Partner-specific prediction of RNA-binding residues in proteins: A critical assessment

RNA-protein interactions play essential roles in regulating gene expression. While some RNA-protein interactions are “specific”, that is, the RNA-binding proteins preferentially bind to particular RNA sequence or structural motifs, others are “non-RNA specific.” Deciphering the protein-RNA recognition code is essential for comprehending the functional implications of these interactions and for developing new therapies for many diseases. Because of the high cost of experimental determination of protein-RNA interfaces, there is a need for computational methods to identify RNA-binding residues in proteins. While most of the existing computational methods for predicting RNA-binding residues in RNA-binding proteins are oblivious to the characteristics of the partner RNA, there is growing interest in methods for partner-specific prediction of RNA binding sites in proteins. In this work, we assess the performance of two recently published partner-specific protein-RNA interface prediction tools, PS-PRIP, and PRIdictor, along with our own new tools. Specifically, we introduce a novel metric, RNA-specificity metric (RSM), for quantifying the RNA-specificity of the RNA binding residues predicted by such tools. Our results show that the RNA-binding residues predicted by previously published methods are oblivious to the characteristics of the putative RNA binding partner. Moreover, when evaluated using partner-agnostic metrics, RNA partner-specific methods are outperformed by the state-of-the-art partner-agnostic methods. We conjecture that either (a) the protein-RNA complexes in PDB are not representative of the protein-RNA interactions in nature, or (b) the current methods for partner-specific prediction of RNA-binding residues in proteins fail to account for the differences in RNA partner-specific versus partner-agnostic protein-RNA interactions, or both.     

Classifying RNA-binding proteins based on electrostatic properties

Protein structure can provide new insight into the biological function of a protein and can enable the design of better experiments to learn its biological roles. Moreover, deciphering the interactions of a protein with other molecules can contribute to the understanding of the protein's function within cellular processes. In this study, we apply a machine learning approach for classifying RNA-binding proteins based on their three-dimensional structures. The method is based on characterizing unique properties of electrostatic patches on the protein surface. Using an ensemble of general protein features and specific properties extracted from the electrostatic patches, we have trained a support vector machine (SVM) to distinguish RNA-binding proteins from other positively charged proteins that do not bind nucleic acids. Specifically, the method was applied on proteins possessing the RNA recognition motif (RRM) and successfully classified RNA-binding proteins from RRM domains involved in protein-protein interactions. Overall the method achieves 88% accuracy in classifying RNA-binding proteins, yet it cannot distinguish RNA from DNA binding proteins. Nevertheless, by applying a multiclass SVM approach we were able to classify the RNA-binding proteins based on their RNA targets, specifically, whether they bind a ribosomal RNA (rRNA), a transfer RNA (tRNA), or messenger RNA (mRNA). Finally, we present here an innovative approach that does not rely on sequence or structural homology and could be applied to identify novel RNA-binding proteins with unique folds and/or binding motifs.     

动植物中RNA结合蛋白的研究进展

RNA结合蛋白(RNA-binding proteins)在转录后基因表达调节中起着重要的作用,它通过和RNA相互作用来调节细胞的功能。RNA结合蛋白参与RNA剪接、多聚腺苷化作用、序列编辑、RNA转运、维持RNA的稳定和降解、细胞内定位和翻译控制等RNA代谢的各个方面。主要介绍了RNA结合蛋白的结构、靶标RNA及RNA结合蛋白在动植物和疾病中的研究。     

Ribonomic approaches to study the RNA-binding proteome

Gene expression is controlled through a complex interplay among mRNAs, non-coding RNAs and RNA-binding proteins (RBPs), which all assemble along with other RNA-associated factors in dynamic and functional ribonucleoprotein complexes (RNPs). To date, our understanding of RBPs is largely limited to proteins with known or predicted RNA-binding domains. However, various methods have been recently developed to capture an RNA of interest and comprehensively identify its associated RBPs. In this review, we discuss the RNA-affinity purification methods followed by mass spectrometry analysis (AP-MS); RBP screening within protein libraries and computational methods that can be used to study the RNA-binding proteome (RBPome).     

