Occluded molecular surface: analysis of protein packing

Lessons on protein structure from interleukin-4: All disulfides are not created equal

Interleukin-4 (IL-4) is a hematopoietic cytokine composed by a four-helix bundle stabilized by an antiparallel beta-sheet and three disulfide bonds: Cys3-Cys127, Cys24-Cys65, and Cys46-Cys99. IL-4 is involved in several immune responses associated to infection, allergy, autoimmunity, and cancer. Besides its physiological relevance, IL-4 is often used as a "model" for protein design and engineering. Hence, to understand the role of each disulfide in the structure and dynamics of IL-4, we carried out several spectroscopic analyses (circular dichroism [CD], fluorescence, nuclear magnetic resonance [NMR]), and molecular dynamics (MD) simulations on wild-type IL-4 and four IL-4 disulfide mutants. All disulfide mutants showed loss of structure, altered interhelical angles, and looser core packings, showing that all disulfides are relevant for maintaining the overall fold and stability of the four-helix bundle motif, even at very low pH. In the absence of the disulfide connecting both protein termini Cys3-Cys127, C3T-IL4 showed a less packed protein core, loss of secondary structure (~9%) and fast motions on the sub-nanosecond time scale (lower S2 order parameters and larger τc correlation time), especially at the two protein termini, loops, beginning of helix A and end of helix D. In the absence of Cys24-Cys65, C24T-IL4 presented shorter alpha-helices (14% loss in helical content), altered interhelical angles, less propensity to form the small anti-parallel beta-sheet and increased dynamics. Simultaneously deprived of two disulfides (Cys3-Cys127 and Cys24-Cys65), IL-4 formed a partially folded "molten globule" with high 8-anilino-1-naphtalenesulphonic acid-binding affinity and considerable loss of secondary structure (~50%decrease), as shown by the far UV-CD, NMR, and MD data.

Statistical Analysis of the Role of Cavity Flexibility in Thermostability of Proteins

Conventional statistical investigations have primarily focused on the comparison of the simple one-dimensional characteristics of protein cavities, such as number, surface area, and volume. These studies have failed to discern the crucial distinctions in cavity properties between thermophilic and mesophilic proteins that contribute to protein thermostability. In this study, the significance of cavity properties, i.e., flexibility and location, in protein thermostability was investigated by comparing structural differences between homologous thermophilic and mesophilic proteins. Three dimensions of protein structure were categorized into three regions (core, boundary, and surface) and a comparative analysis of cavity properties using this structural index was conducted. The statistical analysis revealed that cavity flexibility is closely related to protein thermostability. The core cavities of thermophilic proteins were less flexible than those of mesophilic proteins (averaged B' factor values, -0.6484 and -0.5111), which might be less deleterious to protein thermostability. Thermophilic proteins exhibited fewer cavities in the boundary and surface regions. Notably, cavities in mesophilic proteins, across all regions, exhibited greater flexibility than those in thermophilic proteins (>95% probability). The increased flexibility of cavities in the boundary and surface regions of mesophilic proteins, as opposed to thermophilic proteins, may compromise stability. Recent protein engineering investigations involving mesophilic xylanase and protease showed results consistent with the findings of this study, suggesting that the manipulation of flexible cavities in the surface region can enhance thermostability. Consequently, our findings suggest that a rational or computational approach to the design of flexible cavities in surface or boundary regions could serve as an effective strategy to enhance the thermostability of mesophilic proteins.

Interleukin-2 superkines by computational design

While computational engineering of therapeutic proteins is a desirable goal, in practice the optimization of protein–protein interactions requires substantial experimental intervention. We present here a computational approach that focuses on stabilizing core protein structures rather than engineering the protein–protein interface. Using this approach, we designed thermostabilized interleukin-2 (IL-2) variants that bind tightly to their receptor without experimental optimization, mimicking the properties of the yeast-display engineered IL-2 variant “super-2.” Our results suggest that structure-guided stabilization may be a general method for in silico affinity maturation of protein–protein interactions.

Affinity maturation of protein–protein interactions is an important approach in the development of therapeutic proteins such as cytokines. Typical experimental strategies involve targeting the cytokine-receptor interface with combinatorial libraries and then selecting for higher-affinity variants. Mutations to the binding scaffold are usually not considered main drivers for improved affinity. Here we demonstrate that computational design can provide affinity-enhanced variants of interleukin-2 (IL-2) “out of the box” without any requirement for interface engineering. Using a strategy of global IL-2 structural stabilization targeting metastable regions of the three-dimensional structure, rather than the receptor binding interfaces, we computationally designed thermostable IL-2 variants with up to 40-fold higher affinity for IL-2Rβ without any library-based optimization. These IL-2 analogs exhibited CD25-independent activities on T and natural killer (NK) cells both in vitro and in vivo, mimicking the properties of the IL-2 superkine “super-2” that was engineered through yeast surface display [A. M. Levin et al., Nature, 484, 529–533 (2012)]. Structure-guided stabilization of cytokines is a powerful approach to affinity maturation with applications to many cytokine and protein–protein interactions.

Voronoia 4-ever

We present an updated version of the Voronoia service that enables fully automated analysis of the atomic packing density of macromolecules. Voronoia combines previous efforts to analyse 3D protein and RNA structures into a single service, combined with state-of-the-art online visualization. Voronoia uses the Voronoi cell method to calculate the free space between neighbouring atoms to estimate van der Waals interactions. Compared to other methods that derive van der Waals interactions by calculating solvent-free surfaces, it explicitly considers volume or packing defects. Large internal voids refer either to water molecules or ions unresolved by X-ray crystallography or cryo-EM, cryptic ligand binding pockets, or parts of a structural model that require further refinement. Voronoia is, therefore mainly used for functional analyses of 3D structures and quality assessments of structural models. Voronoia 4-ever updates the database of precomputed packing densities of PDB entries, allows uploading multiple structures, adds new filter options and facilitates direct access to the results through intuitive display with the NGL viewer. Voronoia is available at: htttp://proteinformatics.org/voronoia.

