Simulation of phosphorylation-induced structural changes for DNA polymerase β, using molecular modelling

DNA polymerase beta is a 39 kDa enzyme that comprises two major domains, a 31 kDa domain responsible for the polymerase activity and an 8 kDa domain, which bind ssDNA and has a dRP Lyase activity. DNA polymerase beta was shown to be phosphorylated in vitro with Protein Kinase C at serines 44 and 55, resulting in loss of its polymerase enzymic activity, but not its ability to bind ss DNA. The enzyme has a potential phosphorylation site at tyrosine 250, reported on www.phosphosite.org. We set out to simulate potential phosphorylation-induced structural changes for DNA polymerase beta using molecular modelling algorithm (CHARMM22 and AMBER for serines and tyrosine, respectively) and published structural coordinates for the enzyme (pdb: 2FMS). Simulations with CHARMM22 were for 10 ns and with AMBER for 100 ns. RMSF plot of Ser44P showed highest deviation. RMSF plot of Ser55P and Ser44/55P showed a similar fluctuation pattern except at aa 1-75, the major part for ssDNA binding and dRP lyase domains. Radius for gyration (Rg) for apo_enzyme and Ser55P were consistent throughout the simulation. The Rg of Ser44P and Ser44/55P fluctuated throughout the simulation. Phosphorylation at Ser44, Ser55 and Ser44/55 induced major conformational fluctuations from apo_enzyme at the ssDNA binding domain and dNTP selection domain. Tyr250P enzyme was noticeably more dynamic than reference. The results may explain the Protein Kinase C phosphorylation-induced loss of enzymic activity reported in vitro . Data also suggest that phosphorylation at tyrosine 250 is worthy of being investigated. Such studies may pave the way for simulating variations in the atomic structure for the other post-translationally modified (e.g. acetylated) form(s) for this and other enzymes/proteins and be used as an indicator that a potential modification may have structural (and functional) consequences.