PLMLA: prediction of lysine methylation and lysine acetylation by combining multiple features

Post-translational lysine methylation and acetylation are two major modifications of lysine residues. They play critical roles in various biological processes, especially in gene regulation. Identification of protein methylation and acetylation sites would be a foundation for understanding their modification dynamics and molecular mechanism. This work presents a method called PLMLA that incorporates protein sequence information, secondary structure and amino acid properties to predict methylation and acetylation of lysine residues in whole protein sequences. We apply an encoding scheme based on grouped weight and position weight amino acid composition to extract sequence information and physicochemical properties around lysine sites. The prediction accuracy for methyllysine and acetyllysine are 83.02% and 83.08%, respectively. Feature analysis reveals that methyllysine is likely to occur at the coil region and acetyllysine prefers to occur at the helix region of protein. The upstream residues away from the central site may be close to methylated lysine in three-dimensional structure and have a significant influence on methyllysine, while the positively charged residues may have a significant influence on acetyllysine. The online service is available at http://bioinfo.ncu.edu.cn/inquiries_PLMLA.aspx.     

综述: 非组蛋白甲基化修饰的研究进展

非组蛋白的赖氨酸和精氨酸残基上的甲基化修饰已经被证明是一种普遍的蛋白质翻译后修饰方式,在生命活动中发挥重要作用．甲基化修饰方式的多样性以及它们与其他修饰之间的交互作用(crosstalk)复杂但精细地调控了基因表达、蛋白质活性及稳定性、DNA复制及基因组稳定性、RNA加工等多种功能．本文将对非组蛋白的甲基化修饰特征进行总结,归纳近些年来已报道的甲基化修饰酶、修饰位点及这些位点的生物学功能,并将特别阐述不同蛋白质修饰之间的交互作用,概述鉴定非组蛋白甲基化修饰的方法．     

Prediction of ubiquitin proteins using artificial neural networks, hidden markov model and support vector machines

Ubiquitin functions to regulate protein turnover in a cell by closely regulating the degradation of specific proteins. Such a regulatory role is very important, and thus I have analyzed the proteins that are ubiquitin-like, using an artificial neural network, support vector machines and a hidden Markov model (HMM). The methods were trained and tested on a set of 373 ubiquitin proteins and 373 non-ubiquitin proteins, obtained from Entrez protein database. The artificial neural network and support vector machine are trained and tested using both the physicochemical properties and PSSM matrices generated from PSI-BLAST, while in the HMM based method direct sequences are used for training-testing procedures. Further, the performance measures of the methods are calculated for test sequences, i.e. accuracy, specificity, sensitivity and Matthew's correlation coefficients of the methods are calculated. The highest accuracy of 90.2%, specificity of 87.04% and sensitivity of 94.08% was achieved using the support vector machine model with PSSM matrices. While accuracies of 86.82%, 83.37%, 80.18% and 72.11% were obtained for the support vector machine with physicochemical properties, neural network with PSSM matrices, neural networks with physicochemical properties, and hidden Markov model, respectively. As the accuracy for SVM model is better both using physicochemical properties and the PSSM matrices, it is concluded that kernel methods such as SVM outperforms neural networks and hidden Markov models.     

SUMOhydro: a novel method for the prediction of sumoylation sites based on hydrophobic properties

Sumoylation is one of the most essential mechanisms of reversible protein post-translational modifications and is a crucial biochemical process in the regulation of a variety of important biological functions. Sumoylation is also closely involved in various human diseases. The accurate computational identification of sumoylation sites in protein sequences aids in experimental design and mechanistic research in cellular biology. In this study, we introduced amino acid hydrophobicity as a parameter into a traditional binary encoding scheme and developed a novel sumoylation site prediction tool termed SUMOhydro. With the assistance of a support vector machine, the proposed method was trained and tested using a stringent non-redundant sumoylation dataset. In a leave-one-out cross-validation, the proposed method yielded an excellent performance with a correlation coefficient, specificity, sensitivity and accuracy equal to 0.690, 98.6%, 71.1% and 97.5%, respectively. In addition, SUMOhydro has been benchmarked against previously described predictors based on an independent dataset, thereby suggesting that the introduction of hydrophobicity as an additional parameter could assist in the prediction of sumoylation sites. Currently, SUMOhydro is freely accessible at http://protein.cau.edu.cn/others/SUMOhydro/.     