Comparison of the pH- and thermally-induced fluctuations of a therapeutic antibody Fab fragment by molecular dynamics simulation

Successful development of protein therapeutics depends critically on achieving stability under a range of conditions. A deeper understanding of the drivers of instability across different stress conditions, will enable the engineering of more robust protein scaffolds. We compared the impacts of low pH and high temperature stresses on the structure of a humanized antibody fragment (Fab) A33, using atomistic molecular dynamics simulations, using a recent 2.5 Å crystal structure. This revealed that low-pH induced the loss of native contacts in the domain C_L. By contrast, thermal stress led to 5-7% loss of native contacts in all four domains, and simultaneous loss of >30% of native contacts in the V_L-V_H and C_L-C_H interfaces. This revealed divergent destabilising pathways under the two different stresses. The underlying cause of instability was probed using FoldX and Rosetta mutation analysis, and packing density calculations. These agreed that mutations in the C_L domain, and C_L-C_H1 interface have the greatest potential for stabilisation of Fab A33. Several key salt bridge losses underpinned the conformational change in C_L at low pH, whereas at high temperature, salt bridges became more dynamic, thus contributing to an overall destabilization. Lastly, the unfolding events at the two stress conditions exposed different predicted aggregation-prone regions (APR) to solvent, which would potentially lead to different aggregation mechanisms. Overall, our results identified the early stages of unfolding and stability-limiting regions of Fab A33, and the V_H and C_L domains as interesting future targets for engineering stability to both pH- and thermal-stresses simultaneously.

Prediction of impacts of mutations on protein structure and interactions: SDM, a statistical approach, and mCSM, using machine learning

Next-generation sequencing methods have not only allowed an understanding of genome sequence variation during the evolution of organisms but have also provided invaluable information about genetic variants in inherited disease and the emergence of resistance to drugs in cancers and infectious disease. A challenge is to distinguish mutations that are drivers of disease or drug resistance, from passengers that are neutral or even selectively advantageous to the organism. This requires an understanding of impacts of missense mutations in gene expression and regulation, and on the disruption of protein function by modulating protein stability or disturbing interactions with proteins, nucleic acids, small molecule ligands, and other biological molecules. Experimental approaches to understanding differences between wild-type and mutant proteins are most accurate but are also time-consuming and costly. Computational tools used to predict the impacts of mutations can provide useful information more quickly. Here, we focus on two widely used structure-based approaches, originally developed in the Blundell lab: site-directed mutator (SDM), a statistical approach to analyze amino acid substitutions, and mutation cutoff scanning matrix (mCSM), which uses graph-based signatures to represent the wild-type structural environment and machine learning to predict the effect of mutations on protein stability. Here, we describe DUET that uses machine learning to combine the two approaches. We discuss briefly the development of mCSM for understanding the impacts of mutations on interfaces with other proteins, nucleic acids, and ligands, and we exemplify the wide application of these approaches to understand human genetic disorders and drug resistance mutations relevant to cancer and mycobacterial infections. STATEMENT FOR A BROADER AUDIENCE: Genetic or somatic changes in genes can lead to mutations in human proteins, which give rise to genetic disorders or cancer, or to genes of pathogens leading to drug resistance. Computer software described here, using statistical approaches or machine learning, uses the information from genome sequencing of humans and pathogens, together with experimental or modeled 3D structures of gene products, the proteins, to predict impacts of mutations in genetic disease, cancer and drug resistance.

Protein stability engineering insights revealed by domain-wide comprehensive mutagenesis

The accurate prediction of protein stability upon sequence mutation is an important but unsolved challenge in protein engineering. Large mutational datasets are required to train computational predictors, but traditional methods for collecting stability data are either low-throughput or measure protein stability indirectly. Here, we develop an automated method to generate thermodynamic stability data for nearly every single mutant in a small 56-residue protein. Analysis reveals that most single mutants have a neutral effect on stability, mutational sensitivity is largely governed by residue burial, and unexpectedly, hydrophobics are the best tolerated amino acid type. Correlating the output of various stability-prediction algorithms against our data shows that nearly all perform better on boundary and surface positions than for those in the core and are better at predicting large-to-small mutations than small-to-large ones. We show that the most stable variants in the single-mutant landscape are better identified using combinations of 2 prediction algorithms and including more algorithms can provide diminishing returns. In most cases, poor in silico predictions were tied to compositional differences between the data being analyzed and the datasets used to train the algorithm. Finally, we find that strategies to extract stabilities from high-throughput fitness data such as deep mutational scanning are promising and that data produced by these methods may be applicable toward training future stability-prediction tools.

SDM: a server for predicting effects of mutations on protein stability

Here, we report a webserver for the improved SDM, used for predicting the effects of mutations on protein stability. As a pioneering knowledge-based approach, SDM has been highlighted as the most appropriate method to use in combination with many other approaches. We have updated the environment-specific amino-acid substitution tables based on the current expanded PDB (a 5-fold increase in information), and introduced new residue-conformation and interaction parameters, including packing density and residue depth. The updated server has been extensively tested using a benchmark containing 2690 point mutations from 132 different protein structures. The revised method correlates well against the hypothetical reverse mutations, better than comparable methods built using machine-learning approaches, highlighting the strength of our knowledge-based approach for identifying stabilising mutations. Given a PDB file (a Protein Data Bank file format containing the 3D coordinates of the protein atoms), and a point mutation, the server calculates the stability difference score between the wildtype and mutant protein. The server is available at http://structure.bioc.cam.ac.uk/sdm2

G Protein-Coupled Receptors Contain Two Conserved Packing Clusters

G protein-coupled receptors (GPCRs) have evolved a seven-transmembrane helix framework that is responsive to a wide range of extracellular signals. An analysis of the interior packing of family A GPCR crystal structures reveals two clusters of highly packed residues that facilitate tight transmembrane helix association. These clusters are centered on amino acid positions 2.47 and 4.53, which are highly conserved as alanine and serine, respectively. Ala2.47 mediates the interaction between helices H1 and H2, while Ser4.53 mediates the interaction between helices H3 and H4. The helical interfaces outside of these clusters are lined with residues that are more loosely packed, a structural feature that facilitates motion of helices H5, H6, and H7, which is required for receptor activation. Mutation of the conserved small side chain at position 4.53 within packing cluster 2 is shown to disrupt the structure of the visual receptor rhodopsin, whereas sites in packing cluster 1 (e.g., positions 1.46 and 2.47) are more tolerant to mutation but affect the overall stability of the protein. These findings reveal a common structural scaffold of GPCRs that is important for receptor folding and activation.

Genomes, structural biology and drug discovery: combating the impacts of mutations in genetic disease and antibiotic resistance

For over four decades structural biology has been used to understand the mechanisms of disease, and structure-guided approaches have demonstrated clearly that they can contribute to many aspects of early drug discovery, both computationally and experimentally. Structure can also inform our understanding of impacts of mutations in human genetic diseases and drug resistance in cancers and infectious diseases. We discuss the ways that structural insights might be useful in both repurposing off-licence drugs and guide the design of new molecules that might be less susceptible to drug resistance in the future.

