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.     

Isolation of a subtilisin-like serine protease gene (MyxSubtSP) from spores of Myxobolus cerebralis, the causative agent of whirling disease

Proteases play important roles in parasite life cycles and host-parasite interactions. They are pathogenesis factors of many pathogenic organisms and are hence potential targets for chemotherapeutic treatment of disease. We identified a subtilisin-like serine protease gene, MyxSubtSP, expressed by Myxobolus cerebralis. After PCR with subtilisin-like serine protease primers, the gene was cloned, sequenced and aligned against the NCBI database. Its corresponding amino acid sequence included the putative conserved domains of Peptidase_S8, subtilase family and AprE, subtilisin-like serine proteases. Rapid amplification of 5' and 3' cDNA ends (RACE) was used to generate the full length (1385 bp) gene, with a 429 bp open reading frame. The gene encompasses coding regions for a catalytic triad formed by Asp-74, His-100 and Ser-110.     

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.     

Challenges in structural modeling of RNA-protein interactions

In the past few years, the number of RNA-binding proteins (RBP) and RNA-RBP interactions has increased significantly. Here, we review recent developments in the methodology for protein-RNA and protein–protein complex structure modeling with deep learning and co-evolution, as well as discuss the challenges and opportunities for building a reliable approach for protein-RNA complex structure modelling. Protein Data bank (PDB) and Cross-linking immunoprecipitation (CLIP) data could be combined together and used to infer 2D geometry of protein-RNA interactions by deep learning.     

Prediction of RNA-binding proteins from primary sequence by a support vector machine approach

Elucidation of the interaction of proteins with different molecules is of significance in the understanding of cellular processes. Computational methods have been developed for the prediction of protein-protein interactions. But insufficient attention has been paid to the prediction of protein-RNA interactions, which play central roles in regulating gene expression and certain RNA-mediated enzymatic processes. This work explored the use of a machine learning method, support vector machines (SVM), for the prediction of RNA-binding proteins directly from their primary sequence. Based on the knowledge of known RNA-binding and non-RNA-binding proteins, an SVM system was trained to recognize RNA-binding proteins. A total of 4011 RNA-binding and 9781 non-RNA-binding proteins was used to train and test the SVM classification system, and an independent set of 447 RNA-binding and 4881 non-RNA-binding proteins was used to evaluate the classification accuracy. Testing results using this independent evaluation set show a prediction accuracy of 94.1%, 79.3%, and 94.1% for rRNA-, mRNA-, and tRNA-binding proteins, and 98.7%, 96.5%, and 99.9% for non-rRNA-, non-mRNA-, and non-tRNA-binding proteins, respectively. The SVM classification system was further tested on a small class of snRNA-binding proteins with only 60 available sequences. The prediction accuracy is 40.0% and 99.9% for snRNA-binding and non-snRNA-binding proteins, indicating a need for a sufficient number of proteins to train SVM. The SVM classification systems trained in this work were added to our Web-based protein functional classification software SVMProt, at http://jing.cz3.nus.edu.sg/cgi-bin/svmprot.cgi. Our study suggests the potential of SVM as a useful tool for facilitating the prediction of protein-RNA interactions.     

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.     

Characterization of cDNA clones encoding mouse proteinase 3 (myeloblastine) and cathepsin G

 Serine proteases are the most abundant granule constituents of several major hematopoietic cell lineages. Due to their high abundance and their strict tissue specificity they have become important phenotypic cell markers used for studies of various aspects of hematopietic cell development. Using a polymerase chain reaction (PCR)-based strategy for the isolation of trypsin-related serine proteases, we were able to isolate cDNAs for two of the major neutrophil and monocyte serine proteases in the mouse, cathepsin G and mouse protease 3 (myeloblastin). The internal PCR fragments were used as probes to screen a mouse mast cell cDNA library and a cDNA library originating from a mouse monocytic cell line (WEHI-274.1). Full-length cDNAs for mouse cathepsin G and proteinase 3 were isolated and their complete sequences were determined. Northern blot analysis revealed expression of cathepsin G in immature cells of the monocyte macrophage lineage but also in the connective tissue mast cell line MTC. Proteinase 3 was expressed in several cell lines of myelo-monocytic origin and in one B-cell line, but not in any of the other cell lines tested. The isolation of cDNAs for mouse cathepsin G and mouse proteinase 3, together with the previous characterization of the gene for mouse N-elastase, and the entire or partial amino acid sequences for porcine azurocidine, equine N-elastase and proteinase 3, rat, dog, and rabbit cathepsin Gs in evolutionary relatively distantly related mammalian species, indicates that these four members of the serine protease family have been maintained for more than 100 million years of mammalian evolution. This latter finding indicates a strong evolutionary pressure to maintain specific immune functions associated with these neutrophil and monocyte proteases. All amino acid positions of major importance for the cleavage site selection have also been fully conserved between mouse and human proteinase 3 and a few minor changes have occurred between mouse and human cathepsin G. Received: 3 August 1996 / Revised: 24 February 1997     

