Quantifying sequence and structural features of protein–RNA interactions

Increasing awareness of the importance of protein–RNA interactions has motivated many approaches to predict residue-level RNA binding sites in proteins based on sequence or structural characteristics. Sequence-based predictors are usually high in sensitivity but low in specificity; conversely structure-based predictors tend to have high specificity, but lower sensitivity. Here we quantified the contribution of both sequence- and structure-based features as indicators of RNA-binding propensity using a machine-learning approach. In order to capture structural information for proteins without a known structure, we used homology modeling to extract the relevant structural features. Several novel and modified features enhanced the accuracy of residue-level RNA-binding propensity beyond what has been reported previously, including by meta-prediction servers. These features include: hidden Markov model-based evolutionary conservation, surface deformations based on the Laplacian norm formalism, and relative solvent accessibility partitioned into backbone and side chain contributions. We constructed a web server called aaRNA that implements the proposed method and demonstrate its use in identifying putative RNA binding sites.     

Predicting protein–protein interactions from sequence using correlation coefficient and high-quality interaction dataset

Identifying protein–protein interactions (PPIs) is critical for understanding the cellular function of the proteins and the machinery of a proteome. Data of PPIs derived from high-throughput technologies are often incomplete and noisy. Therefore, it is important to develop computational methods and high-quality interaction dataset for predicting PPIs. A sequence-based method is proposed by combining correlation coefficient (CC) transformation and support vector machine (SVM). CC transformation not only adequately considers the neighboring effect of protein sequence but describes the level of CC between two protein sequences. A gold standard positives (interacting) dataset MIPS Core and a gold standard negatives (non-interacting) dataset GO-NEG of yeast Saccharomyces cerevisiae were mined to objectively evaluate the above method and attenuate the bias. The SVM model combined with CC transformation yielded the best performance with a high accuracy of 87.94% using gold standard positives and gold standard negatives datasets. The source code of MATLAB and the datasets are available on request under smgsmg@mail.ustc.edu.cn.     

Alternative modulation of protein–protein interactions by small molecules

1. Download : Download high-res image (97KB)
2. Download : Download full-size image

     

Evolution of acyl-ACP thioesterases and β-ketoacyl-ACP synthases revealed by protein–protein interactions

The fatty acid synthase (FAS) is a conserved primary metabolic enzyme complex capable of tolerating cross-species engineering of domains for the development of modified and overproduced fatty acids. In eukaryotes, acyl-acyl carrier protein thioesterases (TEs) off-load mature cargo from the acyl carrier protein (ACP), and plants have developed TEs for short/medium-chain fatty acids. We showed that engineering plant TEs into the green microalga Chlamydomonas reinhardtii does not result in the predicted shift in fatty acid profile. Since fatty acid biosynthesis relies on substrate recognition and protein–protein interactions between the ACP and its partner enzymes, we hypothesized that plant TEs and algal ACP do not functionally interact. Phylogenetic analysis revealed major evolutionary differences between FAS enzymes, including TEs and ketoacyl synthases (KSs), in which the former is present only in some species, whereas the latter is present in all, and has a common ancestor. In line with these results, TEs appeared to be selective towards their ACP partners, whereas KSs showed promiscuous behavior across bacterial, plant, and algal species. Based on phylogenetic analyses, in silico docking, in vitro mechanistic cross-linking, and in vivo algal engineering, we propose that phylogeny can predict effective interactions between ACPs and partner enzymes.     

Computational design of novel protein–protein interactions – An overview on methodological approaches and applications

Protein–protein interactions (PPIs) govern numerous cellular functions in terms of signaling, transport, defense and many others. Designing novel PPIs poses a fundamental challenge to our understanding of molecular interactions. The capability to robustly engineer PPIs has immense potential for the development of novel synthetic biology tools and protein-based therapeutics. Over the last decades, many efforts in this area have relied purely on experimental approaches, but more recently, computational protein design has made important contributions. Template-based approaches utilize known PPIs and transplant the critical residues onto heterologous scaffolds. De novo design instead uses computational methods to generate novel binding motifs, allowing for a broader scope of the sites engaged in protein targets. Here, we review successful design cases, giving an overview of the methodological approaches used for templated and de novo PPI design.     

Weak protein–protein interactions revealed by immiscible filtration assisted by surface tension

Biological mechanisms are often mediated by transient interactions between multiple proteins. The isolation of intact protein complexes is essential to understanding biochemical processes and an important prerequisite for identifying new drug targets and biomarkers. However, low-affinity interactions are often difficult to detect. Here, we use a newly described method called immiscible filtration assisted by surface tension (IFAST) to isolate proteins under defined binding conditions. This method, which gives a near-instantaneous isolation, enables significantly higher recovery of transient complexes compared to current wash-based protocols, which require reequilibration at each of several wash steps, resulting in protein loss. The method moves proteins, or protein complexes, captured on a solid phase through one or more immiscible-phase barriers that efficiently exclude the passage of nonspecific material in a single operation. We use a previously described polyol-responsive monoclonal antibody to investigate the potential of this new method to study protein binding. In addition, difficult-to-isolate complexes involving the biologically and clinically important Wnt signaling pathway were isolated. We anticipate that this simple, rapid method to isolate intact, transient complexes will enable the discoveries of new signaling pathways, biomarkers, and drug targets.     

