High-throughput computational structure-based characterization of protein families: START domains and implications for structural genomics

Abstract

SkyLine, a high-throughput homology modeling pipeline tool, detects and models true sequence homologs to a given protein structure. Structures and models are stored in SkyBase with links to computational function annotation, as calculated by MarkUs. The SkyLine/SkyBase/MarkUs technology represents a novel structure-based approach that is more objective and versatile than other protein classification resources. This structure-centric strategy provides a multi-dimensional organization and coverage of protein space at the levels of family, function, and genome. The concept of "modelability", the ability to model sequences on related structures, provides a reliable criterion for membership in a protein family ("leverage") and underlies the unique success of this approach. The overall procedure is illustrated by its application to START domains, which comprise a Biomedical Theme for the Northeast Structural Genomics Consortium as part of the Protein Structure Initiative. START domains are typically involved in the non-vesicular transport of lipids. While 19 experimentally determined structures are available, the family, whose evolutionary hierarchy is not well determined, is highly sequence diverse, and the ligand-binding potential of many family members is unknown. The SkyLine/SkyBase/MarkUs approach provides significant insights and predicts: (1) many more family members (approximately 4,000) than any other resource; (2) the function for a large number of unannotated proteins; (3) instances of START domains in genomes from which they were thought to be absent; and (4) the existence of two types of novel proteins, those containing dual START domain and those containing N-terminal START domains.

Figure 1

Flowchart of the SkyLine pipeline. SkyLine [6] starts with a single PDB structure as input. The sequence of the structure is used as a seed in PSI-BLAST profile searches for homologs. Models are built for all non-redundant sequences detected using the input structure as the template and the sequence alignments derived from the PSI-BLAST profiles. Structure evaluation programs discern the reliable models, which comprise family members associated with the input structure. The models and associated information are stored in SkyBase, a web-accessible database.

Figure 2

Overall scheme of the computational structure-based characterization of protein families. START domain sequences and there models, calculated in the SkyLine analysis of a protein structure (PDB id 1LN1, [27]) are retrieved according to a variety of search parameters, including sequence and modeling features. Each model can be manipulated and visualized in a Jmol window, side-by-side with its Verify3D evaluation profile [37]. Links to MarkUs [13] provide access to previously calculated annotation results and the opportunity for further function analysis.

Figure 3

Analysis of the lipid-binding pocket of the human phosphatidylcholine transfer protein (PC-TP, PDB id 1LN1, [27]). A. Conservation of residues lining the pocket is calculated and displayed by ConSurf [33]. B. Electrostatic surface potential is calculated and imaged with GRASP [34]. C. Volume plot of the ligand binding pocket is generated by VOIDOO [35]. D. The PC-TP ligand DLP (1,2-dilinoleoyl-dn-glycero-3-phosphocholine) is shown in molecular representation. Residues lining the ligand-binding pocket and predicted to be functionally important are delineated in C and D.

Figure 4

Analysis of the proposed ligand-binding pocket of the Bacillus halodurans protein BH1534 (PDB id, 1XN5, [24]). A. Conservation of residues lining the pocket is calculated and displayed by ConSurf [33]. B. Curvature and conservation of the pocket surface is calculated with ConSurf [33]. C. Electrostatic surface potential is calculated and imaged with GRASP [34]. D and E. Volume plots of the putative ligand binding pocket were generated by SURFNET [36].