Reduced protease activity in transformed rice cell suspension cultures expressing a proteinase inhibitor

In this study, we synthesized a synthetic serine proteinase inhibitor II gene (sPI-II) that harbored the chymotrypsin and trypsin inhibitor domains of the PI-II gene from Nicotiana alata. In an effort to reduce protease activity in a rice cell suspension culture, we first synthesized sPI-II using overlap PCR and then introduced the gene into a rice calli (Oryza sativa L. cv. Dongin) by particle bombardment-mediated transformation. The sPI-II gene was under the control of a rice alpha-amylase 3D promoter induced by sugar starvation. To verify the integration and expression of the sPI-II gene in the transformed rice cells, we employed genomic DNA PCR amplification and Northern blot analysis, respectively. The relative protease activity of the transformed cell suspension culture was reduced to approximately 23% when compared to the non-transformed culture. This indicates that a transformed suspension culture system expressing a proteinase inhibitor, may be a useful tool to protect against recombinant protein losses resulting from extracellular proteases.     

New insights into the interaction of ribosomal protein L1 with RNA

The RNA-binding ability of ribosomal protein L1 is of profound interest, since L1 has a dual function as a ribosomal structural protein that binds rRNA and as a translational repressor that binds its own mRNA. Here, we report the crystal structure at 2.6 A resolution of ribosomal protein L1 from the bacterium Thermus thermophilus in complex with a 38 nt fragment of L1 mRNA from Methanoccocus vannielii. The conformation of RNA-bound T.thermophilus L1 differs dramatically from that of the isolated protein. Analysis of four copies of the L1-mRNA complex in the crystal has shown that domain II of the protein does not contribute to mRNA-specific binding. A detailed comparison of the protein-RNA interactions in the L1-mRNA and L1-rRNA complexes identified amino acid residues of L1 crucial for recognition of its specific targets on the both RNAs. Incorporation of the structure of bacterial L1 into a model of the Escherichia coli ribosome revealed two additional contact regions for L1 on the 23S rRNA that were not identified in previous ribosome models.     

Co-labeling using in situ PCR: a review.

In situ amplification permits the histological localization of low-copy DNA and RNA targets. However, in many instances it would be useful to know the specific phenotype of the target-containing cell or to ascertain the distribution of a different nucleic acid sequence in the same tissue section. This review describes a methodology that allows co-in situ localization of two nucleic acid targets or a DNA/RNA sequence and a protein in paraffin-embedded, formalin-fixed tissue. The key variable for detection of low-copy RNA targets by RT in situ PCR is optimal protease digestion to permit cDNA target-specific incorporation of the reporter nucleotide. This is achieved via inactivation of nonspecific DNA synthesis by overnight DNase digestion. The key variable for immunohistochemical localization of proteins is to determine the effect of protease digestion on the antigen-based signal intensity. Background for DNA targets by in situ hybridization or, for targets present in 1-10 copies per cell, PCR ISH is dependent primarily on probe concentration and the stringency of the post-hybridization wash. Radioactive 3H-labeled nucleotides permit an excellent distinction with colorimetric signals for co-localization, although two distinct chromogens can in many instances allow successful localization of two different targets.     

