Você está na página 1de 34

Annu. Rev. Biochem. 1999. 68:425458 Copyright 1999 by Annual Reviews. All rights reserved.

CELLULAR AND MOLECULAR BIOLOGY OF THE AQUAPORIN WATER CHANNELS


Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Mario Borgnia,1 Sren Nielsen,2 Andreas Engel,3 and Peter Agre1


1Departments of Biological Chemistry and Medicine, Johns Hopkins School of Medicine, Baltimore, Maryland 21205-2185; 2Department of Cell Biology, Institute of Anatomy, University of Aarhus, Aarhus, Denmark; 3Maurice-Mller Institute, University of Basel, Basel, Switzerland; e-mail: pagre@jhmi.edu

Key Words
I Abstract

water transport, channel structure, gene family

The high water permeability characteristic of mammalian red cell membranes is now known to be caused by the protein AQP1. This channel freely permits movement of water across the cell membrane, but it is not permeated by other small, uncharged molecules or charged solutes. AQP1 is a tetramer with each subunit containing an aqueous pore likened to an hourglass formed by obversely arranged tandem repeats. Cryoelectron microscopy of reconstituted AQP1 membrane crystals has revealed the three-dimensional structure at 36 . AQP1 is distributed in apical and basolateral membranes of renal proximal tubules and descending thin limbs as well as capillary endothelia. Ten mammalian aquaporins have been identied in water-permeable tissues and fall into two groupings. Orthodox aquaporins are waterselective and include AQP2, a vasopressin-regulated water channel in renal collecting duct, in addition to AQP0, AQP4, and AQP5. Multifunctional aquaglyceroporins AQP3, AQP7, and AQP9 are permeated by water, glycerol, and some other solutes. Aquaporins are being dened in numerous other species including amphibia, insects, plants, and microbials. Members of the aquaporin family are implicated in numerous physiological processes as well as the pathophysiology of a wide range of clinical disorders.

CONTENTS Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426 Discovery of Aquaporin-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427 Measurement of Membrane Water Permeability . . . . . . . . . . . . . . . . . . . . . . . . 428 AQP1 Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430 Tissue Distribution of AQP1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 434 Mammalian Aquaporin Homologs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437

0066-4154/99/0701-0425$08.00

425

426

BORGNIA ET AL

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Orthodox Aquaporins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437 Vasopressin-Regulated AQP2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437 Low-Permeability AQP0 and AQP6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 441 Brain AQP4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 441 Apical Membrane AQP5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443 AQP8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443 Aquaglyceroporins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 444 Multifunctional AQP3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 444 AQP7 and AQP9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445 Complex Tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446 Nonmammalian Aquaporins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448 Amphibian and Insect Aquaporins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448 Plant Aquaporins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448 Microbial Aquaporins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449 Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 452

INTRODUCTION
Because water is the major component of all living cells, the ability to absorb and release water must be considered a fundamental property of life. Cell membranes are exquisitely selective barriers that control the solute composition of the enclosed compartments by regulating the entry of ions; small, uncharged solutes; and water into cells. In addition, cells organized in epithelial tissues have apical and basolateral membranes, which constitute serial barriers that regulate the transepithelial movement of solutes and water, thereby contributing to the homeostasis of multicellular organisms. The identication and characterization of the molecular entities responsible for the function of biological membranes have been a goal of investigators over the past half century. Electrophysiologic measurements of cells and isolated patches of membrane revealed the properties of ion channels, which are expressed in extremely low copy numbers. Biochemical purication and functional reconstitution techniques contributed greatly to the understanding of active transport, such as the Na+/K+-ATPase for which the 1997 Nobel Prize in Chemistry was awarded to Jens Christian Skou of the University of Aarhus. Nevertheless, purication procedures are vulnerable to loss of activity owing to protein denaturation or disassociation of subunits. More recently, development of cloning techniques and expression systems has permitted direct assessment of the functions of transporters and channels in biological membranes. At the beginning of this decade, cDNAs encoding numerous transporters had already been isolated, but although expression of many pumps, carriers, exchangers, and channels was possible, the molecular identities of membrane water transporters remained unknown. Water is known to diffuse through lipid bilayers, so all membranes are at least somewhat water permeable. Until recently, a widely held misconception assumed that simple diffusion accounts for all water movement through biological membranes and that specic water channels are not necessary. Nevertheless, detailed

AQUAPORIN WATER CHANNELS

427

observations from multiple laboratories indicated that specialized membrane water transport molecules must exist in tissues with intrinsically high water permeability [reviewed by Finkelstein (1)]. The basal water permeability of red cell membranes is much higher than that observed for other cell types and articial lipid bilayers. The activation energy of this process (Ea5 kcal mol-1) is equivalent to the diffusion of water in solution [reviewed by Solomon (2)], indicating that water moves through aqueous pathways across these membranes. Reversible inhibition by HgCl2 and a subset of organomercurials suggested the proteinaceous nature of the pathway [reviewed by Macey (3)]. Further evidence of protein involvement in water transport was provided by the observation that some epithelial tissues exhibit changes in water permeability on a timescale that is not compatible with changes in lipid composition. Because of its relevance to human disease, water transport is best dened in the mammalian kidney, where the nephron is the functional unit. Water permeability has been well characterized in all nephron segments [reviewed by Knepper et al (4)]. Renal proximal tubules and descending thin limbs of Henles loop are known to have constitutively high water permeability, allowing for the reabsorption of ~150 liters per day in the average adult human. The ascending thin limbs and thick limbs are relatively impermeable to water and empty into renal collecting ducts. This tissue is extremely important in clinical water balance, because collecting-duct water permeability is low unless stimulated with vasopressin (antidiuretic hormone). Toad urinary bladder serves a function similar to the collecting duct in amphibians and was used as an early model of vasopressinregulated water permeability. Stimulation of the basolateral membrane of this epithelium with antidiuretic hormone produces an increase in water permeability in the apical membrane, which coincides with the redistribution of intracellular particles (aggrephores) to the cell surface (5, 6). Attempts to ascribe water permeability to known membrane proteins, to isolate putative water channel proteins from native tissues, or to isolate water channel cDNAs by expression cloning were all unsuccessful [reviewed by Agre et al (7)]. This may be explained by the ubiquity of water, a simple molecule that is obviously not amenable to chemical modication such as introduction of chemical cross-linking groups or labels. While HgCl2 inhibits membrane water channels, the agent reacts with free sulfhydryls in many proteins, and no specic inhibitors are known. In addition, the relatively high diffusional permeability of membrane bilayers creates a high background permeability, which has apparently frustrated efforts to clone by functional expression.

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

DISCOVERY OF AQUAPORIN-1
The recognized characteristics of membrane water channels led to chance identication of the rst known water channel. In the process of isolating the 32-kDa bilayer-spanning polypeptide component of the red cell Rh blood group antigen, a 28-kDa polypeptide, which stained poorly with Coomassie, was initially assumed

428

BORGNIA ET AL

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

to represent a proteolytic fragment of Rh (8). Attempts to raise antiserum in rabbits resulted in a strong immunoreaction exclusively with the 28-kDa band, which prompted its biochemical characterization (9). Initial studies demonstrated that the protein is composed of hydrophobic amino acids and exists in two forms: a nonglycosylated 28-kDa polypeptide and a 40- to 60-kDa N-glycosylated polypeptide. The unusual detergent in solubility of the 28-kDa band that remained in the pellet when red cell membrane vesicles were solubilized in Nlauroylsarcosine facilitated purication and biochemical characterization. Most notable were the observations that the core polypeptides of the 28- and 40- to 60kDa bands are identical and exist as an oligomer with physical dimensions of a tetramer, the N- and C-termini are intracellular, and the N-terminal amino acid sequence is unrelated to the Rh polypeptide (10). This sequence permitted cDNA cloning from an erythroid library (11) and recognition that the deduced amino acid sequence is related to a functionally undened family of putative membrane channel proteins, including major intrinsic protein (MIP) of lens (12). Also of note, radiation inactivation studies of water permeability by renal vesicles yielded a target size of 30 kDa (13). Because the 28-kDa polypeptide was found to be abundant in red cells and renal proximal tubules and descending thin limbs (9), it was suggested by the late John C. Parker of the University of North Carolina at Chapel Hill that this protein may be the sought-after water channel. Although this protein was rst known as CHIP28 (channel-like integral protein of 28 kDa), the need for a functionally relevant name was recognized. The name aquaporin was rst conceived at an informal gathering in Baltimore. After recognition of related proteins with similar functions, this name was formally proposed for the emerging family of water channels now known as the aquaporins (7). Now designated aquaporin-1 (AQP1), the Human Genome Nomenclature Committee has embraced this nomenclature for all related proteins (14).

