reasonable, compared to other uncertainties. © 2014 Wiley Periodicals, Inc.     

From protein domains to drug candidates-natural products as guiding principles in the design and synthesis of compound libraries

In the continuing effort to find small molecules that alter protein function and ultimately might lead to new drugs, combinatorial chemistry has emerged as a very powerful tool. Contrary to original expectations that large libraries would result in the discovery of many hit and lead structures, it has been recognized that the biological relevance, design, and diversity of the library are more important. As the universe of conceivable compounds is almost infinite, the question arises: where is a biologically validated starting point from which to build a combinatorial library? Nature itself might provide an answer: natural products have been evolved to bind to proteins. Recent results in structural biology and bioinformatics indicate that the number of distinct protein families and folds is fairly limited. Often the same structural domain is used by many proteins in a more or less modified form created by divergent evolution. Recent progress in solid-phase organic synthesis has enabled the synthesis of combinatorial libraries based on the structure of complex natural products. It can be envisioned that natural-product-based combinatorial synthesis may permit hit or lead compounds to be found with enhanced probability and quality.     

Fast search algorithms for computational protein design

One of the main challenges in computational protein design (CPD) is the huge size of the protein sequence and conformational space that has to be computationally explored. Recently, we showed that state‐of‐the‐art combinatorial optimization technologies based on Cost Function Network (CFN) processing allow speeding up provable rigid backbone protein design methods by several orders of magnitudes. Building up on this, we improved and injected CFN technology into the well‐established CPD package Osprey to allow all Osprey CPD algorithms to benefit from associated speedups. Because Osprey fundamentally relies on the ability of to produce conformations in increasing order of energy, we defined new strategies combining CFN lower bounds, with new side‐chain positioning‐based branching scheme. Beyond the speedups obtained in the new ‐CFN combination, this novel branching scheme enables a much faster enumeration of suboptimal sequences, far beyond what is reachable without it. Together with the immediate and important speedups provided by CFN technology, these developments directly benefit to all the algorithms that previously relied on the DEE/ combination inside Osprey* and make it possible to solve larger CPD problems with provable algorithms. © 2016 Wiley Periodicals, Inc.     

An object-oriented library for computational protein design

Recent advances in computational protein design have established it as a viable technique for the rational generation of stable protein sequences, novel protein folds, and even enzymatic activity. We present a new and object-oriented library of code, written specifically for protein design applications in C(++), called EGAD Library. The modular fashion in which this library is written allows developers to tailor various energy functions and minimizers for a specific purpose. It also allows for the generation of novel protein design applications with a minimal amount of code investment. It is our hope that this will permit labs that have not considered protein design to apply it to their own systems, thereby increasing its potential as a tool in biology. We also present various uses of EGAD Library: in the development of Interaction Viewer, a PyMOL plug-in for viewing interactions between protein residues; in the repacking of protein cores; and in the prediction of protein-protein complex stabilities.     

Inclusion of inversion symmetry in centroid molecular dynamics: a possible avenue to recover quantum coherence

Inversion symmetry is included in the operator formulation of the centroid molecular dynamics (CMD). This work involves the development of a symmetry-adapted CMD (SA-CMD), here particularly for symmetrization and antisymmetrization projections. A symmetry-adapted quasidensity operator, as defined by Blinov and Roy [J. Chem. Phys. 115, 7822 (2001)], is employed to obtain the centroid representation of quantum mechanical operators. Numerical examples are given for a single particle confined to one-dimensional symmetric quartic and symmetric double-well potentials. Two SA-CMD simulations are performed separately for both projections, and centroid position autocorrelation functions are obtained. For each projection, the quality of the approximation as well as the accuracy are similar to those of regular CMD. It is shown that individual trajectories from two separate SA-CMD simulations can be properly combined to recover trajectories for Boltzmann statistics. Position autocorrelation functions are compared to the exact quantum mechanical ones. This explicit account of inversion symmetry provides a qualitative improvement on the conventional CMD approach and allows the recovery of some quantum coherence.     

