Computational methods to identify novel methyltransferases
BMC Bioinformatics volume 10, Article number: P7 (2009)
1.2% of the yeast genes are estimated to encode enzymes that catalyze the transfer of a methyl group from S-adenosylmethionine (AdoMet) to protein, nucleic acid, lipid, and small molecule substrates . These enzymes function in biosynthesis, regulating metabolic pathways, and controlling gene expression, including writing the histone code. BLAST and MEME/MAST analysis using the amino acid sequence of motifs have previously generated a list of putative Class I methyltransferases . Recently we have used a combination of a new search algorithm and structural information to refine this analysis . This study utilizes these updated methods of identifying motifs and scanning the proteome to predict new members of the different families of methyltransferases in different organisms. These new members may function in novel pathways or new modes of regulation.
Materials and methods
Advanced hidden Markov models (HMM) profiles, predicted secondary structures, and solved crystal structures are used to identify the AdoMet-binding motifs of the different families of methyltransferases [1, 3]. To generate a list of putative methyltransferases, we used both our newly developed program "Multiple Motif Scanning" [3, 4] and HHpred . Sequence similarity networks are then used to predict the probable substrates for the putative methyltransferases . Additionally, several of the candidate methyltransferases were incubated with radioactive AdoMet to reveal binding by detection of the radioactive protein-ligand via SDS-PAGE separation .
The putative list of methyltransferases for S. cerevisiae among four of the methyltransferases families are italicized (see Table 1). Known methyltransferases are shown for only the SET and SPOUT families. Several putative methyltransferases are found to bind AdoMet through UV-crosslinking experiments (designated * in Table 1). This approach validated previously suggested putative enzymes and additionally identified several new candidates . Extending this analysis to the human proteome surprisingly reveals little expansion of family members (Figure 1). Our goal is to enhance the functional identification of novel methyltransferases by providing lists of the best candidates for biochemical analyses.
Petrossian TC, Clarke SG: Bioinformatic identification of novel methyltransferases. Epigenomics 2009, in press.
Katz JE, Dlakić M, Clarke S: Automated identification of putative methyltransferaess from genomic open reading frames. Mol Cell Proteomics 2003, 2: 525–540.
Petrossian TC, Clarke SG: Multiple Motif Scanning to identify methyltransferases from the yeast proteome. Mol. Cell. Proteomics 2009, 8: 1516–1526. 10.1074/mcp.M900025-MCP200
Multiple Motif Scanning[http://www.chem.ucla.edu/files/MotifSetup.Zip]
Söding J, Biegert A, Lupas AN: The HHpred interactive server for protein homology detection and structure prediction. Nucleic Acids Res 2005, 33: W244-W248. 10.1093/nar/gki408
This research was supported by the National Institutes of Health Grant GM026020 and the Office of Science (BER), U.S. Department of Energy, Grant No. DE-FG02-06ED64270. T.C.P. was supported by the UCLA Chemistry-Biology Interface Training Grant GM008496.
About this article
Cite this article
Petrossian, T.C., Clarke, S.G. Computational methods to identify novel methyltransferases. BMC Bioinformatics 10 (Suppl 13), P7 (2009). https://doi.org/10.1186/1471-2105-10-S13-P7
- Hide Markov Model
- Yeast Gene
- Control Gene Expression
- Similarity Network