Measurement of Membrane Water Permeability


Unlike ion conductances, which may be measured electrophysiologically, or solute transport, which may be measured with radioactive substrates, the movement of water across cell membranes posed a unique experimental challenge. Karl Windhager of Cornell Medical School rst proposed the Xenopus oocyte expression system to search for water channel RNAs, because these cells are known to exhibit remarkably low membrane water permeability (15, 16). When oocytes injected with cRNA for the 28-kDa polypeptide were analyzed (17), they exhibited remarkably high osmotic water permeability (Pf ~200 10-4 cm s-1) and rapidly exploded in hypotonic buffer, whereas control oocytes injected with water alone exhibited less than one-tenth this permeability (Figure 1). The oocyte studies demonstrated that AQP1 behaved like water channels in native cell membranes (17). The osmotically induced swelling of oocytes expressing AQP1 exhibited low activation energy and was reversibly inhibited by HgCl2. Moreover, AQP1 oocytes failed to demonstrate any measurable increase in ion conductance. Although these early studies demonstrated swelling of oocytes, it was predicted that the water permeation of

AQUAPORIN WATER CHANNELS

429

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

AQP1 is actually bidirectional, with the direction of water ow determined by the orientation of the osmotic gradient. This has been recently demonstrated with a highly rened oocyte system that permits rapid change of extracellular buffer, thereby permitting swelling or shrinking of AQP1 oocytes (18). To conrm that the interpretation of the oocyte studies was correct, highly puried AQP1 protein from human red cells was reconstituted with pure phospholipid into proteoliposomes, which were compared with simple vesicles (liposomes) by rapid transfer to hyperosmotic buffer (19) while the change in volume was measured by monitoring quenching of internal carboxyuorescein. These studies led to an estimate of the unit water permeability (pf ~3 109 water molecules subunit-1 s-1). Moreover, the water permeability is reversibly inhibited by HgCl 2 and exhibits a low activation energy (Ea5 kcal mol-1). Several of these studies have been conrmed by using red cell membranes partially depleted of other proteins (20), and attempts to demonstrate permeation by other small solutes or even protons showed AQP1 to be water selective (21). Together these studies indicated that AQP1 is both necessary and sufcient to explain the wellrecognized membrane water permeability of the red cell.

Figure 1 Water permeability of aquaporin-1 expressed in Xenopus oocytes. (left) When transferred to hypo-osmolar buffer for 2 min, control oocytes exhibit negligible water permeability. (right) Under the same conditions, oocytes previously injected with aquaporin-1 (AQP1) cRNA rapidly swell and explode. Reproduced and modied with permission (175).

430

BORGNIA ET AL

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Although most workers agree that AQP1 is a constitutively active, waterselective pore that permits movement of water in the direction of higher osmotic strength, some investigators have questioned this. Permeation by glycerol has been seen in preshrunk oocytes and may represent opening of an unidentied leak pathway (22). In support of this, a rened oocyte system has been developed to measure shrinkage or swelling of oocytes and conrmed a small degree of glycerol permeation, although the biological signicance remains unclear (18). A different group of investigators reported that forskolin induced a cation current in AQP1 expressing oocytes (23); however, several other scientic groups have failed to reproduce this effect (24). Small changes in water permeability by oocytes expressing a bovine homolog of AQP1 have also been ascribed to vasopressin and atrial natriuretic peptide, but the signicance is uncertain (25). Likewise, secretin-induced membrane trafcking has been noted in isolated cholangiocytes (26), but this awaits conrmation by immunoelectron microscopy. Recently, the permeation of AQP1 by CO2 has been evaluated. Rates of pH change are about 40% higher in oocytes expressing AQP1 (27) than in control oocytes. Although the background permeation of lipid bilayers by carbon dioxide, oxygen, nitric oxide, and other gases may be high, the potential physiological relevance of AQP1 permeation by gases warrants more study (28). Permeation of AQP1 by methanol and ethanol is also suspected; however, the high lipid permeation by each of these agents has so far precluded meaningful transport studies (BL Smith, LS King & P Agre, unpublished study). Thus, although the evidence that AQP1 functions as a water channel is incontrovertible, the possibility of yet undiscovered transport functions cannot be excluded.

AQP1 Structure
The availability of pure AQP1 protein in milligram quantities and the simple functional assay in oocytes led to rapid advances in the understanding of the molecular structure of AQP1. Hydropathy analysis of the deduced amino acid sequence of AQP1 led to the prediction that the protein resides primarily within the lipid bilayer (11), a feature in agreement with initial studies of red cell AQP1 (9, 10). As previously described for the homolog major intrinsic protein from lens (MIP, now referred to as AQP0), the polypeptide contains an internal repeat, the N- and the C-terminal halves are sequence related, and each contains the signature motif Asn-Pro-Ala (NPA) (29, 30). When evaluated by hydropathy analysis, six bilayer-spanning domains are apparent; however, loops B and E exhibit signicant hydrophobicity. Critical to the topology is the location of loop C, which connects the two halves of the molecule. By studying truncated fragments of the protein, other investigators proposed that AQP1 has four bilayer-spanning domains (31), but this has subsequently been repostulated as possibly representing an immature form of the protein (32). A system was adapted for analyzing the AQP1 structure after minimally perturbing the molecule by adding BamHI restriction sites into the cDNA for insertion

AQUAPORIN WATER CHANNELS

431

of DNA encoding a 29-residue epitope of the E1 coronavirus, as suggested by Carolyn Machamer (personal communication). Most of the recombinants were functional. Using anti-E1 and vectorial proteolysis demonstrated that loop C resides at the extracellular surface of the oocytes, conrming the obverse symmetry of the Nand C-terminal halves of the molecule (33; Figure 2). Several other studies of AQP1 expressed in oocytes led to the observations that Cys-189 is the site of mercurial inhibition (34), which has been conrmed by other workers (35). Thus, loop E has been implicated as a structural component of the aqueous pathway. Further pursuit of this line of inquiry led to the recognition that loops B and E were functionally essential for water permeability, leading to the hourglass model (36), in which these domains overlap midway between the leaets of the bilayer, creating a constitutively open, narrow aqueous pathway (Figure 3). Although the oligomerization of AQP1 is still not understood in detail, all of the studies indicate that the protein is a tetramer composed of functionally independent aqueous pores (36, 37). In conjunction with the studies noted above, analyses of AQP1 protein from human red cell membranes provided detailed molecular insights. Rotary and unidirectionally shadowed freeze fracture electron microscopic analyses of reconstituted AQP1 conrmed the proposed tetrameric assembly of AQP1 (21, 38). Initial spectroscopic studies of the puried, reconstituted AQP1 were interpreted as showing that the protein was approximately half alpha helix and half beta structure (39), leading to the prediction that AQP1 is a beta barrel structure (40). This has been contradicted by other investigators, who used attenuated total reection Fourier transform infrared (FTIR) spectroscopy of highly puried red cell AQP1 reconstituted into membrane crystals. Direct comparison was made to bacteriorhodopsin in purple membranes and reconstituted porin OmpF, because the structure of these proteins had already been dened at high resolution (41). These studies demonstrated the lack of beta structure in AQP1 and indicated the existence of alpha helices tilted at 2127. Recently, uorescently labeled antibody specic for the extracellular domain of loop A (Co antigen) was used to measure the rate of lateral mobility of the AQP1 molecule in red cell membranes after photobleaching and the rate of redistribution after microdeformation (42). The studies indicate that the lateral mobility of AQP1 in red cell membranes is regulated by passive steric hindrance from the membrane skeleton, which provides signicant constraint in the unperturbed cell membranes and is released when freed from the membrane skeleton. By reconstituting the highly puried red cell AQP1 into membranes under controlled conditions, membrane crystals were produced with AQP1 in highly uniform lattices with p4221 symmetry. The reconstituted membranes appeared as at sheets or as large, resealed vesicles in which the AQP1 protein was found to fully retain water permeability when measured by quenching of carboxyuorescein uorescence after rapid change of osmolality (43). Thus the opportunity to dene the structure of AQP1 in a biologically active state became possible. These and other electron microscopic studies by multiple groups rapidly permitted the elucidation of the protein at increasing levels of resolution. By

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

432

BORGNIA ET AL

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Figure 2 Membrane topology of aquaporin-1 subunit. The Asn-Pro-Ala signature motifs are indicated (NPA). Shaded areas represent sequence-related repeats in obverse orientation. Sites of N-linked glycan and surface polymorphism are represented. Reproduced and modied with permission (52.).

performing high-resolution electron microscopic evaluation of negatively stained membranes at a series of tilts, a preliminary three-dimensional view was obtained that showed one side of the tetramer to contain subunits widely separated around a central stain-lled depression (the extracellular surface) or closely surrounding a central area (cytoplasmic surface) (44). By cryoelectron microscopy, higher-resolution images were achieved revealing the presence of multiple bilayer-spanning domains (4547). Removal of surface carbohydrate and freeze-drying/metal shadowing electron microcopy as well as atomic force microscopy further dened the orientation and extramembrane dimensions of AQP1 (48). Electron crystallography of cryopreserved specimens at extremely low temperatures using a liquid helium-cooled stage and tilts of 60 revealed the presence of AQP1 tetramers with individual subunits

AQUAPORIN WATER CHANNELS

433

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Figure 3 Hourglass model for aquaporin-1 membrane topology. (top) Each aquaporin-1 (AQP1) subunit contains six bilayer-spanning domains composed of two obversely symmetrical structures (TM1-3, hemipore 1, and TM4-6, hemipore 2). (bottom) When NPA motifs in loops B and E are juxtaposed, they form a single aqueous channel spanning the bilayer (the hourglass) anked by the mercury-sensitive residue (C189). Reproduced and modied with permission (36).

434

BORGNIA ET AL

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

containing six bilayer-spanning alpha helices viewed with resolution of 6 (49). Moreover, this approach partially resolved the extracellular and intracellular connecting loops. Together with the surface topography recorded by atomic force microscopy, this dened the subunit as a right-handed helix bundle, which was then novel among membrane proteins. These studies have permitted threedimensional visualization of the AQP1 subunit, which contains an intrasubunit structure that is strikingly similar to the proposed hourglass (Figure 4). Other investigators have achieved similar observations (50, 51). Although the handedness of the bundle was initially disputed and the limitations on tilting precluded denition of the connecting loops, a general agreement now exists regarding the structure of AQP1. The current challenge is to merge structural understandings derived from molecules with unambiguous landmarks [reviewed by Heymann et al (52)]. It is plausible that the structure of aquaporins will be understood at levels comparable to that for bacteriorhodopsin.

