The electroneutral sodium bicarbonate cotransporter 3 (NBC3) coimmunoprecipitates from renal lysates with the vacuolar H+-ATPase. In renal type A and B intercalated cells, NBC3 colocalizes with the vacuolar H+-ATPase. The involvement of the COOH termini of NBC3 and the 56-kDa subunit of the proton pump in the interaction of these proteins was investigated. The intact and modified COOH termini of NBC3 and the 56-kDa subunit of the proton pump were synthesized, coupled to Sepharose beads, and used to pull down kidney membrane proteins. Both the 56- and the 70-kDa subunits of the proton pump, as well as a PDZ domain containing protein Na+/H+ exchanger regulatory factor 1 (NHERF-1), were bound to the intact 18 amino acid NBC3 COOH terminus. A peptide truncated by five COOH-terminal amino acids did not bind these proteins. Replacement of the COOH-terminal leucine with glycine blocked binding of both the proton pump subunits but did not affect binding of NHERF-1. The 18 amino acid COOH terminus of the 56-kDa subunit of the proton pump bound NHERF-1 and NBC3, but the truncated and modified peptide did not. A complex of NBC3, the 56-kDa subunit of the proton pump, and NHERF-1 was identified in rat kidney. The data indicate that the COOH termini of NBC3 and the 56-kDa subunit of the vacuolar proton pump are PDZ-interacting motifs that are necessary for the interaction of these proteins. NHERF-1 is involved in the interaction of NBC3 and the vacuolar proton pump.
sodium bicarbonate cotransporters (NBCs) play an important role in transepithelial transport and pH regulation (1, 2, 4, 6, 7, 29,30). All functionally characterized NBCs mediate either electrogenic or electroneutral transport of sodium and bicarbonate. We have cloned and functionally characterized the first electroneutral sodium bicarbonate cotransporter, NBC3 (SLC4A7) (26), which, unlike the electrogenic sodium-bicarbonate cotransporters, kNBC1 (2, 6, 30), pNBC1 (1), and NBC4 (34), mediates 4,4′-diisothiocyanostilbene-2,2′-disulfonic acid (DIDS)-insensitive transport of these ions.
NBC3 has recently been colocalized with the vacuolar H+-ATPase on the apical membrane of type A intercalated cells (29) and on the basolateral membrane of type B intercalated cells in the kidney (20) and on the apical membrane of clear cells in epididymis (28). Their coimmunoprecipitation demonstrated the close association of these two transporters in rabbit kidney (29). These results indicated that NBC3 and the vacuolar proton pump might interact in vivo. The vacuolar proton pump is a heteromeric protein that mediates active H+ transport and consists of at least 11 different polypeptides that comprise two domains: the peripheral domain (V1), responsible for ATP hydrolysis (located in the cytoplasm), and the integral domain (V0) located in the plasma membrane (12). To determine the subunit(s) of the proton pump that might interact with NBC3, we analyzed its primary structure and the subunits of the vacuolar H+-ATPase. We have recently reported (27) that the COOH terminus of NBC3 contains a putative PDZ-interacting motif, ETSL, which should bind to class I PDZ domains. In addition, it has recently been reported (5) that the COOH terminus of the 56-kDa subunit of the vacuolar H+-ATPase also has a consensus PDZ-interacting motif, DTAL, which should also bind to class I PDZ domains. The finding that the COOH termini of both NBC3 and the 56-kDa subunit of the H+-ATPase have putative PDZ-interacting motifs is a potential clue to the mechanism of interaction of these two transport proteins, which might involve one or more PDZ domain-containing proteins (PDZ proteins). PDZ-interacting motifs are short COOH-terminal protein sequences that specifically interact with the PDZ-binding domains (usually ∼70–90 amino acid sequences) of PDZ proteins (33, 37, 38). PDZ domains consist of six β-strands and two α-helices and usually contain the conserved motif Gly-Leu-Gly-Phe (GLGF). PDZ interactions are required for receptor clustering, signal transduction, protein targeting and retention, and modulating the function of proteins. Not all amino acid residues in COOH-terminal PDZ-interacting motifs are important for the interactions with PDZ domains. X-ray crystallography and binding studies indicate that the last five COOH-terminal amino acids are most important (8, 24,38). Furthermore, amino acids at the 0 and −2 positions are essential for normal PDZ-mediated binding (15, 38). The PDZ-interacting motifs recognized by class I PDZ domains at position −2 should have serine or threonine and at the 0 position a hydrophobic amino acid (valine, isoleucine, or leucine) (16). It has recently been shown that the amino acid residues at −1 and −3 positions are also important for the interactions with PDZ domains (10, 16, 19, 35, 40).
