Prediction of protein secondary structure from amino acid sequence

The conformational parametersP _k for each amino acid species (j=1–20) of sequential peptides in proteins are presented as the product ofP _i,k , wherei is the number of the sequential residues in thekth conformational state (k=-helix,-sheet,-turn, or unordered structure). Since the average parameter for ann-residue segment is related to the average probability of finding the segment in the kth state, it becomes a geometric mean of (P _k )_av=(P _i,k ) ^1/n with amino acid residuei increasing from 1 ton. We then used ln(P_k)_av to convert a multiplicative process to a summation, i.e., ln(P _k ) _av =(1/n)P _i,k (i=1 ton) for ease of operation. However, this is unlike the popular Chou-Fasman algorithm, which has the flaw of using the arithmetic mean for relative probabilities. The Chou-Fasman algorithm happens to be close to our calculations in many cases mainly because the difference between theirP _k and our InP _k is nearly constant for about one-half of the 20 amino acids. When stronger conformation formers and breakers exist, the difference become larger and the prediction at the N- and C-terminal-helix or-sheet could differ. If the average conformational parameters of the overlapping segments of any two states are too close for a unique solution, our calculations could lead to a different prediction.     

Prediction of protein domain boundaries from statistics of appearance of amino acid residues

A database of 452 two-domain proteins with less than 25% homology was constructed. One half of the database was used to obtain statistics on the appearance of amino acid residues at domain boundaries. Small and hydrophilic residues (proline, glycine, asparagine, glutamic acid, arginine, etc.) occurred more often at domain boundaries than in total proteins. Hydrophobic residues (tryptophan, methionine, phenylalanine, etc.) were rarer at domain boundaries than in total proteins. Probability scales of amino acid appearance in boundary-flanking regions were constructed with these statistics and used to predict the domain boundaries in proteins of the other half of the database. The probability scale obtained by averaging the appearance of amino acids over an 8-residue region (±4 residues from the real domain boundaries) yielded the best results: domain boundaries were predicted within 40 residues of the real boundary in 57% of proteins and within 20 residues of the real boundary in 41% of proteins. The probability scale was used to predict the domain boundaries in proteins with unknown structures (CASP6).     

Cooperative alpha-helix unfolding in a protein-DNA complex from hydrogen-deuterium exchange

We present experimental evidence for a cooperative unfolding transition of an alpha-helix in the lac repressor headpiece bound to a symmetric variant of the lac operator, as inferred from hydrogen-deuterium (H-D) exchange experiments monitored by NMR spectroscopy. In the EX1 limit, observed exchange rates become pH-independent and exclusively sensitive to local structure fluctuations that expose the amide proton HN to exchange. Close to this regime, we measured decay rates of individual backbone HN signals in D2O, and of their mutual HN-HN NOE by time-resolved two-dimensional (2D) NMR experiments. The data revealed correlated exchange at the center of the lac headpiece recognition helix, Val20-Val23, and suggested that the correlation breaks down at Val24, at the C terminus of the helix. A lower degree of correlation was observed for the exchange of Val9 and Ala10 at the center of helix 1, while no correlation was observed for Val38 and Glu39 at the center of helix 3. We conclude that HN exchange in the recognition helix and, to some extent, in helix 1 is a cooperative event involving the unfolding of these helices, whereas the HN exchange in helix 3 is dominated by random local structure fluctuations.     

Rapid protein domain assignment from amino acid sequence using predicted secondary structure

The elucidation of the domain content of a given protein sequence in the absence of determined structure or significant sequence homology to known domains is an important problem in structural biology. Here we address how successfully the delineation of continuous domains can be accomplished in the absence of sequence homology using simple baseline methods, an existing prediction algorithm (Domain Guess by Size), and a newly developed method (DomSSEA). The study was undertaken with a view to measuring the usefulness of these prediction methods in terms of their application to fully automatic domain assignment. Thus, the sensitivity of each domain assignment method was measured by calculating the number of correctly assigned top scoring predictions. We have implemented a new continuous domain identification method using the alignment of predicted secondary structures of target sequences against observed secondary structures of chains with known domain boundaries as assigned by Class Architecture Topology Homology (CATH). Taking top predictions only, the success rate of the method in correctly assigning domain number to the representative chain set is 73.3%. The top prediction for domain number and location of domain boundaries was correct for 24% of the multidomain set (+/-20 residues). These results have been put into context in relation to the results obtained from the other prediction methods assessed.     