Tissue Distribution of AQP1


Well before recognition of its function, the red cell AQP1 protein was known to be expressed at high levels in the proximal convoluted tubules and descending thin limbs of kidney (9). This observation was conrmed with polyclonal rabbit antiserum (53) and was dened in rat and human kidney with afnity-puried immunoglobulin specic for the N- and C-terminal domains of AQP1 (54, 55; Figure 5). In all studies, AQP1 is constitutively present in the apical plasma membranes (including the brush borders) and in basolateral membranes as well. Quantitation of AQP1 at these sites veried that AQP1 alone could serve as the conduit for the huge volumes of water reabsorbed in the proximal nephron (56). Other immunohistochemical studies have demonstrated AQP1 in multiple capillary endothelia throughout the body (5759), including the renal vasorecta (60). AQP1 is also abundant in peribronchiolar capillary endothelium, where expression is induced by glucocorticoids (58, 61) acting through the classic glucocorticoid response elements in the proximal gene promoter (62). In addition, AQP1 has been dened in multiple water-permeable epithelia including choroid plexus (site of cerebrospinal uid secretion) at multiple eye tissues, including ciliary epithelium, lens epithelium, and corneal endothelium (57, 63, 64), and hepatobiliary epithelium (65). Developmental expression of AQP1 at these sites has been found to be complex in the rat model, with transient expression in some tissues prior to birth, expression in other tissues subsequent to birth, and constitutive life-long expression in still other tissues (61, 66, 67). Although AQP1 was initially considered to be an essential gene product, humans were identied who totally lack the protein. Owing to their linkage to human chromosome 7p14, the idea emerged that the Co blood group antigen and the AQP1 protein are the same, and Co was shown to be an Ala/Val polymorphism at the extracellular surface of red cell AQP1 (68; Figure 2). The exceedingly rare Co null individuals were then found to have mutations in the AQP1 gene (69). These Co null individuals are all women who developed

AQUAPORIN WATER CHANNELS

435

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Figure 4 Stereoscopic view of aquaporin-1 at 6 resolution as established by cryoelectron microscopy. (top) Six tilted, bilayer-spanning a helices (numbered) in a single subunit when viewed from inside the tetramer. Loop B (green) and loop E (orange) are indicated. (bottom) Same as top when viewed from opposite side (180o). Modied and reproduced with permission (49).

anti-Co during pregnancy. Although the exceedingly rare blood group phenotype makes them impossible to match for blood transfusion, it was surprising that none of them exhibited any other obvious clinical phenotype, and it is uncertain whether this is explained by naturally occurring backup systems or the existence of compensating mutations. The recent development of an AQP1

436

BORGNIA ET AL

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Figure 5 Light- and electron-microscopic study of rat tissues stained with antiaquaporin-1. (top) Proximal tubule is shown. (bottom) Descending thin limbs of Henles loop are shown. Lumen (LU) and basement membrane (BM) are indicated.

AQUAPORIN WATER CHANNELS

437

gene knockout has revealed that renal water reabsorption in mice is very dependent upon AQP1 protein, because the AQP1 null animals became hyperosmolar after uid restriction (70). Detailed classical physiological evaluations of the isolated proximal tubules from the AQP1 null mice were undertaken to establish that transmembrane water permeability was reduced by 80% and led to the observation of a compensatory reduction in glomerular ltration (71). Mice tolerate the defect well until deprived of water. Evaluation of AQP1 null humans is now being planned.

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

MAMMALIAN AQUAPORIN HOMOLOGS


Just as the Rosetta Stone permitted translation of ancient Egyptian hieroglyphics into European languages, the rst functional denition of one member of a protein family often reveals the functions of the other members. This has certainly been the case for the aquaporins, and the homology cloning approach is now being undertaken by multiple laboratories which have aimed their efforts at expanding the list of members of the aquaporin family. This is most frequently undertaken by amplication with the polymerase chain reaction using degenerate oligonucleotide primers (72). So far, all members of the family are believed to be water channels, all contain structural motifs similar to AQP1, but each has structural features apparently needed for functional and regulatory differences. At least 10 mammalian aquaporins are now known (Figure 6), and they form at least two subgroups: water-selective channels (orthodox aquaporins) and channels permeated by water, glycerol, and other small molecules (aquaglyceroporins). Given the large potential for confusion, the Human Genome Organization has established an Aquaporin Nomenclature System for designation of novel family members (14), and updates are available on the Internet (URL: http://www.gene.ucl.ac.uk/nomenclature).

ORTHODOX AQUAPORINS Vasopressin-Regulated AQP2


Renal collecting ducts are known to have water permeability that is regulated by vasopressin. Absence of AQP1 from collecting ducts (9) predicted the existence of additional members of the aquaporin family (17). By using the homology cloning approach, Fushimi et al (73) isolated the cDNA encoding a renal collecting-duct water channel, now designated as AQP2. The combined efforts of several groups established the major signicance of this homolog [reviewed by Knepper & Inoue (74)]. Expression of AQP2 is restricted to the principal cells of the renal collecting duct (73, 75, 76). Unlike AQP1, which is constitutively present at the plasma membrane, AQP2 appears to be regulated by vasopressin through short-term exocytosis to the plasma membrane, as well as long-term biosynthetic mechanisms.

438

BORGNIA ET AL

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Figure 6 Phylogenetic comparisons of aquaporins. (top) Mammalian aquaporins and Escherichia coli AqpZ and GlpF are shown. (bottom) Compendium of microbial aquaporins is shown with reference to mammalian aquaporins and aquaglyceroporins. Generated using ClustalW (177).

Vasopressin has long been known to cause the redistribution of intracellular vesicles to the apical surface of principal cells, the membrane shuttle mechanism (77). This hypothesis was supported by immunolocalization of AQP2 in sections of rat kidney after treatment with vasopressin (76, 78, 79). Proof for the hypothesis was achieved in a classic study of isolated rat renal collecting ducts by immunoelectron microscopy (Figure 7). In response to vasopressin, intracellular vesicles containing

AQUAPORIN WATER CHANNELS

439

AQP2 were found to redistribute to the cell surface, and, after removal of the agent, AQP2 reappeared in intracellular vesicles; corresponding increases and reductions of water permeability were measured in the isolated tubules (80). The mechanism of vasopressin action is through a basolateral membrane V2 receptor coupled to activation of adenylyl cyclase and phosphorylation of the C-terminus of the AQP2 protein. The vesicles are then targeted to the apical membrane via the vesicle-targeting proteins syntaxin-4 and synaptobrevin-2. Although the molecular details of this pathway are being evaluated (81), it is still unclear whether phosphorylation of sites on all four subunits of AQP2 is needed or whether other proteins in the vesicle such as the urea transporters will be carried along for the ride [reviewed by Knepper & Inoue (74)]. Participation of a heterotrimeric member of the Gi protein family is required for membrane trafcking in renal epithelial cells (82). The long-term mechanisms for regulation of AQP2 biosynthesis and removal are also being evaluated. Identication of a cyclic-AMP regulatory element in the 5 anking DNA of AQP2 supports a role for transcriptional regulation (83, 84). Downregulation in the postantidiuretic state is not well understood, but escape from vasopressin is known to involve reduced expression of AQP2 protein (85). The clinical importance of AQP2 has also attracted increasing interest, and it is generally believed that the AQP2 protein may be involved in most imbalances of water metabolism. Clinical problems featuring impaired renal water reabsorption are associated with reduced AQP2. Diabetes Insipidus (DI) results from inadequate levels of vasopressin and leads to secretion of large volumes of dilute urine, even as the individual becomes dehydrated. Nephrogenic Diabetes Insipidus (NDI) is a disorder that occurs when vasopressin levels are not reduced, but the kidney fails to respond to the hormone; mutations in genes encoding renal vasopressin V2 receptors have been demonstrated in many patients with the X-linked disorder [reviewed by Fujiwara et al (86)]. A family with recessively inherited NDI and normal V2 receptors was found to have functionally disruptive mutations in the AQP2 gene (87). Multiple AQP2 mutations in this series of patients with recessively inherited NDI have been shown to correspond to residues in the aqueous pore domains (88), and abnormal trafcking has been found in cell cultures to improve with chemical agents believed to work as chaperones (89). A single family with dominantly inherited NDI was recently identied with a mutation in the C-terminus of AQP2 which cooligomerizes with the product of the normal allele, thereby restricting membrane trafcking of both polypeptides (90). Thus multiple distinct mechanisms have been dened as causes for different forms of this disease. Secondary decreases in AQP2 expression may be common. Bipolar disorder (also known as manic depressive disorder) is a potentially serious form of mental illness that affects ~1% of the US population but is known to be more frequently encountered among individuals with marked creativity (including biochemists and other research scientists) (90a). Lithium is widely prescribed for treatment of bipolar disorder, but polyuria is a problematic side effect. This side-effect is explained by a striking reduction (90%) in AQP2 expression caused by lithium

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

440

BORGNIA ET AL

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Figure 7 Immunogold electron microscopy of aquaporin-2 in principal cells from isolated rat kidney-collecting ducts perfused in the presence (top) or absence (bottom) of vasopressin (AVP).

treatment as measured in rats (91). Other forms of polyuria have also been evaluated, and reduced AQP2 expression was also documented after reversal of urinary obstruction, a situation often complicating prostate surgery (92), in hypokalemiainduced polyuria (93), as a consequence of the nephrotic syndrome, and is suspected in nocturnal enuresis (94).

