QSE: A new 3-D solvent exposure measure for the analysis of protein structure

Solvent exposure of amino acids measures how deep residues are buried in tertiary structure of proteins, and hence it provides important information for analyzing and predicting protein structure and functions. Existing methods of calculating solvent exposure such as accessible surface area, relative accessible surface area, residue depth, contact number, and half-sphere exposure still have some limitations. In this article, we propose a novel solvent exposure measure named quadrant-sphere exposure (QSE) based on eight quadrants derived from spherical neighborhood. The proposed measure forms a microenvironment around Cα atom as a sphere with a radius of 13??, and subdivides it into eight quadrants according to a rectangular coordinate system constructed based on geometric relationships of backbone atoms. The number of neighboring Cα atoms whose labels are the same is given as the QSE value of the center Cα atom at hand. As evidenced by histograms that show very different distributions for different structure configurations, the proposed measure captures local properties that are characteristic for a residue's eight-directional neighborhood within a sphere. Compared with other measures, QSE provides a different view of solvent exposure, and provides information that is specific for different tertiary structure. As the experimental results show, QSE measure can potentially be used in protein structure analysis and predictions.     

Simple solvation potential for coarse-grained models of proteins

We formulate a simple solvation potential based on a coarsed-grained representation of amino acids with two spheres modeling the C(alpha) atom and an effective side-chain centroid. The potential relies on a new method for estimating the buried area of residues, based on counting the effective number of burying neighbors in a suitable way. This latter quantity shows a good correlation with the buried area of residues computed from all atom crystallographic structures. We check the discriminatory power of the solvation potential alone to identify the native fold of a protein from a set of decoys and show the potential to be considerably selective.     

Easy method to predict solvent accessibility from multiple protein sequence alignments

An easy and uncomplicated method to predict the solvent accessibility state of a site in a multiple protein sequence alignment is described. The approach is based on amino acid exchange and compositional preference matrices for each of three accessibility states: buried, exposed, and intermediate. Calculations utilized a modified version of the 3D―ali databank, a collection of multiple sequence alignments anchored through protein tertiary structural superpositions. The technique achieves the same accuracy as much more complex methods and thus provides such advantages as computational affordability, facile updating, and easily understood residue substitution patterns useful to biochemists involved in protein engineering, design, and structural prediction. The program is available from the authors; and, due to its simplicity, the algorithm can be readily implemented on any system. For a given alignment site, a hand calculation can yield a comparative prediction. Proteins 32:190–199, 1998. © 1998 Wiley-Liss, Inc.     

Amino acid function relates to its embedded protein microenvironment: A study on disulfide‐bridged cystine

In our previous study, we have shown that the microenvironments around conserved amino acids are also conserved in protein families (Bandyopadhyay and Mehler, Proteins 2008; 72:646–659). In this study, we have hypothesized that amino acids perform similar functions when embedded in a certain type of protein microenvironment. We have tested this hypothesis on the microenvironments around disulfide‐bridged cysteines from high‐resolution protein crystal structures. Although such cystines mainly play structural role in proteins, in certain enzymes they participate in catalysis and redox reactions. We have performed and report a functional annotation of enzymatically active cystines to their respective microenvironments. Three protein microenvironment clusters were identified: (i) buried‐hydrophobic, (ii) exposed‐hydrophilic, and (iii) buried‐hydrophilic. The buried‐hydrophobic cluster encompasses a small group of 22 redox‐active cystines, mostly in alpha‐helical conformations in a –C‐x‐x‐C‐ motif from the Oxido‐reductase enzyme class. All these cystines have high strain energy and near identical microenvironments. Most of the active cystines in hydrolase enzyme class belong to buried hydrophilic microenvironment cluster. In total there are 34 half‐cystines detected in buried hydrophilic cluster from hydrolases, as a part of enzyme active site. Even within the buried hydrophilic cluster, there is clear separation of active half‐cystines between surface exposed part of the protein and protein interior. Half‐cystines toward the surface exposed region are higher in number compared to those in protein interior. Apart from cystines at the active sites of the enzymes, many more half‐cystines were detected in buried hydrophilic cluster those are part of the microenvironment of enzyme active sites. However, no active half‐cystines were detected in extremely hydrophilic microenvironment cluster, that is, exposed hydrophilic cluster, indicating that total exposure of cystine toward the solvent is not favored for enzymatic reactions. Although half‐cystines in exposed‐hydrophilic clusters occasionally stabilize enzyme active sites, as a part of their microenvironments. Analysis performed in this work revealed that cystines as a part of active sites in specific enzyme families or folds share very similar protein microenvironment regions, despite of their dissimilarity in protein sequences and position specific sequence conservations. Proteins 2016; 84:1576–1589. © 2016 Wiley Periodicals, Inc.     