Pareto optimization in computational protein design with multiple objectives

The optimization for function in computational design requires the treatment of, often competing, multiple objectives. Current algorithms reduce the problem to a single objective optimization problem, with the consequent loss of relevant solutions. We present a procedure, based on a variant of a Pareto algorithm, to optimize various competing objectives in protein design that allows reducing in several orders of magnitude the search of the solution space. Our methodology maintains the diversity of solutions and provides an iterative way to incorporate automatic design methods in the design of functional proteins. We have applied our systematic procedure to design enzymes optimized for both catalysis and stability. However, this methodology can be applied to any computational chemistry application requiring multi-objective combinatorial optimization techniques.     

Statistical analysis of computational docking of large compound data bases to distinct protein binding sites

The results of 16 docking simulations with rigid receptor sites and flexible ligands (∼60,000 compounds in each case) are statistically analyzed and compared. Different combinations of binding sites, scoring functions, and compound collections are used in these calculations. The docking scores are not randomly distributed over the scoring range; they follow Gaussian distributions (regardless of the binding sites), scoring functions, or screened compounds. If the docking sites are small, the Gaussian distributions are positively skewed. Peaks of the Gaussian distributions are populated with compounds having similar scores but different sizes and binding modes. These findings have implications for compound selection via computational docking. ©1999 John Wiley & Sons, Inc. J Comput Chem 20: 1634–1643, 1999     

A computational workflow for the design of irreversible inhibitors of protein kinases

Design of irreversible inhibitors is an emerging and relatively less explored strategy for the design of protein kinase inhibitors. In this paper, we present a computational workflow that was specifically conceived to assist such design. The workflow takes the form of a multi-step procedure that includes: the creation of a database of already known reversible inhibitors of protein kinases, the selection of the most promising scaffolds that bind one or more desired kinase templates, the modification of the scaffolds by introduction of chemically reactive groups (suitable cysteine traps) and the final evaluation of the reversible and irreversible protein–ligand complexes with molecular dynamics simulations and binding free energy predictions. Most of these steps were automated. In order to prove that this is viable, the workflow was tested on a database of known inhibitors of ERK2, a protein kinase possessing a cysteine in the ATP site. The modeled ERK2-ligand complexes and the values of the estimated binding free energies of the putative ligands provide useful indicators of their aptitude to bind reversibly and irreversibly to the protein kinase. Moreover, the computational data are used to rank the ligands according to their computed binding free energies and their ability to bind specific protein residues in the reversible and irreversible complexes, thereby providing a useful decision-making tool for each step of the design. In this work we present the overall procedure and the first proof of concept results.     

Studies toward soraphen A: an aldol-metathesis avenue to the macrocyclic framework

[reaction: see text] We describe a convergent approach to soraphen A, 1, that involves coupling of two fragments by an aldol condensation-olefin metathesis sequence. This route permits rapid access to congeners of 1.     

Reference energy extremal optimization: a stochastic search algorithm applied to computational protein design

We adapt a combinatorial optimization algorithm, extremal optimization (EO), for the search problem in computational protein design. This algorithm takes advantage of the knowledge of local energy information and systematically improves on the residues that have high local energies. Power-law probability distributions are used to select the backbone sites to be improved on and the rotamer choices to be changed to. We compare this method with simulated annealing (SA) and motivate and present an improved method, which we call reference energy extremal optimization (REEO). REEO uses reference energies to convert a problem with a structured local-energy profile to one with more random profile, and extremal optimization proves to be extremely efficient for the latter problem. We show in detail the large improvement we have achieved using REEO as compared to simulated annealing and discuss a number of other heuristics we have attempted to date.     

Mini-review: computational structure-based design of inhibitors that target protein surfaces

Finding drugs that inhibit protein-protein interactions is usually difficult. While computer-aided design is used widely to facilitate the drug discovery process for protein targets with well-defined binding pockets, its application to the design of inhibitors targeting a protein surface is very limited. In this mini-review we address two aspects of this issue: firstly, we overview the current state of design methodology for inhibitors specifically targeting protein surfaces, and secondly, we briefly outline recent advances in computational methods for structure-based drug design. These methods are closely related to protein docking and protein recognition, the difference being that in ligand design, ligands are built on a fragment-by-fragment basis. A novel scheme of computational combinatorial ligand design developed for the design of inhibitors that interfere with protein-protein interaction is described in detail. Current applications and limitations of this methodology, as well as its future prospects, are discussed.     