AQUAPORIN WATER CHANNELS

441

Secondary increases in AQP2 expression may also be common in clinical problems with excessive renal water reabsorption. Congestive heart failure is a serious complication of coronary vascular disease, and patients often succumb to refractory pulmonary edema caused by the retention of water. Two different research groups have independently identied increased expression of AQP2 in rat models of congestive heart failure (95, 96). Increased expression of AQP2 may also explain uid retention known to complicate pregnancy (97), as well as uid retention in cirrhosis and the syndrome of inappropriate antidiuretic hormone (98). Although not yet conrmed, AQP2 may be an effective target for pharmacologic intervention in some of these disorders.
Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Low-Permeability AQP0 and AQP6


Before the molecular discovery of AQP1, the homolog major intrinsic protein of lens, MIP (now referred to as AQP0) was cloned but not functionally dened (12). More recently, this polypeptide was shown to confer water permeability that is an order of magnitude lower than that of AQP1 and is resistant to inhibition by mercurials (99101). AQP0 is expressed exclusively in lens ber cells, where it is suspected to play a role in the removal of water. Naturally occurring AQP0 mouse mutants known as cat mice suffer congenital cataracts as a dominantly inherited disorder (102). Another sequence-related protein, AQP6, has been recognized in the kidney (103) but has not been functionally dened. Like AQP0, AQP6 exhibits extremely low inherent water permeability. Recently, AQP6 has been shown to reside within intracellular vesicles in collecting-duct intercalated cells where it is believed to participate in acid secretion (104). The curiously low water permeability exhibited by AQP0 and AQP6 is unexplained and may reect the need for an activation step.

Brain AQP4
Other aquaporins have been identied and are being dened by their sites of expression; AQP4 is the predominant member in brain (105, 106). Residing at the perivascular margin of astroglial cells (Figure 8), AQP4 may function as an exit port for excess brain water, which can be lethal in cerebral edema (107). AQP4 is also present in glial lamellae surrounding vasopressin secretory neurons, where it has been suggested to function as an osmoreceptor (107), and in ependymal cells lining the cerebrospinal uid lled cavities (108). In addition, the distribution of AQP4 has been dened in retina and optic nerve, where the protein is particularly abundant in Mller cell end feet adjacent to the vitreous body and vascular endothelium (109). Recently, AQP4 has been demonstrated in fast-twitch skeletal muscle in rats (110). Because fast-twitch bers accumulate high concentrations of lactate, the presence of AQP4 may permit the rapid water ux needed to restore osmotic equilibrium. Of particular interest is the unexplained observation that AQP4 is reduced in fast-twitch bers in a mouse model of Duchennes muscular dystrophy (110).

442

BORGNIA ET AL

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Figure 8 (top) Immunogold electron microscopy of rat cerebellum cells reacted with anti-aquaporin-4. (bottom) Stereoscopic image of astrocyte end-feet in freeze fracture replica; reproduced with permission (115). Figure provided by J Rash, Colorado State University.

AQUAPORIN WATER CHANNELS

443

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Unlike the other aquaporins, AQP4 exists in two forms owing to the alternate use of two translation initiation sites (111). Also unlike most aquaporins, AQP4 is not inhibited by mercury (105, 106) and was reported to have a higher unit permeability (112). Disruption of the mouse AQP4 gene has been reported to result in a defect in renal concentration (113). Based on their apparent absence in tissues from AQP4 null mice, it has been proposed that AQP4 is the molecular basis of square arrays within astroglial membranes (114). Direct anti-AQP4 labeling of freeze-fracture replicas from brain and spinal cord tissues demonstrated AQP4 in the majority of square arrays (Figure 8, bottom; 115). Although it was previously reported that protein kinases A and C do not modulate the water permeability of aquaporins 05 (112), this is now being questioned, because it has recently been reported that protein kinase C activated by phorbol diesters produces a 90% reduction in water permeability by AQP4 (116). This suggests that AQP4 may be actively regulated and may reversibly open and close the blood brain barrier to the movement of water, a process which may be of signicance in various clinical problems of brain ischemia.

Apical Membrane AQP5


The fth aquaporin was isolated from salivary gland cDNA (117). AQP5 resides at the apical membranes of type 1 alveolar pneumocytes, as well as in a subset of salivary and lachrymal glands where it presumably regulates airway humidication and the release of saliva and tears (118). The expression in type 1 alveolar pneumocytes is interesting, because these cells have a large surface area, which is attened around the basement membrane surrounding the alveolar capillary epithelia, the site where water is presumed to enter the bloodstream during freshwater drownings [see review by King & Agre (119)]. Moreover, type 1 alveolar epithelial cells exhibit unusually high water permeability (120). The origin of type 1 alveolar pneumocytes is believed to be the columnar type 2 cells that secrete surfactant but lack AQP5 (61, 118). In support of this hypothesis, cultured type 2 cells exposed to keratinocyte growth factor transform into type 1 cells and express high levels of AQP5 (121). AQP5 may prove useful as an agent for gene therapy. Patients suffering from carcinomas of the tongue and larynx are often cured by surgery combined with radiation therapy. Unfortunately, these patients often become severely debilitated by radiation-induced damage to salivary glands, making it difcult to chew and swallow without aspiration. The gene encoding rat AQP5 has been engineered into an adenoviral vector, which is being evaluated as a potential gene transfer to salivary glands lacking functional aquaporins (122).

AQP8
The cDNA encoding AQP8 was isolated from testis (123) and pancreas and liver (124), but the mRNA is also found in colon, salivary glands, and other tissues. Whereas AQP8 was found by two groups to be permeable only to water (123,

444

BORGNIA ET AL

124), a similar cDNA cloned by another group encodes an aquaporin which is putatively permeable to water and urea (125). Unfortunately, the discrepancy in transport functions is not yet resolved.

AQUAGLYCEROPORINS Multifunctional AQP3


Some members of the aquaporin family are permeated by water, glycerol, and sometimes other small solutes. Using the homology cloning approach, three scientic groups isolated cDNAs encoding AQP3, a member of the aquaporin family with a somewhat different biophysical function (126128). Whereas previously cloned aquaporins were permeable only to water, AQP3 was noted to be genetically closer to the known Escherichia coli glycerol transport protein GlpF (Figure 6), and all three groups noted that AQP3 shares this feature and is also somewhat permeable to urea. Urea transport is surprising, because discovery of urea transport molecules seemingly obviates the need for a single transport molecule to carry both urea and water (129). Although one group reported that AQP3 is impermeable to water and the glycerol transport is inhibited by diisothiocyanodisulfonic stilbene, this has been corrected (112). The structural explanation for how AQP3 may permit transport of water and glycerol is not agreed upon; however, primary sequence analysis reveals specic differences in aquaglyceroporins, such as a motif (GLYY) in loop C and an aspartate residue after the second NPA motif (NPARD) instead of the motif found in orthodox aquaporins (NPARS) (Figure 9). The possibility that the molecule contains independent water- and glycerol-transporting domains is supported by selective inhibitors (130), whereas other investigators have demonstrated the existence of a single pore that permits the ow of water or glycerol (131). Although still unsettled, a single pathway for water plus glycerol seems more likely to be correct. Nevertheless, it is clear that AQP3 is selectively permeated by glycerol and water, because the rened oocyte system that permits highly sensitive determination of swelling and shrinking has shown that AQP3 is markedly permeable to glycerol but not urea (18). The potential clinical signicance of AQP3 is also not yet understood. Nevertheless, the abundant expression of this polypeptide at the basolateral membranes of principal cells in the collecting duct suggests a role in renal water reabsorption (108, 132, 133). It should be noted that, although studies of AQP2 have led to its recognition as the vasopressin-regulated aquaporin, AQP3 may also be regulated at a biosynthetic level quite similar to AQP2. AQP3 is also expressed in lower levels at multiple other sites, but its presence in airways is considerable (108). AQP3 is strikingly abundant in nasopharyngeal epithelium, and a role in mucosal secretions and allergic rhinitis is suspected (61, 118). In addition, low levels of AQP3 have been detected in red blood cells where it may facilitate the uptake of glycerol utilized for long-term cryopreservation (134).
Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

AQUAPORIN WATER CHANNELS

445

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Figure 9 Sequence alignment of human aquaporin-1 and aquaporin-3 compared with Escherichia coli AqpZ and GlpF. Generated using ClustalW (177).

AQP7 and AQP9


Multiple other aquaglyceroporins are now being identied (Figure 6). A cDNA encoding AQP7 was identied from mRNA found in rat testis and this gene was shown to be expressed in spermatids and seminiferous tubules (135). Presumably, the existence of AQP7 will provide a port for water and glycerol as a carbon source for mature sperm metabolism, and bifunctional aquaglyceroporin may permit the replacement of water by glycerol, as is used for long-term cryopreservation of sperm. At the same time, another group of investigators cloned a highly related cDNA from human adipose tissue which they regrettably selfdesignated as AQP8 and AQP9 (136). This sequence has been temporarily designated AQP7L by the Human Genome Nomenclature Committee, because it may represent either the human homolog of AQP7, an alternatively spliced form of AQP7, or the product of a closely related human gene. Existence of an aquaglyceroporin in fat suggests involvement in glycerol export during lipolysis, an important and novel function for this polypeptide. Another member of the family designated AQP9 has been cloned from leucocytes and is ascribed to transport urea but not glycerol (137). Curiously, another group has cloned a highly related cDNA from rat liver and found it to be permeated by a wide variety of hydrophylic solutes (138). Although it is not yet resolved, these investigators have probably reported the human and rat orthologs

446

BORGNIA ET AL

of the same gene, AQP9. The major functional differences are not yet explained, however the solute permeability studies involved extensive characterization (138), so it seems very likely that AQP9 is permeated by glycerol, urea, and a range of small uncharged osmolytes.