Prediction of protein folding class from amino acid composition

An empirical relation between the amino acid composition and three-dimensional folding pattern of several classes of proteins has been determined. Computer simulated neural networks have been used to assign proteins to one of the following classes based on their amino acid composition and size: (1) 4α-helical bundles, (2) parallel (α/β)₈ barrels, (3) nucleotide binding fold, (4) immunoglobulin fold, or (5) none of these. Networks trained on the known crystal structures as well as sequences of closely related proteins are shown to correctly predict folding classes of proteins not represented in the training set with an average accuracy of 87%. Other folding motifs can easily be added to the prediction scheme once larger databases become available. Analysis of the neural network weights reveals that amino acids favoring prediction of a folding class are usually over represented in that class and amino acids with unfavorable weights are underrepresented in composition. The neural networks utilize combinations of these multiple small variations in amino acid composition in order to make a prediction. The favorably weighted amino acids in a given class also form the most intramolecular interactions with other residues in proteins of that class. A detailed examination of the contacts of these amino acids reveals some general patterns that may help stabilize each folding class. © 1993 Wiley-Liss, Inc.     

Patterns in ionizable side chain interactions in protein structures

In a selected set of 44 high-resolution, non-homologous protein structures, the intramolecular hydrogen bonds or salt bridges formed by ionizable amino acid side chains were identified and analyzed. The analysis was based on the investigation of several properties of the involved residues such as their solvent exposure, their belonging to a certain secondary structural element, and their position relative to the N- and C-termini of their respective structural element. It was observed that two-thirds of the interactions made by basic or acidic side chains are hydrogen bonds to polar uncharged groups. In particular, the majority (78%) of the hydrogen bonds between ionizable side chains and main chain polar groups (sch:mch bonds) involved at least one buried atom, and in 42% of the cases both interacting atoms were buried. In α-helices, the sch:mch bonds observed in the proximity of the C- and N-termini show a clear preference for acidic and basic side chains, respectively. This appears to be due to the partial charges of peptide group atoms at the termini of α-helices, which establish energetically favorable electrostatic interactions with side chain carrying opposite charge, at distances even greater than 4.5 Å. The sch:mch interactions involving ionizable side chains that belong either to β-strands or to the central part of α-helices are based almost exclusively on basic residues. This results from the presence of main chain carbonyl oxygen atoms in the protein core which have unsatisfied hydrogen bonding capabilities.     

Prediction of protein secondary structure content using amino acid composition and evolutionary information

Knowing protein structure and inferring its function from the structure are one of the main issues of computational structural biology, and often the first step is studying protein secondary structure. There have been many attempts to predict protein secondary structure contents. Previous attempts assumed that the content of protein secondary structure can be predicted successfully using the information on the amino acid composition of a protein. Recent methods achieved remarkable prediction accuracy by using the expanded composition information. The overall average error of the most successful method is 3.4%. Here, we demonstrate that even if we only use the simple amino acid composition information alone, it is possible to improve the prediction accuracy significantly if the evolutionary information is included. The idea is motivated by the observation that evolutionarily related proteins share the similar structure. After calculating the homolog-averaged amino acid composition of a protein, which can be easily obtained from the multiple sequence alignment by running PSI-BLAST, those 20 numbers are learned by a multiple linear regression, an artificial neural network and a support vector regression. The overall average error of method by a support vector regression is 3.3%. It is remarkable that we obtain the comparable accuracy without utilizing the expanded composition information such as pair-coupled amino acid composition. This work again demonstrates that the amino acid composition is a fundamental characteristic of a protein. It is anticipated that our novel idea can be applied to many areas of protein bioinformatics where the amino acid composition information is utilized, such as subcellular localization prediction, enzyme subclass prediction, domain boundary prediction, signal sequence prediction, and prediction of unfolded segment in a protein sequence, to name a few.     