Breton et al. (5) have shown that both the 56- and the 31-kDa subunit of the vacuolar proton pump are pulled down from the rat homogenate by Na+/H+ exchanger regulatory factor 1 (NHERF-1). This protein has been shown to possess two class I PDZ domains with similar but not identical specificity (40). NHERF-1 has been immunolocalized to the basolateral membrane of type B intercalated cells, where it has been suggested to play an important role in targeting the vacuolar H+-ATPase (5). NHERF-1 was not detected in type A intercalated cells (5), suggesting that it is not involved in the interaction between NBC3 and the vacuolar proton pump and that other proteins might be involved in the targeting of the H+-ATPase to the apical membrane in this cell type.
Based on these considerations, we hypothesized that the COOH termini of NBC3 and the 56-kDa subunit of the proton pump have PDZ-interacting motifs that are involved in the binding of NBC3 to the vacuolar H+-ATPase. Furthermore, because PDZ domains are not found in NBC3 or any of the proton pump subunits, NBC3 and the vacuolar H+-ATPase would likely interact indirectly through PDZ protein(s). Given the recently reported interaction between the 56-kDa subunit and NHERF-1 (5), an additional goal was to determine whether NHERF-1 is involved in the interaction between NBC3 and the vacuolar proton pump.
MATERIALS AND METHODS
NBC3 COOH-terminal peptide synthesis and immobilization.
It has been shown that the most critical amino acid residues in the COOH-terminal PDZ-interacting motifs are at positions 0 and −2 (38). The importance of residues at −3 and −1 has also been reported recently (38, 40). If the COOH terminus of NBC3 is a PDZ-interacting motif that is necessary for binding of the vacuolar proton pump, the mutations of the amino acids at these positions should decrease the binding of the vacuolar proton pump. On the basis of these considerations, three peptides were synthesized:1) the last 18 COOH-terminal amino acids of NBC3 (ISFEDEPRKKYVDAETSL), 2) the same peptide lacking five COOH-terminal amino acids (ISFEDEPRKKYVD), and 3) the last 18 COOH-terminal amino acids of NBC3 with the COOH-terminal leucine changed to glycine (ISFEDEPRKKYVDAETSG). The peptides were coupled to Sepharose 4B beads (Amersham Biosciences, Piscataway, NJ) through NH2-terminal cysteine.
The vacuolar proton pump 56-kDa subunit COOH-terminal peptide synthesis and immobilization.
Three of the 56-kDa subunits of the vacuolar proton pump COOH-terminal peptides were synthesized and coupled to Sepharose 4B beads similar to NBC3 peptides: 1) the last 18 amino acid COOH terminus (EFYSREGALQDLAPDTAL), 2) the same peptide lacking five COOH-terminal amino acids (EFYSREGALQDLA), and 3) the peptide with the COOH-terminal leucine changed to glycine (EFYSREGALQDLAPDTAG).
Rat kidney membrane protein isolation.
