E0032

Two- and Three-dimensional Crystallization of Membrane Proteins. Werner Kühlbrandt, Max-Planck-Inst. für Biophysik, Marie-Curie-Str. 13-15, D-60439 Frankfurt am Main, Germany, kuehlbrandt@mpibp-frankfurt.mpg.de

The majority of membrane proteins are difficult to isolate in quantities required for biochemical, biophysical or structural studies. High-level expression and structure determination tend to be problematical, especially in the case of eukaryotic membrane proteins.

Since most membrane proteins are too large for NMR spectroscopy, structure determination has relied mainly on diffraction methods. X-ray crystallography of three-dimensional crystals and electron crystallography of two-dimensional crystals have both yielded high-resolution structures. The majority of the ~30 different high-resolution structures known to date have been determined by x-ray crystallography, suggesting that 3D crystallisation is getting more predictable. Apart from the general detergent requirement, the main parameters for 3D crystallisation are the same as for soluble proteins, although the need for specific lipids or small amphiphiles adds complexity. The use of antibody fragments to increase the hydrophilic surface available for crystal contacts may prove to be a general method for growing diffraction-quality crystals, but the production of suitable Fab or Fv-fragments is rate-limiting.

As a rule, membrane proteins form 2D crystals more readily than 3D crystals. 2D crystallisation conditions of different proteins tend to be similar. Highly ordered 2D crystals that are large enough for electron diffraction appear to be the exception, but even imperfect 2D crystals yield structures at 6-8 Å resolution quite readily, because electron imaging avoids the phase problem. With other information from biochemical studies, mutagenesis, multiple sequence alignment, or high-resolution structures of related proteins, a high-quality map at this resolution can be sufficient for building an atomic model. Two examples will be discussed: (1) The plasma membrane proton ATPase from Neurospora and (2) the c-subunit ring of the Fo-ATP synthase from Ilyobacter tartaricus.