The Floodgates Open for Aquaporins

WET SERENDIPITY

From the start, aquaporins have defied expectations. Agre and his group originally set out to purify an unrelated protein from red blood cells, but their attention was caught by a mysterious, highly abundant protein contaminant, which turned out to be AQP1. Because researchers used the red blood cell extensively as a model experimental system, the discovery of a major new protein in its membrane was highly unexpected, he says. After purifying it and cloning the cDNA, their first clue that it might be a channel was its amino acid sequence, which predicted six membrane-spanning α-helices. Moreover, sequence homology showed that it was related to MIP26, a major lens protein involved in fluid uptake, as well as to a protein in plants whose expression is induced by water deprivation.

Based on these clues, Agre and his group set out to determine whether they had isolated a water channel, using a somewhat unusual experimental system: the exploding Xenopus oocyte. Normal oocytes survive quite well in a low-salt environment. But the presence of a water channel in the oocyte membrane spells trouble, because water can then rush down the osmotic gradient into the high-salt oocyte interior. Eventually, the oocyte swells and bursts due to water influx. Agre's group injected oocytes with mRNA transcribed from their cDNA clone, and numerous videos of badly damaged oocytes confirmed their suspicions that they had identified a water channel protein.³

STRUCTURE OF A SIEVE

These newly discovered channels displayed some intriguing properties. For example, while highly permeable to water, aquaporins prevent other small molecules, such as urea, ions, or even protons, from passing through. The desire to understand how such rapid permeation could be coupled with such high specificity led to a "friendly rivalry" among structural biologists to solve structure, says Bing Jap, Lawrence Berkeley National Laboratory. Two high-resolution versions resulted. Robert Stroud's group at the University of California, San Francisco, crystallized the glycerol facilitator, GlpF, which transports both glycerol and water, from Escherichia coli.⁴ A year later, Jap's team produced a high-resolution structure for AQP1.⁵ Most recently, Stroud's group has published the structure of AQPZ, also from E. coli.⁶

These structures show that aquaporins consist of a right-handed bundle of six highly tilted, membrane-spanning α-helices. The pore's hour-glass shape has wide extracellular and intracellular vestibules, joined by a long narrow region that acts as the selectivity filter. In native tissues, four aquaporin monomers team up to form a tetramer, but each monomer contains a separate water pathway.

The high-resolution AQP1 structure also contained a bonus: four almost perfectly resolved water molecules caught in the act of traversing the pore. According to Jap, water is not usually visible in protein crystal structures unless it is very tightly bound. These waters were thus bound in a highly specific fashion that reflected the channel's mechanism. "From the structure itself, you already know exactly how the channel works," he says.

The pore's high selectivity arises from two features. First, it is very narrow, excluding all but the tiniest molecules, and ensuring that ions cannot enter unless they shed their bulky hydration shells. Second, it is highly hydrophobic, with the exception of a few strategically placed hydrophilic sites where water molecules can form hydrogen bonds. The relatively long distances between these hydrophilic sites prevent ions, including protons, from traversing the pore's entire length. This high hydrophobicity also provides rapid permeation, since it discourages water molecules from lingering inside for very long, says Jap.

SUBSTRATES, SUBSTRATES EVERYWHERE

Perhaps the most controversial proposed new role for aquaporin is Yale University researcher Walter Boron's* finding that AQP1 can conduct CO₂.⁷ In a scenario reminiscent of aquaporin's original discovery, other investigators have questioned the need for a specific CO₂ channel, since most membranes are highly gas permeable. Boron admits that under normal conditions, CO₂ membrane permeation is not "a limiting factor" in red blood cells where AQP1 resides. However, he points out that CO₂ permeation could become a limiting factor under stress conditions, such as during peak athletic performance or for those living at high altitudes. These ideas have been given credence by the recent finding that the tobacco aquaporin, NtAQP1, also conducts CO₂.⁸ "Photosynthesis is CO₂-limited," he points out, supporting a physiological relevance for this function.

Bert de Groot, who studies aquaporin mechanisms together with colleague Helmut Grubmüller at the Max Planck Institute for Biophysical Chemistry, Göttingen, Germany, writes in an E-mail: "Based on its size and physiochemical properties, it is well conceivable that CO₂ might permeate aquaporin [-like] pores." He further suggests that the great need for CO₂ in plants might then explain why

"plant genes encode for hundreds of aquaporins, whereas mammals only have something on the order of 10."

Ventures into this bustling city of membrane proteins already have revealed surprises. Further exploration may show that these multipurpose channels are even more pervasive. As Harvard's Brown remarks, "Perhaps the name aquaporin was a little hasty." Only time will tell.

Megan Stephan dr.meganstephan@att.net is a freelance writer in Cheshire, Conn.

*The author performs research in a Yale University laboratory that is not affiliated with Dr. Boron's work.