Publications

Abstract not provided.

Sandia National Laboratories and DSM Innovation, Inc. collaborated on the investigation of the structure and function of cellulases from thermophilic fungi. Sandia's role was to use its expertise in protein structure determination and X-ray crystallography to solve the structure of these enzymes in their native state and in their substrate and product bound states. Sandia was also tasked to work with DSM to use the newly solved structure to, using computational approaches, analyze enzyme interactions with both bound substrate and bound product; the goal being to develop approaches for rationally designing improved cellulases for biomass deconstruction. We solved the structures of five cellulases from thermophilic fungi. Several of these were also solved with bound substrate/product, which allowed us to predict mutations that might enhance activity and stability.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

The innate immune system represents our first line of defense against microbial pathogens, and in many cases is activated by recognition of pathogen cellular components (dsRNA, flagella, LPS, etc.) by cell surface membrane proteins known as toll-like receptors (TLRs). As the initial trigger for innate immune response activation, TLRs also represent a means by which we can effectively control or modulate inflammatory responses. This proposal focused on TLR4, which is the cell-surface receptor primarily responsible for initiating the innate immune response to lipopolysaccharide (LPS), a major component of the outer membrane envelope of gram-negative bacteria. The goal was to better understand TLR4 activation and associated membrane proximal events, in order to enhance the design of small molecule therapeutics to modulate immune activation. Our approach was to reconstitute the receptor in biomimetic systems in-vitro to allow study of the structure and dynamics with biophysical methods. Structural studies were initiated in the first year but were halted after the crystal structure of the dimerized receptor was published early in the second year of the program. Methods were developed to determine the association constant for oligomerization of the soluble receptor. LPS-induced oligomerization was observed to be a strong function of buffer conditions. In 20 mM Tris pH 8.0 with 200 mM NaCl, the onset of receptor oligomerization occurred at 0.2 uM TLR4/MD2 with E coli LPS Ra mutant in excess. However, in the presence of 0.5 uM CD14 and 0.5 uM LBP, the onset of receptor oligomerization was observed to be less than 10 nM TLR4/MD2. Several methods were pursued to study LPS-induced oligomerization of the membrane-bound receptor, including CryoEM, FRET, colocalization and codiffusion followed by TIRF, and fluorescence correlation spectroscopy. However, there approaches met with only limited success.

Abstract not provided.

Abstract not provided.

New challenges in knowledge extraction include interpreting and classifying data sets while simultaneously considering related information to confirm results or identify false positives. We discuss a data fusion algorithmic framework targeted at this problem. It includes separate base classifiers for each data type and a fusion method for combining the individual classifiers. The fusion method is an extension of current ensemble classification techniques and has the advantage of allowing data to remain in heterogeneous databases. In this paper, we focus on the applicability of such a framework to the protein phosphorylation prediction problem. © Springer Science+Business Media, LLC 2008.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Motivation: Identifying protein enzymatic or pharmacological activities are important areas of research in biology and chemistry. Biological and chemical databases are increasingly being populated with linkages between protein sequences and chemical structures. There is now sufficient information to apply machine - learning techniques to predict interactions between chemicals and proteins at a genome scale. Current machine-learning techniques use as input either protein sequences and structures or chemical information. We propose here a method to infer protein - chemical interactions using heterogeneous input consisting of both protein sequence and chemical information. Results: Our method relies on expressing proteins and chemicals with a common cheminformatics representation. We demonstrate our approach by predicting whether proteins can catalyze reactions not present in training sets. We also predict whether a given drug can bind a target, in the absence of prior binding information for that drug and target. Such predictions cannot be made with current machine - learning techniques requiring binding information for individual reactions or individual targets. © 2007 The Author(s).

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

The goal of this project was to examine the protein-protein docking problem, especially as it relates to homology-based structures, identify the key bottlenecks in current software tools, and evaluate and prototype new algorithms that may be developed to improve these bottlenecks. This report describes the current challenges in the protein-protein docking problem: correctly predicting the binding site for the protein-protein interaction and correctly placing the sidechains. Two different and complementary approaches are taken that can help with the protein-protein docking problem. The first approach is to predict interaction sites prior to docking, and uses bioinformatics studies of protein-protein interactions to predict theses interaction site. The second approach is to improve validation of predicted complexes after docking, and uses an improved scoring function for evaluating proposed docked poses, incorporating a solvation term. This scoring function demonstrates significant improvement over current state-of-the art functions. Initial studies on both these approaches are promising, and argue for full development of these algorithms.

Membrane proteins make up a diverse and important subset of proteins for which structural information is limited. In this study, chemical cross-linking and mass spectrometry were used to explore the structure of the G-protein-coupled photoreceptor bovine rhodopsin in the dark-state conformation. All experiments were performed in rod outer segment membranes using amino acid 'handles' in the native protein sequence and thus minimizing perturbations to the native protein structure. Cysteine and lysine residues were covalently cross-linked using commercially available reagents with a range of linker arm lengths. Following chemical digestion of cross-linked protein, cross-linked peptides were identified by accurate mass measurement using liquid chromatography-fourier transform mass spectrometry and an automated data analysis pipeline. Assignments were confirmed and, if necessary, resolved, by tandem MS. The relative reactivity of lysine residues participating in cross-links was evaluated by labeling with NHS-esters. A distinct pattern of cross-link formation within the C-terminal domain, and between loop I and the C-terminal domain, emerged. Theoretical distances based on cross-linking were compared to inter-atomic distances determined from the energy-minimized X-ray crystal structure and Monte Carlo conformational search procedures. In general, the observed cross-links can be explained by re-positioning participating side-chains without significantly altering backbone structure. One exception, between C3 16 and K325, requires backbone motion to bring the reactive atoms into sufficient proximity for cross-linking. Evidence from other studies suggests that residues around K325 for a region of high backbone mobility. These findings show that cross-linking studies can provide insight into the structural dynamics of membrane proteins in their native environment.

A deterministic algorithm for enumeration of transmembrane protein folds is presented. Using a set of sparse pairwise atomic distance constraints (such as those obtained from chemical cross-linking, FRET, or dipolar EPR experiments), the algorithm performs an exhaustive search of secondary structure element packing conformations distributed throughout the entire conformational space. The end result is a set of distinct protein conformations, which can be scored and refined as part of a process designed for computational elucidation of transmembrane protein structures.