COMPLEX TISSUES
Organs including kidney, airways, eye, and brain have tissues with complex expression patterns involving multiple different aquaporins. Precise subcellular locations are being elucidated by high-resolution light and electron microscopy, and, generally, it is found that individual aquaporin homologs are expressed at unique intracellular sites. Because of its major importance in transport, the mammalian kidney has been studied extensively, and six aquaporins have been documented [reviewed by Knepper et al (4)]. Most data suggest that these aquaporins function together to provide transcellular water ow. In principal cells of the collecting duct, AQP2 trafcs to the apical membrane, where water enters from the lumen (Figure 7), whereas AQP3 and AQP4 reside at the basolateral membranes in the outer medulla (139) and the inner medulla (132), providing cellular exit ports to the interstitium. Multiple aquaporins reside in complex distributions within lung and airways (Figure 10). Pulmonary tissues develop a large capacity for uid absorption at the time of birth when they presumably contribute to the reabsorption of amniotic uid, humidify the airways, and generate airway secretions [reviewed by Matthay et al (140)]. To accomplish this, complex developmental expression patterns have evolved (61, 141) contributing to different phases of developmental water permeability (142). These may be explained by highly specic distribution patterns of individual aquaporins (118). For example, AQP5 is found in the apical membrane of type 1 alveolar epithelium, and AQP1 is found in the underlying capillary endothelium (Figure 10A). In the airways, AQP1 is found predominantly in peribronchiolar capillary endothelium. AQP3 is restricted to basal cells, whereas AQP4 resides exclusively in tall columnar cells reaching the surface of the airways (Figure 10B). In secretory glands, AQP5 resides in the apical surface, whereas AQP3 or AQP4 is found in basolateral membranes of some acinar cells (Figure 10C). Together, these aquaporins permit water entry through the base and exit from the surface, with the driving force provided by osmotic gradients created by solute channels and transporters. The distribution of aquaporins in the eye is also complex (143; Figure 10 D). AQP0 is present only in lens ber cells, whereas AQP1 is present in aqueous humor secretory epithelium, lens epithelium, corneal keratocytes, and endothelium. AQP3 is present in bulbar conjunctival epithelium, and AQP5 is present in lachrymal gland and corneal epithelium. AQP4 is present in complex distributions within end feet of ocular glial cells including retinal Mller cells and brous astrocytes of the optic nerve (109).

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

AQUAPORIN WATER CHANNELS

447

Respiratory Airspace

A
P2

B
AQP4

Goblet Cell

Alveolar Airspace
H2O
AQP5 P1

Surface Layer

AQP3
Basal Layer Basement Membrane

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Basement Membrane Capillary

AQP1

H2 O
conjunctiva AQP1

Fibroblasts

Capillary

C Secretory Gland
AQP3

AQP5 corneal epithelium lacrimal gland

AQP1 nonpigmented epithelium


(ciliary and iris)

lens epithelium corneal endothelium

MIP (AQP0) lens fiber cells

AQP4 H 2O
Basement Membrane

trabecular meshwork AQP3 AQP4 retinal glia (Mller cells)

AQP5

Figure 10 Representations of multiple aquaporins in complex tissues. (A) Alveolus. (B) Airway. (C) Secretory gland. (D) Eye. Reproduced with permission (118, 143).

Distinct aquaporin expressions are also found in brain tissues. AQP1 is found in spinal uid secreting epithelium of choroid plexus (57), and AQP4 may participate in reabsorption, because it resides in ependymal cells lining the ventricles. Also, the presence of AQP4 in the pericapillary membrane of astroglial cells suggests a role in elimination of water from the brain, and its presence in glial

448

BORGNIA ET AL

lamellae surrounding vasopressin-secretory membranes suggests a role in osmosensation (107).

NONMAMMALIAN AQUAPORINS
DNA sequences encoding aquaporin proteins have been found throughout nature. A selection of nonmammalian homologs and their anticipated biological signicance are summarized here.
Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Amphibian and Insect Aquaporins


Aquaporins are being identied in numerous species that slither, hop, and buzz. Amphibia exhibit special needs for aquaporins because they may live in very wet environments. Early studies of toad bladder provided some of the rst insights that water channels may exist (5). Homology cloning has yielded a constitutively active member of the family in toad bladder with high homology to mammalian AQP1 and the absence of vasopressin regulatory features (144, 145), suggesting that other homologs may remain undiscovered. Other workers, using similar methods, have identied a constitutively active homolog in frogs that is suspected to mediate adaptations in salt environments by producing large changes in skin water permeability by altered levels of expression (146). Insects contain genes encoding proteins with homology to the aquaporins. Although it has not been shown to transport water, Drosophila Big Brain encodes a polypeptide with an amino acid sequence related to that of the aquaporins, and the expression appears critical to the developmental divergence of epiderm and neuroderm (147). Another insect homolog, AQPcic, has been identied in the lter chamber of the sap-sucking insect Cicadella viridis, where the protein permits transfer of excessive water from midgut to Malpighian tubules (148).

Plant Aquaporins
Being rooted, plants are entirely dependent on their local environments for water, so it is no surprise that numerous homologs known as green aquaporins are being identied. For example, Arabidopsis thaliana, a small relative of the mustard plant, was recently found to have 23 different aquaporin homologs, and the total number is expected to be far higher (149). Plant homologs are more closely related to the water-selective aquaporins than to aquaglyceroporins. Two basic subgroups exist, and their developmental expression is tightly regulated. Plasma membrane intrinsic proteins (PIPs) reside in the plasma membrane, where they mediate cellular water uptake and release; tonoplast intrinsic protein (TIPs) reside in the intracellular vacuole (tonoplast), where they mediate cellular turgor [reviewed by Maurel (150)]. A fascinating array of physiological processes are now being ascribed to plant aquaporins. -TIP was the rst plant homolog evaluated in the oocyte system, where it confers water permeability similar to mammalian AQP1 (151), and its

AQUAPORIN WATER CHANNELS

449

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

expression is most notable within the stem. The benet of hybrid vigor is well recognized in plants. To promote genetic diversity, plants have developed a mechanism to inhibit self-pollination involving a receptor protein kinase in the stigmata of owers. Mutation in mod, a gene encoding an aquaporin homolog in the crucifer Brassica, led to the hypothesis that water transfer from stigma to pollen is a critical checkpoint in pollination (152). During seed germination, rehydration is believed to osmotically alter the integrity of the vacuole, the site where the aquaporin homolog -TIP is expressed (153). Closure of leaf guard cells regulates the escape of water during transpiration, where two sunower aquaporins reside (SunTIP17 and SunTIP 20). SunTIP17 mRNA markedly increased during stomatal closure, presumably facilitating water escape needed for stomatal opening (154). Expression of the homolog TUR in pea plants was found to be upregulated by water deprivation (155). Reduced expression of the Arabidopsis root homolog PIP1b was accomplished by overexpression of antisense constructs; although the mature plants at rst appeared normal, root arborization was increased vefold (Figure 11; 156). Regulation of the PIP1b promoter is induced by blue light or application of gibberellic acid (157). Nematodal infestation of tobacco roots induces the appearance of a feeding site containing large expression of the aquaporin homolog TobRB7 (158). An aquaporin homolog Nod26, encoded by a legume gene, is expressed in the symbiosome surrounding nitrogen-xing bacteria. Although water permeability was demonstrated, Nod26mediated transport of small osmolytes has been proposed, but it remains to be established whether this is the result of other molecular transporters that reside in the same membranes (159).

Microbial Aquaporins
The two branches of the aquaporin family are each represented by a homolog in the model bacterium Escherichia coli (Figure 6). Soon after the cloning of AQP1, its similarity with the sequence of GlpF, the glycerol facilitator of E. coli, was recognized. Owing to the apparent simplicity of their membrane system, however, specic water channels were not suspected to be necessary in bacteria, yeast, and unicellular algae. While searching for homologs in bovine ocular tissues, a novel cDNA was detected by polymerase chain reaction (160). The isolated sequence, designated aqpZ proved to be an E. coli DNA contaminant to the library. Sequence analysis of aqpZ shows a close relationship to the aquaporins (161). Expression in amphibian oocytes indicated that AqpZ is a water-selective channel (160). More recently, functional reconstitution of afnity-puried, histidine-tagged AqpZ conrmed its selectivity for water and did not reveal permeation by urea, glycerol, or other small uncharged molecules (MJ Borgnia & P Agre, submitted for publication). Biochemical analysis showed that detergent solubilized AqpZ is arranged in a tightly packed tetramer. Two-dimensional crystals of the puried protein have been obtained, and structural studies using electron-crystallography and atomic force microscopy demonstrated a close structural resemblance to AQP1 (P Ringler & A Engel, unpublished data). The

450

BORGNIA ET AL

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Figure 11 Arabidopsis thaliana plants with antisense inhibition of aquaporin PIP1b (left) and control plant (right). Reproduced with permission (156). Photo provided by R Kaldenhoff, University of Wurtzburg.

physiological importance of AqpZ is apparent when the bacteria are cultured in hypo-osmolar medium and at maximum growth rates (162). The second branch is represented in E. coli by GlpF. Owing to its role in sugar metabolism in bacteria, the glycerol facilitator of E. coli had been functionally characterized long before AQP1 was discovered (163, 164). This protein facilitates the movement of glycerol across the inner membrane of E. coli. Ten years later the DNA sequence encoding for this protein was identied (165). The GlpF sequence is closely related to the aquaglyceroporins (Figure 9), and expression in Xenopus oocytes showed an increase in glycerol permeability by a porelike mechanism (160, 166). Although one group of investigators found no water transport (166), other scientists detected a small increase over the basal water permeability in oocytes injected with glpF mRNA (160). The activity of the glycerol facilitator in E. coli appears to be affected by changes in the lipid composition of the membrane (167). In a recent study aimed at revealing the molecular basis of antimonite resistance in E. coli, disruption in glpF was observed (168). Resistant bacteria exhibited reduced accumulation of this heavy metal. The predominant form of antimonite at a neutral pH, Sb(OH)3, may be recognized by the protein as the inorganic mimic of glycerol. The bacterial aquaporins share a 31% sequence identity including multiple sequence signatures of the family (161). In addition, each protein displays the characteristics of its subgroup (Figure 9). Comparison of GlpF and AqpZ protein may explain the structural basis of the functional differ-

