Benchmarking of TASSER in the ab initio limit

A significant number of protein sequences in a given proteome have no obvious evolutionarily related protein in the database of solved protein structures, the PDB. Under these conditions, ab initio or template-free modeling methods are the sole means of predicting protein structure. To assess its expected performance on proteomes, the TASSER structure prediction algorithm is benchmarked in the ab initio limit on a representative set of 1129 nonhomologous sequences ranging from 40 to 200 residues that cover the PDB at 30% sequence identity and which adopt alpha, alpha + beta, and beta secondary structures. For sequences in the 40-100 (100-200) residue range, as assessed by their root mean square deviation from native, RMSD, the best of the top five ranked models of TASSER has a global fold that is significantly close to the native structure for 25% (16%) of the sequences, and with a correct identification of the structure of the protein core for 59% (36%). In the absence of a native structure, the structural similarity among the top five ranked models is a moderately reliable predictor of folding accuracy. If we classify the sequences according to their secondary structure content, then 64% (36%) of alpha, 43% (24%) of alpha + beta, and 20% (12%) of beta sequences in the 40-100 (100-200) residue range have a significant TM-score (TM-score > or = 0.4). TASSER performs best on helical proteins because there are less secondary structural elements to arrange in a helical protein than in a beta protein of equal length, since the average length of a helix is longer than that of a strand. In addition, helical proteins have shorter loops and dangling tails. If we exclude these flexible fragments, then TASSER has similar accuracy for sequences containing the same number of secondary structural elements, irrespective of whether they are helices and/or strands. Thus, it is the effective configurational entropy of the protein that dictates the average likelihood of correctly arranging the secondary structure elements.     

Assembly and crystallization of the complex between the human T cell coreceptor CD8alpha homodimer and HLA-A2.

A strategy for overexpression in Escherichia coli of the extracellular immunoglobulin domain of human CD8alpha was devised using codon usage alterations in the 5' region of the gene, designed so as to prevent the formation of secondary structures in the mRNA. A fragment of CD8alpha, comprising residues 1-120 of the mature protein, excluding the signal peptide and the membrane-proximal stalk region, was recovered from bacterial inclusion bodies and refolded to produce a single species of homodimeric, soluble receptor. HLA-A2 heavy chain, beta2-microglobulin and a synthetic peptide antigen corresponding to the pol epitope from HIV-1 were also expressed in E. coli, refolded and purified. CD8alpha/HLA-A2 complexes were formed in solution and by co-crystallization with a stoichiometry of one CD8alpha alpha dimer to one HLA-A2-peptide unit.     

Identification and characterization of key substructures involved in the early folding events of a (beta/alpha)8-barrel protein as studied by experimental and computational methods

A number of studies have examined the structural properties of late folding intermediates of (beta/alpha)8-barrel proteins involved in tryptophan biosynthesis, whereas there is little information available about the early folding events of these proteins. To identify the contiguous polypeptide segments important to the folding of the (beta/alpha)8-barrel protein Escherichia coli N-(5'-phosphoribosyl)anthranilate isomerase, we structurally characterized fragments and circularly permuted forms of the protein. We also simulated thermal unfolding of the protein using molecular dynamics. Our fragmentation experiments demonstrate that the isolated (beta/alpha)(1-4)beta5 fragment is almost as stable as the full-length protein. The far and near-UV CD spectra of this fragment are indicative of native-like secondary and tertiary structures. Structural analysis of the circularly permutated proteins shows that if the protein is cleaved within the two N-terminal betaalpha modules, the amount of secondary structure is unaffected, whereas, when cleaved within the central (beta/alpha)(3-4)beta5 segment, the protein simply cannot fold. An ensemble of the denatured structures produced by thermal unfolding simulations contains a persistent local structure comprised of beta3, beta4 and beta5. The presence of this three-stranded beta-barrel suggests that it may be an important early-stage folding intermediate. Interactions found in (beta/alpha)(3-4)beta5 may be essential for the early events of ePRAI folding if they provide a nucleation site that directs folding.     

