A new method for evaluating the specificity of indirect readout in protein-DNA recognition.

Proteins recognize a specific DNA sequence not only through direct contact (direct readout) with base pairs but also through sequence-dependent conformation and/or flexibility of DNA (indirect readout). However, it is difficult to assess the contribution of indirect readout to the sequence specificity. What is needed is a straightforward method for quantifying its contributions to specificity. Using Bayesian statistics, we derived the probability of a particular sequence for a given DNA structure from the trajectories of molecular dynamics (MD) simulations of DNAs containing all possible tetramer sequences. Then, we quantified the specificity of indirect readout based on the information entropy associated with the probability. We tested this method with known structures of protein-DNA complexes. This method enabled us to correctly predict those regions where experiments suggested the involvement of indirect readout. The results also indicated new regions where the indirect readout mechanism makes major contributions to the recognition. The present method can be used to estimate the contribution of indirect readout without approximations to the distributions in the conformational ensembles of DNA, and would serve as a powerful tool to study the mechanism of protein-DNA recognition.

CCRXP: exploring clusters of conserved residues in protein structures.

Conserved residues forming tightly packed clusters have been shown to be energy hot spots in both protein-protein and protein-DNA complexes. A number of analyses on these clusters of conserved residues (CCRs) have been reported, all pointing to a crucial role that these clusters play in protein function, especially protein-protein and protein-DNA interactions. However, currently there is no publicly available tool to automatically detect such clusters. Here, we present a web server that takes a coordinate file in PDB format as input and automatically executes all the steps to identify CCRs in protein structures. In addition, it calculates the structural properties of each residue and of the CCRs. We also present statistics to show that CCRs, determined by these procedures, are significantly enriched in 'hot spots' in protein-protein and protein-RNA complexes, which supplements our more detailed similar results on protein-DNA complexes. We expect that CCRXP web server will be useful in studies of protein structures and their interactions and selecting mutagenesis targets. The web server can be accessed at http://ccrxp.netasa.org.

Prediction of dinucleotide-specific RNA-binding sites in proteins.

BACKGROUND: Regulation of gene expression, protein synthesis, replication and assembly of many viruses involve RNA-protein interactions. Although some successful computational tools have been reported to recognize RNA binding sites in proteins, the problem of specificity remains poorly investigated. After the nucleotide base composition, the dinucleotide is the smallest unit of RNA sequence information and many RNA-binding proteins simply bind to regions enriched in one dinucleotide. Interaction preferences of protein subsequences and dinucleotides can be inferred from protein-RNA complex structures, enabling a training-based prediction approach. RESULTS: We analyzed basic statistics of amino acid-dinucleotide contacts in protein-RNA complexes and found their pairing preferences could be identified. Using a standard approach to represent protein subsequences by their evolutionary profile, we trained neural networks to predict multiclass target vectors corresponding to 16 possible contacting dinucleotide subsequences. In the cross-validation experiments, the accuracies of the optimum network, measured as areas under the curve (AUC) of the receiver operating characteristic (ROC) graphs, were in the range of 65-80%. CONCLUSIONS: Dinucleotide-specific contact predictions have also been extended to the prediction of interacting protein and RNA fragment pairs, which shows the applicability of this method to predict targets of RNA-binding proteins. A web server predicting the 16-dimensional contact probability matrix directly from a user-defined protein sequence was implemented and made available at: http://tardis.nibio.go.jp/netasa/srcpred.

Analysis of electric moments of RNA-binding proteins: implications for mechanism and prediction.