AQUAPORIN WATER CHANNELS

451

ences between the aquaporin family branches. Unfortunately, the functional characterization of GlpF is incomplete. Dependence of GlpF activity on the lipid environment indicates that solute transport studies undertaken in a heterologous membrane such as the Xenopus oocyte should be viewed with caution. Further functional and structural studies of puried GlpF are needed. The functional signicance of the need for the aquaporin-aquaglyceroporin dualism in E. coli may provide insight into their roles in higher organisms. Since the initial cloning of glpF from Escherichia coli, several glycerol facilitator proteins have been identied by homology cloning in other bacteria. These are clearly divided into two groups, one bearing similarity with the E. coli sequence, and the other with the Bacillus subtilis glpF (Figure 6, bottom). Interestingly, despite being evolutionarily closer to E. coli, the gram-negative bacterium Haemophilus inuenzae is represented by a glycerol facilitator protein in each of these groups. In contrast, only one AqpZ protein homolog was described by homology cloning in Shigella exneri, a close relative of E. coli (G Calamita, unpublished data). The homologous sequence from Synechococcus sp., a cyanobacterium, does not belong in any of the other groups and is closer to mammalian AQP8. As new microbial genome-sequencing projects are completed, novel bacterial aquaporin homologs are recognized (Figure 6, bottom). Putative water specic aquaporins were found in the genomes of the archaebacteria Archaeoglobus fulgidus and Methanobacterium thermoautotrophicum and in the cyanobacterium Synechocystis sp. Bacterial aquaglyceroporins were conrmed or discovered in other species. Genome sequencing provides the opportunity to identify those species in which either one or both genes are represented. With the exception of E. coli, no bacterium has yet been shown to contain genes encoding representatives of both aquaporin and aquaglyceroporin subfamilies. As for many members of the aquaporin family, the biophysical properties of most bacterial homologs are yet to be demonstrated; however, the experimental data obtained to date are in close agreement with the classication based on primary structure. However, expression and characterization of the multiple variants offered by nature may contribute to the understanding of the molecular mechanics of these proteins. The Saccharomyces genome contains two open reading frames related to aquaglyceroporins and two others related to aquaporins (Figure 6, bottom). FPS1 has previously been characterized to encode a glycerol transporter, which is apparently closed in hyperosmolar environments and may function to release the osmolyte (169). Deletion of an N-terminal domain of this protein resulted in the loss of osmotic regulation of the function of the channel, which remained constitutively open. These data suggest that the N-terminal domain of FPS1 may be involved in channel gating (170). The rst yeast aquaporin (AQY1) from laboratory strains was found to encode a nonfunctional molecule. However, the wild-type strains contain functional molecules, and gene disruption confers competitive advantages for growth under laboratory conditions (171). Vestiges of a second yeast homolog, obliterated by a deletion with a frame shift, were found in the genomic sequence. The frame shift in the laboratory strain was conrmed by cloning of the gene from wild-type

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

452

BORGNIA ET AL

yeast cells. The product, designated AQY2, has not yet been shown to function (M Bonhivers & P Agre, unpublished data). The physiological role of yeast homologs is yet uncertain. The duplication of genes encoding both glycerol transporters and aquaporins may be related to an ancient genome duplication in Saccharomyces cerevisiae. Likewise, other unicellular eukaryotic species contain aquaporin homologs. Dictyostelium discoideum expresses WacA, an aquaporin homolog expressed in prespore cells under tight developmental control; however, disruption of the gene did not provide a clear phenotype (172).

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

PERSPECTIVES
Although the molecular identity of water channels began with the discovery of AQP1 just 8 years ago, the rapid progress has been extremely gratifying. Nevertheless, numerous issues remain unsettled and warrant additional study. The structural understanding of aquaporins is so far based on cryoelectron microscopic analysis of two-dimensional membrane crystals formed from puried red cell AQP1. Structural resolution at 36 reveals ample details of the bilayer spanning domains and the central hourglass (Figure 4); however, the cytoplasmic and extracellular connecting loops as well as the atomic structure of the hourglass remain incompletely dened. Whereas additional renements may be possible with this technique, the isolation of AQP1 from human red cells may be less optimal than purication from heterologous expression systems such as yeast or bacteria cells. Such systems may also permit expression of mutant forms of the protein, so that epitopes can be added to provide unambiguous landmarks. Similarly, the structural basis of glycerol transport by the aquaglyceroporins, bacterial GlpF or mammalian AQP3, has not been explored. In addition, although preparation of three-dimensional crystals for X-ray diffraction has so far proven impractical, this may become feasible with a heterologous source of aquaporins. The list of aquaporins is growing rapidly. Ten mammalian homologs were quickly identied, and new mammalian aquaporins are being identied at the rate of two per year. Because numerous plant processes seem to operate by hydraulics, plants apparently have a greater need for aquaporins than mammals, and the list of recognized plant aquaporins is approximately threefold the mammalian list. Just as sequencing of the entire genomes from yeasts and other microorganisms led to recognition of microbial aquaporin homologs, the human genome must certainly contain genes for several mammalian aquaporins. Thus, the challenge will be to dene the ontogeny and the sites of expression. At the same time, it cannot be assumed that only aquaporins can serve as water channels, because genetically unrelated cotransporter proteins such as the sodiumwater transporter were shown to carry water along with the specied solutes (173). Additionally, the already recognized primary and secondary roles of aquaporins in human pathophysiology suggest that other clinical problems such as glaucoma, brain edema, or body temperature regulation may involve members of

AQUAPORIN WATER CHANNELS

453

this protein family. Likewise, plant aquaporins may be useful in generating pestor drought-resistant organisms. It is therefore considered likely that the cellular and molecular biology of the aquaporins is far from completely understood. ACKNOWLEDGMENTS This work was supported by grants from the National Institutes of Health, the Cystic Fibrosis Foundation, the Swiss National Foundation for Scientic Research, the Maurice E. Mller Foundation, the Novo Nordic Foundation, the Karen Elise Jensen Foundation, the Danish Medical Research Council, the Biomembrane Research Center at University of Aarhus, and a Postdoctoral Fellowship from Human Frontier Science Organization. An earlier version of this review was published in abridged form and has been modied with permission (174). LITERATURE CITED
1. Finkelstein A. 1987. Water Movement Through Lipid Bilayers, Pores, and Plasma Membranes. Theory and Reality. New York: Wiley 2. Solomon AK. 1968. J. Gen. Physiol. 51: S33564 3. Macey RI. 1984. Am. J. Physiol. 246: C195203 4. Knepper MA, Wade JB, Terris J, Ecelbarger CA, Marples D, et al. 1996. Kidney Int. 49:171217 5. Bourguet J, Chevalier J, Hugon JS. 1976. Biophys. J. 16:62739 6. Brown D, Orci L. 1983. Nature 302: 25355 7. Agre P, Preston GM, Smith BL, Jung JS, Raina S, et al. 1993. Am. J. Physiol. 265: F46376 8. Agre P, Saboori AM, Asimos A, Smith BL. 1987. J. Biol. Chem. 262:17497503 9. Denker BM, Smith BL, Kuhajda FP, Agre P. 1988. J. Biol. Chem. 263:1563442 10. Smith BL, Agre P. 1991. J. Biol. Chem. 266: 640715 11. Preston GM, Agre P. 1991. Proc. Natl. Acad. Sci. USA 88:1111014 12. Gorin MB, Yancey SB, Cline J, Revel JP, Horwitz J. 1984. Cell 39:4959 13. van Hoek AN, Hom ML, Luthjens LH, de Jong MD, Dempster JA, van Os CH. 1991. J. Biol. Chem. 266:1663335 14. Agre P. 1997. Biol. Cell. 89:25557 15. Fischbarg J, Kuang KY, Vera JC, Arant S, Silverstein SC, et al. 1990. Proc. Natl. Acad. Sci. USA 87:324447 16. Zhang RB, Logee KA, Verkman AS. 1990. J. Biol. Chem. 265:1537578 17. Preston GM, Carroll TP, Guggino WB, Agre P. 1992. Science 256:38587 18. Meinild AK, Klaerke DA, Zeuthen T. 1998. J. Biol. Chem. 273:3244651 19. Zeidel ML, Ambudkar SV, Smith BL, Agre P. 1992. Biochemistry 31:743640 20. van Hoek AN, Verkman AS. 1992. J. Biol. Chem. 267:1826769 21. Zeidel ML, Nielsen S, Smith BL, Ambudkar SV, Maunsbach AB, Agre P. 1994. Biochemistry 33:160615 22. Abrami L, Tacnet F, Ripoche P. 1995. Eur. J. Physiol. 430:44758 23. Yool AJ, Stamer WD, Regan JW. 1996. Science 273:121618 24. Agre P, Lee MD, Devidas S, Guggino WB, Sasaki S, et al. 1997. Science 275: 149092 25. Patil RV, Saito I, Yang X, Wax MB. 1997. Exp. Eye Res. 64:2039 26. Marinelli RA, Pham L, Agre P, LaRusso NF. 1997. J. Biol. Chem. 272:1298488 27. Nakhoul NL, Davis BA, Romero MF, Boron WF. 1998. Am. J. Physiol. 274: C54348

