One of the basic members of the family of cell-membrane water channels is a protein called aquaporin 1 (AQP1). Late in 2001, Berkeley Lab researchers announced that they had determined the structure of AQP1 to a resolution of 2.2 angstroms, or 22 hundred-billionths of a meter.

	Biophysicist Bing Jap with the water-channel protein aquaporin 1 (AQP1). Jap led a team of researchers that determined the structure of AQPI, revealing the means by which the protein transports water through the cell membrane. The image shows structural views of the four subunits of AQPI as seen from outside the cell.

Such high resolution reveals the elegantly simple means by which AQP1 can transport water through the cell membrane at a high rate while effectively blocking everything else--even individual protons, the nuclei of hydrogen atoms.

"Membrane proteins are a very large class of proteins; some 30 percent of the genes in the human genome code for them. But they are notoriously difficult to crystallize, and only a few structures have been solved at very high resolution," says biophysicist Bing Jap, who led a team from Berkeley Lab's Life Sciences Division in the difficult and painstaking crystallization of AQP1.

The crystallized AQP1 was from bovine red blood cells, closely similar to that in human and other cells. The team liberated enough protein from "gallons of blood" to make 0.2-millimeter crystals, suitable for x-ray crystallography at beamline 5.0.2 of the Advanced Light Source.

The structure of AQP1 had previously been solved to a resolution of about 4 angstroms using electron-microscope crystallography, which can use even smaller crystals. But at this resolution it is impossible to see individual water molecules, and vital features were left out or mistakenly characterized.

"AQP1 is interesting because it is so specific for water" says Jap. "The key question was how it achieves this specificity. Theorists had come up with lots of ideas, but before we saw the structure in high resolution, nobody knew how it was accomplished."

The structure of one of the four subunits of the water-channel protein AQP1, embedded in the cell membrane. Each unit contains an independent water pore (marked by blue dots). A wide "vestibule" outside the cell (top) leads to the pore, which is about 2.8 angstroms across at its narrowest, just wide enough to admit water molecules. The pore widens into another vestibule inside the cell (bottom).

Architecturally, AQP1 is an assembly of four units, each with three major structural features: each has an entrance, or "vestibule," on the outside of the cell envelope, connected to a similar vestibule inside the cell by a long, narrow pore.

"The secret of AQP1's specificity is two-fold: it selects for size and for chemical nature," Jap says. "There is a very narrow constriction in the pore, which admits no molecule bigger than water. To keep out molecules smaller than water there is also a chemical filter, formed by the specific orientation and distribution of the amino acid residues lining the pore."

Molecules attempting to enter the channel are chemically bound to water molecules, which are stripped away as the pore narrows; charged species are therefore left with net electrical charge. "The filter strongly rejects charged molecules or ions, even as small as single protons," Jap explains.

The unique distribution of amino acid residues along the pore wall also accounts for the channel's ability to move water quickly, explains Peter Walian, a member of the team that solved the structure. "It's a schizophrenic environment, half hydrophilic and half hydrophobic"--that is, half water-loving and half water-fearing. "Water molecules readily get in because of the hydrophilic sites, but the hydrophobic regions prevent them from binding too frequently."

Thus water and only water flows freely in and out of the cell through AQP1's pores, the direction of flow depending only on changing relative pressure inside and outside the cell. "It's a beautiful mechanism," Walian remarks. "It's remarkable that nobody thought of it before now."

"This is what structural biology is for," Jap says. "It shows us how extremely simple nature's solutions can be."

-- Paul Preuss