Impact of germline and somatic missense variations on drug binding sites

Advancements in next-generation sequencing (NGS) technologies are generating a vast amount of data. This exacerbates the current challenge of translating NGS data into actionable clinical interpretations. We have comprehensively combined germline and somatic nonsynonymous single-nucleotide variations (nsSNVs) that affect drug binding sites in order to investigate their prevalence. The integrated data thus generated in conjunction with exome or whole-genome sequencing can be used to identify patients who may not respond to a specific drug because of alterations in drug binding efficacy due to nsSNVs in the target protein's gene. To identify the nsSNVs that may affect drug binding, protein–drug complex structures were retrieved from Protein Data Bank (PDB) followed by identification of amino acids in the protein–drug binding sites using an occluded surface method. Then, the germline and somatic mutations were mapped to these amino acids to identify which of these alter protein–drug binding sites. Using this method we identified 12 993 amino acid–drug binding sites across 253 unique proteins bound to 235 unique drugs. The integration of amino acid–drug binding sites data with both germline and somatic nsSNVs data sets revealed 3133 nsSNVs affecting amino acid–drug binding sites. In addition, a comprehensive drug target discovery was conducted based on protein structure similarity and conservation of amino acid–drug binding sites. Using this method, 81 paralogs were identified that could serve as alternative drug targets. In addition, non-human mammalian proteins bound to drugs were used to identify 142 homologs in humans that can potentially bind to drugs. In the current protein–drug pairs that contain somatic mutations within their binding site, we identified 85 proteins with significant differential gene expression changes associated with specific cancer types. Information on protein–drug binding predicted drug target proteins and prevalence of both somatic and germline nsSNVs that disrupt these binding sites can provide valuable knowledge for personalized medicine treatment. A web portal is available where nsSNVs from individual patient can be checked by scanning against DrugVar to determine whether any of the SNVs affect the binding of any drug in the database.

Structure-function analysis of hepatitis C virus envelope glycoproteins E1 and E2

Hepatitis C virus (HCV) is the leading cause of chronic liver disease in humans. The envelope proteins of HCV are potential candidates for vaccine development. The absence of three-dimensional (3D) structures for the functional domain of HCV envelope proteins [E1.E2] monomer complex has hindered overall understanding of the virus infection, and also structure-based drug design initiatives. In this study, we report a 3D model containing both E1 and E2 proteins of HCV using the recently published structure of the core domain of HCV E2 and the functional part of E1, and investigate immunogenic implications of the model. HCV [E1.E2] molecule is modeled by using aa205-319 of E1 to aa421-716 of E2. Published experimental data were used to further refine the [E1.E2] model. Based on the model, we predict 77 exposed residues and several antigenic sites within the [E1.E2] that could serve as vaccine epitopes. This study identifies eight peptides which have antigenic propensity and have two or more sequentially exposed amino acids and 12 singular sites are under negative selection pressure that can serve as vaccine or therapeutic targets. Our special interest is 285FLVGQLFTFSPRRHW299 which has five negatively selected sites (L286, V287, G288, T292, and G303) with three of them sequential and four amino acids exposed (F285, L286, T292, and R296). This peptide in the E1 protein maps to dengue envelope vaccine target identified previously by our group. Our model provides for the first time an overall view of both the HCV envelope proteins thereby allowing researchers explore structure-based drug design approaches.

Thermostabilization of Bacillus subtilis lipase A by minimizing the structural deformation caused by packing enhancement

Enzyme thermostabilization is a critical research topic due to potential industrial benefits. Among the various reasons to increase enzyme thermostability, enhancement of residual packing at the core of the enzyme structure has been commonly accepted as a successful strategy. However, structural changes that occur with residual packing enhancement may decrease enzyme activity. In this study, a strategy to minimize structural deformation by calculating the overlapping packing volume of a single-point mutation followed by applying a double-point mutation was suggested. Four double mutants, A38V_K23A, A75V_T83A, G80A_N106A, and G172A_V100A, were selected for the in vitro experiment; three of the four showed enhancements in both thermostability and catalytic activity. In particular, G80A_N106A showed 2.78 times higher catalytic activity compared with wild type.

Amino acid conservation and interactions in rhodopsin: probing receptor activation by NMR spectroscopy

Rhodopsin is a classical two-state G protein-coupled receptor (GPCR). In the dark, its 11-cis retinal chromophore serves as an inverse agonist to lock the receptor in an inactive state. Retinal-protein and protein-protein interactions have evolved to reduce the basal activity of the receptor in order to achieve low dark noise in the visual system. In contrast, absorption of light triggers rapid isomerization of the retinal, which drives the conversion of the receptor to a fully active conformation. Several specific protein-protein interactions have evolved that maintain the lifetime of the active state in order to increase the sensitivity of this receptor for dim-light vision in vertebrates. In this article, we review the molecular interactions that stabilize rhodopsin in the dark-state and describe the use of solid-state NMR spectroscopy for probing the structural changes that occur upon light-activation. Amino acid conservation provides a guide for those interactions that are common in the class A GPCRs as well as those that are unique to the visual system. This article is part of a Special Issue entitled: Retinal Proteins - You can teach an old dog new tricks.

Retinitis pigmentosa mutants provide insight into the role of the N-terminal cap in rhodopsin folding, structure, and function

Autosomal dominant retinitis pigmentosa (ADRP) mutants (T4K, N15S, T17M, V20G, P23A/H/L, and Q28H) in the N-terminal cap of rhodopsin misfold when expressed in mammalian cells. To gain insight into the causes of misfolding and to define the contributions of specific residues to receptor stability and function, we evaluated the responses of these mutants to 11-cis-retinal pharmacological chaperone rescue or disulfide bond-mediated repair. Pharmacological rescue restored folding in all mutants, but the purified mutant pigments in all cases were thermo-unstable and exhibited abnormal photobleaching, metarhodopsin II decay, and G protein activation. As a complementary approach, we superimposed this panel of ADRP mutants onto a rhodopsin background containing a juxtaposed cysteine pair (N2C/D282C) that forms a disulfide bond. This approach restored folding in T4K, N15S, V20G, P23A, and Q28H but not T17M, P23H, or P23L. ADRP mutant pigments obtained by disulfide bond repair exhibited enhanced stability, and some also displayed markedly improved photobleaching and signal transduction properties. Our major conclusion is that the N-terminal cap stabilizes opsin during biosynthesis and contributes to the dark-state stability of rhodopsin. Comparison of these two restorative approaches revealed that the correct position of the cap relative to the extracellular loops is also required for optimal photochemistry and efficient G protein activation.