Analysis of secondary structure in proteins by chemical cross-linking coupled to MS

Chemical cross-linking is an attractive technique for the study of the structure of protein complexes due to its low sample consumption and short analysis time. Furthermore, distance constraints obtained from the identification of cross-linked peptides by MS can be used to construct and validate protein models. If a sufficient number of distance constraints are obtained, then determining the secondary structure of a protein can allow inference of the protein's fold. In this work, we show how the distance constraints obtained from cross-linking experiments can identify secondary structures within the protein sequence. Molecular modeling of alpha helices and beta sheets reveals that each secondary structure presents different cross-linking possibilities due to the topological distances between reactive residues. Cross-linking experiments performed with amine reactive cross-linkers with model alpha helix containing proteins corroborated the molecular modeling predictions. The cross-linking patterns established here can be extended to other cross-linkers with known lengths for the determination of secondary structures in proteins.     

Prediction of the tertiary structure of the alpha-subunit of tryptophan synthase

The tertiary structure of the alpha-subunit of tryptophan synthase was proposed using a combination of experimental data and computational methods. The vacuum-ultraviolet circular dichroism spectrum was used to assign the protein to the alpha/beta-class of supersecondary structures. The two-domain structure of the alpha-subunit (Miles et al.: Biochemistry 21:2586, 1982; Beasty and Matthews: Biochemistry 24:3547, 1985) eliminated consideration of a barrel structure and focused attention on a beta-sheet structure. An algorithm (Cohen et al.: Biochemistry 22:4894, 1983) was used to generate a secondary structure prediction that was consistent with the sequence data of the alpha-subunit from five species. Three potential secondary structures were then packed into tertiary structures using other algorithms. The assumption of nearest neighbors from second-site revertant data eliminated 97% of the possible tertiary structures; consideration of conserved hydrophobic packing regions on the beta-sheet eliminated all but one structure. The native structure is predicted to have a parallel beta-sheet flanked on both sides by alpha-helices, and is consistent with the available data on chemical cross-linking, chemical modification, and limited proteolysis. In addition, an active site region containing appropriate residues could be identified as well as an interface for beta 2-subunit association. The ability of experimental data to facilitate the prediction of protein structure is discussed.     

NMR structure of the second intracellular loop of the alpha 2A adrenergic receptor: evidence for a novel cytoplasmic helix

A major, unresolved question in signal transduction by G protein coupled receptors (GPCRs) is to understand how, at atomic resolution, a GPCR activates a G protein. A step toward answering this question was made with the determination of the high-resolution structure of rhodopsin; we now know the intramolecular interactions that characterize the resting conformation of a GPCR. To what degree does this structure represent a structural paradigm for other GPCRs, especially at the cytoplasmic surface where GPCR-G protein interaction occurs and where the sequence homology is low among GPCRs? To address this question, we performed NMR studies on approximately 35-residue-long peptides including the critical second intracellular loop (i2) of the alpha 2A adrenergic receptor (AR) and of rhodopsin. To stabilize the secondary structure of the peptide termini, 4-12 residues from the adjacent transmembrane helices were included and structures determined in dodecylphosphocholine micelles. We also characterized the effects on an alpha 2A AR peptide of a D130I mutation in the conserved DRY motif. Our results show that in contrast to the L-shaped loop in the i2 of rhodopsin, the i2 of the alpha 2A AR is predominantly helical, supporting the hypothesis that there is structural diversity within GPCR intracellular loops. The D130I mutation subtly modulates the helical structure. The spacing of nonpolar residues in i2 with helical periodicity is a predictor of helical versus loop structure. These data should lead to more accurate models of the intracellular surface of GPCRs and of receptor-mediated G protein activation.     

3D structure and significance of the GPhiXXG helix packing motif in tetramers of the E1beta subunit of pyruvate dehydrogenase from the archeon Pyrobaculum aerophilum.

