A framework for the reconstitution of membrane proteins

Over the last decade, several artificial devices, imitating functionalities found in nature, have emerged in the field of synthetic biology. Often they resemble cellular vesicles which carry out a defined function and where molecular transport is mediated via specific membrane proteins. This work describes the creation of a framework for the reconstitution of membrane proteins into synthetic membranes. The study of membrane proteins in terms of their structure (e.g. protein crystallization) and their detailed functionality requires the isolation and re-insertion into a non-native environment. A process called reconstitution which is considered delicate. Beside the commonly used phospholipids, which are part of the natural cell membrane, a membrane environment can be created by the use of amphiphilic block copolymers. Driven by self-assembly, these molecules can be used as a platform for nano-devices, as they can be decorated with active molecular compounds and the resulting membrane can incorporate membrane proteins. Various factors and their interplay and dependencies affect the outcome of the reconstitution of membrane proteins into synthetic membranes. Identifying the key factors and predicting their effect a priori is a challenging task. Reliable and systematic approaches are available for lipid based systems but, up to now, not for polymeric ones. A well-established method in the fields of chemical and process engineering is design of experiments. This statistical tool provides a way to do experimental planning systematically and assess the effects and interactions of factors on a measurable response. Within this thesis, this framework was applied to the reconstitution of the light-driven proton pump proteorhodopsin into membranes for the first time. As proteorhodopsin provides a vectorial transport of protons across a membrane, its orientation is critical for its use as an energy generator in a synthetic system. Six factors were studied: the polarization of the membrane, the pH value during reconstitution, the lipid to protein ratio, the salt concentration in the buffer, the amount of detergent used and the effect of the addition of the ionophore valinomycin. Two insertion pathways were identified for proteorhodopsin: i) charge assisted and ii) detergent mediated. Both of them result in functional proteoliposomes which exhibit the formation of a proton gradient upon illumination. The conditions of the reconstitution decide which path will be taken, as detergent concentrations around 0.5 % will induce the detergent mediated pathway and the combination of a polarized membrane together with higher detergent concentrations around 1 % will induce the charge assisted pathway. It is noteworthy that this study provides evidence that the detergent mediated one is dominant, as at 0.5 % detergent, an increased membrane charge does not affect the result. Transferring the knowledge gained towards polymeric systems, the second part of this study aims to investigate and compare the reconstitution of proteorhodopsin into polymer and lipid vesicles. As data from successful reconstitutions into polymersomes is rare, a lipid based system was used as a benchmark. Similar to the former chapter presented here, statistical modeling takes a significant part. Efficient one-step screening and optimization designs were employed to examine the assembly process of both membrane types together with proteorhodopsin. It could be revealed that both systems react differently to changing parameter combinations. The assembly of proteopolymersomes has stronger pH dependency compared to proteoliposomes and the addition of detergent does not show the membrane saturation effect known from liposomes. Probing the resulting proteovesicles for proton pumping activities, it was revealed that their performance is comparable, even though polymer membranes are not able to host the same numbers of proteorhodopsin molecules as lipid ones. Due to the applied statistical modeling, the derived equations could be used for mathematical optimization which predicted a set of parameters for reconstitution which are predicted to yield large, uniform and highly functional proteovesicles. Indeed, the results obtained from the verification of these factor settings were close to the predictions. The study provides experimental and modeling evidence for different reconstitution mechanisms depending on the membrane type. By making use of them, proteorhodopsin can be used to provide energy in an artificially created vesicular environment. Depending on the desired application, the membrane base can be composed of biocompatible lipids or robust block copolymers, providing a novel flexibility to researchers. Altogether, this thesis serves as an example of thoroughly designed procedure which fulfills the requirements of reproducibility and predictability. It can pave the way for creation of a toolbox which makes the expansion into the field of hybrid materials (lipid/polymer/protein) as well as more complex systems as molecular factories possible.