Position-specific chemical modification and quantitative proteomics disclose protein orientation adsorbed on silica nanoparticles

We describe a method for determining the orientation of cytochrome c, RNase A, and lysozyme on silica nanoparticles (SNPs) using chemical modification combined with proteolysis-mass spectrometry. The proteins interacted with SNPs through preferential adsorption sites, which are dependent on SNP diameter; 4 nm SNPs induce greater structural stabilization than 15 nm particles, presumably due to greater surface curvature of the former. These results suggest that nanoparticle size and protein structure influence protein orientation on SNPs.

MAISTAS: a tool for automatic structural evaluation of alternative splicing products

Motivation: Analysis of the human genome revealed that the amount of transcribed sequence is an order of magnitude greater than the number of predicted and well-characterized genes. A sizeable fraction of these transcripts is related to alternatively spliced forms of known protein coding genes. Inspection of the alternatively spliced transcripts identified in the pilot phase of the ENCODE project has clearly shown that often their structure might substantially differ from that of other isoforms of the same gene, and therefore that they might perform unrelated functions, or that they might even not correspond to a functional protein. Identifying these cases is obviously relevant for the functional assignment of gene products and for the interpretation of the effect of variations in the corresponding proteins.

Results: Here we describe a publicly available tool that, given a gene or a protein, retrieves and analyses all its annotated isoforms, provides users with three-dimensional models of the isoform(s) of his/her interest whenever possible and automatically assesses whether homology derived structural models correspond to plausible structures. This information is clearly relevant. When the homology model of some isoforms of a gene does not seem structurally plausible, the implications are that either they assume a structure unrelated to that of the other isoforms of the same gene with presumably significant functional differences, or do not correspond to functional products. We provide indications that the second hypothesis is likely to be true for a substantial fraction of the cases.

Availability:http://maistas.bioinformatica.crs4.it/.

Contact:anna.tramontano@uniromal.it

Structural and functional roles of small group-conserved amino acids present on helix-H7 in the β(2)-adrenergic receptor

Sequence analysis of the class A G protein-coupled receptors (GPCRs) reveals that most of the highly conserved sites are located in the transmembrane helices. A second level of conservation exists involving those residues that are conserved as a group characterized by small and/or weakly polar side chains (Ala, Gly, Ser, Cys, Thr). These positions can have group conservation levels of up to 99% across the class A GPCRs and have been implicated in mediating helix-helix interactions in membrane proteins. We have previously shown that mutation of group-conserved residues present on transmembrane helices H2-H4 in the β(2)-adrenergic receptor (β(2)-AR) can influence both receptor expression and function. We now target the group-conserved sites, Gly315(7.42) and Ser319(7.46), on H7 for structure-function analysis. Replacing Ser319(7.46) with smaller amino acids (Ala or Gly) did not influence the ability of the mutant receptors to bind to the antagonist dihydroalprenolol (DHA) but resulted in ~15-20% agonist-independent activity. Replacement of Ser319(7.46) with the larger amino acid leucine lowered the expression of the S319L mutant and its ability to bind DHA. Both the G315A and G315S mutants also exhibited agonist-independent signaling, while the G315L mutant did not show specific binding to DHA. These data indicate that Gly315(7.42) and Ser319(7.46) are stabilizing β(2)-AR in an inactive conformation. We discuss our results in the context of van der Waals interactions of Gly315(7.42) with Trp286(6.48) and hydrogen bonding interactions of Ser319(7.46) with amino acids on H1-H2-H7 and with structural water.

Coding potential of the products of alternative splicing in human

Background

Analysis of the human genome has revealed that as much as an order of magnitude more of the genomic sequence is transcribed than accounted for by the predicted and characterized genes. A number of these transcripts are alternatively spliced forms of known protein coding genes; however, it is becoming clear that many of them do not necessarily correspond to a functional protein.

Results

In this study we analyze alternative splicing isoforms of human gene products that are unambiguously identified by mass spectrometry and compare their properties with those of isoforms of the same genes for which no peptide was found in publicly available mass spectrometry datasets. We analyze them in detail for the presence of uninterrupted functional domains, active sites as well as the plausibility of their predicted structure. We report how well each of these strategies and their combination can correctly identify translated isoforms and derive a lower limit for their specificity, that is, their ability to correctly identify non-translated products.

Conclusions

The most effective strategy for correctly identifying translated products relies on the conservation of active sites, but it can only be applied to a small fraction of isoforms, while a reasonably high coverage, sensitivity and specificity can be achieved by analyzing the presence of non-truncated functional domains. Combining the latter with an assessment of the plausibility of the modeled structure of the isoform increases both coverage and specificity with a moderate cost in terms of sensitivity.

Comparison of NMR and crystal structures for the proteins TM1112 and TM1367

NMR structures of the proteins TM1112 and TM1367 solved by the JCSG in solution at 298 K could be superimposed with the corresponding crystal structures at 100 K with r.m.s.d. values of <1.0 Å for the backbone heavy atoms. For both proteins the structural differences between multiple molecules in the asymmetric unit of the crystals correlated with structural variations within the bundles of conformers used to represent the NMR solution structures. A recently introduced JCSG NMR structure-determination protocol, which makes use of the software package UNIO for extensive automation, was further evaluated by comparison of the TM1112 structure obtained using these automated methods with another NMR structure that was independently solved in another PSI center, where a largely interactive approach was applied.

The NMR structures of the TM1112 and TM1367 proteins from Thermotoga maritima in solution at 298 K were determined following a new protocol which uses the software package UNIO for extensive automation. The results obtained with this novel procedure were evaluated by comparison with the crystal structures solved by the JCSG at 100 K to 1.83 and 1.90 Å resolution, respectively. In addition, the TM1112 solution structure was compared with an NMR structure solved by the NESG using a conventional largely interactive methodology. For both proteins, the newly determined NMR structure could be superimposed with the crystal structure with r.m.s.d. values of <1.0 Å for the backbone heavy atoms, which provided a starting platform to investigate local structure variations, which may arise from either the methods used or from the different chemical environments in solution and in the crystal. Thereby, these comparative studies were further explored with the use of reference NMR and crystal structures, which were computed using the NMR software with input of upper-limit distance constraints derived from the molecular models that represent the results of structure determination by NMR and by X-ray diffraction, respectively. The results thus obtained show that NMR structure calculations with the new automated UNIO software used by the JCSG compare favorably with those from a more labor-intensive and time-intensive interactive procedure. An intriguing observation is that the ‘bundles’ of two TM1112 or three TM1367 molecules in the asymmetric unit of the crystal structures mimic the behavior of the bundles of 20 conformers used to represent the NMR solution structures when comparing global r.m.s.d. values calculated either for the polypeptide backbone, the core residues with solvent accessibility below 15% or all heavy atoms.