BACKGROUND: Protein-RNA interactions play important role in many biological processes such as gene regulation, replication, protein synthesis and virus assembly. Although many structures of various types of protein-RNA complexes have been determined, the mechanism of protein-RNA recognition remains elusive. We have earlier shown that the simplest electrostatic properties viz. charge, dipole and quadrupole moments, calculated from backbone atomic coordinates of proteins are biased relative to other proteins, and these quantities can be used to identify DNA-binding proteins. Closely related, RNA-binding proteins are investigated in this study. In particular, discrimination between various types of RNA-binding proteins, evolutionary conservation of these bulk electrostatic features and effect of conformational changes by complex formation are investigated. Basic binding mechanism of a putative RNA-binding protein (HI1333 from Haemophilus influenza) is suggested as a potential application of this study. RESULTS: We found that similar to DNA-binding proteins (DBPs), RNA-binding proteins (RBPs) also show significantly higher values of electric moments. However, higher moments in RBPs are found to strongly depend on their functional class: proteins binding to ribosomal RNA (rRNA) constitute the only class with all three of the properties (charge, dipole and quadrupole moments) being higher than control proteins. Neural networks were trained using leave-one-out cross-validation to predict RBPs from control data as well as pair-wise classification capacity between proteins binding to various RNA types. RBPs and control proteins reached up to 78% accuracy measured by the area under the ROC curve. Proteins binding to rRNA are found to be best distinguished (AUC = 79%). Changes in dipole and quadrupole moments between unbound and bound structures were small and these properties are found to be robust under complex formation. CONCLUSIONS: Bulk electric moments of proteins considered here provide insights into target recognition by RNA-binding proteins, as well as ability to recognize one type of RBP from others. These results help in understanding the mechanism of protein-RNA recognition, and identifying RNA-binding proteins.

Genetic algorithm optimization in drug design QSAR: Bayesian-regularized genetic neural networks (BRGNN) and genetic algorithm-optimized support vectors machines (GA-SVM).

Many articles in "in silico" drug design implemented genetic algorithm (GA) for feature selection, model optimization, conformational search, or docking studies. Some of these articles described GA applications to quantitative structure-activity relationships (QSAR) modeling in combination with regression and/or classification techniques. We reviewed the implementation of GA in drug design QSAR and specifically its performance in the optimization of robust mathematical models such as Bayesian-regularized artificial neural networks (BRANNs) and support vector machines (SVMs) on different drug design problems. Modeled data sets encompassed ADMET and solubility properties, cancer target inhibitors, acetylcholinesterase inhibitors, HIV-1 protease inhibitors, ion-channel and calcium entry blockers, and antiprotozoan compounds as well as protein classes, functional, and conformational stability data. The GA-optimized predictors were often more accurate and robust than previous published models on the same data sets and explained more than 65% of data variances in validation experiments. In addition, feature selection over large pools of molecular descriptors provided insights into the structural and atomic properties ruling ligand-target interactions.

A generalized conformational energy function of DNA derived from molecular dynamics simulations.

Proteins recognize DNA sequences by two different mechanisms. The first is direct readout, in which recognition is mediated by direct interactions between the protein and the DNA bases. The second is indirect readout, which is caused by the dependence of conformation and the deformability of the DNA structure on the sequence. Various energy functions have been proposed to evaluate the contribution of indirect readout to the free-energy changes in complex formations. We developed a new generalized energy function to estimate the dependence of the deformability of DNA on the sequence. This function was derived from molecular dynamics simulations previously conducted on B-DNA dodecamers, each of which had one possible tetramer sequence embedded at its center. By taking the logarithm of the probability distribution function (PDF) for the base-step parameters of the central base-pair step of the tetramer, its ability to distinguish the native sequence from random ones was superior to that with the previous method that approximated the energy function in harmonic form. From a comparison of the energy profiles calculated with these two methods, we found that the harmonic approximation caused significant errors in the conformational energies of the tetramers that adopted multiple stable conformations.

Proteochemometric recognition of stable kinase inhibition complexes using topological autocorrelation and support vector machines.