Rat kidney membranes were isolated using differential centrifugation (29). Briefly, 2 g of rat kidneys were disrupted at 0°C in a glass homogenizer with 100 ml of buffer A [20 mM Tris · HCl, pH 7.5, containing 140 mM NaCl, 1 mM phenyl methyl sulfonylfluoride, 1 mM EDTA, 1 μg/ml pepstatin, 1 μg/ml leupeptin, and 1 μg/ml aprotinin (proteinase inhibitors were purchased from Roche, Indianapolis, IN)]. The homogenate was centrifuged at 300 g for 10 min, and the supernatant was centrifuged at 4,000 g for 10 min. The supernatant was centrifuged at 150,000 g for 2 h. The pellet was solubilized in 30 ml of buffer A containing 0.1% Triton X-100 (Sigma, St. Louis, MO) for 15 h at 0°C. After centrifugation at 150,000 g for 1 h, the supernatant was used as a source of rat renal membrane proteins.
Pull-down experiments using rat renal membrane proteins.
The rat renal membrane protein solution (5 ml, ∼2 mg/ml) was incubated at 0°C for 5 h with NBC3 or the 56-kDa subunit of the vacuolar proton pump COOH-terminal peptides coupled to 0.2 ml of Sepharose 4B beads. After incubation at 0°C for 5 h, the beads were washed ten times in 5 ml of buffer A containing 0.1% Triton X-100. Proteins were eluted from the beads using 0.2 ml 1× Laemmli buffer and analyzed by sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting.
Identification of a complex of NHERF-1 with NBC3 and the proton pump.
The rat kidney membranes (100–200 mg) were solubilized in 5 ml of BugBuster HT reagent (Novagen, Madison, WI) containing 1 mM phenylmethylsulfonyl fluoride, 1 mM EDTA, 1 μg/ml pepstatin, 1 μg/ml leupeptin, and 1 μg/ml aprotinin for 1 h at 0°C. The homogenate was centrifuged at 18,000 g for 20 min, and the supernatant was passed through a Sephadex G-25 (Amersham Biosciences) column (2 × 3 cm) equilibrated with buffer Acontaining 0.1% Triton X-100. The eluate was mixed with ∼0.5 mg of recombinant GST-NHERF-1 and incubated for 16 h at 4°C, and then 200 μl of glutathione-Sepharose beads (Amersham Biosciences) were added to the solution and incubated at 4°C for 1 h. The beads were washed three times with buffer A containing 0.1% Triton X-100 and five times with PBS. The proteins were eluted from the beads in 0.5 ml of 10 mM DTT in PBS and resolved on Criterion IEF ready gels (Bio-Rad, Hercules, CA), transferred to polyvinylidene difluoride (PVDF) membrane, and analyzed by Western blotting.
Expression and purification of GST-NHERF-1.
NHERF-1 coding sequence inserted in XhoI-NotI site of pGEX-4T vector (Amersham Biosciences) was used for expression GST-NHERF-1 construct in Escherichia coli BL21(DE3)pLysS cells (Amersham Biosciences). The fusion GST-NHERF-1 protein was extracted from ∼5 g of E. coli cells using 50 ml of BugBuster HT reagent containing 1 mM phenylmethylsulfonyl fluoride, 1 mM EDTA, 1 μg/ml pepstatin, 1 μg/ml leupeptin, and 1 μg/ml aprotinin. After centrifugation at 18,000 g for 20 min, the supernatant was dialyzed overnight against buffer Acontaining 0.1% Triton X-100 and incubated with 1 ml of glutathione-Sepharose beads (Amersham Bioscience) for 1 h. The beads were washed 10 times with PBS, and the proteins were eluted from the beads in 5 ml of 10 mM DTT in PBS according to the manufacturer's protocol. The GST-NHERF-1 eluted from the column was ∼99% pure, which was confirmed by SDS-PAGE followed by Coomassie staining and Western blotting using antibody specific for NHERF-1 and GST (Amersham Biosciences).
SDS-PAGE and Western blotting.