Prediction of protein solvent accessibility using support vector machines

A Support Vector Machine learning system has been trained to predict protein solvent accessibility from the primary structure. Different kernel functions and sliding window sizes have been explored to find how they affect the prediction performance. Using a cut-off threshold of 15% that splits the dataset evenly (an equal number of exposed and buried residues), this method was able to achieve a prediction accuracy of 70.1% for single sequence input and 73.9% for multiple alignment sequence input, respectively. The prediction of three and more states of solvent accessibility was also studied and compared with other methods. The prediction accuracies are better than, or comparable to, those obtained by other methods such as neural networks, Bayesian classification, multiple linear regression, and information theory. In addition, our results further suggest that this system may be combined with other prediction methods to achieve more reliable results, and that the Support Vector Machine method is a very useful tool for biological sequence analysis.     

Comparison of straight chain and cyclic unnatural amino acids embedded in the core of staphylococcal nuclease.

We have determined by X-ray crystallography the structures of several variants of staphylococcal nuclease with long flexible straight chain and equivalent length cyclic unnatural amino acid side chains embedded in the protein core. The terminal atoms in the straight side chains are not well defined by the observed electron density even though they remain buried within the protein interior. We have previously observed this behavior and have suggested that it may arise from the addition of side-chain vibrational and oscillational motions with each bond as a side chain grows away from the relatively rigid protein main chain and/or the population of multiple rotamers (Wynn R, Harkins P, Richards FM. Fox RO. 1996. Mobile unnatural amino acid side chains in the core of staphylococcal nuclease. Protein Sci 5:1026-1031). Reduction of the number of degrees of freedom by cyclization of a side chain would be expected to constrain these motions. These side chains are in fact well defined in the structures described here. Over-packing of the protein core results in a 1.0 A shift of helix 1 away from the site of mutation. Additionally, we have determined the structure of a side chain containing a single hydrogen to fluorine atom replacement on a methyl group. A fluorine atom is intermediate in size between methyl group and a hydrogen atom. The fluorine atom is observed in a single position indicating it does not rotate like methyl hydrogen atoms. This change also causes subtle differences in the packing interactions.     

Predicting surface exposure of amino acids from protein sequence

The amino acid residues on a protein surface play a key role in interaction with other molecules, determined many physical properties, and constrain the structure of the folded protein. A database of monomeric protein crystal structures was used to teach computer-simulated neural networks rules for predicting surface exposure from local sequence. These trained networks are able to correctly predict surface exposure for 72% of residues in a testing set using a binary model, (buried/exposed) and for 54% of residues using a ternary model (buried/intermediate/exposed). In the ternary model, only 11% of the exposed residues are predicted as buried and only 5% of the buried residues are predicted as exposed. Also, since the networks are able to predict exposure with a quantitative confidence estimate, it is possible to assign exposure for over half of the residues in a binary model with greater than 80% accuracy. Even more accurate predictions are obtained by making a consensus prediction of exposure for a homologous family. The effect of the local environment of an amino acid on its accessibility, though smaller than expected, is significant and accounts for the higher success rate of prediction than obtained with previously used criteria. In the absence of a three-dimensional structure, the ability to predict surface accessibility of amino acids directly from the sequence is a valuable tool in choosing sites of chemical modification or specific mutations and in studies of molecular interaction.     

Statistical analysis of protein structures suggests that buried ionizable residues in proteins are hydrogen bonded or form salt bridges

It is well known that nonpolar residues are largely buried in the interior of proteins, whereas polar and ionizable residues tend to be more localized on the protein surface where they are solvent exposed. Such a distribution of residues between surface and interior is well understood from a thermodynamic point: nonpolar side chains are excluded from the contact with the solvent water, whereas polar and ionizable groups have favorable interactions with the water and thus are preferred at the protein surface. However, there is an increasing amount of information suggesting that polar and ionizable residues do occur in the protein core, including at positions that have no known functional importance. This is inconsistent with the observations that dehydration of polar and in particular ionizable groups is very energetically unfavorable. To resolve this, we performed a detailed analysis of the distribution of fractional burial of polar and ionizable residues using a large set of ?2600 nonhomologous protein structures. We show that when ionizable residues are fully buried, the vast majority of them form hydrogen bonds and/or salt bridges with other polar/ionizable groups. This observation resolves an apparent contradiction: the energetic penalty of dehydration of polar/ionizable groups is paid off by favorable energy of hydrogen bonding and/or salt bridge formation in the protein interior. Our conclusion agrees well with the previous findings based on the continuum models for electrostatic interactions in proteins. Proteins 2011; © 2011 Wiley‐Liss, Inc.     