Structural characterization of the organic solvent-stable cholesterol oxidase from Chromobacterium sp. DS-1

Cholesterol oxidase is of significant commercial interest as it is widely used as a biosensor for the detection of cholesterol in clinical samples, blood serum and food. Increased stability of this enzyme with regards to temperature and different solvent conditions are of great importance to the reliability and versatility of its applications. We here report the crystal structure of the cholesterol oxidase of Chromobacterium sp. DS-1 (CHOLOX). In contrast to other previously characterized cholesterol oxidases, this enzyme retains high activity in organic solvents and detergents at temperatures above 85 degrees C despite its mesophilic origin. With the availability of one other homologous oxidase of known three-dimensional structure, a detailed comparison of its sequence and structure was performed to elucidate the mechanisms of stabilization. In contrast to factors that typically contribute to the stability of thermophilic proteins, the structure of CHOLOX exhibits a larger overall cavity volume, less charged residues and less salt bridge interactions. Moreover, the vast majority of residue substitutions were found on or near the protein's solvent exposed surface. We propose that the engineering of enhanced stability may also be accomplished through selective engineering of the protein periphery rather than by redesigning its entire core.

Structure of the lethal phage pinhole

Perhaps the simplest of biological timing systems, bacteriophage holins accumulate during the phage morphogenesis period and then trigger to permeabilize the cytoplasmic membrane with lethal holes; thus, terminating the infection cycle. Canonical holins form very large holes that allow nonspecific release of fully-folded proteins, but a recently discovered class of holins, the pinholins, make much smaller holes, or pinholes, that serve only to depolarize the membrane. Here, we interrogate the structure of the prototype pinholin by negative-stain transmission electron-microscopy, cysteine-accessibility, and chemical cross-linking, as well as by computational approaches. Together, the results suggest that the pinholin forms symmetric heptameric structures with the hydrophilic surface of one transmembrane domain lining the surface of a central channel approximately 15 A in diameter. The structural model also suggests a rationale for the prehole state of the pinholin, the persistence of which defines the duration of the viral latent period, and for the sensitivity of the holin timing system to the energized state of the membrane.

Relating protein conformational changes to packing efficiency and disorder

Changes in protein conformation play key roles in facilitating various biochemical processes, ranging from signaling and phosphorylation to transport and catalysis. While various factors that drive these motions such as environmental changes and binding of small molecules are well understood, specific causative effects on the structural features of the protein due to these conformational changes have not been studied on a large scale. Here, we study protein conformational changes in relation to two key structural metrics: packing efficiency and disorder. Packing has been shown to be crucial for protein stability and function by many protein design and engineering studies. We study changes in packing efficiency during conformational changes, thus extending the analysis from a static context to a dynamic perspective and report some interesting observations. First, we study various proteins that adopt alternate conformations and find that tendencies to show motion and change in packing efficiency are correlated: residues that change their packing efficiency show larger motions. Second, our results suggest that residues that show higher changes in packing during motion are located on the changing interfaces which are formed during these conformational changes. These changing interfaces are slightly different from shear or static interfaces that have been analyzed in previous studies. Third, analysis of packing efficiency changes in the context of secondary structure shows that, as expected, residues buried in helices show the least change in packing efficiency, whereas those embedded in bends are most likely to change packing. Finally, by relating protein disorder to motions, we show that marginally disordered residues which are ordered enough to be crystallized but have sequence patterns indicative of disorder show higher dislocation and a higher change in packing than ordered ones and are located mostly on the changing interfaces. Overall, our results demonstrate that between the two conformations, the cores of the proteins remain mostly intact, whereas the interfaces display the most elasticity, both in terms of disorder and change in packing efficiency. By doing a variety of tests, we also show that our observations are robust to the solvation state of the proteins.

Conservation of side-chain dynamics within a protein family

The question of protein dynamics and its relevance to function is currently a topic of great interest. Proteins are particularly dynamic at the side-chain level on the time scale of picoseconds to nanoseconds. Here, we present a comparison of NMR-monitored side-chain motion between three PDZ domains of approximately 30% sequence identity and show that the side-chain dynamics display nontrivial conservation. Methyl (2)H relaxation was carried out to determine side-chain order parameters (S(2)), which were found to be more similar than naively expected from sequence, local packing, or a combination of the two. Thus, the dynamics of a rather distant homologue appears to be an excellent predictor of a protein's side-chain dynamics and, on average, better than current structure-based methods. Fast side-chain dynamics therefore display a high level of organization associated with global fold. Beyond simple conservation, the analysis herein suggests that the pattern of side-chain flexibility has significant contributions from nonlocal elements of the PDZ fold, such as correlated motions, and that the conserved dynamics may directly support function.

Voronoia: analyzing packing in protein structures

The packing of protein atoms is an indicator for their stability and functionality, and applied in determining thermostability, in protein design, ligand binding and to identify flexible regions in proteins. Here, we present Voronoia, a database of atomic-scale packing data for protein 3D structures. It is based on an improved Voronoi Cell algorithm using hyperboloid interfaces to construct atomic volumes, and to resolve solvent-accessible and -inaccessible regions of atoms. The database contains atomic volumes, local packing densities and interior cavities calculated for 61 318 biological units from the PDB. A report for each structure summarizes the packing by residue and atom types, and lists the environment of interior cavities. The packing data are compared to a nonredundant set of structures from SCOP superfamilies. Both packing densities and cavities can be visualized in the 3D structures by the Jmol plugin. Additionally, PDB files can be submitted to the Voronoia server for calculation. This service performs calculations for most full-atomic protein structures within a few minutes. For batch jobs, a standalone version of the program with an optional PyMOL plugin is available for download. The database can be freely accessed at: http://bioinformatics.charite.de/voronoia.

Outer membrane phospholipase a dimer stability does not correlate to occluded surface area

Despite the key roles of oligomeric membrane proteins (MPs) in many known cellular pathways, the principles governing their oligomer stability are not well-understood. Previous work with the alpha-helical MPs bacteriorhodopsin (bR) and glycophorin A (GpA) shows that lost buried surface area linearly correlated with perturbations in protein stability. Although this is a significant discovery, the predictive power of this correlation is limited by the data. Because both bR and GpA have alpha-helical secondary structural motifs, it is unclear whether this correlation would be observed for MPs with a beta-barrel motif. We addressed this question by measuring the thermodynamic consequences of interfacial amino acid changes at the dimer interface of beta-barrel MP outer membrane phospholipase A (OMPLA). We created sequence variants to reduce the contact surface area of the OMPLA dimer interface by introducing single-alanine substitutions and used sedimentation equilibrium analytical ultracentrifugation to determine the dimerization free energies for these variants. The integrity of each variant was verified by two functional assays: specific activity and resistance to thermal denaturation, which showed that structural changes were restricted to the local environment. Using this information, we calculated the anticipated packing defects due to side chain deletion and compared this to the free energy perturbations for each residue. Contrary to the findings with bR and GpA, our study found no correlation between the contact surface area lost and the perturbations to OMPLA dimer stability. We conclude that van der Waals packing may not be a strong predictor of stability for all membrane proteins.