Expression and processing of human rhinovirus type 14 polypeptide precursors in Escherichia coli maxicells

Human rhinovirus serotype-14 (HRV-14) cDNA, encompassing 87.9% of the coding region, was subcloned in an Escherichia coli expression vector, generating plasmid pKCC101. HRV-14 polypeptides encoded by pKCC101 were synthesized in E. coli maxicells. Pulse-chase experiments with pKCC110, a smaller derivative of pKCC101 containing the protease 3C coding region, have clearly demonstrated the proteolysis of a 55-kDa precursor to several polypeptides, including a doublet with the expected size of protease 3C (20 kDa). The proteolysis of the 55-kDa precursor polypeptide was prevented by ZnCl2, a known inhibitor of picornavirus 3C proteases. Results with a derivative of pKCC110 (pKCC115) which is partially deleted for the protease 3C sequence, support the idea that the doublet proteins are specified by the protease 3C coding region. Taken together, our investigations indicate that the precursor form of protease 3C must be responsible for its own cleavage.     

Balanced-PCR amplification allows unbiased identification of genomic copy changes in minute cell and tissue samples

Analysis of genomic DNA derived from cells and fresh or fixed tissues often requires whole genome amplification prior to microarray screening. Technical hurdles to this process are the introduction of amplification bias and/or the inhibitory effects of formalin fixation on DNA amplification. Here we demonstrate a balanced-PCR procedure that allows unbiased amplification of genomic DNA from fresh or modestly degraded paraffin-embedded DNA samples. Following digestion and ligation of a target and a control genome with distinct linkers, the two are mixed and amplified in a single PCR, thereby avoiding biases associated with PCR saturation and impurities. We demonstrate genome-wide retention of allelic differences following balanced-PCR amplification of DNA from breast cancer and normal human cells and genomic profiling by array-CGH (cDNA arrays, 100 kb resolution) and by real-time PCR (single gene resolution). Comparison of balanced-PCR with multiple displacement amplification (MDA) demonstrates equivalent performance between the two when intact genomic DNA is used. When DNA from paraffin-embedded samples is used, balanced PCR overcomes problems associated with modest DNA degradation and produces unbiased amplification whereas MDA does not. Balanced-PCR allows amplification and recovery of modestly degraded genomic DNA for subsequent retrospective analysis of human tumors with known outcomes.     

Isolation and molecular characterization of cathepsin L-like cysteine protease cDNAs from western flower thrips (Frankliniella occidentalis)

Cysteine proteases are predominant in thrips guts (TGs) and, therefore, a suitable target for selecting effective protease inhibitors against western flower thrips (Frankliniella occidentalis). We report the isolation of four full-length cysteine protease cDNA clones from thrips in a two-step PCR approach with degenerate oligonucleotides designed on conserved cathepsin L domains. At the deduced amino acid level, the clones possessed all functional and structural characteristics of cathepsin L, and showed high mutual identity and strong similarity with cathepsin L-like cysteine proteases from other insects and arthropods. Southern analysis indicated that a family of four closely related and 10-12 less-related genes encode the cathepsin L-like cysteine proteases in the thrips genome. Partial sequencing of genomic DNA demonstrated the presence of three introns in the coding DNA.     

Quantitative monitoring of the mRNA expression pattern of the TGF-beta-isoforms (beta 1, beta 2, beta 3) during transdifferentiation of hepatic stellate cells using a newly developed real-time SYBR Green PCR

Current methods to determine the mRNA of the TGF-beta-isoforms, beta 1, beta 2, and beta 3, are not sensitive enough to detect small alterations in the expression levels. Therefore, we established a SYBR Green I-based real-time quantitative PCR procedure with fragment-specific standards. The advantage of gene-specific quantification is the possibility to be abstain from the need to compare results with a house-keeping gene having a different sequence and PCR efficiency. Reproducibility of the results and analytical variances of the real-time PCR assays were tested. In transdifferentiating rat hepatic stellate cells (HSC) the TGF-beta 1-mRNA was found to be the predominant isoform expressed followed by TGF-beta 3 and low amounts of TGF-beta 2-mRNA. An alteration of the TGF-beta 1,-beta 2, and -beta 3 ratio during HSC transdifferentiation could not be detected. Furthermore, the GAPDH mRNA expression varied during HSC activation, and thus is not recommended as a standard in real-time PCR quantifications.     