As part of a structural genomics project, we have determined the 2.0 A structure of the E1beta subunit of pyruvate dehydrogenase from Pyrobaculum aerophilum (PA), a thermophilic archaeon. The overall fold of E1beta from PA is closely similar to the previously determined E1beta structures from humans (HU) and P. putida (PP). However, unlike the HU and PP structures, the PA structure was determined in the absence of its partner subunit, E1alpha. Significant structural rearrangements occur in E1beta when its E1alpha partner is absent, including rearrangement of several secondary structure elements such as helix C. Helix C is buried by E1alpha in the HU and PP structures, but makes crystal contacts in the PA structure that lead to an apparent beta(4) tetramer. Static light scattering and sedimentation velocity data are consistent with the formation of PA E1beta tetramers in solution. The interaction of helix C with its symmetry-related counterpart stabilizes the tetrameric interface, where two glycine residues on the same face of one helix create a packing surface for the other helix. This GPhiXXG helix-helix interaction motif has previously been found in interacting transmembrane helices, and is found here at the E1alpha-E1beta interface for both the HU and PP alpha(2)beta(2) tetramers. As a case study in structural genomics, this work illustrates that comparative analysis of protein structures can identify the structural significance of a sequence motif.     

Crystal structure of DFA0005 complexed with alpha-ketoglutarate: a novel member of the ICL/PEPM superfamily from alkali-tolerant Deinococcus ficus

The crystal structure of the DFA0005 protein complexed with alpha-ketoglutarate (AKG) from an alkali-tolerant bacterium Deinococcus ficus has been determined to a resolution of 1.62 A. The monomer forms an incomplete alpha7/beta8 barrel with a protruding alpha8 helix that interacts extensively with another subunit to form a stable dimer of two complete alpha8/beta8 barrels. The dimer is further stabilized by four glycerol molecules situated at the interface. One unique AKG ligand binding pocket per subunit is detected. Fold match using the DALI and SSE servers identifies DFA0005 as belonging to the isocitrate lyase/phosphoenolpyruvate mutase (ICL/PEPM) superfamily. However, further detailed structural and sequence comparison with other members in this superfamily and with other families containing AKG ligand indicate that DFA0005 protein exhibits considerable distinguishing features of its own and can be considered a novel member in this ICL/PEPM superfamily.     

Selection and characterization of small random transmembrane proteins that bind and activate the platelet-derived growth factor beta receptor

Growth factor receptors are typically activated by the binding of soluble ligands to the extracellular domain of the receptor, but certain viral transmembrane proteins can induce growth factor receptor activation by binding to the receptor transmembrane domain. For example, homodimers of the transmembrane 44-amino acid bovine papillomavirus E5 protein bind the transmembrane region of the PDGF beta receptor tyrosine kinase, causing receptor dimerization, phosphorylation, and cell transformation. To determine whether it is possible to select novel biologically active transmembrane proteins that can activate growth factor receptors, we constructed and identified small proteins with random hydrophobic transmembrane domains that can bind and activate the PDGF beta receptor. Remarkably, cell transformation was induced by approximately 10% of the clones in a library in which 15 transmembrane amino acid residues of the E5 protein were replaced with random hydrophobic sequences. The transformation-competent transmembrane proteins formed dimers and stably bound and activated the PDGF beta receptor. Genetic studies demonstrated that the biological activity of the transformation-competent proteins depended on specific interactions with the transmembrane domain of the PDGF beta receptor. A consensus sequence distinct from the wild-type E5 sequence was identified that restored transforming activity to a non-transforming poly-leucine transmembrane sequence, indicating that divergent transmembrane sequence motifs can activate the PDGF beta receptor. Molecular modeling suggested that diverse transforming sequences shared similar protein structure, including the same homodimer interface as the wild-type E5 protein. These experiments have identified novel proteins with transmembrane sequences distinct from the E5 protein that can activate the PDGF beta receptor and transform cells. More generally, this approach may allow the creation and identification of small proteins that modulate the activity of a variety of cellular transmembrane proteins.     

The solution structure of the pH-induced monomer of dynein light-chain LC8 from Drosophila

The structure of Drosophila LC8 pH-induced monomer has been determined by NMR spectroscopy using the program AutoStructure. The structure at pH 3 and 30 degrees C is similar to the individual subunits of mammalian LC8 dimer with the exception that a beta strand, which crosses between monomers to form an intersubunit beta-sheet in the dimer, is a flexible loop with turnlike conformations in the monomer. Increased flexibility in the interface region relative to the rest of the protein is confirmed by dynamic measurements based on (15)N relaxation. Comparison of the monomer and dimer structures indicates that LC8 is not a domain swapped dimer.     

Reconstruction of protein conformations from estimated positions of the C alpha coordinates.