A specific cholesterol binding site is established by the 2.8 A structure of the human beta2-adrenergic receptor

The role of cholesterol in eukaryotic membrane protein function has been attributed primarily to an influence on membrane fluidity and curvature. We present the 2.8 A resolution crystal structure of a thermally stabilized human beta(2)-adrenergic receptor bound to cholesterol and the partial inverse agonist timolol. The receptors pack as monomers in an antiparallel association with two distinct cholesterol molecules bound per receptor, but not in the packing interface, thereby indicating a structurally relevant cholesterol-binding site between helices I, II, III, and IV. Thermal stability analysis using isothermal denaturation confirms that a cholesterol analog significantly enhances the stability of the receptor. A consensus motif is defined that predicts cholesterol binding for 44% of human class A receptors, suggesting that specific sterol binding is important to the structure and stability of other G protein-coupled receptors, and that this site may provide a target for therapeutic discovery.

Structure of a protein-DNA complex essential for DNA protection in spores of Bacillus species

The DNA-binding alpha/beta-type small acid-soluble proteins (SASPs) are a major factor in the resistance and long-term survival of spores of Bacillus species by protecting spore DNA against damage due to desiccation, heat, toxic chemicals, enzymes, and UV radiation. We now report the crystal structure at 2.1 A resolution of an alpha/beta-type SASP bound to a 10-bp DNA duplex. In the complex, the alpha/beta-type SASP adopt a helix-turn-helix motif, interact with DNA through minor groove contacts, bind to approximately 6 bp of DNA as a dimer, and the DNA is in an A-B type conformation. The structure of the complex provides important insights into the molecular details of both DNA and alpha/beta-type SASP protection in the complex and thus also in spores.

Fold assessment for comparative protein structure modeling

Accurate and automated assessment of both geometrical errors and incompleteness of comparative protein structure models is necessary for an adequate use of the models. Here, we describe a composite score for discriminating between models with the correct and incorrect fold. To find an accurate composite score, we designed and applied a genetic algorithm method that searched for a most informative subset of 21 input model features as well as their optimized nonlinear transformation into the composite score. The 21 input features included various statistical potential scores, stereochemistry quality descriptors, sequence alignment scores, geometrical descriptors, and measures of protein packing. The optimized composite score was found to depend on (1) a statistical potential z-score for residue accessibilities and distances, (2) model compactness, and (3) percentage sequence identity of the alignment used to build the model. The accuracy of the composite score was compared with the accuracy of assessment by single and combined features as well as by other commonly used assessment methods. The testing set was representative of models produced by automated comparative modeling on a genomic scale. The composite score performed better than any other tested score in terms of the maximum correct classification rate (i.e., 3.3% false positives and 2.5% false negatives) as well as the sensitivity and specificity across the whole range of thresholds. The composite score was implemented in our program MODELLER-8 and was used to assess models in the MODBASE database that contains comparative models for domains in approximately 1.3 million protein sequences.

Structural adaptation to low temperatures − analysis of the subunit interface of oligomeric psychrophilic enzymes

Enzymes from psychrophiles show higher catalytic efficiency in the 0-20 degrees C temperature range and often lower thermostability in comparison with meso/thermophilic homologs. Physical and chemical characterization of these enzymes is currently underway in order to understand the molecular basis of cold adaptation. Psychrophilic enzymes are often characterized by higher flexibility, which allows for better interaction with substrates, and by a lower activation energy requirement in comparison with meso/thermophilic counterparts. In their tertiary structure, psychrophilic enzymes present fewer stabilizing interactions, longer and more hydrophilic loops, higher glycine content, and lower proline and arginine content. In this study, a comparative analysis of the structural characteristics of the interfaces between oligomeric psychrophilic enzyme subunits was carried out. Crystallographic structures of oligomeric psychrophilic enzymes, and their meso/thermophilic homologs belonging to five different protein families, were retrieved from the Protein Data Bank. The following structural parameters were calculated: overall and core interface area, characterization of polar/apolar contributions to the interface, hydrophobic contact area, quantity of ion pairs and hydrogen bonds between monomers, internal area and total volume of non-solvent-exposed cavities at the interface, and average packing of interface residues. These properties were compared to those of meso/thermophilic enzymes. The results were analyzed using Student's t-test. The most significant differences between psychrophilic and mesophilic proteins were found in the number of ion pairs and hydrogen bonds, and in the apolarity of their subunit interface. Interestingly, the number of ion pairs at the interface shows an opposite adaptation to those occurring at the monomer core and surface.

A tetrapeptide fragment-based design method results in highly stable artificial proteins

Computational protein design has progressed rapidly over the last years. A number of design methods have been proposed and tested. In this paper, we report the successful application of a fragment-based method for protein design. The method uses statistical information on tetrapeptide backbone conformations. The previously published artificial fold of TOP 7 (Kuhlman et al., Science, 2003; 302:1364-1368) was chosen as template. A series of polypeptide sequences were created that were predicted to fold into this target structure. Two of the designed proteins, M5 and M7, were expressed and characterized by fluorescence spectroscopy, circular dichroism and NMR. They showed the hallmarks of well-ordered tertiary structure as well as cooperative folding/unfolding transitions. Furthermore, the two novel proteins were found to be highly stable against temperature and denaturant-induced unfolding.

Computational analysis and identification of amino acid sites in dengue E proteins relevant to development of diagnostics and vaccines

We have identified 72 completely conserved amino acid residues in the E protein of major groups of the Flavivirus genus by computational analyses. In the dengue species we have identified 12 highly conserved sequence regions, 186 negatively selected sites, and many dengue serotype-specific negatively selected sites. The flavivirus-conserved sites included residues involved in forming six disulfide bonds crucial for the structural integrity of the protein, the fusion motif involved in viral infectivity, and the interface residues of the oligomers. The structural analysis of the E protein showed 19 surface-exposed non-conserved residues, 128 dimer or trimer interface residues, and regions, which undergo major conformational change during trimerization. Eleven consensus T(h)-cell epitopes common to all four dengue serotypes were predicted. Most of these corresponded to dengue-conserved regions or negatively selected sites. Of special interest are six singular sites (N(37), Q(211), D(215), P(217), H(244), K(246)) in dengue E protein that are conserved, are part of the predicted consensus T(h)-cell epitopes and are exposed in the dimer or trimer. We propose these sites and corresponding epitopic regions as potential candidates for prioritization by experimental biologists for development of diagnostics and vaccines that may be difficult to circumvent by natural or man-made alteration of dengue virus.