Ready gels from Bio-Rad were used for SDS-PAGE. For Western blotting, proteins were electrotransferred from the gel to PVDF membrane (Amersham Biosciences). Nonspecific binding was blocked by incubation of the membrane in TBST (20 mM Tris · HCl, pH 7.5, 140 mM NaCl, and 0.1% Tween 20) containing 5% dry milk (Bio-Rad). The primary antibodies specific for the 56- and the 70-kDa subunits of the proton pump and to NBC3 were used at a dilution 1:1,000. NHERF-1-specific antibody was used at a dilution of 1:3,000. Secondary horseradish peroxidase-conjugated mouse anti-rabbit antibody (Jackson ImmunoResearch, West Grove, PA) was used at a dilution of 1:10,000. Bands were visualized using an ECL kit and ECL hyperfilm from Amersham Biosciences.
A peptide PQDTEADTAL (amino acids 504 to 513 in the 56-kDa subunit of bovine vacuolar H+-ATPase) and a peptide SHITGGDIYGIVNEN (amino acids 146 to 160 in the 70-kDa subunit of bovine vacuolar H+-ATPase) coupled through an NH2-terminal cysteine to keyhole limpet hemocyanin were used to raise in rabbits polyclonal antibodies specific for the 56- and 70-kDa subunits of the vacuolar proton pump, respectively. The antibodies were affinity-purified against the peptides used for immunization. Both antibodies detected polypeptides of the expected size when used for Western blotting of rat kidney proteins (Fig.1). The NBC3-specific antibody C1 (28, 29) was raised against 18 COOH-terminal amino acid of human NBC3. It was successfully used for NBC3 detection in Western blotting and immunocytochemistry in different mammalian species (27-29). The NHERF-1-specific antibody raised in rabbit against recombinant NHERF-1 (5) was a gift from Dr. Vijaya Ramesh.
All chemicals if not specifically indicated were purchased from Fisher (Los Angeles, CA) or Sigma and were ACS or higher quality.
Figure 2 demonstrates a typical Coomassie-stained gel depicting rat kidney proteins, bound to each of the COOH-terminal NBC3 peptides. All peptides were able to bind rat kidney proteins, but the yield and pattern differed. The unmodified COOH-terminal tail of NBC3 gave the highest yield of the proteins, whereas less total protein was bound to the truncated peptide and to the peptide in which the COOH-terminal leucine was substituted to glycine. SDS-PAGE of the proteins bound to the unmodified peptide revealed bands ranging from ∼10–500 kDa (the approximate sizes of the bands were 500, 350, 200, 72, 60, 55, 50, 38, 35, and 28 kDa), some of which potentially represented subunits of the vacuolar proton pump (∼72, ∼60, ∼55, and ∼35 kDa). Several bands had sizes distinct from the known size of the proton pump subunits. The ∼500- and ∼60-kDa bands were not detected if two other NBC3 peptides were used, and the amount of the ∼350- and ∼72-kDa bands were significantly less. There were additional bands detected by silver staining ranging from ∼10 to ∼300 kDa, whose amounts were also less, if the modified or truncated peptides were used (data not shown).
Further experiments were done to determine whether the 56-kDa subunit of the vacuolar proton pump was pulled down by the COOH terminus of NBC3. As shown in Fig. 3 A, the 56-kDa subunit was present if the unmodified NBC3 peptide was used (lane 1). If the truncated (lane 2) or modified (lane 3) NBC3 peptides were used, the 56-kDa subunit was not detected. In addition, the 70-kDa subunit of the vacuolar proton pump, known to directly bind to 56-kDa subunit (11), was pulled down by the unmodified NBC3 COOH terminus (Fig. 3 B,lane 1) but not with the truncated (lane 2) or modified (lane 3) peptide. The intact 18 amino acid COOH terminus of the 56-kDa subunit also pulled down NBC3, but the truncated or modified peptides did not (data not shown). These findings demonstrate that both NBC3 and the 56-kDa subunit COOH termini are necessary for their complexation and that both COOH termini act as PDZ-interacting motifs.
