News and Publications

Three-Dimensional Architecture of Membrane Protein Channels

A.K. Mitra, G. Ren, P. Melnyk

We are interested in understanding the structural basis of channeling or transporting of solutes across the cell membrane by membrane proteins. For this purpose, we use high-resolution electron crystallography to reveal 3-dimensional structure and modulations in structure related to function in the "native" lipid-bilayer environment.

In prokaryotes and eukaryotes, rapid transport of water, a fundamental physiologic process necessary for homeostasis, is facilitated by aquaporin channels. We are determining the high-resolution structure of the archetype aquaporin 1 (AQP1) from human erythrocytes to understand how this protein facilitates selective, bidirectional water transport in certain water-permeable cells. We have generated 2-dimensional crystals by reconstitutiing AQP1 into synthetic lipid bilayers. Electron diffraction patterns recorded from unstained, frozen-hydrated AQP1 crystals display reflections to 2.8 Å.

We have calculated a 3.7-Å-resolution density map of AQP1 projected onto the membrane plane by analyzing images and diffraction patterns recorded from untilted crystals. Using data recorded from crystals tilted (up to 45°) in the microscope, we generated a 7-Å-resolution 3-dimensional density map. This map reveals that AQP1 is composed of 6 tilted, membrane-spanning -helices. These helices form a barrel in the membrane surrounding a vestibular region that presumably encloses the aqueous pathway. The 3-dimensional map shows a local, pseudo 2-fold axis of symmetry located in the plane and near the center of the bilayer. The N- and C-terminal halves of AQP1 are homologous; therefore, coupling of these repeats in the polypeptide sequence and the observed intramolecular symmetry in the structure provide an elegant solution to the problem of bidirectional transport across the bilayer. We are now accumulating crystallographic data at a higher tilt (~55°) to extend the resolution; the goal is an atomic model of AQP1.

Mercurial compounds such as mercuric chloride and P4-(chloromercuri)benzene sulfonate reversibly inhibit water transport in many aquaporins. In the case of AQP1, the inhibition is due to binding of the compounds to a cysteine residue (Cys189) at the extracellular face of AQP1 and represents the only known way to block the channel. We are examining mercurial-labeled AQP1 crystals and are applying difference Fourier analysis to understand the structural correlates of the inhibition and define the location of the aqueous pathway.

To investigate the similarity and diversity in the structural principles that define transport of solutes, we are pursuing 2-dimensional crystallization for structural studies on a number of other channels and transporters. This project includes collaborative studies with A. Verkman, University of California, San Francisco, on aquaporin 4, which has the highest osmotic water permeability of members of the aquaporin family, and studies with D. Roberts, University of Tennessee, on Nod26, a channel from soybean nodules that transports both cations and anions. We are also collaborating with M. Hermodson, Purdue University, in structural studies on recombinant Escherichia coli ribose transporter, a member of the so-called ATP-binding cassette transporters that include the cystic fibrosis transmembrane regulator and the multidrug resistance P-glycoprotein. As a step toward dissecting the structure-function correlates of the transport system, we are focusing on the membrane-spanning polypeptide component of ribose transporter.

In addition to integral membrane proteins, we are interested in a class of soluble proteins that insert into membranes and form channels in vivo by a receptor-mediated process and in vitro at acidic pH. In collaboration with J. Collier, Harvard University, we are using anthrax toxin, a tripartite toxin, as a model of this interesting phenomenon. To understand the process leading to the activation of anthrax toxin, we are using single-particle image analysis to study the mode of binding of the lethal factor to soluble and membrane-integrated PA63 antigen of anthrax toxin. Research on such secreted proteins will provide knowledge about the structural dynamics involved in the translocation of soluble proteins into a low-dielectric hydrophobic milieu, leading to a possible insight into the general question of how proteins insert into membranes.

PUBLICATIONS

Cheng, A., van Hoek, A.N., Yeager, M., Verkman, A.S., Mitra A.K. Three-dimensional organization of a human water channel. Nature 387:627, 1997.

Yeager, M., Unger, V.M., Mitra, A.K. Three-dimensional structure of membrane proteins determined by two-dimensional crystallization, electron cryo-microscopy and image analysis. Methods Enzymol., in press.