Predicting protein-protein binding sites in membrane proteins

Background  

Many integral membrane proteins, like their non-membrane counterparts, form either transient or permanent multi-subunit complexes in order to carry out their biochemical function. Computational methods that provide structural details of these interactions are needed since, despite their importance, relatively few structures of membrane protein complexes are available.     

Predicting human splicing branchpoints by combining sequence-derived features and multi-label learning methods

Background

Alternative splicing is the critical process in a single gene coding, which removes introns and joins exons, and splicing branchpoints are indicators for the alternative splicing. Wet experiments have identified a great number of human splicing branchpoints, but many branchpoints are still unknown. In order to guide wet experiments, we develop computational methods to predict human splicing branchpoints.

Results

Considering the fact that an intron may have multiple branchpoints, we transform the branchpoint prediction as the multi-label learning problem, and attempt to predict branchpoint sites from intron sequences. First, we investigate a variety of intron sequence-derived features, such as sparse profile, dinucleotide profile, position weight matrix profile, Markov motif profile and polypyrimidine tract profile. Second, we consider several multi-label learning methods: partial least squares regression, canonical correlation analysis and regularized canonical correlation analysis, and use them as the basic classification engines. Third, we propose two ensemble learning schemes which integrate different features and different classifiers to build ensemble learning systems for the branchpoint prediction. One is the genetic algorithm-based weighted average ensemble method; the other is the logistic regression-based ensemble method.

Conclusions

In the computational experiments, two ensemble learning methods outperform benchmark branchpoint prediction methods, and can produce high-accuracy results on the benchmark dataset.

     

RNA-containing nuclear binding sites for glucocorticoid-receptor complexes

RNase A treatment of HeLa cell nuclei causes a time- and concentration-dependent release of dexamethasone-receptor complexes. If nuclei are incubated in the absence of enzyme, only 60% of RNase-releasable complexes can be detected. Sucrose density gradient analysis of nuclear extracts shows that receptor complexes released by RNase treatment sediment at 3.6 S, whereas complexes obtained from untreated nuclei sediment between 7 and 3.6 S. Our results show that a fraction of dexamethasone-receptor complexes retained by HeLa cell nuclei is located in binding sites involving RNA.     

Determining binding sites in protein–nucleic acid complexes by cross-saturation

Cross-saturation experiments have been shown to give accurate information regarding the interacting surfaces in protein-protein and protein-RNA complexes. The rate of magnetization transfer depends on a number of factors including geometry, spin topology, and effective correlation times. To assess the influence of these variables on such experiments, and to determine the range of applicability of the technique, we have simulated the time-course of magnetization transfer across the interface in a variety of protein-nucleic acid complexes (434 Cro, SRY, MetJ and U1A), each having a different binding geometry. The simulations have been carried out primarily to provide information about the experimentally accessible targets for selective saturation, such as the anomeric protons and the imino protons of the nucleic acid. Saturation of either of these groups of signals leads to partial excitation throughout the nucleic acid molecule, and the resulting transfer of saturation to the labelled protein moiety can be readily detected by the reduction in intensity of particular peaks in the HSQC spectrum of the protein. The simulations show that information can be obtained about the residues in contact with the nucleic acid without any need for deuteration. Experimental cross-saturation data have been obtained from the Mbp1-DNA complex and interpreted in conjunction with the simulations to map out the binding surface in detail.     

Fluorescence of proflavine--DNA complexes: heterogeneity of binding sites

Measurements of the relative quantum yield of fluorescence of proflavine bound to DNA as a function of the number of bound dyes per nucleotide and the ionic strength allow the determination of the binding constants and respective number of the two types of sites previously postulated. It is demonstrated that 2–3% of the base pairs form sites where the dye is strongly bound and fluoresces normally while in the other set of sites the binding constant is 3–4 times weaker and the fluorescence completely quenched. Comparison with complexes of Pro with double stranded polynucleotides poly (A + U), poly (I + C), poly(G + C), confirm that the strong binding sites correspond to A-T-rich regions of the DNA while the quenched sites seem to require the presence of a neighboring guanine. The role of charge transfer in quenching of fluorescence and mutagnic action is considered. An original method for the determination of free dye and bound dye, based upon the use of an external quencher is described in the Appendix.     