Evaluation of single CpG sites as proxies of CpG island methylation states at the genome scale

Methylation of a CpG island is a faithful marker of silencing of its associated gene. Different approaches report the methylation status of a CpG island based on the determination of one or a few CpG sites by assuming the homogeneity of methylation along the element. This strategy is frequently applied in both locus-specific and genome-wide studies, but often without a validation of the representativeness of the interrogated CpG site compared with the whole element. We have evaluated the predictive informativeness of the HpaII sites located in CpG islands using data from high-resolution methylome maps, which offer the possibility to assess the methylation homogeneity of each CpG island and to determine the reporter accuracy of single sites as surrogate markers. An excellent correlation was observed between the HpaII and CpG island methylation levels (r > 0.93). At the qualitative level, the predictive sensitivity of HpaII was >95% with >92% specificity for methylated CpG islands and >90% sensitivity with >95% specificity for unmethylated CpG islands. This analysis provides a global validation framework for strategies based on the use of the methylation-sensitive HpaII restriction enzyme.     

Prediction of Protein Phosphorylation Sites by Using the Composition of k-Spaced Amino Acid Pairs

As one of the most widespread protein post-translational modifications, phosphorylation is involved in many biological processes such as cell cycle, apoptosis. Identification of phosphorylated substrates and their corresponding sites will facilitate the understanding of the molecular mechanism of phosphorylation. Comparing with the labor-intensive and time-consuming experiment approaches, computational prediction of phosphorylation sites is much desirable due to their convenience and fast speed. In this paper, a new bioinformatics tool named CKSAAP_PhSite was developed that ignored the kinase information and only used the primary sequence information to predict protein phosphorylation sites. The highlight of CKSAAP_PhSite was to utilize the composition of k-spaced amino acid pairs as the encoding scheme, and then the support vector machine was used as the predictor. The performance of CKSAAP_PhSite was measured with a sensitivity of 84.81%, a specificity of 86.07% and an accuracy of 85.43% for serine, a sensitivity of 78.59%, a specificity of 82.26% and an accuracy of 80.31% for threonine as well as a sensitivity of 74.44%, a specificity of 78.03% and an accuracy of 76.21% for tyrosine. Experimental results obtained from cross validation and independent benchmark suggested that our method was very promising to predict phosphorylation sites and can be served as a useful supplement tool to the community. For public access, CKSAAP_PhSite is available at http://59.73.198.144/cksaap_phsite/.     

Protein arginine methyltransferase 1-directed methylation of Kaposi sarcoma-associated herpesvirus latency-associated nuclear antigen

The Kaposi sarcoma-associated herpesvirus (KSHV) latency-associated nuclear antigen (LANA) is a multifunctional protein with roles in gene regulation and maintenance of viral latency. Post-translational modification of LANA is important for functional diversification. Here, we report that LANA is subject to arginine methylation by protein arginine methyltransferase 1 in vitro and in vivo. The major arginine methylation site in LANA was mapped to arginine 20. This site was mutated to either phenylalanine (bulky hydrophobic, constitutive methylated mimetic) or lysine (positively charged, non-arginine methylatable) residues. The significance of the methylation in LANA function was examined in both the isolated form and in the context of the viral genome through the generation of recombinant KSHV. In addition, authentic LANA binding sites on the KSHV episome in naturally infected cells were identified using a whole genome KSHV tiling array. Although mutation of the methylation site resulted in no significant difference in KSHV LANA subcellular localization, we found that the methylation mimetic mutation resulted in augmented histone binding in vitro and increased LANA occupancy at identified LANA target promoters in vivo. Moreover, a cell line carrying the methylation mimetic mutant KSHV showed reduced viral gene expression relative to controls both in latency and in the course of reactivation. These results suggest that residue 20 is important for modulation of a subset of LANA functions and properties of this residue, including the hydrophobic character induced by arginine methylation, may contribute to the observed effects.     