Identifying protein–protein interaction sites in transient complexes with temperature factor, sequence profile and accessible surface area

Transient protein–protein interactions play a vital role in many biological processes, such as cell regulation and signal transduction. A nonredundant dataset of 130 protein chains extracted from transient complexes was used to analyze the features of transient interfaces. It was found that besides the two well-known features, sequence profile and accessible surface area (ASA), the temperature factor (B-factor) can also reflect the differences between interface and the rest of protein surface. These features were utilized to construct support vector machine (SVM) classifiers to identify interaction sites. The results of threefold cross-validation on the nonredundant dataset show that when B-factor was used as an additional feature, the prediction performance can be improved significantly. The sensitivity, specificity and correlation coefficient were raised from 54 to 62%, 41 to 45% and 0.20 to 0.29, respectively. To further illustrate the effectiveness of our method, the classifiers were tested with an independent set of 53 nonhomologous protein chains derived from benchmark 2.0. The sensitivity, specificity and correlation coefficient of the classifier based on the three features were 63%, 45% and 0.33, respectively. It is indicated that our classifiers are robust and can be applied to complement experimental techniques in studying transient protein–protein interactions.     

Membrane phosphoinositides and protein–membrane interactions

Proteins with polybasic clusters bind to negatively charged phosphoinositides at the cell membrane. In this review, I have briefly discussed the types of phosphoinositides naturally found on membrane surfaces and how they recruit protein complexes for carrying out the process of signal transduction. A large number of researchers from around the world are now focusing their attention on protein–membrane binding, as these interactions have started to offer us a much better insight into the process of cell signaling. The main areas discussed in this brief review article include the phosphoinositide binding specificities of proteins and the role of their lipid binding in signaling processes downstream of membrane recruitment.     

In vivo imaging of protein–protein and RNA–protein interactions using novel far-red fluorescence complementation systems

Imaging of protein–protein and RNA–protein interactions in vivo, especially in live animals, is still challenging. Here we developed far-red mNeptune-based bimolecular fluorescence complementation (BiFC) and trimolecular fluorescence complementation (TriFC) systems with excitation and emission above 600 nm in the ‘tissue optical window’ for imaging of protein–protein and RNA–protein interactions in live cells and mice. The far-red mNeptune BiFC was first built by selecting appropriate split mNeptune fragments, and then the mNeptune-TriFC system was built based on the mNeptune-BiFC system. The newly constructed mNeptune BiFC and TriFC systems were verified as useful tools for imaging protein–protein and mRNA–protein interactions, respectively, in live cells and mice. We then used the new mNeptune-TriFC system to investigate the interactions between human polypyrimidine-tract-binding protein (PTB) and HIV-1 mRNA elements as PTB may participate in HIV mRNA processing in HIV activation from latency. An interaction between PTB and the 3′long terminal repeat region of HIV-1 mRNAs was found and imaged in live cells and mice, implying a role for PTB in regulating HIV-1 mRNA processing. The study provides new tools for in vivo imaging of RNA–protein and protein–protein interactions, and adds new insight into the mechanism of HIV-1 mRNA processing.     

Peptides and peptidomimetics as regulators of protein–protein interactions

1. Download : Download high-res image (114KB)
2. Download : Download full-size image

     

Transmembrane helix–helix interactions are modulated by the sequence context and by lipid bilayer properties

Folding of polytopic transmembrane proteins involves interactions of individual transmembrane helices, and multiple TM helix–helix interactions need to be controlled and aligned to result in the final TM protein structure. While defined interaction motifs, such as the GxxxG motif, might be critically involved in transmembrane helix–helix interactions, the sequence context as well as lipid bilayer properties significantly modulate the strength of a sequence specific transmembrane helix–helix interaction. Structures of 11 transmembrane helix dimers have been described today, and the influence of the sequence context as well as of the detergent and lipid environment on a sequence specific dimerization is discussed in light of the available structural information. This article is part of a Special Issue entitled: Protein Folding in Membranes.     

Mapping protein–protein interactions by double-REDOR-filtered magic angle spinning NMR spectroscopy

REDOR-based experiments with simultaneous ¹H–¹³C and ¹H?¹⁵N dipolar dephasing are explored for investigating intermolecular protein–protein interfaces in complexes formed by a U–¹³C,¹⁵N-labeled protein and its natural abundance binding partner. The application of a double-REDOR filter (dREDOR) results in a complete dephasing of proton magnetization in the U–¹³C,¹⁵N-enriched molecule while the proton magnetization of the unlabeled binding partner is not dephased. This retained proton magnetization is then transferred across the intermolecular interface by ¹H–¹³C or ¹H–¹⁵N cross polarization, permitting to establish the residues of the U–¹³C,¹⁵N-labeled protein, which constitute the binding interface. To assign the interface residues, this dREDOR-CPMAS element is incorporated as a building block into ¹³C–¹³C correlation experiments. We established the validity of this approach on U–¹³C,¹⁵N-histidine and on a structurally characterized complex of dynactin’s U–¹³C,¹⁵N-CAP-Gly domain with end-binding protein 1 (EB1). The approach introduced here is broadly applicable to the analysis of intermolecular interfaces when one of the binding partners in a complex cannot be isotopically labeled.