The researchers used the Macromolecular Crystallography Facility at Beamline 5.0.2 to solve the structure of GlpF, crystallized with three glycerol molecules conveniently passing through each of four channels. Their model was based on multiple isomorphous replacement and anomalous dispersion studies, yielding R_cryst and R_free values of 19.7 and 22.3, respectively.

The overall structure of GlpF comprises four parallel channels arranged with fourfold symmetry about a central axis that runs perpendicular to the membrane bilayer. Each channel is formed by six membrane-spanning alpha helices and two half-membrane-spanning alpha helices that also form a hydrophobic outer shell. The complete molecule's amino acid sequence can be divided into two halves that show about 20% homology, each containing an arginine-proline-alanine (NPA) sequence near its center. In the three-dimensional structure, the two halves are related by a quasi-twofold axis running through the center of the bilayer, perpendicular to the fourfold axis. A tight link between the two was found to be formed by the NPA sequences, the proline ring of each being clasped between the other's proline and neighboring alanine residues.

The structure of the channel revealed much about GlpF's selectivity mechanism. A constriction just wide enough to accommodate linear alditols such as glycerol in single file begins near the quasi-twofold axis, 8 Å toward the periplasmic side (the side nearer the outside of the cell) and ends on the cytoplasmic surface. Inside this "selectivity filter," a hydrophobic corner is formed by the rings of a tryptophan and a phenylalanine residue. The carbon backbones of the glycerol molecules become wedged against this corner so that their oxygens come near residues with which they can form successive hydrogen bonds. Thus, passage through the channel depends not only on narrowness but also on a molecule's ability to become polarized along the correct axis. Because some of the successive hydrogen-bonding residues are hydrogen bond donors and others are acceptors, the molecule passing through must have atoms that can act both as acceptors and as donors. This criterion facilitates the passage of glycerol but leaves ions grumbling at the door.

GlpF Structure Shows Aquaporin Action

The family of proteins known as aquaporins handles a surprisingly diverse set of duties. Be they in plants, animals, bacteria, or yeast, though, all relate to the job of allowing the flow of specific small molecules across cell membranes while giving ions the cold shoulder. In humans, defects in these proteins have been implicated in such disparate troubles as kidney disease (nephrogenic diabetes insipidus), salivary gland disfunction (Sjogren's syndrome), and altered water transport in the brain and the lungs. As an aquaporin found in E. coli bacteria, GlpF would seem to be an uninformatively distant relation to the aquaporins found in humans. Not so, as it turns out. Any protein can be thought of as a sequence of amino acids that folds upon itself in a very precise way to build a functional structure. The amino acid sequences that determine the shapes of the various aquaporins have a high percentage of overlap from one to the other. Thus, solving the structure of one provides a tremendous leap toward understanding them all. With a finely detailed (2.2-Å) view of GlpF, scientists now can understand how one member of the family does its job and begin to surmise how the others do theirs.

A view along the length of the "selectivity filter," (left) showing the hydrogen bonding (dotted lines). Radial lines indicate hydrophobic areas. The arrows indicate the cross sections through the channel, shown at right.

Research conducted by D. Fu, A. Libson, L.J.W. Miercke, C. Weitzman, P. Nollert, J. Krucinski, and R.M. Stroud (University of California, San Francisco).

Research funding: National Institutes of Health, Human Frontiers Research Science Organization. Operation of the ALS is supported by the U.S. Department of Energy, Office of Basic Energy Sciences.

Publication about this research: D. Fu, A. Libson, L.J.W. Miercke, C. Weitzman, P. Nollert, J. Krucinski, and R.M. Stroud, "Structure of a glycerol-conducting channel and the basis for its selectivity," Science 290, 481-486 (2000).

ALSNews Vol. 177, May 23, 2001