Predicting protein-ligand binding sites based on an improved geometric algorithm

Knowledge of protein-ligand binding sites is very important for structure-based drug designs. To get information on the binding site of a targeted protein with its ligand in a timely way, many scientists tried to resort to computational methods. Although several methods have been released in the past few years, their accuracy needs to be improved. In this study, based on the combination of incremental convex hull, traditional geometric algorithm, and solvent accessible surface of proteins, we developed a novel approach for predicting the protein-ligand binding sites. Using PDBbind database as a benchmark dataset and comparing the new approach with the existing methods such as POCKET, Q-SiteFinder, MOE-SiteFinder, and PASS, we found that the new method has the highest accuracy for the Top 2 and Top 3 predictions. Furthermore, our approach can not only successfully predict the protein-ligand binding sites but also provide more detailed information for the interactions between proteins and ligands. It is anticipated that the new method may become a useful tool for drug development, or at least play a complementary role to the other existing methods in this area.     

Lectin binding sites of glycoproteins as revealed by lectin-gold-silver and lectin-peroxidase-diaminobenzidine methods

Summary For the precise histochemical detection of lectin binding sites of glycoproteins, the results obtained by lectin-gold-silver (LT-G-S) staining methods have been systematicaly compared with those revealed by alternative techniques of lectin-peroxidase-diaminobenzidine (LT-PO-DAB) reactions in a series of organs from different mammalian species.Ricinus communis agglutinin-I and concanavalin A were the lectins used in the present study. In the tissues subjected to the LT-G-S procedures, reactive tissue structures exhibited positive reactions of varying intensities of black. The results of control staining for the LT-G-S methods substantiated the view that the reaction products demonstrated the precise lectin binding sites of glycoproteins. The staining images obtained by the LT-PO-DAB techniques were not necessarily correlated precisely with those revealed by the LT-G-S procedures, and unavoidable background staining in pale brownish shades was noted in the majority of LT-G-S negative tissue structures. In view of these results, the LT-G-S staining methods employed in the present study are believed to be a reliable technique for the precise localization of saccharide residues of glycoproteins in light microscopy.     

Identification of functional PDZ domain binding sites in several human proteins

TIP-15 was previously identified as a cellular protein that can bind to the C-terminal end of the HTLV-1 Tax protein via its two PDZ domains. The sequence of the N-terminal part of TIP-15 is identical to that of the synaptic protein PSD-95. Both proteins are likely to be produced from the same gene by alternative splicing. Whereas expression of the PSD-95 mRNA was detected only with brain RNAs, that of TIP-15 was detected with RNAs from thymus, brain, skeletal muscle and Jurkat cells. The TIP-15 protein exhibits an apparent molecular weight of 40 kD and is weakly expressed in T cell lines. A two-hybrid screen performed with TIP-15 as bait revealed the presence of a PDZ binding site (PDZ-BS) in the following proteins: Lysyl tRNA synthetase, 6-phosphogluconolactonase (6-GPL), Stress-activated protein kinase 3 (SAPK3), NET-1, Diacylglycerol kinase zeta, MTMR1, MCM7, and hSec8. The sequence at the C-terminal ends of these proteins matches the X-S/T-X-V-COOH consensus previously defined for PDZ-BSs, with the exception of 6-GPL and SAPK3 which include a leucine as the C-terminal residue. For Lysyl tRNA synthetase, NET1, MTMR1 and hSec8, binding to TIP-15 was confirmed by co-immunoprecipitation experiments performed with the extracts of transfected COS7 cells. These results show the existence of functional PDZ-BSs in these proteins, but future studies will be necessary to establish whether or not TIP-15 represents a physiological partner. The significance of the presence of a PDZ-BS in these various proteins is discussed with respect to their function.     

