Ion channels are transmembrane proteins that control large fluxes of ions across otherwise impermeable membranes. They can be seen as naturally occurring nanotubes whose properties can be controlled very precisely using the tools of molecular biology. Since they are current conducting proteins they are suitable candidates to function as intermediates between the a-biotic and biotic worlds. A number of requirements have to be fulfilled to apply channel proteins in devices: the ion channels should be embedded in high resistance membranes (GU), the membranes should be robust and the functionality of the channels should be under external control of the user. Using the bacterial ion channel OmpF as a model system some of these issues are being addressed. The approach we are using is to develop robust membranes either by cross-linking of embedded proteins or by providing a solid membrane support that allows functional incorporation of membrane proteins. Functionality is being manipulated by site-directed mutagenesis, sometimes in combination with chemical modification as well as by a random mutagenesis approach. Results so far show that it is possible to modify the selectivity of the channel in a rational way, and work is underway to control this property using external signals such as light.