Predicting RNA-binding sites from the protein structure based on electrostatics, evolution and geometry

An RNA-binding protein places a surface helix, β-ribbon, or loop in an RNA helix groove and/or uses a cavity to accommodate unstacked bases. Hence, our strategy for predicting RNA-binding residues is based on detecting a surface patch and a disparate cleft. These were generated and scored according to the gas-phase electrostatic energy change upon mutating each residue to Asp⁻/Glu⁻ and each residue's relative conservation. The method requires as input the protein structure and sufficient homologous sequences to define each residue's relative conservation. It yields as output a priority list of surface patch residues followed by a backup list of surface cleft residues distant from the patch residues for experimental testing of RNA binding. Among the 69 structurally non-homologous proteins tested, 81% possess a RNA-binding site with at least 70% of the maximum number of true positives in randomly generated patches of the same size as the predicted site; only two proteins did not contain any true RNA-binding residues in both predicted regions. Regardless of the protein conformational changes upon RNA-binding, the prediction accuracies based on the RNA-free/bound protein structures were found to be comparable and their binding sites overlapped as long as there are no disordered RNA-binding regions in the free structure that are ordered in the corresponding RNA-bound protein structure.     

Identifying RNA-binding residues based on evolutionary conserved structural and energetic features

Increasing numbers of protein structures are solved each year, but many of these structures belong to proteins whose sequences are homologous to sequences in the Protein Data Bank. Nevertheless, the structures of homologous proteins belonging to the same family contain useful information because functionally important residues are expected to preserve physico-chemical, structural and energetic features. This information forms the basis of our method, which detects RNA-binding residues of a given RNA-binding protein as those residues that preserve physico-chemical, structural and energetic features in its homologs. Tests on 81 RNA-bound and 35 RNA-free protein structures showed that our method yields a higher fraction of true RNA-binding residues (higher precision) than two structure-based and two sequence-based machine-learning methods. Because the method requires no training data set and has no parameters, its precision does not degrade when applied to ‘novel’ protein sequences unlike methods that are parameterized for a given training data set. It was used to predict the ‘unknown’ RNA-binding residues in the C-terminal RNA-binding domain of human CPEB3. The two predicted residues, F430 and F474, were experimentally verified to bind RNA, in particular F430, whose mutation to alanine or asparagine nearly abolished RNA binding. The method has been implemented in a webserver called DR_bind1, which is freely available with no login requirement at http://drbind.limlab.ibms.sinica.edu.tw.     

Structure, backbone dynamics and interactions with RNA of the C-terminal RNA-binding domain of a mouse neural RNA-binding protein, Musashi1

Musashi1 is an RNA-binding protein abundantly expressed in the developing mouse central nervous system. Its restricted expression in neural precursor cells suggests that it is involved in the regulation of asymmetric cell division. Musashi1 contains two ribonucleoprotein (RNP)-type RNA-binding domains (RBDs), RBD1 and RBD2. Our previous studies showed that RBD1 alone binds to RNA, while the binding of RBD2 is not detected under the same conditions. Joining of RBD2 to RBD1, however, increases the affinity to greater than that of RBD1 alone, indicating that RBD2 contributes to RNA-binding. We have determined the three-dimensional solution structure of the C-terminal RBD (RBD2) of Musashi1 by NMR. It folds into a compact alpha beta structure comprising a four-stranded antiparallel beta-sheet packed against two alpha-helices, which is characteristic of RNP-type RBDs. Special structural features of RBD2 include a beta-bulge in beta2 and a shallow twist of the beta-sheet. The smaller 1H-15N nuclear Overhauser enhancement values for the residues of loop 3 between beta2 and beta3 suggest that this loop is flexible in the time-scale of nano- to picosecond order. The smaller 15N T2 values for the residues around the border between alpha2 and the following loop (loop 5) suggest this region undergoes conformational exchange in the milli- to microsecond time-scale. Chemical shift perturbation analysis indicated that RBD2 binds to an RNA oligomer obtained by in vitro selection under the conditions for NMR measurements, and thus the nature of the weak RNA-binding of RBD2 was successfully characterized by NMR, which is otherwise difficult to assess. Mainly the residues of the surface composed of the four-stranded beta-sheet, loops and C-terminal region are involved in the interaction. The appearance of side-chain NH proton resonances of arginine residues of loop 3 and imino proton resonances of RNA bases upon complex formation suggests the formation of intermolecular hydrogen bonds. The structural arrangement of the rings of the conserved aromatic residues of beta2 and beta3 is suitable for stacking interaction with RNA bases, known to be one of the major protein-RNA interactions, but a survey of the perturbation data suggested that the stacking interaction is not ideally achieved in the complex, which may be related to the weaker RNA-binding of RBD2.     

RNA-binding activities of the different domains of a spinach chloroplast ribonucleoprotein.