We next determined whether NHERF-1 was part of the complex of NBC3 and the proton pump. The COOH terminus of NBC3 was used in pull-down experiments. As shown in Fig.4 A, lane 1, a strong band of NHERF-1 was detected on Western blot, suggesting that NHERF-1 is present in the protein complex(s) containing the 56- and the 70-kDa subunits of the proton pump. The peptide in which the leucine residue was substituted with glycine also bound NHERF-1 (Fig.4 A, lane 2). The truncated peptide did not pull down NHERF-1 (Fig. 4 A, lane 3).
Breton et al. (5) have shown that NHERF-1 binds the 56-kDa subunit of the vacuolar proton pump and that a peptide containing the last 10 COOH-terminal amino acids of the 56-kDa subunit blocks this binding. They suggested that this binding took place because of a PDZ interaction between these proteins. In our experiments, when the 18 amino acid COOH-terminal tail of the 56-kDa subunit of the proton pump was used in pull-down experiments, NHERF-1 was detected (Fig.4 B, lane 1). In contrast to NBC3, replacement of the extreme COOH-terminal leucine to glycine significantly decreased NHERF-1 binding (Fig. 4 B, lane 3) similarly to truncation of the last five COOH-terminal amino acids (Fig. 4 B, lane 2). These data demonstrate that the COOH terminus of the 56-kDa subunit of the proton pump is, in fact, a PDZ-interacting motif.
To confirm that NHERF-1 is part of a complex of NBC3 and the proton pump, we used a GST-NHERF-1 construct to assess the in vitro binding of rat renal membrane proteins to NHERF-1 under nondenaturing conditions. The rat renal membranes were solubilized under nondenaturing conditions releasing individual proteins and protein complexes. The GST-NHERF-1 protein complexes were attached to glutathione-Sepharose beads and analyzed by SDS-PAGE and isoelectrofocusing. Both NBC3 and the 56-kDa subunit were detected by SDS-PAGE and Western blotting (Fig.5 A), indicating that both proteins bind directly/indirectly to NHERF-1 under these conditions. These results suggested, but did not demonstrate, that NBC3, the H+-ATPase, and NHERF-1 were components of the same complex. To determine whether such a complex exists, the protein complexes were eluted from glutathione-Sepharose beads under nondenaturing conditions, separated using isoelectrofocusing, and analyzed by Western blotting. The same protein band was detected using NHERF-1, NBC3, and the 56-kDa subunit-specific antibodies (Fig. 5 B, lanes 1–3), indicating that these proteins were components of the same protein complex. This band was not detected if GST-NHERF-1 was not added (Fig. 5 B, lanes 4–6).
The results in this study indicate that the COOH termini of NBC3 and the 56-kDa subunit of the vacuolar proton pump are necessary for the interaction of these proteins. Furthermore, both COOH termini act as PDZ-interacting motifs, which is necessary for the interaction of NBC3 with the vacuolar proton pump.
Our results also indicate that NHERF-1, a PDZ protein, which contains two type I PDZ domains and plays an important role in regulating the function of the apical Na+/H+ exchanger NHE3 (5), is part of a complex of NBC3 and the vacuolar proton pump (Fig. 6). Breton et al. (5) have shown that both the 56- and the 31-kDa subunits of the vacuolar proton pump are pulled down from the rat kidney homogenate by GST-NHERF-1, and a peptide containing the last 10 COOH-terminal amino acids of the 56-kDa subunit blocks this binding. Immunoprecipitation of NHERF-1 from the rat kidney homogenate revealed the 31-kDa subunit of the vacuolar proton pump. They suggested that the COOH terminus of the 56-kDa subunit of the proton pump binds to NHERF-1 through a PDZ interaction with its PDZ domains. In our experiments, the binding of NHERF-1 to the 56-kDa subunit of the vacuolar proton pump was blocked by a truncation of the last five COOH-terminal amino acids and by a modification of the COOH-terminal leucine in the 56-kDa subunit to glycine, which is typical for PDZ interactions. Therefore, our data indicate that the COOH terminus of the 56-kDa subunit of the proton pump is a PDZ-interacting motif.