Sites of covalent modification in Trg, a sensory transducer of Escherichia coli

The Trg protein mediates chemotactic response of Escherichia coli to the attractants ribose and galactose. Like other transducers, Trg is a transmembrane protein that undergoes post-translational covalent modification. The modifications are hydrolysis (deamidation) of certain glutamine side chains to create glutamate residues and methylation of specific glutamates to form carboxyl methyl esters. Analysis of radiolabeled, tryptic peptides by high performance liquid chromatography and gas-phase sequencing allowed direct identification of the modified residues of Trg. The protein has 5 methyl-accepting residues. Four, at positions 304, 310, 311, and 318, are contained in a 23-residue tryptic peptide ending in lysine. The fifth, at position 500, is within a 25-residue tryptic peptide ending in arginine. At two sites, 311 and 318, glutamines are deamidated to create methyl-accepting glutamates. There is not a required order of modification among the sites. However, there is a substantial preference for methylation on the arginine peptide and, among sites on the lysine peptide, for the middle pair. Comparison of sequences surrounding modified residues identified in this work for Trg and previously for Tsr and Tar suggests a consensus sequence for methyl-accepting sites of Ala/Ser-Xaa-Xaa-Glu-Glu*-Xaa-Ala/OH-Ala-OH/Ala, where OH signifies Ser or Thr and the asterick marks the site of modification.     

Methylation of H3-lysine 79 is mediated by a new family of HMTases without a SET domain

The N-terminal tails of core histones are subjected to multiple covalent modifications, including acetylation, methylation, and phosphorylation. Similar to acetylation, histone methylation has emerged as an important player in regulating chromatin dynamics and gene activity. Histone methylation occurs on arginine and lysine residues and is catalyzed by two families of proteins, the protein arginine methyltransferase family and the SET-domain-containing methyltransferase family. Here, we report that lysine 79 (K79) of H3, located in the globular domain, can be methylated. K79 methylation occurs in a variety of organisms ranging from yeast to human. In budding yeast, K79 methylation is mediated by the silencing protein DOT1. Consistent with conservation of K79 methylation, DOT1 homologs can be found in a variety of eukaryotic organisms. We identified a human DOT1-like (DOT1L) protein and demonstrated that this protein possesses intrinsic H3-K79-specific histone methyltransferase (HMTase) activity in vitro and in vivo. Furthermore, we found that K79 methylation level is regulated throughout the cell cycle. Thus, our studies reveal a new methylation site and define a novel family of histone lysine methyltransferase.     

Role of lysine residues of plasma lipoproteins in high affinity binding to cell surface receptors on human fibroblasts.

The low density lipoprotein (LDL) cell surface receptors on human fibroblasts grown in culture bind specific plasma lipoproteins, initiating a series of events which regulate intracellular cholesterol metabolism. Specificity for the interaction with the receptors resides with the protein moieties of the lipoproteins, specifically with the B and E apoproteins of LDL and certain high density lipoproteins (HDLc HDLl), respectively. It was previously established that the amino acid arginine is a functionally significant residue in or near the recognition sites on the B and E apoproteins and that modification of this residue abolishes the ability of these apolipoproteins to bind to the receptor. The present study indicates that lysine residues are also involved in the lipoprotein-receptor interaction. Chemical modification of 15% of the lysine residues of LDL by carbamylation with cyanate or 20% by acetoacetylation with diketene prevents the LDL from competitively displacing unmodified 125I-LDL from the high affinity receptor sites or from binding directly to the receptor. Moreover, quantitative reversal of the aceto-acetylation of the lysine residues of LDL by hydroxylamine treatment regenerates the lysyl residues and reestablishes greater than 90% of the original binding activity of the LDL. The reversibility of this reaction establishes that the loss of binding activity which follows lysine modification is not due to an irreversible alteration of the LDL or HDLc but is probably due to an alteration of a property of the recognition site associated with specific lysine residues. While acetoacetylation and carbamylation neutralize the positive charge on the epsilon-amino group of lysine, reductive methylation selectively modifies lysine residues of LDL and HDLc without altering the positive charge, yet abolishes their ability to bind to the receptor. Preservation of the charge but loss of binding activity following reductive methylation of the lipoproteins suggests that the specificity of the recognition site does not reside simply with the presence of positive charges but depends on other more specific properties of the site determined by the presence of a limited number of the lysine (and arginine) residues. The precise role of lysine remains to be defined, but its function may be to establish and maintain the conformation of the recognition site or the alignment of reactive residues, or both, or to chemically react, through its epsilon-amino group, with the receptor (hydrogen bond formation would be such a possibility).     