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

454

BORGNIA ET AL 48. Walz T, Tittmann P, Fuchs KH, Muller DJ, Smith BL, et al. 1996. J. Mol. Biol. 264: 90718 49. Walz T, Hirai T, Murata K, Heymann JB, Mitsuoka K, et al. 1997. Nature 387: 62427 50. Li HL, Lee S, Jap BK. 1997. Nat. Struct. Biol. 4:26365 51. Cheng A, van Hoek AN, Yeager M, Verkman AS, Mitra AK. 1997. Nature 387: 62730 52. Heymann JB, Agre P, Engel A. 1998. J. Struct. Biol. 121:191206 53. Sabolic I, Valenti G, Verbavatz JM, van Hoek AN, Verkman AS, et al. 1992. Am. J. Physiol. 263:C122533 54. Nielsen S, Smith BL, Christensen EI, Knepper MA, Agre P. 1993. J. Cell Biol. 120:37183 55. Maunsbach AB, Marples D, Chin E, Ning G, Bondy C, et al. 1997. J. Am. Soc. Nephrol. 8:114 56. Maeda Y, Smith BL, Agre P, Knepper MA. 1995. J. Clin. Invest. 95:42228 57. Nielsen S, Smith BL, Christensen EI, Agre P. 1993. Proc. Natl. Acad. Sci. USA 90:727579 58. King LS, Nielsen S, Agre P. 1996. J. Clin. Invest. 97:218391 59. Schnitzer JE, Oh P. 1996. Am. J. Physiol. 270:H41622 60. Pallone TL, Kishore BK, Nielsen S, Agre P, Knepper MA. 1997. Am. J. Physiol. 272:F58796 61. King LS, Nielsen S, Agre P. 1997. Am. J. Physiol. 273:C154148 62. Moon C, King LS, Agre P. 1997. Am. J. Physiol. 273:C156270 63. Hasegawa H, Lian SC, Finkbeiner WE, Verkman AS. 1994. Am. J. Physiol. 266: C893903 64. Stamer WD, Snyder RW, Smith BL, Agre P, Regan JW. 1994. Invest. Ophthalmol. Vis. Sci. 35:386772 65. Roberts SK, Yano M, Ueno Y, Pham L, Alpini G, et al. 1994. Proc. Natl. Acad. Sci. USA 91:1300913

28. Reuss L. 1998. Am. J. Physiol. 274: C29798 29. Wistow GJ, Pisano MM, Chepelinsky AB. 1991. Trends Biochem. Sci. 16:17071 30. Pao GM, Wu LF, Johnson KD, Hofte H, Chrispeels MJ, et al. 1991. Mol. Microbiol. 5:3337 31. Skach WR, Shi LB, Calayag MC, Frigeri A, Lingappa VR, Verkman AS. 1994. J. Cell Biol. 125:80315 32. Shi LB, Skach WR, Ma T, Verkman AS. 1995. Biochemistry 34:825056 33. Preston GM, Jung JS, Guggino WB, Agre P. 1994. J. Biol. Chem. 269:166873 34. Preston GM, Jung JS, Guggino WB, Agre P. 1993. J. Biol. Chem. 268:1720 35. Zhang R, van Hoek AN, Biwersi J, Verkman AS. 1993. Biochemistry 32:293841 36. Jung JS, Preston GM, Smith BL, Guggino WB, Agre P. 1994. J. Biol. Chem. 269: 1464854 37. Shi LB, Skach WR, Verkman AS. 1994. J. Biol. Chem. 269:1041722 38. Verbavatz JM, Brown D, Sabolic I, Valenti G, Ausiello DA, et al. 1993. J. Cell Biol. 123:60518 39. van Hoek AN, Wiener M, Bicknese S, Miercke L, Biwersi J, Verkman AS. 1993. Biochemistry. 32:1184756 40. Fischbarg J, Li J, Cheung M, Czegledy F, Iserovich P, Kuang K. 1995. J. Membr. Biol. 143:17788 41. Cabiaux V, Oberg KA, Pancoska P, Walz T, Agre P, Engel A. 1997. Biophys. J. 73:40617 42. Cho MR, Knowles DW, Smith BL, Moulds JJ, Agre P, et al. 1999. Biophys. J. 76:113644 43. Walz T, Smith BL, Zeidel ML, Engel A, Agre P. 1994. J. Biol. Chem. 269:158386 44. Walz T, Smith BL, Agre P, Engel A. 1994. EMBO J. 13:298593 45. Walz T, Typke D, Smith BL, Agre P, Engel A. 1995. Nat. Struct. Biol. 2:73032 46. Jap BK, Li HL. 1995. J. Mol. Biol. 251: 41320 47. Mitra AK, van Hoek AN, Wiener MC, Verkman AS, Yeager M. 1995. Nat. Struct. Biol. 2:72629

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

AQUAPORIN WATER CHANNELS 66. Bondy C, Chin E, Smith BL, Preston GM, Agre P. 1993. Proc. Natl. Acad. Sci. USA 90:45004 67. Smith BL, Baumgarten R, Nielsen S, Raben D, Zeidel ML, Agre P. 1993. J. Clin. Invest. 92:203541 68. Smith BL, Preston GM, Spring FA, Anstee DJ, Agre P. 1994. J. Clin. Invest. 94:104349 69. Preston GM, Smith BL, Zeidel ML, Moulds JJ, Agre P. 1994. Science 265:158587 70. Ma T, Yang B, Gillespie A, Carlson EJ, Epstein CJ, Verkman AS. 1998. J. Biol. Chem. 273:429699 71. Schnermann J, Chou CL, Ma T, Traynor T, Knepper MA, Verkman AS. 1998. Proc. Natl. Acad. Sci. USA 95:966064 72. Agre P, Mathai JC, Smith BL, Preston GM. 1999. Methods Enzymol. 294:55072 73. Fushimi K, Uchida S, Hara Y, Hirata Y, Marumo F, Sasaki S. 1993. Nature 361: 54952 74. Knepper MA, Inoue T. 1997. Curr. Opin. Cell. Biol. 9:56064 75. Nielsen S, DiGiovanni SR, Christensen EI, Knepper MA, Harris HW. 1993. Proc. Natl. Acad. Sci. USA 90:1166367 76. Sabolic I, Katsura T, Verbavatz JM, Brown D. 1995. J. Membr. Biol. 143: 16575 77. Wade JB, Stetson DL, Lewis SA. 1981. Ann. NY Acad. Sci. 372:10617 78. DiGiovanni SR, Nielsen S, Christensen EI, Knepper MA. 1994. Proc. Natl. Acad. Sci. USA 91:898488 79. Yamamoto T, Sasaki S, Fushimi K, Kawasaki K, Yaoita E, et al. 1995. Exp. Nephrol. 3:193201 80. Nielsen S, Chou CL, Marples D, Christensen EI, Kishore BK, Knepper MA. 1995. Proc. Natl. Acad. Sci. USA 92: 101317 81. Katsura T, Gustafson CE, Ausiello DA, Brown D. 1997. Am. J. Physiol. 272: F81722 82. Valenti G, Procino G, Liebenhoff U, Frigeri A, Benedetti PA, et al. 1998. J. Biol. Chem. 273:2262734

455

83. Hozawa S, Holtzman EJ, Ausiello DA. 1996. Am. J. Physiol. 270:C1695702 84. Matsumura Y, Uchida S, Rai T, Sasaki S, Marumo F. 1997. J. Am. Soc. Nephrol. 8: 86167 85. Ecelbarger CA, Nielsen S, Olson BR, Murase T, Baker EA, et al. 1997. J. Clin. Invest. 99:185263 86. Fujiwara TM, Morgan K, Bichet DG. 1995. Annu. Rev. Med. 46:33143 87. Deen PM, Verdijk MA, Knoers NV, Wieringa B, Monnens LA, et al. 1994. Science 264:9295 88. van Lieburg AF, Verdijk MA, Knoers VV, van Essen AJ, Proesmans W, et al. 1994. Am. J. Hum. Genet. 55:64852 89. Tamarappoo BK, Verkman AS. 1998. J. Clin. Invest. 101:225767 90. Mulders SM, Bichet DG, Rijss JP, Kamsteeg EJ, Arthus MF, et al. 1998. J. Clin. Invest. 102:5766 90a. Jamison KR. 1995. Sci. Am. 272:6267 91. Marples D, Christensen S, Christensen EI, Ottosen PD, Nielsen S. 1995. J. Clin. Invest. 95:183845 92. Frokiaer J, Marples D, Knepper MA, Nielsen S. 1996. Am. J. Physiol. 270: F65768 93. Marples D, Frokiaer J, Dorup J, Knepper MA, Nielsen S. 1996. J. Clin. Invest. 97: 196068 94. Frokiaer J, Nielsen S. 1997. Scand. J. Urol. Nephrol. Suppl. 183:3132 95. Nielsen S, Terris J, Andersen D, Ecelbarger C, Frokiaer J, et al. 1997. Proc. Natl. Acad. Sci. USA 94:545055 96. Xu DL, Martin PY, Ohara M, St. John J, Pattison T, et al. 1997. J. Clin. Invest. 99: 15005 97. Ohara M, Martin PY, Xu DL, St. John J, Pattison TA, et al. 1998. J. Clin. Invest. 101:107683 98. Fujita N, Ishikawa SE, Sasaki S, Fujisawa G, Fushimi K, et al. 1995. Am. J. Physiol. 269:F92631 99. Mulders SM, Preston GM, Deen PM, Guggino WB, van Os CH, Agre P. 1995. J. Biol. Chem. 270:901016