Visualization of several binding sites for basic fibroblast growth factor (FGF-2) on fibroblasts by photoaffinity labeling: Evidence for intracellular complexes

The internalization of basic fibroblast growth factor (FGF-2) was studied in Chinese hamster lung fibroblasts (CCL39). Recombinant FGF-2 was derivatized with a photoactivable agent, N-hydroxysuccinimidyl-4-azido-benzoate (HSAB), iodinated, and used to visualize intracellular FGF-2-affinity-labeled molecules after internalization at 37°C. Iodinated HSAB-FGF-2 maintained the properties of natural FGF-2 such as affinity for heparin, binding to Bek and Flg receptors, interaction with high- and low-affinity binding sites, and reinitiating of DNA synthesis in CCL39 cells. Affinity-labeling experiments at 4°C with ¹²⁵I-HSAB-FGF-2 led to the detection of several FGF-cell surface complexes with apparent molecular mass of 80, 100, 125, 150, 170–180, 220, 260, and about 320 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), whereas two specific bands at 80 and 130–160 kDa were obtained using the homobifunctional cross-linking reagent, disuccinimidyl suberate. When the cells, preincubated with ¹²⁵I-HSAB-FGF-2 at 4°C and then washed, were shifted to 37°C, irradiation of the internalized labeled FGF-2 led to detection of a similar but fainted profile with one major specific band at 80 kDa. Heparitinase II treatment of the cells reduced binding of ¹²⁵I-HSAB-FGF-2 to its cell surface sites by 80% and internalization by 55%, indicating the involvement of heparan sulfate proteoglycans in these processes. Among the heparitinase-sensitive bands was the 80-kDa complex. © 1996 Wiley-Liss, Inc.     

Flexible interaction of Drosophila Smad complexes with bipartite binding sites

A subset of BMP-responsive enhancer elements are characterized by pairing of a GC-rich Smad1 binding site and an SBE-type Smad4 binding site. Such paired, or bipartite, sites are in some cases just 5 bp apart and thus might be contacted by a single Smad1-Smad4 complex. Other potential pairings are separated as much as 60 bp but it is not known whether such longer distances can be spanned by a Smad1-Smad4 complex, indeed binding of native Smad1-Smad4 complexes to any of these bipartite elements has yet to be reported. Here we report that a complex of the homologous Drosophila Smad proteins, Mad and Medea, is capable of concerted binding to GC-rich and SBE sites separated by as much as 20 bp. The wider the separation, the more severely binding affinity was reduced by shortening of the linker region that tethers the DNA binding domain of Medea. In contrast, length of the Mad linker did not affect the allowed distance between paired sites, rather it contributes specifically to Mad contact with the GC-rich site. Finally, we show that Smad1 and Smad4 can participate in binding to bipartite sites.     

Discovering amino acid patterns on binding sites in protein complexes

Discovering amino acid (AA) patterns on protein binding sites has recently become popular. We propose a method to discover the association relationship among AAs on binding sites. Such knowledge of binding sites is very helpful in predicting protein-protein interactions. In this paper, we focus on protein complexes which have protein-protein recognition. The association rule mining technique is used to discover geographically adjacent amino acids on a binding site of a protein complex. When mining, instead of treating all AAs of binding sites as a transaction, we geographically partition AAs of binding sites in a protein complex. AAs in a partition are treated as a transaction. For the partition process, AAs on a binding site are projected from three-dimensional to two-dimensional. And then, assisted with a circular grid, AAs on the binding site are placed into grid cells. A circular grid has ten rings: a central ring, the second ring with 6 sectors, the third ring with 12 sectors, and later rings are added to four sectors in order. As for the radius of each ring, we examined the complexes and found that 10Å is a suitable range, which can be set by the user. After placing these recognition complexes on the circular grid, we obtain mining records (i.e. transactions) from each sector. A sector is regarded as a record. Finally, we use the association rule to mine these records for frequent AA patterns. If the support of an AA pattern is larger than the predetermined minimum support (i.e. threshold), it is called a frequent pattern. With these discovered patterns, we offer the biologists a novel point of view, which will improve the prediction accuracy of protein-protein recognition. In our experiments, we produced the AA patterns by data mining. As a result, we found that arginine (arg) most frequently appears on the binding sites of two proteins in the recognition protein complexes, while cysteine (cys) appears the fewest. In addition, if we discriminate the shape of binding sites between concave and convex further, we discover that patterns {arg, glu, asp} and {arg, ser, asp} on the concave shape of binding sites in a protein more frequently (i.e. higher probability) make contact with {lys} or {arg} on the convex shape of binding sites in another protein. Thus, we can confidently achieve a rate of at least 78%. On the other hand {val, gly, lys} on the convex surface of binding sites in proteins is more frequently in contact with {asp} on the concave site of another protein, and the confidence achieved is over 81%. Applying data mining in biology can reveal more facts that may otherwise be ignored or not easily discovered by the naked eye. Furthermore, we can discover more relationships among AAs on binding sites by appropriately rotating these residues on binding sites from a three-dimension to two-dimension perspective. We designed a circular grid to deposit the data, which total to 463 records consisting of AAs. Then we used the association rules to mine these records for discovering relationships. The proposed method in this paper provides an insight into the characteristics of binding sites for recognition complexes.     