Dissecting the roles of local packing density and longer‐range effects in protein sequence evolution

What are the structural determinants of protein sequence evolution? A number of site‐specific structural characteristics have been proposed, most of which are broadly related to either the density of contacts or the solvent accessibility of individual residues. Most importantly, there has been disagreement in the literature over the relative importance of solvent accessibility and local packing density for explaining site‐specific sequence variability in proteins. We show that this discussion has been confounded by the definition of local packing density. The most commonly used measures of local packing, such as contact number and the weighted contact number, represent the combined effects of local packing density and longer‐range effects. As an alternative, we propose a truly local measure of packing density around a single residue, based on the Voronoi cell volume. We show that the Voronoi cell volume, when calculated relative to the geometric center of amino‐acid side chains, behaves nearly identically to the relative solvent accessibility, and each individually can explain, on average, approximately 34% of the site‐specific variation in evolutionary rate in a data set of 209 enzymes. An additional 10% of variation can be explained by nonlocal effects that are captured in the weighted contact number. Consequently, evolutionary variation at a site is determined by the combined effects of the immediate amino‐acid neighbors of that site and effects mediated by more distant amino acids. We conclude that instead of contrasting solvent accessibility and local packing density, future research should emphasize on the relative importance of immediate contacts and longer‐range effects on evolutionary variation. Proteins 2016; 84:841–854. © 2016 Wiley Periodicals, Inc.     

Accurate prediction of solvent accessibility using neural networks-based regression

Accurate prediction of relative solvent accessibilities (RSAs) of amino acid residues in proteins may be used to facilitate protein structure prediction and functional annotation. Toward that goal we developed a novel method for improved prediction of RSAs. Contrary to other machine learning-based methods from the literature, we do not impose a classification problem with arbitrary boundaries between the classes. Instead, we seek a continuous approximation of the real-value RSA using nonlinear regression, with several feed forward and recurrent neural networks, which are then combined into a consensus predictor. A set of 860 protein structures derived from the PFAM database was used for training, whereas validation of the results was carefully performed on several nonredundant control sets comprising a total of 603 structures derived from new Protein Data Bank structures and had no homology to proteins included in the training. Two classes of alternative predictors were developed for comparison with the regression-based approach: one based on the standard classification approach and the other based on a semicontinuous approximation with the so-called thermometer encoding. Furthermore, a weighted approximation, with errors being scaled by the observed levels of variability in RSA for equivalent residues in families of homologous structures, was applied in order to improve the results. The effects of including evolutionary profiles and the growth of sequence databases were assessed. In accord with the observed levels of variability in RSA for different ranges of RSA values, the regression accuracy is higher for buried than for exposed residues, with overall 15.3-15.8% mean absolute errors and correlation coefficients between the predicted and experimental values of 0.64-0.67 on different control sets. The new method outperforms classification-based algorithms when the real value predictions are projected onto two-class classification problems with several commonly used thresholds to separate exposed and buried residues. For example, classification accuracy of about 77% is consistently achieved on all control sets with a threshold of 25% RSA. A web server that enables RSA prediction using the new method and provides customizable graphical representation of the results is available at http://sable.cchmc.org.     

Nearest-neighbor effects and structural preferences in dipeptides are a function of the electronic properties of amino acid side-chains

The electronic properties of amino acid side-chains are emerging as an important factor in the preference for secondary structure in proteins. These properties have not been fully characterized, nor has their role in the behavior of peptides been explored in any detail. The present studies sought to evaluate several possibilities: 1) that hydrophilicity can be expressed solely in electronic terms, 2) that substituent effects of side-chains extend across the peptide bond, and (3) nearest-neighbor effects in dipeptides correlate with secondary structural preferences. Quantum mechanics (QM) calculations were used to define the electronic properties of individual amino acids and dipeptides. It was found that the hydrophilicity of an amino acid side-chain can be accurately represented as a function of the electron densities of its component atoms. In addition, the nature of an amino acid in the second position of a dipeptide affects the electronic properties (Mulliken populations and electron densities) of the main-chain atoms of the first residue. Certain electronic features of the dipeptides strongly correlated with propensity for secondary structure. Specifically, Mulliken population data at the Calpha atom and N atom predicted preference for alpha-helices versus coil and strand conformations, respectively. Analysis of dipeptides arrayed in either helical or extended structures revealed lengthening of main-chain bonds in the alpha-helical conformations. A thorough characterization of the electronic properties of amino acids and short peptide segments may provide a better understanding of the forces that determine secondary structure in proteins.     