NMR Dynamics Distinguish Between Hard and Soft Hydrophobic Cores in the DNA-binding Domain of PhoB and Demonstrate Different Roles of the Cores in Binding to DNA

The transcription factor PhoB contains an N-terminal regulatory domain and a C-terminal DNA-binding/transactivation domain. The DNA-binding/transactivation domain alone can bind specifically to DNA and consequently activate transcription. It consists of an N-terminal four-stranded beta-sheet and a winged helix domain, containing a three-helix bundle and a C-terminal beta-hairpin. The second and third helices, together with the beta-hairpin, contact DNA and the loop between the second and third helices is responsible for the transactivation. Here, we have examined the backbone and side-chain dynamics of the DNA-binding domain in its DNA-free and bound forms by NMR. The side-chain dynamics identified two apparent hydrophobic cores: one, a soft hydrophobic core, shows inherently flexible dynamics on the pico-to nanosecond timescale and maintains the DNA-binding and transactivation surfaces; the other is a hard hydrophobic core formed between the N-terminal beta-sheet and the three-helix bundle, which maintains the other non-functional surface. Upon binding to DNA, the flexibility of the soft core decreases but remains more flexible than the hard core. The winged helix domain itself has inherent flexibility in the DNA-binding and transactivation functions. However, the back surface of both functional surfaces seems to be covered by the N-terminal beta-sheet in order to mask a possible function arising from the inherent flexibility of the winged helix domain.

A Model for the Recognition of Protein Kinases Based on the Entropy of 3D van der Waals Interactions

The study and prediction of kinase function (kinomics) is of major importance for proteome research due to the widespread distribution of kinases. However, the prediction of protein function based on the similarity between a functionally annotated 3D template and a query structure may fail, for instance, if a similar protein structure cannot be identified. Alternatively, function can be assigned using 3D-structural empirical parameters. In previous studies, we introduced parameters based on electrostatic entropy (Proteins 2004, 56, 715) and molecular vibration entropy (Bioinformatics 2003, 19, 2079) but ignored other important factors such as van der Waals (vdw) interactions. In the work described here, we define 3D-vdw entropies (degrees theta(k)) and use them for the first time to derive a classifier for protein kinases. The model classifies correctly 88.0% of proteins in training and more than 85.0% of proteins in validation studies. Principal components analysis of heterogeneous proteins demonstrated that degrees theta(k) codify information that is different to that described by other bulk or folding parameters. In additional validation experiments, the model recognized 129 out of 142 kinases (90.8%) and 592 out of 677 non-kinases (87.4%) not used above. This study provides a basis for further consideration of degrees theta(k) as parameters for the empirical search for structure-function relationships.

Lipid chain selectivity by outer membrane phospholipase A

Outer membrane phospholipase A (OMPLA) is a unique, integral membrane enzyme found in Gram-negative bacteria and is an important virulence factor for pathogens such as Helicobacter pylori. This broad-specificity lipase degrades a variety of lipid substrates, and it plays a direct role in adjusting the composition and permeability of bacterial membranes under conditions of stress. Interestingly, OMPLA shows little preference for the lipid headgroup and, instead, the length of the hydrophobic acyl chain is the strongest determinant for substrate selection by OMPLA, with the enzyme strongly preferring substrates with chains equal to or longer than 14 carbon atoms. The question remains as to how a hydrophobic protein like OMPLA can achieve this specificity, particularly when the shorter chains can be accommodated in the binding pocket. Using a series of sulfonyl fluoride inhibitors with various lengths of acyl chain, we show here that the thermodynamics of substrate-induced OMPLA dimerization are guided by the acyl chain length, demonstrating that OMPLA uses a unique biophysical mechanism to select its phospholipid substrate.

Complete predicted three-dimensional structure of the facilitator transmembrane protein and hepatitis C virus receptor CD81: conserved and variable structural domains in the tetraspanin superfamily

Tetraspanins are a superfamily of transmembrane proteins implicated in cellular development, motility, and activation through their interactions with a large range of proteins and with specific membrane microdomains. The complete three-dimensional structure of the tetraspanin CD81 has been predicted by molecular modeling and from the crystallographic structure of the EC2 large extracellular domain. Periodicity of sequence conservation, homology modeling, secondary structure prediction, and protein docking were used. The transmembrane domain appears organized as a four-stranded left-handed coiled coil directly connecting to two helices of the EC2. A smaller extracellular loop EC1 contains a small largely hydrophobic beta-strand that packs in a conserved hydrophobic groove of the EC2. The palmitoylable intracellular N-terminal segment forms an amphipathic membrane-parallel helix. Structural variability occurs mainly in an hypervariable subdomain of the EC2 and in intracellular regions. Therefore, the variable interaction selectivity of tetraspanins originates both from sequence variability within structurally conserved domains and from the occurrence of small structurally variable domains. In CD81 and other tetraspanins, the numerous membrane-exposed aromatic residues are asymmetrically clustered and protrude on one side of the transmembrane domain. This may represent a functional specialization of these two sides for interactions with cholesterol, proteins, or membrane microdomains.

Energetics of outer membrane phospholipase A (OMPLA) dimerization

Outer membrane phospholipase A (OMPLA) is a widely conserved transmembrane enzyme found in Gram-negative bacteria, and it is implicated in the virulence of a number of pathogenic organisms. The regulation of the protein's phospholipase activity is not well understood despite the existence of a number of high resolution structures. Previous biochemical studies have demonstrated that dimerization of OMPLA is a prerequisite for its phospholipase activity, and it has been shown in vitro that this dimerization is dependent on calcium and substrate binding. Therefore, to fully understand the regulation of OMPLA, it is necessary to understand the stability of the protein dimer and the extent to which it is influenced by its effector molecules. We have used sedimentation equilibrium analytical ultracentrifugation to dissect the energetics of Escherichia coli OMPLA dimerization in detergent micelles. We find that calcium contributes relatively little stability to the dimer, while interactions with the substrate acyl chain are the predominant force in stabilizing the dimeric conformation of the enzyme. The resulting thermodynamic cycle suggests that interactions between effector molecules are additive. These energetic measurements not only provide insight into the activation of OMPLA, but they also represent the first quantitative investigation of the association energetics of a transmembrane beta-barrel. This thermodynamic study allows us to begin to address the differences between protein-protein interfaces in transmembrane proteins with a helical fold to those of a beta-barrel fold and to more fully understand the forces involved in membrane protein interactions.

