Geometry Dynamics of α-Helices in Different Class I Major Histocompatibility Complexes

Figure 15

Root mean square deviations. Root mean square deviations of TCR/pMHC systems B4402, B4403, and B4405. Superposition of successive frames was done with respect to protein backbone of the first frame of the simulation (nonprogressive fitting) and RMSD was calculated between protein backbones. Row TCR/pMHC shows that the whole protein system, TCR/pMHC, was fitted to itself and RMSD calculated for the whole protein. Row MHC shows that TCR was fitted to itself and RMSD calculated for TCR. Row TCR shows that MHC was fitted to itself and RMSD calculated for MHC. Results for B4405 indicate that 250 ns of simulation time does not suffice to sample the whole phase space, which is a common finding for such large proteins. RMSD time courses for B4402 and B4403 do not explicitly indicate nonstationary behaviour. They indeed show slower and smaller growth of RMSD over time than does B4405, also indicating their stability as a molecule (despite two point mutations introduced). As noted before, the present work intends to delineate techniques for modelling geometries of MHC components and does not aim at statistical comparisons between the motions of different HLA alleles. Ergodicity is hence not a vital issue; see the discussion in Schreiner et al. [45].