Amyloidogenic sequences in native protein structures

Numerous short peptides have been shown to form β‐sheet amyloid aggregates in vitro. Proteins that contain such sequences are likely to be problematic for a cell, due to their potential to aggregate into toxic structures. We investigated the structures of 30 proteins containing 45 sequences known to form amyloid, to see how the proteins cope with the presence of these potentially toxic sequences, studying secondary structure, hydrogen‐bonding, solvent accessible surface area and hydrophobicity. We identified two mechanisms by which proteins avoid aggregation: Firstly, amyloidogenic sequences are often found within helices, despite their inherent preference to form β structure. Helices may offer a selective advantage, since in order to form amyloid the sequence will presumably have to first unfold and then refold into a β structure. Secondly, amyloidogenic sequences that are found in β structure are usually buried within the protein. Surface exposed amyloidogenic sequences are not tolerated in strands, presumably because they lead to protein aggregation via assembly of the amyloidogenic regions. The use of α‐helices, where amyloidogenic sequences are forced into helix, despite their intrinsic preference for β structure, is thus a widespread mechanism to avoid protein aggregation.     

Thermodynamic effects of replacements of Pro residues in helix interiors of maltose-binding protein

Introduction of Pro residues into helix interiors results in protein destabilization. It is currently unclear if the converse substitution (i.e., replacement of Pro residues that naturally occur in helix interiors would be stabilizing). Maltose-binding protein is a large 370-amino acid protein that contains 21 Pro residues. Of these, three nonconserved residues (P48, P133, and P159) occur at helix interiors. Each of the residues was replaced with Ala and Ser. Stabilities were characterized by differential scanning calorimetry (DSC) as a function of pH and by isothermal urea denaturation studies as a function of temperature. The P48S and P48A mutants were found to be marginally more stable than the wild-type protein. In the pH range of 5-9, there is an average increase in T(m) values of P48A and P48S of 0.4 degrees C and 0.2 degrees C, respectively, relative to the wild-type protein. The other mutants are less stable than the wild type. Analysis of the effects of such Pro substitutions in MBP and in three other proteins studied to date suggests that substitutions are more likely to be stabilizing if the carbonyl group i-3 or i-4 to the mutation site is not hydrogen bonded in the wild-type protein.     

Mass spectrometric measurement of protein amide hydrogen exchange rates of apo- and holo-myoglobin.

Measurement of backbone amide hydrogen exchange rates can provide detailed information concerning protein structure, dynamics, and interactions. Although nuclear magnetic resonance is typically used to provide these data, its use is restricted to lower molecular weight proteins that are soluble at millimolar concentrations. Not subject to these limitations is a mass spectrometric approach for measuring deuterium incorporation into proteins that are subsequently proteolyzed by pepsin; the resulting peptide masses are measured using a flowing-fast atom bombardment ionization source (Zhang Z, Smith DL, 1993, Protein Sci 2:522-531). In the current study, amide deuterium incorporation for intact apo- and holo-myoglobin was measured using liquid chromatography coupled directly to an electrospray ionization (LC/MS) source. Electrospray ionization provided a more complete coverage of the protein sequence and permitted the measurement of deuterium incorporation into intact proteins. Tandem mass spectrometry was used to rapidly identify the peptic peptides. It was found that within 30 s, the amides in apo-myoglobin were 47% deuterated, whereas holo-myoglobin was 12% deuterated. Peptic digestion and LC/MS demonstrated that regions represented by peptic peptides encompassing positions 1-7, 12-29, and 110-134 were not significantly altered by removal of the heme. Likewise, destabilized regions were identified within positions 33-106 and 138-153.     

