Expression and purification of Aquaporin-6 in different systems : comparison of cell-free, Semliki forest virus, and "Pichia pastoris" expression systems

Water is the major component of all cells. Due to its polarity water is transported via selective, high-capacity water channels, referred to as aquaporins (AQPs). AQPs are conserved throughout lower organisms, the animal and plant kingdoms Driven by the osmotic gradient, water passes the AQPs in a bidirectional manner. Strikingly, most AQPs have a high water selectivity, thus preventing cotransport of protons and hydronium ions. Today, we possess a profound understanding of water transport that was initially based on the cloning of the first AQPs in the early 1990s and the first atomic structure of human aquaporin-1 (AQP1) in 2000. Compared to other human AQPs, AQP6 has a unique distribution and a distinct function. It is an intracellular channel that is gated and permeated by water and anions. In contrast to many (mammalian) AQPs, AQP6 has not been characterized in a purified form. The initial aim of this work was to determine parameters for the expression and purification of AQP6, using an expression system, which is able to produce AQP6 in adequate amounts and quality for functional and crystallographic studies. In order to identify the best system in terms of sample quantity and quality, overexpression in several systems was investigated. Initial expression trials in E. coli have not shown expression of AQP6. Furthermore, it is known that AQP6 is cytotoxic due to its ion channel functionality. Therefore, a cell-free expression system was initially considered, which additionally offers several approaches for membrane proteins. As an open system, cell-free expression allows different detergents, lipids und liposomes to be explored for enhancing protein quality and quantity. Moreover, the toxic effects of an open anion channel can be ignored in a cell-free environment. Due to these considerations, the expression of AQP6 was carried out using a cell-free system, despite the fact that there is no high-resolution structure of a mammalian protein expressed in a cell-free system available, yet. Another approach was the expression in mammalian cells using the Semliki Forest virus expression system, which has been used for expression of G-protein coupled receptors in the past. It offers the most native environment for expression of mammalian membrane proteins, since expression is performed in mammalian cells. With a stringent control of the pH in the expression media, cytotoxic effects of AQP6 were circumvented. Expression in Pichia pastoris was another option, because AQPs expressed in P. pastoris have been successfully crystallized. The expression of a non-functional AQP6 mutant was a way to evaluate the implications of the opened ion channel in terms of expression, when comparing it to the wild type AQP6. In principle the overexpression of AQP6 was possible in all three systems (cell-free, SFV-BHK cells and P. pastoris). In this work expression and solubilization conditions of AQP6 were evaluated extensively for the mentioned expression systems and initially low expression and purification yields were increased to amounts allowing reconstitution into liposomes in order to perform activity measurements. In the cell-free system reasonable expression yields could be improved further after adding a N-terminal sequence tag to the open reading frame of the AQP6 gene. Quality of purified protein was determined by analytical size exclusion chromatography and single particle transmission electron microscopy of negatively stained samples. Both methods revealed the presence of AQP6 tetramers, indicating the purification of AQP6 in a functional confirmation. Reconstitution experiments of AQP6 were carried out in order to demonstrate functional expression. Nevertheless, reconstitution experiments, using different lipids of artificial and natural sources, revealed that immunodetection of AQP6 in liposome fractions after ultracentrifugation was not a proof for correct reconstitution. In the case of cell-free expressed AQP6, in absence or in presence of detergent, AQP6 was detected in the fractions of liposomes, but electron micrographs of freeze fractured liposomes showed no protein reconstitution. However, precipitated protein, associated to the liposomes, was visible. Freeze fracture electron microscopy in combination with immunodetection was the method of choice for demonstrating protein reconstitution. Water transport measurements by stopped-flow light scattering were carried out in order to investigate channel properties. AQP6 was expressed either in cell-free systems or in BHK cells, purified and reconstituted into liposomes. Rapid mixing of proteo-liposomes with sucrose buffer resulted, due to the osmotic pressure, in outward directed water flow and vesicle shrinkage, which was detected by changes in the intensity of scattered light. In addition, measurements in presence of Hg2+, which was reported to be an activator of water permeability of AQP6, were executed.