Nickel-mediated N-arylation with arylboronic acids: an avenue to chan-lam coupling

An efficient use of NiCl(2)·6H(2)O, for the cross-coupling of arylboronic acids with various N-nucleophiles, has been demonstrated. The method is practical and offers an alternative to the corresponding Cu-mediated Chan-Lam process for the construction of the C-N bond.     

An improved pairwise decomposable finite-difference Poisson-Boltzmann method for computational protein design

Our goal is to develop accurate electrostatic models that can be implemented in current computational protein design protocols. To this end, we improve upon a previously reported pairwise decomposable, finite difference Poisson-Boltzmann (FDPB) model for protein design (Marshall et al., Protein Sci 2005, 14, 1293). The improvement involves placing generic sidechains at positions with unknown amino acid identity and explicitly capturing two-body perturbations to the dielectric environment. We compare the original and improved FDPB methods to standard FDPB calculations in which the dielectric environment is completely determined by protein atoms. The generic sidechain approach yields a two to threefold increase in accuracy per residue or residue pair over the original pairwise FDPB implementation, with no additional computational cost. Distance dependent dielectric and solvent-exclusion models were also compared with standard FDPB energies. The accuracy of the new pairwise FDPB method is shown to be superior to these models, even after reparameterization of the solvent-exclusion model.     

Exploring experimental sources of multiple protein conformations in structure-based drug design

Receptor flexibility must be incorporated into structure-based drug design in order to portray a more accurate representation of a protein in solution. Our approach is to generate pharmacophore models based on multiple conformations of a protein and is very similar to solvent mapping of hot spots. Previously, we had success using computer-generated conformations of apo human immunodeficiency virus-1 protease (HIV-1p). Here, we examine the use of an NMR ensemble versus a collection of crystal structures, and we compare back to our previous study based on computer-generated conformations. To our knowledge, this is the first direct comparison of an NMR ensemble and a collection of crystal structures to incorporate protein flexibility in structure-based drug design. To provide an accurate comparison between the experimental sources, we used bound structures for our multiple protein structure (MPS) pharmacophore models. The models from an NMR ensemble and a collection of crystal structures were both able to discriminate known HIV-1p inhibitors from decoy molecules and displayed superior performance over models created from single conformations of the protein. Although the active-site conformations were already predefined by bound ligands, the use of MPS allows us to overcome the cross-docking problem and generate a model that does not simply reproduce the chemical characteristics of a specific ligand class. We show that there is more structural variation between 28 structures in an NMR ensemble than 90 crystal structures bound to a variety of ligands. MPS models from both sources performed well, but the model determined using the NMR ensemble appeared to be the most general yet accurate representation of the active site. This work encourages the use of NMR models in structure-based design.     

Toward the computational design of diastereomeric resolving agents: an experimental and computational study of 1-phenylethylammonium-2-phenylacetate derivatives

The crystal structures, including two new polymorphs, of three diastereomerically related salt pairs formed by (R)-1-phenylethylammonium (1) with (S&R)-2-phenylpropanoate (2), (S&R)-2-phenylbutyrate (3), and (S&R)-mandelate (4) ions were characterized by low-temperature single crystal or powder X-ray diffraction. Thermal, solubility, and solution calorimetry measurements were used to determine the relative stabilities of the salt pairs and polymorphs. These were qualitatively predicted by lattice energy calculations combining realistic models for the dominant intermolecular electrostatic interactions and ab initio calculations for the ions' conformational energies due to the distortion of their geometries by the crystal packing forces. Crystal structure prediction studies were also performed for the highly polymorphic diastereomeric salt pair (R)-1-phenylethylammonium-(S&R)-2-phenylbutyrate (1-3) in an attempt to predict the separation efficiency without relying on experimental information. This joint experimental and computational investigation provides a stringent test for the reliability of lattice modeling approaches to explain the origins of chiral resolution via diastereomer formation (Pasteurian resolution). The further developments required for the computational screening of single-enantiomer resolving agents to achieve optimal chiral separation are discussed.     

RAFT polymerization: an avenue to functional polymeric micelles for drug delivery

RAFT (Reversible addition fragmentation chain transfer) polymerization is demonstrated as a versatile tool to obtain functional and crosslinked micelles for drug delivery purposes.