Assignment of polar states for protein amino acid residues using an interaction cluster decomposition algorithm and its application to high resolution protein structure modeling

We have developed a new method (Independent Cluster Decomposition Algorithm, ICDA) for creating all-atom models of proteins given the heavy-atom coordinates, provided by X-ray crystallography, and the pH. In our method the ionization states of titratable residues, the crystallographic mis-assignment of amide orientations in Asn/Gln, and the orientations of OH/SH groups are addressed under the unified framework of polar states assignment. To address the large number of combinatorial possibilities for the polar hydrogen states of the protein, we have devised a novel algorithm to decompose the system into independent interacting clusters, based on the observation of the crucial interdependence between the short range hydrogen bonding network and polar residue states, thus significantly reducing the computational complexity of the problem and making our algorithm tractable using relatively modest computational resources. We utilize an all atom protein force field (OPLS) and a Generalized Born continuum solvation model, in contrast to the various empirical force fields adopted in most previous studies. We have compared our prediction results with a few well-documented methods in the literature (WHATIF, REDUCE). In addition, as a preliminary attempt to couple our polar state assignment method with real structure predictions, we further validate our method using single side chain prediction, which has been demonstrated to be an effective way of validating structure prediction methods without incurring sampling problems. Comparisons of single side chain prediction results after the application of our polar state prediction method with previous results with default polar state assignments indicate a significant improvement in the single side chain predictions for polar residues.     

Determinants of protein hydrogen exchange studied in equine cytochrome c.

The exchange of a large number of amide hydrogens in oxidized equine cytochrome c was measured by NMR and compared with structural parameters. Hydrogens known to exchange through local structural fluctuations and through larger unfolding reactions were separately considered. All hydrogens protected from exchange by factors greater than 10(3) are in defined H-bonds, and almost all H-bonded hydrogens including those at the protein surface were measured to exchange slowly. H-exchange rates do not correlate with H-bond strength (length) or crystallographic B factors. It appears that the transient structural fluctuation necessary to bring an exchangeable hydrogen into H-bonding contact with the H-exchange catalyst (OH(-)-ion) involves a fairly large separation of the H-bond donor and acceptor, several angstroms at least, and therefore depends on the relative resistance to distortion of immediately neighboring structure. Accordingly, H-exchange by way of local fluctuational pathways tends to be very slow for hydrogens that are neighbored by tightly anchored structure and for hydrogens that are well buried. The slowing of buried hydrogens may also reflect the need for additional motions that allow solvent access once the protecting H-bond is separated, although it is noteworthy that burial in a protein like cytochrome c does not exceed 4 angstroms. When local fluctuational pathways are very slow, exchange can become dominated by a different category of larger, cooperative, segmental unfolding reactions reaching up to global unfolding.     

Combining the GOR V algorithm with evolutionary information for protein secondary structure prediction from amino acid sequence

We have modified and improved the GOR algorithm for the protein secondary structure prediction by using the evolutionary information provided by multiple sequence alignments, adding triplet statistics, and optimizing various parameters. We have expanded the database used to include the 513 non-redundant domains collected recently by Cuff and Barton (Proteins 1999;34:508-519; Proteins 2000;40:502-511). We have introduced a variable size window that allowed us to include sequences as short as 20-30 residues. A significant improvement over the previous versions of GOR algorithm was obtained by combining the PSI-BLAST multiple sequence alignments with the GOR method. The new algorithm will form the basis for the future GOR V release on an online prediction server. The average accuracy of the prediction of secondary structure with multiple sequence alignment and full jack-knife procedure was 73.5%. The accuracy of the prediction increases to 74.2% by limiting the prediction to 375 (of 513) sequences having at least 50 PSI-BLAST alignments. The average accuracy of the prediction of the new improved program without using multiple sequence alignments was 67.5%. This is approximately a 3% improvement over the preceding GOR IV algorithm (Garnier J, Gibrat JF, Robson B. Methods Enzymol 1996;266:540-553; Kloczkowski A, Ting K-L, Jernigan RL, Garnier J. Polymer 2002;43:441-449). We have discussed alternatives to the segment overlap (Sov) coefficient proposed by Zemla et al. (Proteins 1999;34:220-223).     