Predicting key long-range interaction sites by B-factors

In this paper, we adopted the bounded support vector machine to locate the key long-range interaction sites by the use of predicted local lowest B-factors. As a result, the key long-range interaction residues can be located based on information of local lowest B-factor sites.     

Predicting protein interaction sites: binding hot-spots in protein-protein and protein-ligand interfaces

MOTIVATION: Protein assemblies are currently poorly represented in structural databases and their structural elucidation is a key goal in biology. Here we analyse clefts in protein surfaces, likely to correspond to binding 'hot-spots', and rank them according to sequence conservation and simple measures of physical properties including hydrophobicity, desolvation, electrostatic and van der Waals potentials, to predict which are involved in binding in the native complex. RESULTS: The resulting differences between predicting binding-sites at protein-protein and protein-ligand interfaces are striking. There is a high level of prediction accuracy (< or =93%) for protein-ligand interactions, based on the following attributes: van der Waals potential, electrostatic potential, desolvation and surface conservation. Generally, the prediction accuracy for protein-protein interactions is lower, with the exception of enzymes. Our results show that the ease of cleft desolvation is strongly predictive of interfaces and strongly maintained across all classes of protein-binding interface.     

Structure of eukaryotic DNA binding sites for steroid hormone-apolipoprotein A-I complexes

High affinity for DNA and synthetic oligonucleotides was detected for apolipoprotein A-I (ApoA-I) by affinity chromatography, affinity modification, and enzymatic analysis. Competitive inhibition and Southern hybridization showed that the tetrahydrocortisol (THC)-ApoA-I complex specifically bound to high-molecular-weight DNA in regions containing GCC/CGG sequences. The CC(GCC)₃ · GG(CGG)₃ duplex was found to be sensitive to nuclease S1 under the action of the THC-ApoA-I complex. The eukaryotic DNA binding sites for steroid (THC, androsterone)-ApoA-I complexes were found to be involved in the initiation of DNA copying in vitro.     

Improving indicator species analysis by combining groups of sites

Indicator species are species that are used as ecological indicators of community or habitat types, environmental conditions, or environmental changes. In order to determine indicator species, the characteristic to be predicted is represented in the form of a classification of the sites, which is compared to the patterns of distribution of the species found at the sites. Indicator species analysis should take into account the fact that species have different niche breadths: if a species is related to the conditions prevailing in two or more groups of sites, an indicator species analysis undertaken on individual groups of sites may fail to reveal this association. In this paper, we suggest improving indicator species analysis by considering all possible combinations of groups of sites and selecting the combination for which the species can be best used as indicator. When using a correlation index, such as the point‐biserial correlation, the method yields the combination where the difference between the observed and expected abundance/frequency of the species is the largest. When an indicator value index (IndVal) is used, the method provides the set of site‐groups that best matches the observed distribution pattern of the species. We illustrate the advantages of the method in three different examples. Consideration of combinations of groups of sites provides an extra flexibility to qualitatively model the habitat preferences of the species of interest. The method also allows users to cross multiple classifications of the same sites, increasing the amount of information resulting from the analysis. When applied to community types, it allows one to distinguish those species that characterize individual types from those that characterize the relationships between them. This distinction is useful to determine the number of types that maximizes the number of indicator species.