De novo design of membrane protein channels

Abstract

Advances in the field of synthetic biology and de novo protein design come in the aid of the existing methods, and aim to contribute with providing essential answers referring to the sequence-structure- function relationship problem. Advances in the implementation of computational techniques in these research fields, promoted the speed of research, however, the computer-only based studies cannot provide sufficient data, especially due to insufficient real-life-based training information. For this reason, the need of interdisciplinary studies and combining computational techniques with laboratory-based experiments are required. De novo protein design aims to contribute to the design of small building blocks, which can self-associate into known or new structures, stabilising existing scaffolds. Thus, it enables the possibility of creating a library of small building blocks and their influence on structure formation and stability. In this cross-disciplinary study, we aimed to contribute to de novo membrane protein design with a novel, repetitive sequence (CC1), able to associate into an antiparallel homotetrameric helical bundle. In this study it is also proposed an additional computational framework, for de novo membrane protein design, combining Crick parametrisation tools and simulations in implicit and explicit membranes. The design pathway combined minimal and rational approaches, completed by knowledge-based and statistical studies. Molecular dynamics simulations in GROMACS using explicit POPC, POPE and POPG-containing membranes, showed that the antiparallel homotetrameric bundle formed by the repetitive CC1 sequence, appeared to be stable across all types of lipid compositions tested, showing great stability, as predicted by the initial energy score functions. Moreover, in POPE and POPG- explicit membranes, the CC1 antiparallel tetramer showed structural conformations which guide towards a potential mechanosensitive channel-like function. The experiments were compared to controls represented by the designed REAMP tetramer and poly-leucine antiparallel bundle. The CC1 antiparallel homotetramer has also proven biocompatibility, when expressed in E. coli C41 (DE3) cells, having in the structure the mistic protein and the split-variant of the superfolder yellow fluorescent protein, as “internal chaperone”, to overcome the challenges associated with the translocation machinery. The expression levels have shown to be comparable to the REAMP control, having attached the same tags, and the expression levels have shown temperature and medium composition dependence, essential for future considerations. CC1 was also successfully cell-free synthesised. The efficiency of insertion into liposomes of the folded state and the fluorescence of the sfYFP have shown liposome composition dependence. CC1 have shown to also stabilise POPC: DPhPC containing bilayers in droplet-interface bilayers assays. The present work comprises a novel finding in the field of de novo design, highlighting a novel, biocompatible sequence, with a special highlight on the importance of the lipid composition in de novo protein stability and membrane insertion.