Computational prediction of nuclear localization signals and structural characteristics of 91 types of HPV E6 proteins.

ABSTRACT

Human papillomaviruses (HPVs) are small DNA tumor viruses that replicate and assemble exclusively in the nucleus. Thus their proteins, including E6, must carry nuclear localization signals (NLSs) to enter the nucleus. To analyze and to predict the nuclear localization signals and several post translational modifications by bioinformatics analysis, we obtained 91 E6 protein sequences from available databases. To investigate the localization of these sequences, we used Hum-Ploc software. Homology and alignment of sequences were performed by Blast software and Multalin server respectively. Prediction of N-glycosylation and serine, threonine and tyrosine phosphorylation sites of HPV E6 protein sequences was accomplished with NetNGlyc and NetPhos software. Out of 91 types, the NLSs of 29 types were predicted by signal-3L and signal-CF software. We tried to predict the NLSs of remaining HPV E6 proteins according to the homology of the already predicted NLSs. However, because of considerable variation between E6 protein sequences, we could not classify the NLSs in monopartite or bipartite. According to the results, all NLSs of HPV E6 proteins could be assigned to 11 categories. NLSs of several HPV E6 protein sequences were also determined by experimental studies. Overall, different types of HPV E6 protein in same category show approximately similar pattern in post translational modifications such as N-glycosylation and phosphorylation. Some HPV early genes, such as E6, are known to act as oncogenes that promote tumor growth and malignant transformation. Thus more detailed recognition of nuclear localizing sequences and nucleocytoplasmic transport pathway can play a key role in prevention and treatment of HPV infection and related cancers. The results also show that bioinformatics technology can direct and simplify experimental studies.