SET8 recognizes the sequence RHRK20VLRDN within the N terminus of histone H4 and mono-methylates lysine 20

Methylation of lysine 20 in histone H4 has been proven to play important roles in chromatin structure and gene regulation. SET8 is one of the methyltransferases identified to be specific for this modification. In this study, the minimal active SET domain of SET8 has been mapped to the region of amino acids 195-352. This region completely retains the same methylation activity and substrate specificity as the full-length SET8. The SET domain recognizes a stretch of specific amino acid sequence around lysine 20 of H4 for its methylation activity. Methylation assays with N terminus mutants of H4 that contain deletions and single alanine or glutamine substitutions of charged residues revealed that SET8 requires the sequence RHRK20VLRDN for methylation at lysine 20. The individual mutation of any charged residue in this sequence to alanine or glutamine abolished or greatly decreased levels of methylation of lysine 20 of H4 by SET8. Interestingly, mutation of lysine 16 to alanine, arginine, glutamine, or methionine did not affect methylation of lysine 20 by the SET domain. Mass spectrometric analysis of synthesized H4 N-terminal peptides modified by SET8 showed that SET8 selectively mono-methylates lysine 20 of H4. Taken together, our results suggested that the coordination between the amino acid sequence RHRK20VLRDN and the SET domain of SET8 determines the substrate specificity and multiplicity of methylation of lysine 20 of H4.     

Identifying and quantifying in vivo methylation sites by heavy methyl SILAC

Protein methylation is a stable post-translational modification (PTM) with important biological functions. It occurs predominantly on arginine and lysine residues with varying numbers of methyl groups, such as mono-, di- or trimethyl lysine. Existing methods for identifying methylation sites are laborious, require large amounts of sample and cannot be applied to complex mixtures. We have previously described stable isotope labeling by amino acids in cell culture (SILAC) for quantitative comparison of proteomes. In heavy methyl SILAC, cells metabolically convert [(13)CD(3)]methionine to the sole biological methyl donor, [(13)CD(3)]S-adenosyl methionine. Heavy methyl groups are fully incorporated into in vivo methylation sites, directly labeling the PTM. This provides markedly increased confidence in identification and relative quantitation of protein methylation by mass spectrometry. Using antibodies targeted to methylated residues and analysis by liquid chromatography-tandem mass spectrometry, we identified 59 methylation sites, including previously unknown sites, considerably extending the number of in vivo methylation sites described in the literature.     

Systematic Proteomic Analysis of Protein Methylation in Prokaryotes and Eukaryotes Revealed Distinct Substrate Specificity

The studies of protein methylation mainly focus on lysine and arginine residues due to their diverse roles in essential cellular processes from gene expression to signal transduction. Nevertheless, atypical protein methylation occurring on amino acid residues, such as glutamine and glutamic acid, is largely neglected until recently. In addition, the systematic analysis for the distribution of methylation on different amino acids in various species is still lacking, which hinders our understanding of its functional roles. In this study, we deeply explored the methylated sites in three species Escherichia coli, Saccharomyces cerevisiae, and HeLa cells by employing MS‐based proteomic approach coupled with heavy methyl SILAC method. We identify a total of 234 methylated sites on 187 proteins with high localization confidence, including 94 unreported methylated sites on nine different amino acid residues. KEGG and gene ontology analysis show the pathways enriched with methylated proteins are mainly involved in central metabolism for E. coli and S. cerevisiae, but related to spliceosome for HeLa cells. The analysis of methylation preference on different amino acids is conducted in three species. Protein N‐terminal methylation is dominant in E. coli while methylated lysines and arginines are widely identified in S. cerevisiae and HeLa cells, respectively. To study whether some atypical protein methylation has biological relevance in the pathological process in mammalian cells, we focus on histone methylation in diet‐induced obese (DIO) mouse. Two glutamate methylation sites showed statistical significance in DIO mice compared with chow‐fed mice, suggesting their potential roles in diabetes and obesity. Together, these findings expanded the methylome database from microbes to mammals, which will benefit our further appreciation for the protein methylation as well as its possible functions on disease.     