Protein C alpha coordinates are used to accurately reconstruct complete protein backbones and side-chain directions. This work employs potentials of mean force to align semirigid peptide groups around the axes that connect successive C alpha atoms. The algorithm works well for all residue types and secondary structure classes and is stable for imprecise C alpha coordinates. Tests on known protein structures show that root mean square errors in predicted main-chain and C beta coordinates are usually less than 0.3 A. These results are significantly more accurate than can be obtained from competing approaches, such as modeling of backbone conformations from structurally homologous fragments.     

Truncation of the ectodomain of the human insulin receptor results in secretion of a soluble insulin binding protein from transfected CHO cells

The insulin receptor is an integral transmembrane glycoprotein comprised of two alpha-(approximately 135 kDa) and two beta-(approximately 95 kDa) subunits, which is synthesized as a single polypeptide chain precursor (alpha beta). The primary sequence of the human insulin receptor (hIR) protein, deduced from the nucleotide sequence of cloned human placental mRNAs, predicts two large domains (929 and 403 residues) on either side of a single membrane spanning domain (23 residues); each of these major domains has a distinct function (insulin binding and protein/tyrosine kinase activity, respectively). To experimentally test this deduced topology, and to explore the potential for independent domain function by the hIR extracellular domain, we have constructed an expression plasmid encoding an hIR deletion mutant which is truncated 8 residues from the beginning of the predicted transmembrane domain (i.e., 921 residues). This domain of the hIR is in fact processed into alpha- and truncated beta-subunits and secreted with high efficiency from transfected CHO cell lines which express this mutant hIR, and the protein accumulates as an (alpha beta)2 dimer in the medium. This molecule is recognized by a battery of 13 monoclonal antibodies to epitopes on the IR extracellular domain, four of which block insulin binding and two of which require the native conformation of the IR for recognition. Further, this domain binds insulin with an apparent dissociation constant comparable to that of the wild-type hIR. However, the secreted dimer displays a linear Scatchard plot, while that of the wild-type membrane-associated hIR is curvilinear.(ABSTRACT TRUNCATED AT 250 WORDS)     

Urea treatment allows dithiothreitol to release the binding subunit of the insulin receptor from the cell membrane: implications for the structural organization of the insulin receptor

The sequence of the human insulin receptor has only one identifiable transmembrane region which is located in the beta subunit. The structure predicts that the alpha subunit, which binds insulin, is attached to the cell only by disulfide bonds to the beta subunit. However, treatment of membranes with dithiothreitol is ineffective at releasing the alpha subunit. If the receptor structure is unfolded with urea, dithiothreitol is able to release the alpha subunit. These data provided confirmatory evidence that the alpha subunit is not a transmembrane protein.     

Crystal structures of SnoaL2 and AclR: two putative hydroxylases in the biosynthesis of aromatic polyketide antibiotics

SnoaL2 and AclR are homologous enzymes in the biosynthesis of the aromatic polyketides nogalamycin in Streptomyces nogalater and cinerubin in Streptomyces galilaeus, respectively. Evidence obtained from gene transfer experiments suggested that SnoaL2 catalyzes the hydroxylation of the C-1 carbon atom of the polyketide chain. Here we show that AclR is also involved in the production of 1-hydroxylated anthracyclines in vivo. The three-dimensional structure of SnoaL2 has been determined by multi-wavelength anomalous diffraction to 2.5A resolution, and that of AclR to 1.8A resolution using molecular replacement. Both enzymes are dimers in solution and in the crystal. The fold of the enzyme subunits consists of an alpha+beta barrel. The dimer interface is formed by packing of the beta-sheets from the two subunits against each other. In the interior of the alpha+beta barrel a hydrophobic cavity is formed that most likely binds the substrate and harbors the active site. The subunit fold and the architecture of the active site in SnoaL2 and AclR are similar to that of the polyketide cyclases SnoaL and AknH; however, they show completely different quaternary structures. A comparison of the active site pockets of the putative hydroxylases AclR and SnoaL2 with those of bona fide polyketide cyclases reveals distinct differences in amino acids lining the cavity that might be responsible for the switch in chemistry. The moderate degree of sequence similarity and the preservation of the three-dimensional fold of the polypeptide chain suggest that these enzymes are evolutionary related. Members of this enzyme family appear to have evolved from a common protein scaffold by divergent evolution to catalyze reactions chemically as diverse as aldol condensation and hydroxylation.     