Intensive research has been performed on computational design of kinase inhibitors using molecular dynamics simulations, docking and quantitative structure-activity relationship (QSAR) analyses, all of which have their own limitations. In this paper, we report the application of proteochemometrics, a ligand-target modeling approach, to the recognition of stable and unstable kinase-inhibitor complexes using support vector machines (SVM) classifiers. The algorithm consists of creating topological autocorrelation descriptors for kinases and inhibitors and then development of SVM models to relate the feature vectors to the stability class (stable or unstable) of hypothetical protein-inhibitor complexes. The approach based on the autocorrelation features was compared with fragment-based approach and the former was found to outperform the later. The final classifier could recognize 82% of data to be stable or unstable using jackknife type of validation and test set prediction. Analysis of substructure classification showed a very homogeneous behavior of the model on the whole target-ligand space. The predictor is available online at http://gibk21.bse.kyutech.ac.jp/AUTOkinI/SVMpredictor.html.

Indirect readout in drug-DNA recognition: role of sequence-dependent DNA conformation.

DNA-binding drugs have numerous applications in the engineered gene regulation. However, the drug-DNA recognition mechanism is poorly understood. Drugs can recognize specific DNA sequences not only through direct contacts but also indirectly through sequence-dependent conformation, in a similar manner to the indirect readout mechanism in protein-DNA recognition. We used a knowledge-based technique that takes advantage of known DNA structures to evaluate the conformational energies. We built a dataset of non-redundant free B-DNA crystal structures to calculate the distributions of adjacent base-step and base-pair conformations, and estimated the effective harmonic potentials of mean force (PMF). These PMFs were used to calculate the conformational energy of drug-DNA complexes, and the Z-score as a measure of the binding specificity. Comparing the Z-scores for drug-DNA complexes with those for free DNA structures with the same sequence, we observed that in several cases the Z-scores became more negative upon drug binding. Furthermore, the specificity is position-dependent within the drug-bound region of DNA. These results suggest that DNA conformation plays an important role in the drug-DNA recognition. The presented method provides a tool for the analysis of drug-DNA recognition and can facilitate the development of drugs for targeting a specific DNA sequence.

Protein-DNA interactions: structural, thermodynamic and clustering patterns of conserved residues in DNA-binding proteins.

Amino acid residues, which play important roles in protein function, are often conserved. Here, we analyze thermodynamic and structural data of protein-DNA interactions to explore a relationship between free energy, sequence conservation and structural cooperativity. We observe that the most stabilizing residues or putative hotspots are those which occur as clusters of conserved residues. The higher packing density of the clusters and available experimental thermodynamic data of mutations suggest cooperativity between conserved residues in the clusters. Conserved singlets contribute to the stability of protein-DNA complexes to a lesser extent. We also analyze structural features of conserved residues and their clusters and examine their role in identifying DNA-binding sites. We show that about half of the observed conserved residue clusters are in the interface with the DNA, which could be identified from their amino acid composition; whereas the remaining clusters are at the protein-protein or protein-ligand interface, or embedded in the structural scaffolds. In protein-protein interfaces, conserved residues are highly correlated with experimental residue hotspots, contributing dominantly and often cooperatively to the stability of protein-protein complexes. Overall, the conservation patterns of the stabilizing residues in DNA-binding proteins also highlight the significance of clustering as compared to single residue conservation.

A signature-based method for indexing cell cycle phase distribution from microarray profiles.

BACKGROUND: The cell cycle machinery interprets oncogenic signals and reflects the biology of cancers. To date, various methods for cell cycle phase estimation such as mitotic index, S phase fraction, and immunohistochemistry have provided valuable information on cancers (e.g. proliferation rate). However, those methods rely on one or few measurements and the scope of the information is limited. There is a need for more systematic cell cycle analysis methods. RESULTS: We developed a signature-based method for indexing cell cycle phase distribution from microarray profiles under consideration of cycling and non-cycling cells. A cell cycle signature masterset, composed of genes which express preferentially in cycling cells and in a cell cycle-regulated manner, was created to index the proportion of cycling cells in the sample. Cell cycle signature subsets, composed of genes whose expressions peak at specific stages of the cell cycle, were also created to index the proportion of cells in the corresponding stages. The method was validated using cell cycle datasets and quiescence-induced cell datasets. Analyses of a mouse tumor model dataset and human breast cancer datasets revealed variations in the proportion of cycling cells. When the influence of non-cycling cells was taken into account, "buried" cell cycle phase distributions were depicted that were oncogenic-event specific in the mouse tumor model dataset and were associated with patients' prognosis in the human breast cancer datasets. CONCLUSION: The signature-based cell cycle analysis method presented in this report, would potentially be of value for cancer characterization and diagnostics.