An RNA-binding protein of 28 kD (28RNP) has been previously isolated from spinach chloroplasts and was found to be required for 3' end processing of chloroplast mRNAs. The amino acid sequence of 28RNP revealed two approximately 80 amino-acid RNA-binding domains, as well as an acidic and glycine-rich amino terminal domain. Each domain by itself, as well as in combination with other domains, was expressed in bacterial cells and the polypeptides were purified to homogeneity. We have investigated the RNA-binding properties of the different structural domains using UV-crosslinking, saturation binding and competition between the different domains on RNA-binding. It was found that the acidic domain does not bind RNA, but that each of the RNA-binding domains, expressed either individually or together, do bind RNA, although with differing affinities. When either the first or second RNA-binding domain was coupled to the acidic domain, the affinity for RNA was greatly reduced. However, the acidic domain has a positive effect on the binding of the full-length protein to RNA, because the mature protein binds RNA with a better affinity than the truncated protein which lacks the acidic domain. In addition, it was found that a stretch of two or three G residues is enough to mediate binding of the 28RNP, whereas four U residues were insufficient. The implications of the RNA-binding properties of 28RNP to its possible function in the processing of chloroplast RNA is discussed.     

Structural behavior of RNA-binding proteins in free and complex states: ribosomal protein L25 and 5S RNA from Escherichia coli

A novel approach has been proposed to evaluate the steadiness of polar clusters containing RNA-binding sites on the protein surface. The degree of clusterization of RNA-binding polar residues can be a measure of the steadiness of corresponding polar clusters. Ribosomal protein L25 from E. coli forms a complex with a fragment of 5S rRNA by means of two binding sites S1 and S2. We have examined cluster distribution of RNA-contacting polar residues on the protein surface by using the data of two states: complex state (in crystal and solution) and free state (in solution). For the crystal, the extent of clusterization of binding sites S1 and S2 are estimated to be 74.1 and 100%, respectively. For the free state in solution, the degrees of clusterization of these two sites are 22.8 and 68.6%, respectively. Thus, we have obtained a steadiness quantitative measure of two different types of protein sites for binding to RNA: one for the already existing protein binding site, and the other for the RNA-induced protein binding site. It was shown that definite variations of the protein structure in crystal and in solution can be of significant functional meaning. The result could be applied to the structural behavior of numerous protein complexes with double-stranded RNA and DNA.     

A knowledge-based potential function predicts the specificity and relative binding energy of RNA-binding proteins

RNA-protein interactions are fundamental to gene expression. Thus, the molecular basis for the sequence dependence of protein-RNA recognition has been extensively studied experimentally. However, there have been very few computational studies of this problem, and no sustained attempt has been made towards using computational methods to predict or alter the sequence-specificity of these proteins. In the present study, we provide a distance-dependent statistical potential function derived from our previous work on protein-DNA interactions. This potential function discriminates native structures from decoys, successfully predicts the native sequences recognized by sequence-specific RNA-binding proteins, and recapitulates experimentally determined relative changes in binding energy due to mutations of individual amino acids at protein-RNA interfaces. Thus, this work demonstrates that statistical models allow the quantitative analysis of protein-RNA recognition based on their structure and can be applied to modeling protein-RNA interfaces for prediction and design purposes.     

Structural behavior of RNA-binding proteins in the free state and in complex with RNA: <Emphasis Type="Italic">Escherichia coli</Emphasis> ribosomal protein L25 and 5S rRNA

A novel approach was proposed to evaluate the steadiness of polar clusters containing the RNA-binding sites on the protein surface. The degree of clustering of RNA-binding polar residues was used as a measure of the steadiness of the corresponding polar clusters. Escherichia coli ribosomal protein L25 utilizes two binding sites, S1 and S2, to complexate with a 5S rRNA fragment. The cluster distribution of RNA-contacting polar residues on the protein surface was studied using the structural data on the complex (in crystal and in solution) and the free state (in solution). The degree of polar residue clustering in S1 and S2 in crystal was estimated at 71.4 and 100%, respectively. For the free state in solution, the degree of clustering of the two sites was 22.8 and 68.6%, respectively. Thus, the steadiness was quantitatively estimated for the RNA-binding sites of two different types, one preexisting in the protein and the other induced by the RNA structure upon complexation. The difference between the protein structures in crystal and in solution was found to be functionally significant. The results can be extrapolated to numerous complexes of proteins with double-stranded RNA and DNA.     