Revisiting “reverse hydrophobic effect”: Applicable only to coil mutations at the surface

In a seminal paper, Pakula and Sauer (Nature, 1990, 344, 363–364) demonstrated that the increase in side‐chain hydrophobicity has a reverse relationship with protein stability. We have addressed this problem with several examples of mutants that span at different locations in protein structure based on secondary structure and solvent accessibility. We confirmed that the stability change upon single coil mutation at exposed region is reversely correlated with hydrophobicity with a single exception. In addition, we found the existence of such relationship in partially buried coil mutants. The stability of exposed helical mutants is governed by conformational properties. In buried and partially buried helical and strand mutants properties reflecting hydrophobicity have direct relationship with stability, whereas an opposite relationship was obtained with entropy and flexibility. The structural analysis of partially buried/exposed mutants showed that the surrounding residues are important for the stability change upon mutation. These results provide insights to understand the general behavior for the stability of proteins upon amino acid substitutions. © 2009 Wiley Periodicals, Inc. Biopolymers 91: 591–599, 2009. This article was originally published online as an accepted preprint. The “Published Online” date corresponds to the preprint version. You can request a copy of the preprint by emailing the Biopolymers editorial office at biopolymers@wiley.com     

Refined structure of human carbonic anhydrase II at 2.0 Å resolution

The structure of human erythrocytic carbonic anhydrase II has been refined by constrained and restrained structure–factor least-squares refinement at 2.0 Å resolution. The conventional crystallographic R value is 17.3%. Of 167 solvent molecules associated with the protein, four are buried and stabilize secondary structure elements. The zinc ion is ligated to three histidyl residues and one water molecule in a nearly tetrahedral geometry. In addition to the zinc-bound water, seven more water molecules are identified in the active site. Assuming that Glu-106 is deprotonated at pH 8.5, some of the hydrogen bond donor–acceptor relations in the active site can be assigned and are described here in detail. The Oγ1 atom of Thr-199 donates its proton to the Oε1 atom of Glu-106 and can function as a hydrogen bond acceptor only in additional hydrogen bonds.     

Two-stage support vector regression approach for predicting accessible surface areas of amino acids

We address the problem of predicting solvent accessible surface area (ASA) of amino acid residues in protein sequences, without classifying them into buried and exposed types. A two-stage support vector regression (SVR) approach is proposed to predict real values of ASA from the position-specific scoring matrices generated from PSI-BLAST profiles. By adding SVR as the second stage to capture the influences on the ASA value of a residue by those of its neighbors, the two-stage SVR approach achieves improvements of mean absolute errors up to 3.3%, and correlation coefficients of 0.66, 0.68, and 0.67 on the Manesh dataset of 215 proteins, the Barton dataset of 502 nonhomologous proteins, and the Carugo dataset of 338 proteins, respectively, which are better than the scores published earlier on these datasets. A Web server for protein ASA prediction by using a two-stage SVR method has been developed and is available (http://birc.ntu.edu.sg/~ pas0186457/asa.html).     

Amino terminal interaction in the prion protein identified using fusion to green fluorescent protein

In contrast to the well-characterized carboxyl domain, the amino terminal half of the mature cellular prion protein has no defined structure. Here, following fusion of mouse prion protein fragments to green fluorescence protein as a reporter of protein stability, we report extreme variability in fluorescence level that is dependent on the prion fragment expressed. In particular, exposure of the extreme amino terminus in the context of a truncated prion protein molecule led to rapid degradation, whereas the loss of only six amino terminal residues rescued high level fluorescence. Study of the precise endpoints and residue identity associated with high fluorescence suggested a domain within the amino terminal half of the molecule defined by a long-range intramolecular interaction between 23KKRPKP28 and 143DWED146 and dependent upon the anti-parallel beta-sheet ending at residue 169 and normally associated with the structurally defined carboxyl terminal domain. This previously unreported interaction may be significant for understanding prion bioactivity and for structural studies aimed at the complete prion structure.     

Sequence and structure conservation in a protein core

In order to study structural aspects of sequence conservation in families of homologous proteins, we have analyzed structurally aligned sequences of 585 proteins grouped into 128 homologous families. The conservation of a residue in a family is defined as the average residue similarity in a given position of aligned sequences. The residue similarities were expressed in the form of log-odd substitution tables that take into account the environments of amino acids in three-dimensional structures. The protein core is defined as those residues that have less then 7% solvent accessibility. The density of a protein core is described in terms of atom packing, which is investigated as a criterion for residue substitution and conservation. Although there is no significant correlation between sequence conservation and average atom packing around nonpolar residues such as leucine, valine and isoleucine, a significant correlation is observed for polar residues in the protein core. This may be explained by the hydrogen bonds in which polar residues are involved; the better their protection from water access the more stable should be the structure in that position. Proteins 33:358–366, 1998. © 1998 Wiley-Liss, Inc.