MHC class II DR classification based on antigen-binding groove natural selection.

Major histocompatibility complex (MHC) genes are highly polymorphic and play key roles in immune susceptibility and resistance to pathogens. While the immunological and structural functions of several human and murine alleles have been analyzed, little is known about the MHC molecules of other animals. Here, we could classify five mammalian species into three groups (human, cow and dog, and cat and pig) on the basis of DRB nucleotide sequences, synonymous and nonsynonymous mutation rates, and natural selection of individual residues. These observations, along with the locations of the positively and negatively selected residues in three-dimensional DR structures, suggest that the antigen-recognition sites of swine and feline DR molecules have been negatively selected while those of bovine and canine DR molecules have been positively selected. Human DR molecules show evidence of high negative and positive selection. Our observations suggest that MHC-DR molecules are under different selective force depending on each species.

Prediction of mono- and di-nucleotide-specific DNA-binding sites in proteins using neural networks.

BACKGROUND: DNA recognition by proteins is one of the most important processes in living systems. Therefore, understanding the recognition process in general, and identifying mutual recognition sites in proteins and DNA in particular, carries great significance. The sequence and structural dependence of DNA-binding sites in proteins has led to the development of successful machine learning methods for their prediction. However, all existing machine learning methods predict DNA-binding sites, irrespective of their target sequence and hence, none of them is helpful in identifying specific protein-DNA contacts. In this work, we formulate the problem of predicting specific DNA-binding sites in terms of contacts between the residue environments of proteins and the identity of a mononucleotide or a dinucleotide step in DNA. The aim of this work is to take a protein sequence or structural features as inputs and predict for each amino acid residue if it binds to DNA at locations identified by one of the four possible mononucleotides or one of the 10 unique dinucleotide steps. Contact predictions are made at various levels of resolution viz. in terms of side chain, backbone and major or minor groove atoms of DNA. RESULTS: Significant differences in residue preferences for specific contacts are observed, which combined with other features, lead to promising levels of prediction. In general, PSSM-based predictions, supported by secondary structure and solvent accessibility, achieve a good predictability of approximately 70-80%, measured by the area under the curve (AUC) of ROC graphs. The major and minor groove contact predictions stood out in terms of their poor predictability from sequences or PSSM, which was very strongly (>20 percentage points) compensated by the addition of secondary structure and solvent accessibility information, revealing a predominant role of local protein structure in the major/minor groove DNA-recognition. Following a detailed analysis of results, a web server to predict mononucleotide and dinucleotide-step contacts using PSSM was developed and made available at http://sdcpred.netasa.org/ or http://tardis.nibio.go.jp/netasa/sdcpred/. CONCLUSION: Most residue-nucleotide contacts can be predicted with high accuracy using only sequence and evolutionary information. Major and minor groove contacts, however, depend profoundly on the local structure. Overall, this study takes us a step closer to the ultimate goal of predicting mutual recognition sites in protein and DNA sequences.

Evaluation of DNA intramolecular interactions for nucleosome positioning in yeast.

We calculated intramolecular interaction energies of DNA by threading DNA sequences around crystal structures of nucleosomes. The strength of the intramolecular energy oscillations at frequency approximately 10 bps for dinucleotides was in agreement with previous nucleosome models. The intramolecular energy calculated along yeast genome positively correlated with nucleosome positioning experimentally measured.