Prediction of nicotinamide adenine dinucleotide interacting sites based on ensemble support vector machine

Nicotinamide adenine dinucleotide (NAD) plays an important role in cellular metabolism and acts as hydrideaccepting and hydride-donating coenzymes in energy production. Identification of NAD protein interacting sites can significantly aid in understanding the NAD dependent metabolism and pathways, and it could further contribute useful information for drug development. In this study, a computational method is proposed to predict NAD-protein interacting sites using the sequence information and structure-based information. All models developed in this work are evaluated using the 7-fold cross validation technique. Results show that using the position specific scoring matrix (PSSM) as an input feature is quite encouraging for predicting NAD interacting sites. After considering the unbalance dataset, the ensemble support vector machine (SVM), which is an assembly of many individual SVM classifiers, is developed to predict the NAD interacting sites. It was observed that the overall accuracy (Acc) thus obtained was 87.31% with Matthew's correlation coefficient (MCC) equal to 0.56. In contrast, the corresponding rate by the single SVM approach was only 80.86% with MCC of 0.38. These results indicated that the prediction accuracy could be remarkably improved via the ensemble SVM classifier approach.     

Characterization of the Mycobacterium avium subsp. paratuberculosis laminin-binding/histone-like protein (Lbp/Hlp) which reacts with sera from patients with Crohn's disease

Mycobacterium avium subsp. paratuberculosis (Map) causes a chronic enteric disease in ruminants, called paratuberculosis or Johne's disease. The current model proposes that after ingestion by the host, Map crosses the intestinal barrier via internalization by the M cells. Experimental observations suggest, however, that Map may also transcytose the intestinal wall via the enterocytes, but the mechanisms involved in this process remain poorly understood. Cytoadherence assays performed on epithelial cells with Map revealed that the addition of laminin to the cell culture increases adhesion. A Map protein was isolated by heparin-Sepharose chromatography and identified as a laminin-binding protein like. The gene encoding this protein named Lbp/Hlp was identified in the Map genome sequence at locus MAP3024 (annotated Hup B). The deduced Map Lbp/Hlp amino acid sequence reveals 80% identity with that reported for other mycobacteria. The C-terminal domain involved in adhesion is mainly composed of arginine and lysine residues modified by methylation. In vitro tests demonstrated that recombinant Lbp/Hlp binds laminin, heparin, collagen and epithelial cells. Interestingly, we found that this adhesin corresponds to the antigen described as the target of pANCA and serum antibodies of patients with Crohn's disease.     

Using random forest algorithm to predict β-hairpin motifs

A novel method is presented for predicting β-hairpin motifs in protein sequences. That is Random Forest algorithm on the basis of the multi-characteristic parameters, which include amino acids component of position, hydropathy component of position, predicted secondary structure information and value of auto-correlation function. Firstly, the method is trained and tested on a set of 8,291 β-hairpin motifs and 6,865 non-β-hairpin motifs. The overall accuracy and Matthew's correlation coefficient achieve 82.2% and 0.64 using 5-fold cross-validation, while they achieve 81.7% and 0.63 using the independent test. Secondly, the method is also tested on a set of 4,884 β-hairpin motifs and 4,310 non-β-hairpin motifs which is used in previous studies. The overall accuracy and Matthew's correlation coefficient achieve 80.9% and 0.61 for 5-fold cross-validation, while they achieve 80.6% and 0.60 for the independent test. Compared with the previous, the present result is better. Thirdly, 4,884 β-hairpin motifs and 4,310 non-β-hairpin motifs selected as the training set, and 8,291 β-hairpin motifs and 6,865 non-β-hairpin motifs selected as the independent testing set, the overall accuracy and Matthew's correlation coefficient achieve 81.5% and 0.63 with the independent test.