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

456

BORGNIA ET AL 118. Nielsen S, King LS, Christensen BM, Agre P. 1997. Am. J. Physiol. 273: C154961 119. King LS, Agre P. 1996. Annu. Rev. Physiol. 58:61948 120. Dobbs LG, Gonzalez R, Matthay MA, Carter EP, Allen L, Verkman AS. 1998. Proc. Natl. Acad. Sci. USA 95:299196 121. Borok Z, Lubman RL, Danto SI, Zhang XL, Zabski SM, et al. 1998. Am. J. Respir. Cell Mol. Biol. 18:55461 122. Delporte C, OConnell BC, He X, Ambudkar IS, Agre P, Baum BJ. 1996. J. Biol. Chem. 271:2207075 123. Ishibashi K, Kuwahara M, Kageyama Y, Tohsaka A, Marumo F, Sasaki S. 1997. Biochem. Biophys. Res. Commun. 237: 71418 124. Koyama Y, Yamamoto T, Kondo D, Funaki H, Yaoita E, et al. 1997. J. Biol. Chem. 272:3032933 125. Ma T, Yang B, Verkman AS. 1997. Biochem. Biophys. Res. Commun. 240: 32428 126. Ishibashi K, Sasaki S, Fushimi K, Uchida S, Kuwahara M, et al. 1994. Proc. Natl. Acad. Sci. USA 91:626973 127. Echevarria M, Windhager EE, Tate SS, Frindt G. 1994. Proc. Natl. Acad. Sci. USA 91:109971001 128. Ma T, Frigeri A, Hasegawa H, Verkman AS. 1994. J. Biol. Chem. 269:2184549 129. You G, Smith CP, Kanai Y, Lee WS, Stelzner M, Hediger MA. 1993. Nature 365:84447 130. Echevarria M, Windhager EE, Frindt G. 1996. J. Biol. Chem. 271:2507982 131. Kuwahara M, Gu Y, Ishibashi K, Marumo F, Sasaki S. 1997. Biochemistry 36:1397378 132. Terris J, Ecelbarger CA, Marples D, Knepper MA, Nielsen S. 1995. Am. J. Physiol. 269:F77585 133. Ishibashi K, Sasaki S, Fushimi K, Yamamoto T, Kuwahara M, Marumo F. 1997. Am. J. Physiol. 272:F23541 134. Roudier N, Verbavatz JM, Maurel C, Ripoche P, Tacnet F. 1998. J. Biol. Chem. 273:840712

100. Kushmerick C, Rice SJ, Baldo GJ, Haspel HC, Mathias RT. 1995. Exp. Eye Res. 61:35162 101. Chandy G, Zampighi GA, Kreman M, Hall JE. 1997. J. Membr. Biol. 159:2939 102. Shiels A, Bassnett S. 1996. Nat. Genet. 12:21215 103. Ma T, Frigeri A, Skach W, Verkman AS. 1993. Biochem. Biophys. Res. Commun. 197:65459 104. Yasui M, Kwon T-H, Knepper MA, Nielsen S, Agre P. 1999. Proc. Natl. Acad. Sci. USA 96:580813 105. Hasegawa H, Ma T, Skach W, Matthay MA, Verkman AS. 1994. J. Biol. Chem. 269:5497500 106. Jung JS, Bhat RV, Preston GM, Guggino WB, Baraban JM, Agre P. 1994. Proc. Natl. Acad. Sci. USA 91:1305256 107. Nielsen S, Nagelhus EA, Amiry-Moghaddam M, Bourque C, Agre P, Ottersen OP. 1997. J. Neurosci. 17: 17180 108. Frigeri A, Gropper MA, Umenishi F, Kawashima M, Brown D, Verkman AS. 1995. J. Cell Sci. 108:29933002 109. Nagelhus EA, Veruki ML, Torp R, Haug FM, Laake JH, et al. 1998. J. Neurosci. 18:250619 110. Frigeri A, Nicchia GP, Verbavatz JM, Valenti G, Svelto M. 1998. J. Clin. Invest. 102:695703 111. Lu M, Lee MD, Smith BL, Jung JS, Agre P, et al. 1996. Proc. Natl. Acad. Sci. USA 93:1090812 112. Yang B, Verkman AS. 1997. J. Biol. Chem. 272:1614046 113. Chou CL, Ma T, Yang B, Knepper MA, Verkman AS. 1998. Am. J. Physiol. 274:C54954 114. Verbavatz JM, Ma T, Gobin R, Verkman A. 1997. J. Cell Sci. 110:285560 115. Rash JE, Yasumura T, Hudson CS, Agre P, Nielsen S. 1998. Proc. Natl. Acad. Sci. USA 95:1198186 116. Han Z, Wax MB, Patil RV. 1998. J. Biol. Chem. 273:60014 117. Raina S, Preston GM, Guggino WB, Agre P. 1995. J. Biol. Chem. 270: 190812

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

AQUAPORIN WATER CHANNELS 135. Ishibashi K, Kuwahara M, Gu Y, Kageyama Y, Tohsaka A, et al. 1997. J. Biol. Chem. 272:2078286 136. Kuriyama H, Kawamoto S, Ishida N, Ohno I, Mita S, et al. 1997. Biochem. Biophys. Res. Commun. 241:5358 137. Ishibashi K, Kuwahara M, Gu Y, Tanaka Y, Marumo F, Sasaki S. 1998. Biochem. Biophys. Res. Commun. 244:26874 138. Tsukaguchi H, Shayakul C, Berger UV, Mackenzie B, Devidas S, et al. 1998. J. Biol. Chem. 273:2473743 139. Ecelbarger CA, Terris J, Frindt G, Echevarria M, Marples D, et al. 1995. Am. J. Physiol. 269:F66372 140. Matthay MA, Folkesson HG, Verkman AS. 1996. Am. J. Physiol. 270:L487503 141. Yasui M, Zelenin SM, Celsi G, Aperia A. 1997. Am. J. Physiol. 272:F44350 142. Carter EP, Umenishi F, Matthay MA, Verkman AS. 1997. J. Clin. Invest. 100: 107178 143. Hamann S, Zeuthen T, la Cour M, Nagelhus EA, Ottersen OP, et al. 1998. Am. J. Physiol. 274:C133245 144. Siner J, Paredes A, Hosselet C, Hammond T, Strange K, Harris HW. 1996. Am. J. Physiol. 270:C37281 145. Ma T, Yang B, Verkman AS. 1996. Am. J. Physiol. 271:C1699704 146. Abrami L, Gobin R, Berthonaud V, Thanh HL, Chevalier J, et al. 1997. Eur. J. Cell Biol. 73:21521 147. Rao Y, Jan LY, Jan YN. 1990. Nature 345:16367 148. Beuron F, Le Caherec F, Guillam MT, Cavalier A, Garret A, et al. 1995. J. Biol. Chem. 270:1741422 149. Weig A, Deswarte C, Chrispeels MJ. 1997. Plant Physiol. 114:134757 150. Maurel C. 1997. Annu. Rev. Plant Physiol. Plant Mol. Biol. 48:399429 151. Maurel C, Reizer J, Schroeder JI, Chrispeels MJ. 1993. EMBO J. 12: 224147 152. Ikeda S, Nasrallah JB, Dixit R, Preiss S, Nasrallah ME. 1997. Science 276: 156466

457

153. Maurel C, Chrispeels MJ, Lurin C, Tacnet F, Geelen O, et al. 1997. J. Exp. Bot. 48(Suppl. 1):42130 154. Sarda X, Tousch D, Ferrare K, Legrand E, Dupuis JM, et al. 1997. Plant J. 12: 110311 155. Guerrero FD, Jones JT, Mullet JE. 1990. Plant Mol. Biol. 15:1126 156. Kaldenhoff R, Grote K, Zhu JJ, Zimmermann U. 1998. Plant J. 14:12128 157. Kaldenhoff R, Kolling A, Richter G. 1996. J. Photochem. Photobiol. B 36: 35154 158. Opperman CH, Taylor CG, Conkling MA. 1994. Science 263:22123 159. Rivers RL, Dean RM, Chandy G, Hall JE, Roberts DM, Zeidel ML. 1997. J. Biol. Chem. 272:1625661 160. Calamita G, Bishai WR, Preston GM, Guggino WB, Agre P. 1995. J. Biol. Chem. 270:2906366 161. Park JH, Saier MH. 1996. J. Membr. Biol. 153:17180 162. Calamita G, Kempf B, Bonhivers M, Bishai WR, Bremer E, Agre P. 1998. Proc. Natl. Acad. Sci. USA 95:362731 163. Eze MO, McElhaney RN. 1978. J. Gen. Microbiol. 105:23342 164. Heller KB, Lin EC, Wilson TH. 1980. J. Bacteriol. 144:27478 165. Sweet G, Gandor C, Voegele R, Wittekindt N, Beuerle J, et al. 1990. J. Bacteriol. 172:42430 166. Maurel C, Reizer J, Schroeder JI, Chrispeels MJ, Saier MH. 1994. J. Biol. Chem. 269:1186972 167. Truniger V, Boos W. 1993. Res. Microbiol. 144:56574 168. Sanders OI, Rensing C, Kuroda M, Mitra B, Rosen BP. 1997. J. Bacteriol. 179: 336567 169. Luyten K, Albertyn J, Skibbe WF, Prior BA, Ramos J, et al. 1995. EMBO J. 14:136071 170. Tamas MJ, Luyten K, Sutherland FC, Hernandez A, Albertyn J, et al. 1999. Mol. Microbiol. 31:1087104

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

458

BORGNIA ET AL 174. Agre P, Bonhivers M, Borgnia MJ. 1998. J. Biol. Chem. 273:1465962 175. Chrispeels MJ, Agre P. 1994. Trends Biochem. Sci. 19:42125 176. Deleted in proof 177. Thompson JD, Higgins DG, Gibson TJ. 1994. Nucleic Acids Res. 22:467380

171. Bonhivers M, Carbrey JM, Gould SJ, Agre P. 1998. J. Biol. Chem. 273: 2756572 172. Flick KM, Shaulsky G, Loomis WF. 1997. Gene 195:12730 173. Meinild A, Klaerke DA, Loo DD, Wright EM, Zeuthen T. 1998. J. Physiol. 508: 1521

Annu. Rev. Biochem. 1999.68:425-458. Downloaded from www.annualreviews.org by Universidad Complutense de Madrid on 09/24/12. For personal use only.

Você também pode gostar