Sequence-dependent DNA deformability studied using molecular dynamics simulations.

Proteins recognize specific DNA sequences not only through direct contact between amino acids and bases, but also indirectly based on the sequence-dependent conformation and deformability of the DNA (indirect readout). We used molecular dynamics simulations to analyze the sequence-dependent DNA conformations of all 136 possible tetrameric sequences sandwiched between CGCG sequences. The deformability of dimeric steps obtained by the simulations is consistent with that by the crystal structures. The simulation results further showed that the conformation and deformability of the tetramers can highly depend on the flanking base pairs. The conformations of xATx tetramers show the most rigidity and are not affected by the flanking base pairs and the xYRx show by contrast the greatest flexibility and change their conformations depending on the base pairs at both ends, suggesting tetramers with the same central dimer can show different deformabilities. These results suggest that analysis of dimeric steps alone may overlook some conformational features of DNA and provide insight into the mechanism of indirect readout during protein-DNA recognition. Moreover, the sequence dependence of DNA conformation and deformability may be used to estimate the contribution of indirect readout to the specificity of protein-DNA recognition as well as nucleosome positioning and large-scale behavior of nucleic acids.

uORFs, reinitiation and alternative translation start sites in human mRNAs.

It is known that eukaryotic ribosomes are able to translate small ORFs and reinitiate translation at downstream start codons. However, this mechanism is widely considered to be inefficient and it is not commonly taken into account. We compiled a sample of human mRNAs containing small upstream ORFs overlapping with annotated protein coding sequences. Statistical analysis supported the hypothesis on reinitiation of translation at downstream AUG codons and functional significance of potential alternative ORFs. It may be assumed that some 5'UTR-located upstream ORFs can deliver ribosomes to alternative translation starts, and they should be taken into consideration in the prediction of human mRNA coding potential.

DeltaG-based prediction and experimental confirmation of SYCRP1-binding sites on the Synechocystis genome.

DNA-binding sites for SYCRP1, which is a regulatory protein of the cyanobacterium Synechocystissp. PCC6803, were predicted for the whole genome sequence by estimating changes in the binding free energy () for SYCRP1 for those sites. The values were calculated by summing DeltaDeltaG values derived from systematic single base-pair substitution experiments (symmetrical and cooperative binding model). Of the calculated binding sites, 23 sites with a value <3.9kcal.mol(-1) located upstream or between the ORFs were selected as putative binding sites for SYCRP1. In order to confirm whether SYCRP1 actually binds to these binding sites or not, 11 sites with the lowest values were tested experimentally, and we confirmed that SYCRP1 binds to ten of the 11 sites with a DeltaDeltaG(total) value <3.9kcal.mol(-1). The best correlation coefficient between and the observed DeltaDeltaG(total) for binding of SYCRP1 to those sites was 0.78. These results suggest that the DeltaDeltaG values derived from systematic single base-pair experiments may be used to screen for potential binding sites of a regulatory protein in the genome sequence.

ReadOut: structure-based calculation of direct and indirect readout energies and specificities for protein-DNA recognition.

Protein-DNA interactions play a central role in regulatory processes at the genetic level. DNA-binding proteins recognize their targets by direct base-amino acid interactions and indirect conformational energy contribution from DNA deformations and elasticity. Knowledge-based approach based on the statistical analysis of protein-DNA complex structures has been successfully used to calculate interaction energies and specificities of direct and indirect readouts in protein-DNA recognition. Here, we have implemented the method as a webserver, which calculates direct and indirect readout energies and Z-scores, as a measure of specificity, using atomic coordinates of protein-DNA complexes. This server is freely available at http://gibk26.bse.kyutech.ac.jp/jouhou/readout/. The only input to this webserver is the Protein Data Bank (PDB) style coordinate data of atoms or the PDB code itself. The server returns total energy Z-scores, which estimate the degree of sequence specificity of the protein-DNA complex. This webserver is expected to be useful for estimating interaction energy and DNA conformation energy, and relative contributions to the specificity from direct and indirect readout. It may also be useful for checking the quality of protein-DNA complex structures, and for engineering proteins and target DNAs.