Insight into a molecular interaction force supporting peptide backbones and its implication to protein loops and folding

Although not being classified as the most fundamental protein structural elements like α-helices and β-strands, the loop segment may play considerable roles for protein stability, flexibility, and dynamic activity. Meanwhile, the protein loop is also quite elusive; i.e. its interactions with the other parts of protein as well as its own shape-maintaining forces have still remained as a puzzle or at least not quite clear yet. Here, we report a molecular force, the so-called polar hydrogen–π interaction (Hp–π), which may play an important role in supporting the backbones of protein loops. By conducting the potential energy surface scanning calculations on the quasi π-plane of peptide bond unit, we have observed the following intriguing phenomena: (1) when the polar hydrogen atom of a peptide unit is perpendicularly pointing to the π-plane of other peptide bond units, a remarkable Hp–π interaction occurs; (2) the interaction is distance and orientation dependent, acting in a broad space, and belonging to the ‘point-to-plane’ one. The molecular force reported here may provide useful interaction concepts and insights into better understanding the loop’s unique stability and flexibility feature, as well as the driving force of the protein global folding.     

Circular dichroism studies on helical structure preferences of amino acid residues of proteins caused by sodium dodecyl sulfate

The extent of helical structure of 19 intact proteins and of 15 proteins with no disulfide bridges in the absence and presence of 10 mM sodium dodecyl sulfate (SDS) was determined using the curve-fitting method of circular dichroic spectra. The change in helicity caused by the addition of SDS was examined as a function of each amino acid fraction. An increase in the helicity upon the addition of SDS occurred in most of the proteins with no disulfide bridges (C proteins) and containing more than 0.06 Lys fraction. In most of the intact proteins (B proteins), most of which contained disulfide bridges, helicity in SDS decreased with an increase in Lys fraction. The helicity of the C proteins in SDS also tended to increase with an increase in the Leu and Phe fractions, while it decreased with an increase in the Gly fraction. For the helicity of the B proteins in SDS, there was a tendency to increase with increased Asn fraction and decrease with increased His fraction. On the other hand, amino acids were divided into eight groups according to their side-chain properties and the conformational preference for each of the amino acid groups of C proteins was calculated using a simple assumption.     

Temperature-dependence of protein hydrogen bond properties as studied by high-resolution NMR

The temperature-dependence of a large number of NMR parameters describing hydrogen bond properties in the protein ubiquitin was followed over a range from 5 to 65 degrees C. The parameters comprise hydrogen bond (H-bond) scalar couplings, h3JNC', chemical shifts, amide proton exchange rates, 15N relaxation parameters as well as covalent 1JNC' and 1JNH couplings. A global weakening of the h3JNC' coupling with increasing temperature is accompanied by a global upfield shift of the amide protons and a decrease of the sequential 1JNC' couplings. If interpreted as a linear increase of the N...O distance, the change in h3JNC' corresponds to an average linear thermal expansion coefficient for the NH-->O hydrogen bonds of 1.7 x 10(-4)/K, which is in good agreement with overall volume expansion coefficients observed for proteins. A residue-specific analysis reveals that not all hydrogen bonds are affected to the same extent by the thermal expansion. The end of beta-sheet beta1/beta5 at hydrogen bond E64-->Q2 appears as the most thermolabile, whereas the adjacent hydrogen bond I3-->L15 connecting beta-strands beta1 and beta2 is even stabilized slightly at higher temperatures. Additional evidence for the stabilization of the beta1/beta2 beta-hairpin at higher temperatures is found in reduced hydrogen exchange rates for strand end residue V17. This reduction corresponds to a stabilizing change in free energy of 9.7 kJ/mol for the beta1/beta2 hairpin. The result can be linked to the finding that the beta1/beta2 hairpin behaves as an autonomously folding unit in the A-state of ubiquitin under changed solvent conditions. For several amide groups the temperature-dependencies of the amide exchange rates and H-bond scalar couplings are uncorrelated. Therefore, amide exchange rates are not a sole function of the hydrogen bond "strength" as given by the electronic overlap of donors and acceptors, but are clearly dependent on other blocking mechanisms.     

