Systematic Functional Characterization Of Hypothetical Proteins In Synechocystis Sp. PCC6803

Hypothetical proteins whose functions are still unknown pose a challenge not justto genome annotation but also to better understanding of general biology of organisms.Nearly30%of sequenced bacterial genomes annotated as hypothetical and conservedhypothetical proteins. For the photosynthetic bacterium Synechocystis sp. PCC6803,1466of the3179proteins were annotated as conserved hypothetical proteins orhypothetical proteins (46%). Given the large fraction that hypothetical proteinsoccupy in the genome, it is plausible that some of these genes serve critical cellularroles. The goals of our study were to determine which hypothetical protein encodinggenes were expressed and to provide an improved annotation with more functionalinformation. To accomplish these, both computational and experimentalmethodologies were employed.Computational methods like size distribution model and codon usage were used tocalculate the differences between experimentally confirmed proteins and possibleover-annotated proteins. iTRAQ labeled quantitative proteomic analysis combinedwith LC–MS/MS was used to examine the proteins extracted from Synechocystis sp.PCC6803cells across four stress treatment conditions. A total of807hypotheticalproteins were detected in our high quality proteomic datasets, and reminding659proteins were not detected under the tested conditions.Functional inference for hypothetical proteins was performed not only by acombination of traditional computational methods of both homology-andnon-homology-based approaches, such as genome context analysis, protein-proteininteractions analysis, and operon organization analysis, but also by a gene networkconstruction using transcriptomics data to obtain possible functional clues. Usingthese approaches we were able to assign possible functions to210hypotheticalproteins.Preliminary validation of the putative involvement of gene sll0044, sll1418, sll1130and sll1638in photosynthesis was conducted by gene knockout and phenotypicanalysis. The study provided an improved version of genome annotation forSynechocystis PCC6803, and also demonstrated that high-throughput methodologieslike quantitative proteomics and transcriptomics could be a good supplementation to various computational approaches in deciphering function of hypothetical proteins.