Proteolytic specificity of caspases is required to signal the appearance of apoptotic morphology

The caspase family of cysteine proteases is essential for implementation of physiological cell death. Since a wide variety of cellular proteins is cleaved by caspases during apoptosis, it has been predicted that digestion of proteins crucial to maintaining the life of a cell is central to apoptosis. To assess the role of the proteolytic destruction during apoptosis, we introduced the non-specific protease proteinase K into intact cells. This introduction led to extensive digestion of cellular proteins, including physiological caspase-substrates. Caspase-3-like activity was induced rapidly, followed by morphological signs of apoptosis such as membrane blebbing and nuclear condensation. The caspase inhibitor Z-VAD-fmk inhibited the appearance of these morphological changes without reducing the extent of intracellular proteolysis by proteinase K. Loss of integrity of the cell membrane, however, was not blocked by Z-VAD-fmk. This system thus generated conditions of extensive destruction of caspase substrates by proteinase K in the absence of apoptotic morphology. Taken together, these experiments suggest that caspases implement cell death not by protein destruction but by proteolytic activation of specific downstream effector molecules.     

A simple method for the isolation of genomic DNA from mouse tail free of real-time PCR inhibitors

Although real-time PCR is a rapid, quantitative method for the analysis of gene and RNA levels, the presence of inhibitors in samples is an obstacle to its successful use. We have found that genomic DNA isolated from mouse tail tips using a standard proteinase K digestion method caused marked inhibition of real-time PCR. Inhibition was specific for mouse tail DNA since genomic DNA isolated from other tissue sources using the same methodology was readily amplified. We have therefore developed a nonproteinase K DNA isolation method involving the use of Chelex 100 resin. This method produces mouse tail DNA that is free of real-time PCR inhibitors.     

DNA and histone H1 interact with different domains of HMG 1 and 2 proteins.

High mobility group (HMG) proteins 1 and 2 from calf thymus have been digested under structuring conditions (0.35 M NaCl, pH 7.1) with two proteases of different specificities, trypsin and V8. The two proteases give a different but restricted pattern of peptides in a time course digestion study. However, when the interactions of the peptides with DNA are studied by blotting, a closely related peptide from HMG-1 and -2 does not show any apparent binding. This peptide, from the V8 protease digestion, has been isolated by DNA-cellulose chromatography and has the amino acid composition predicted for a fragment containing the two C-terminal domains of the protein, i.e., approximately residues 74-243 for HMG-1. The same peptide shows the only interaction detectable with labelled histone H1. A separate function for the different domains of HMG proteins 1 and 2 is proposed.     

Kinetics of the inhibition of neutrophil proteinases by recombinant elafin and pre-elafin (trappin-2) expressed in Pichia pastoris.

Elafin and its precursor, trappin-2 or pre-elafin, are specific endogenous inhibitors of human neutrophil elastase and proteinase 3 but not of cathepsin G. Both inhibitors belong, together with secretory leukocyte protease inhibitor, to the chelonianin family of canonical protease inhibitors of serine proteases. A cDNA coding either elafin or its precursor, trappin-2, was fused in frame with yeast alpha-factor cDNA and expressed in the Pichia pastoris yeast expression system. Full-length elafin or full-length trappin-2 were secreted into the culture medium with high yield, indicating correct processing of the fusion proteins by the yeast KEX2 signal peptidase. Both recombinant inhibitors were purified to homogeneity from concentrated culture medium by one-step cationic exchange chromatography and characterized by N-terminal amino acid sequencing, Western blot and kinetic studies. Both recombinant elafin and trappin-2 were found to be fast-acting inhibitors of pancreatic elastase, neutrophil elastase and proteinase 3 with k(ass) values of 2-4 x 10(6) m(-1).s(-1), while dissociation rate constants k(diss) were found to be in the 10(-4) s(-1) range, indicating low reversibility of the complexes. The equilibrium dissociation constant K(i) for the interaction of both recombinant inhibitors with their target enzymes was either directly measured for pancreatic elastase or calculated from k(ass) and k(diss) values for neutrophil elastase and proteinase 3. K(i) values were found to be in the 10(-10) molar range and virtually identical for both inhibitors. Based on the kinetic parameters determined here, it may be concluded that both recombinant elafin and trappin-2 may act as potent anti-inflammatory molecules and may be of therapeutic potential in the treatment of various inflammatory lung diseases.