Prediction of prolyl residues in cis-conformation in protein structures on the basis of the amino acid sequence

In proteins most peptide bonds are in trans-conformation: the torsion angle omega = 180 degrees. Only few show cis-conformation in known protein structures (omega = 0 degrees). Most of them are prolyl residues. About 6% of about 4000 prolyl residues are in cis-conformation. Between trans- and cis-prolyl residues significant differences are observed in the surrounding sequences. E.g. there are large amounts of aromatic residues N-terminally in case of cis-prolyl residues, but in the case of trans-prolyl residues more aromatic amino acids occur C-terminally. But in all cases there are only complex patterns which are indicative of cis- and trans-conformation, respectively. Considering the neighbours (+/- 6 residues) of prolyl residues and their physicochemical properties we find 6 different patterns which allow one to assign correctly about 75% of known cis-structured prolyl residues, whereby no false positive one is predicted.     

Multiple conformations of amino acid residues in ribonuclease A

The highly refined 1.26 A structure (R = 0.15) of phosphate-free bovine pancreatic ribonuclease A was modeled with 13 residues having discrete multiple conformations of side chains. These residues are widely distributed over the protein surface, but only one of them, Lys 61, is involved in crystal packing interactions. The discrete conformers have no unusual torsion angles, and their interactions with the solvent and with other atoms of the protein are similar to those residues modeled with a single conformation. For three of the residues--Val 43, Asp 83, and Arg 85--two correlated conformations are found. The observed multiple conformations on the protein surfaces will be of significance in analyzing structure-function relationships and in performing protein engineering.     

Prudent modeling of core polar residues in computational protein design

Hydrogen bond interactions were surveyed in a set of protein structures. Compared to surface positions, polar side-chains at core positions form a greater number of intra-molecular hydrogen bonds. Furthermore, the majority of polar side-chains at core positions form at least one hydrogen bond to main-chain atoms that are not involved in hydrogen bonds to other main-chain atoms. Based on this structural survey, hydrogen bond rules were generated for each polar amino acid for use in protein core design. In the context of protein core design, these prudent polar rules were used to eliminate from consideration polar amino acid rotamers that do not form a minimum number of hydrogen bonds. As an initial test, the core of Escherichia coli thioredoxin was selected as a design target. For this target, the prudent polar strategy resulted in a minor increase in computational complexity compared to a strategy that did not allow polar residues. Dead-end elimination was used to identify global minimum energy conformations for the prudent polar and no polar strategies. The prudent polar strategy identified a protein sequence that was thermodynamically stabilized by 2.5 kcal/mol relative to wild-type thioredoxin and 2.2 kcal/mol relative to a thioredoxin variant whose core was designed without polar residues.     

NdPASA: a novel pairwise protein sequence alignment algorithm that incorporates neighbor-dependent amino acid propensities

Sequence alignment has become one of the essential bioinformatics tools in biomedical research. Existing sequence alignment methods can produce reliable alignments for homologous proteins sharing a high percentage of sequence identity. The performance of these methods deteriorates sharply for the sequence pairs sharing less than 25% sequence identity. We report here a new method, NdPASA, for pairwise sequence alignment. This method employs neighbor-dependent propensities of amino acids as a unique parameter for alignment. The values of neighbor-dependent propensity measure the preference of an amino acid pair adopting a particular secondary structure conformation. NdPASA optimizes alignment by evaluating the likelihood of a residue pair in the query sequence matching against a corresponding residue pair adopting a particular secondary structure in the template sequence. Using superpositions of homologous proteins derived from the PSI-BLAST analysis and the Structural Classification of Proteins (SCOP) classification of a nonredundant Protein Data Bank (PDB) database as a gold standard, we show that NdPASA has improved pairwise alignment. Statistical analyses of the performance of NdPASA indicate that the introduction of sequence patterns of secondary structure derived from neighbor-dependent sequence analysis clearly improves alignment performance for sequence pairs sharing less than 20% sequence identity. For sequence pairs sharing 13-21% sequence identity, NdPASA improves the accuracy of alignment over the conventional global alignment (GA) algorithm using the BLOSUM62 by an average of 8.6%. NdPASA is most effective for aligning query sequences with template sequences whose structure is known. NdPASA can be accessed online at http://astro.temple.edu/feng/Servers/BioinformaticServers.htm.