Alternate interfaces may mediate homomeric and heteromeric assembly in the transmembrane domains of SNARE proteins

The fusion of a vesicle to a target membrane is mediated by temporally and spatially regulated interactions within a set of evolutionarily conserved proteins. Integral to proper fusion is the interaction between proteins originating on both vesicle and target membranes to form a protein bridge between the two membranes, known as the SNARE complex. This protein complex includes the single-pass transmembrane helix proteins: syntaxin and synaptobrevin. Experimental data and amino acid sequence analysis suggest that an interface of interaction is conserved between the transmembrane regions of the two proteins. However, conflicting reports have been presented on the role of the synaptobrevin transmembrane domain in mediating important protein-protein interactions. To address this question, a thermodynamic study was carried out to determine quantitatively the self-association propensities of the transmembrane domains of synaptobrevin and syntaxin. Our results show that the transmembrane domain of synaptobrevin has only a modest ability to self-associate, whereas the transmembrane domain of syntaxin is able to form stable homodimers. Nevertheless, by a single amino acid substitution, synaptobrevin can be driven to dimerize with the same affinity as syntaxin. Furthermore, crosslinking studies show that dimerization of synaptobrevin is promoted by oxidizing agents. Despite the presence of a conserved cysteine residue in the same location as in synaptobrevin, syntaxin dimerization is not promoted by oxidization. This analysis suggests that subtle yet distinct differences are present between the two transmembrane dimer interfaces. A syntaxin/synaptobrevin heterodimer is able to form under oxidizing conditions, and we propose that the interface of interaction for the heterodimer may resemble the homodimer interface formed by the synaptobrevin transmembrane domain. Computational analysis of the transmembrane sequences of syntaxin and synaptobrevin reveal structural models that correlate with the experimental data. These data may provide insight into the role of transmembrane segments in the mechanism of vesicle fusion.

Effect of Protein Backbone Folding on the Stability of Protein−Ligand Complexes

The role played by the degree of folding of protein backbones in explaining the binding energetics of protein-ligand interactions has been studied. We analyzed the protein/peptide interactions in the RNase-S system in which amino acids at two positions of the peptide S have been mutated. The global degree of folding of the protein S correlates in a significant way with the free energy and enthalpy of the protein-peptide interactions. A much better correlation is found with the local contribution to the degree of folding of one amino acid residue: Thr36. This residue is shown to have a destabilizing interaction with Lys41, which interacts directly with peptide S. Another system, consisting of the interactions of small organic molecules with HIV-1 protease was also studied. In this case, the global change in the degree of folding of the protease backbone does not explain the binding energetics of protein-ligand interactions. However, a significant correlation is observed between the free energy of binding and the contribution of two amino acid residues in the HVI-1 protease: Gly49 and Ile66. In general, it was observed that the changes in the degree of folding are not restricted to the binding site of the protein chain but are distributed along the whole protein backbone. This study provides a basis for further consideration of the degree of folding as a parameter for empirical structural parametrizations of the binding energetics of protein folding and binding.

A protein folding degree measure and its dependence on crystal packing, protein size, secondary structure, and domain structural class

Comparing two or more protein structures with respect to their degree of folding is common practice in structural biology despite the fact that there is no scale for a folding degree. Here we introduce a formal definition of a folding degree, capable of quantitative characterization. This enables ordering among protein chains based on their degree of folding. The folding degree of a data set of 152 representative nonhomologous proteins is then studied. We demonstrate that the variation in the folding degree seen for this data set is not due to crystallization artifacts or experimental conditions, such as resolution, refinement protocol, pH, or temperature. A good linear relationship is observed between the folding degree and the percentages of secondary structures in the protein. The folding degree is able to account for the small changes produced in the structure due to crystal packing and temperature. Automating the classification of proteins into their respective structural domain classes, namely mainly-alpha, mainly-beta, and alpha-beta, is also possible.

A systematic search method for the identification of tightly packed transmembrane parallel alpha-helices

Membrane proteins play a major role in number of biological processes such as signaling pathways. The determination of the three-dimensional structure of these proteins is increasingly important for our understanding of their structure-function relationships. Due to the difficulty in isolating membrane proteins for X-ray diffraction studies, computational techniques are being developed to generate the 3D structures of TM domains. Here, we present a systematic search method for the identification of energetically favorable and tightly packed transmembrane parallel alpha-helices. The first step in our systematic search method is the generation of 3D models for pairs of parallel helix bundles with all possible orientations followed by an energy-based filter to eliminate structures with severe non-bonded contacts. Then, a RMS-based filter was used to cluster these structures into families. Furthermore, these dimers were energy minimized using molecular mechanics force field. Finally, we identified the tightly packed parallel alpha-helices by using an interface surface area. To validate our search method, we compared our predicted GlycophorinA dimer structures with the reported NMR structures. With our search method, we are able to reproduce NMR structures of GPA with 0.9A RMSD. In addition, by considering the reported mutational data on GxxxG motif interactions, twenty percent of our predicted dimers are within in the 2.0A RMSD. The dimers obtained from our method were used to generate parallel trimeric and tetramer TM structures of GPA and found that the structure of GPA might exist only in a dimer form as reported earlier.

Perceptron learning of pairwise contact energies for proteins incorporating the amino acid environment

Although a coarse-grained description of proteins is a simple and convenient way to attack the protein folding problem, the construction of a global pairwise energy function which can simultaneously recognize the native folds of many proteins has resulted in partial success. We have sought the possibility of a systematic improvement of this pairwise-contact energy function as we extended the parameter space of amino acids, incorporating local environments of amino acids, beyond a 20 x 20 matrix. We have studied the pairwise contact energy functions of 20 x 20, 60 x 60, and 180 x 180 matrices depending on the extent of parameter space, and compared their effect on the learnability of energy parameters in the context of a gapless threading, bearing in mind that a 20 x 20 pairwise contact matrix has been shown to be too simple to recognize the native folds of many proteins. In this paper, we show that the construction of a global pairwise energy function was achieved using 1006 training proteins of a homology of less than 30%, which include all representatives of different protein classes. After parametrizing the local environments of the amino acids into nine categories depending on three secondary structures and three kinds of hydrophobicity (desolvation), the 16290 pairwise contact energies (scores) of the amino acids could be determined by perceptron learning and protein threading. These could simultaneously recognize all the native folds of the 1006 training proteins. When these energy parameters were tested on the 382 test proteins of a homology of less than 90%, 370 (96.9%) proteins could recognize their native folds. We set up a simple thermodynamic framework in the conformational space of decoys to calculate the unfolded fraction and the specific heat of real proteins. The different thermodynamic stabilities of E.coli ribonuclease H (RNase H) and its mutants were well described in our calculation, agreeing with the experiment.