Exploring the role of cation-pi interactions in glycoproteins lipid-binding proteins and RNA-binding proteins

We have analyzed and compared the influence of cation-pi interactions in glycoproteins (GPs), lipid-binding proteins (LBPs) and RNA-binding proteins (RBPs) in this study. We observed that all the proteins included in the study had profound cation-pi interactions. There is an average of one energetically significant cation-pi interaction for every 71 residues in GPs, for every 58 residues in LBPs and for every 64 residues in RBPs. Long-range contacts are predominant in all the three types of proteins studied. The pair-wise cation-pi interaction energy between the positively charged and aromatic residues shows that Arg-Trp pair energy was the strongest among all six possible pairs in all the three types of proteins studied. There were considerable differences in the preference of cation-pi interacting residues to different secondary structure elements and ASA and these might contribute to differences in biochemical functions of GPs, LBPs and RBPs. It was interesting to note that all the five residues involved in cation-pi interactions were found to have stabilization centers in GPs, LBPs and RBPs. Majority of the cation-pi interacting residues investigated in the present study had a conservation score of 6, the cutoff value used to identify the stabilizing residues. A small percentage of cation-pi interacting residues were also present as stabilizing residues. The cation-pi interaction-forming residues play an important role in the structural stability of in GPs, LBPs and RBPs. The results obtained in this study will be helpful in further understanding the stability, specificity and differences in the biochemical functions of GPs, LBPs and RBPs.     

Structural properties of carnation mottle virus p7 movement protein and its RNA-binding domain

Plant viral movement proteins (MPs) participate actively in the intra- and intercellular movement of RNA plant viruses to such an extent that MP dysfunction impairs viral infection. However, the molecular mechanism(s) of their interaction with cognate nucleic acids are not well understood, partly due to the lack of structural information. In this work, a protein dissection approach was used to gain information on the structural and RNA-binding properties of this class of proteins, as exemplified by the 61-amino acid residue p7 MP from carnation mottle virus (CarMV). Circular dichroism spectroscopy showed that CarMV p7 is an alpha/beta RNA-binding soluble protein. Using synthetic peptides derived from the p7 sequence, we have identified three distinct putative domains within the protein. EMSA showed that the central region, from residue 17 to 35 (represented by peptide p7(17-35)), is responsible for the RNA binding properties of CarMV p7. This binding peptide populates a nascent alpha-helix in water solution that is further stabilized in the presence of either secondary structure inducers, such as trifluoroethanol and monomeric SDS, or RNA (which also changes its conformation upon binding to the peptide). Thus, the RNA recognition appears to occur via an "adaptive binding" mechanism. Interestingly, the amino acid sequence and structural properties of the RNA-binding domain of p7 seem to be conserved among carmoviruses and some other RNA-binding proteins and peptides. The low conserved N terminus of p7 (peptide p7(1-16)) is unstructured in solution. In contrast, the highly conserved C terminus motif (peptide p7(40-61)) adopts a beta-sheet conformation in aqueous solution. Alanine scanning mutagenesis of the RNA-binding motif showed how selected positive charged amino acids are more relevant than others in the RNA binding process and how hydrophobic amino acid side chains would participate in the stabilization of the protein-RNA complex.     

Melanoma RBPome identification reveals PDIA6 as an unconventional RNA-binding protein involved in metastasis

RNA-binding proteins (RBPs) have been relatively overlooked in cancer research despite their contribution to virtually every cancer hallmark. Here, we use RNA interactome capture (RIC) to characterize the melanoma RBPome and uncover novel RBPs involved in melanoma progression. Comparison of RIC profiles of a non-tumoral versus a metastatic cell line revealed prevalent changes in RNA-binding capacities that were not associated with changes in RBP levels. Extensive functional validation of a selected group of 24 RBPs using five different in vitro assays unveiled unanticipated roles of RBPs in melanoma malignancy. As proof-of-principle we focused on PDIA6, an ER-lumen chaperone that displayed a novel RNA-binding activity. We show that PDIA6 is involved in metastatic progression, map its RNA-binding domain, and find that RNA binding is required for PDIA6 tumorigenic properties. These results exemplify how RIC technologies can be harnessed to uncover novel vulnerabilities of cancer cells.     

Solution structures of the double-stranded RNA-binding domains from RNA helicase A

RNA helicase A (RHA) is a highly conserved protein with multifaceted functions in the gene expression of cellular and viral mRNAs. RHA recognizes highly structured nucleotides and catalytically rearranges the various interactions between RNA, DNA, and protein molecules to provide a platform for the ribonucleoprotein complex. We present the first solution structures of the double-stranded RNA-binding domains (dsRBDs), dsRBD1 and dsRBD2, from mouse RHA. We discuss the binding mode of the dsRBDs of RHA, in comparison with the known dsRBD structures in their complexes. Our structural data provide important information for the elucidation of the molecular reassembly mediated by RHA.