ProTherm and ProNIT: thermodynamic databases for proteins and protein-nucleic acid interactions.

ProTherm and ProNIT are two thermodynamic databases that contain experimentally determined thermodynamic parameters of protein stability and protein-nucleic acid interactions, respectively. The current versions of both the databases have considerably increased the total number of entries and enhanced search interface with added new fields, improved search, display and sorting options. As on September 2005, ProTherm release 5.0 contains 17,113 entries from 771 proteins, retrieved from 1497 scientific articles (approximately 20% increase in data from the previous version). ProNIT release 2.0 contains 4900 entries from 273 research articles, representing 158 proteins. Both databases can be queried using WWW interfaces. Both quick search and advanced search are provided on this web page to facilitate easy retrieval and display of the data from these databases. ProTherm is freely available online at http://gibk26.bse.kyutech.ac.jp/jouhou/Protherm/protherm.html and ProNIT at http://gibk26.bse.kyutech.ac.jp/jouhou/pronit/pronit.html.

Classification of protein-DNA complexes based on structural descriptors.

We attempt to classify protein-DNA complexes by using a set of 11 descriptors, mainly characterizing protein-DNA interactions, including the number of atomic contacts at major and minor grooves, conformational deviations from standard B- and A-DNA forms, widths of DNA grooves, GC content, specificity measures of direct and indirect readouts, and buried surface area at the complex interface. The cluster analyses were carried out for a unique set of 62 complexes including a variety of protein motifs, and 7 distinct clusters were revealed from the analyses. We found that some proteins with the same motif are classified into different clusters, whereas different proteins with distinct motifs are classified into the same cluster. These results suggest that the conventional motif-based classification of DNA binding proteins may not necessarily correspond to structural and functional properties of protein-DNA complexes, and that the present classification will help to identify common properties and rules that govern protein-DNA recognition.

AUG_hairpin: prediction of a downstream secondary structure influencing the recognition of a translation start site.

BACKGROUND: The translation start site plays an important role in the control of translation efficiency of eukaryotic mRNAs. The recognition of the start AUG codon by eukaryotic ribosomes is considered to depend on its nucleotide context. However, the fraction of eukaryotic mRNAs with the start codon in a suboptimal context is relatively large. It may be expected that mRNA should possess some features providing efficient translation, including the proper recognition of a translation start site. It has been experimentally shown that a downstream hairpin located in certain positions with respect to start codon can compensate in part for the suboptimal AUG context and also increases translation from non-AUG initiation codons. Prediction of such a compensatory hairpin may be useful in the evaluation of eukaryotic mRNA translation properties. RESULTS: We evaluated interdependency between the start codon context and mRNA secondary structure at the CDS beginning: it was found that a suboptimal start codon context significantly correlated with higher base pairing probabilities at positions 13 - 17 of CDS of human and mouse mRNAs. It is likely that the downstream hairpins are used to enhance translation of some mammalian mRNAs in vivo. Thus, we have developed a tool, AUG_hairpin, to predict local stem-loop structures located within the defined region at the beginning of mRNA coding part. The implemented algorithm is based on the available published experimental data on the CDS-located stem-loop structures influencing the recognition of upstream start codons. CONCLUSION: An occurrence of a potential secondary structure downstream of start AUG codon in a suboptimal context (or downstream of a potential non-AUG start codon) may provide researchers with a testable assumption on the presence of additional regulatory signal influencing mRNA translation initiation rate and the start codon choice. AUG_hairpin, which has a convenient Web-interface with adjustable parameters, will make such an evaluation easy and efficient.