Secondary structure of the human membrane-associated folate binding protein using a joint prediction approach

The secondary structures of the human membrane-associated folate binding protein (FBP) and bovine soluble FBP are assessed by a joint prediction approach that combines neural network models, information theory, homology modeling and the Chou-Fasman methods. Two new profile maps are used to characterize the non-regular secondary structure and to assist in assigning buried and exposed parts of secondary structure: (i) the loop potential profile and (ii) the long range contact profile. Approximately half of human FBP is predicted to form regular secondary structure (alpha-helices-35% or beta-sheets - 12%, excluding the transmembrane helices) and the rest is predicted to form coil, turns or loops. The bovine milk soluble FBP is predicted to have a similar secondary structure as expected because of the high degree of homology between the FBP's. Discriminant analysis predicts two transmembrane segments for the human FBP sequence, one at the amino terminus (a leader sequence) and the other at the carboxy terminus. These predicted transmembrane domains are absent in the bovine milk soluble FBP, further supporting these predictions. The present set of secondary structural predictions for human FBP is obtained by 'consensus' to aid in modeling the super-secondary structure of the protein.     

Computational exploration of the network of sequence flow between protein structures

We investigate small sequence adjustments (of one or a few amino acids) that induce large conformational transitions between distinct and stable folds of proteins. Such transitions are intriguing from evolutionary and protein‐design perspectives. They make it possible to search for ancient protein structures or to design protein switches that flip between folds and functions. A network of sequence flow between protein folds is computed for representative structures of the Protein Data Bank. The computed network is dense, on an average each structure is connected to tens of other folds. Proteins that attract sequences from a higher than expected number of neighboring folds are more likely to be enzymes and alpha/beta fold. The large number of connections between folds may reflect the need of enzymes to adjust their structures for alternative substrates. The network of the Cro family is discussed, and we speculate that capacity is an important factor (but not the only one) that determines protein evolution. The experimentally observed flip from all alpha to alpha + beta fold is examined by the network tools. A kinetic model for the transition of sequences between the folds (with only protein stability in mind) is proposed. Proteins 2010. © 2009 Wiley‐Liss, Inc.     

Recognition and architecture of the framework structure of protein

Based on the concept that the framework structure of a protein is determined by its secondary structure sequence, a new method for recognition and prediction of the structural class is suggested. By use of parameters N(alpha), N(beta), and N(beta(alpha)beta) (the number of alpha-helices, beta-strands, and beta(alpha)beta fragments), one can recognize the structural class with an accuracy higher than 90% when applied to the complete set (standard set) published in October 1995 and the structure data newly released before July 1998 (test set). Furthermore, the framework structures of beta, alpha, and alpha/beta protein are studied. It is found that these structures can be built from some basic units and that their architecture obeys some definite rules. Based on the packing of these basic units a set of rules for the recognition of topologies of the framework structure are worked out. When applied to the 1995 standard set and the 1998 test set the rates of correct recognition are higher than 77%. The simplicity and universality of framework structures are indicated which may be related to the evolutionary conservation of these folds. Proteins 2000;39:9-25.     

Major structural determinants of transmembrane proteins identified by principal component analysis

We identify amino acid characteristics important in determining the secondary structures of transmembrane proteins, and compare them with characteristics important for cytoplasmic proteins. Using information derived from multiple sequence alignments, we perform a principal component analysis (PCA) to identify the directions in the 20-dimensional amino acid frequency space that comprise the most variance within each protein secondary structure. These vectors represent the important position-specific properties of the amino acids for coils, turns, beta sheets, and alpha helices. As expected, the most important axis for most of the datasets was hydrophobicity. Additional axes, distinct from hydrophobicity, are surprising, especially in the case of transmembrane alpha helices, where the effects of aromaticity and beta-branching are the next two most significant characteristics. The axis representing beta-branching also has equal importance in cytoplasmic and transmembrane helices, a finding that contrasts with some experimental results in membrane-like environments. In a further analysis, we examine trends for some of the PCA axes over averaged transmembrane alpha helices, and find interesting results for aromaticity.