Previous studies have shown that the last five amino acid residues in the COOH-terminal PDZ-interacting motifs are essential for normal PDZ-mediated binding (10, 15, 16, 19, 35, 38, 40). In the current study, a substitution of the COOH-terminal leucine to glycine in NBC3 did not affect binding of NHERF-1 but instead blocked the binding of the proton pump. NBC3 and the 56-kDa subunit have different amino acids in positions −1, −3, and −4, which may affect the different binding of their COOH termini to other proteins. Furthermore, although conjectural, our data are compatible with the possibility that the COOH terminus of NBC3 binds to NHERF-1 through both PDZ and non-PDZ interactions, whereas the 56-kDa subunit may bind only via a PDZ-mediated interaction (Fig. 6). Further studies are required to address the question as to whether NBC3 binds to NHERF-1 directly or whether additional PDZ and non-PDZ proteins are involved.
Our data indicate that NHERF-1 is a component of a complex of NBC3 and the proton pump. Whether NBC3 binds directly to NHERF-1 or indirectly through additional proteins is currently unknown. Nevertheless, we can exclude the direct interaction of NBC3 COOH terminus with subunits of the vacuolar proton pump because they do not have PDZ domains, and the PDZ interactions with the COOH termini of both NBC3 and the 56-kDa subunit of the proton pump are necessary for their complexation. It has previously been shown (41) that NHERF-1 is involved in the cAMP-dependent inhibition of the electrogenic sodium bicarbonate cotransporter NBC1 but does not interact with NBC1. Therefore, NBC3 is the first sodium bicarbonate cotransporter that has been shown to form a complex with NHERF-1. Similar to its known role in modulating the function of NHE proteins (42), NHERF-1 may play a role in regulation of NBC3.
We have previously shown that NBC3 colocalizes with the vacuolar proton pump in both type A and type B intercalated cells (28,29). NHERF-1 may play a role in targeting/retention of NBC3, as well as the proton pump to the basolateral membrane in type B intercalated cells. Although the cellular mechanisms involved in targeting of NBC3 and the vacuolar proton pump to the apical membrane in type A intercalated cells are not understood, PDZ-binding proteins other than NHERF-1 may be involved. Some of the proteins detected in the present study (Fig. 2, lane 1) are possible candidates. It is known that PDZ proteins can play a critical role in protein targeting/retention of other membrane proteins. For example, postsynaptic localization of the GLR-1 glutamate receptor (36) and the basolateral localization of the LET-23 tyrosine kinase receptor (31) in C. elegansrequire the expression of PDZ proteins LIN-10 and LIN-7, respectively. In addition, the localization of the Shaker K+channel at the neuromuscular junction is controlled by theDrosophila homolog of PSD-95 (43). NHERF-1 is involved in apical sorting of CFTR (25). Our data indicate that NHERF-1 is a component of the complex between NBC3 and the vacuolar proton pump. Whether other PDZ and non-PDZ proteins form part of this complex is unknown. Examples of this sort have been described in which PDZ proteins can bind directly with other PDZ proteins (22, 23). Known examples of proteins with multiple PDZ-binding domains that have been shown to bind transport proteins are PSD95, which binds potassium channels (13), theDrosophila InaD protein, which binds light-activated ion channels (39), and NHERF-1, which binds NHE3, CFTR, and the 56-kDa subunit of the vacuolar proton pump (5, 25,42). NHERF-1 and NHERF-2 have been shown to oligomerize (22), which may facilitate interaction between different proteins. Furthermore, PDZ proteins can form cysteine disulfide bonds that mediate the formation of multimeric complexes (18). Finally, the formation of multiprotein complexes, containing NBC3 and the vacuolar proton pump, may be dynamically regulated as has been described in the example of the Kir2.3 binding to PSD-95 (21), the interaction of EphB3 receptor tyrosine kinase with the PDZ protein AF-6 (17), and the interaction of the β2-receptor with the PDZ domain of NHERF-1 (15).
The findings in this study are the first demonstration that a member of the bicarbonate transport superfamily has a COOH-terminal PDZ-interacting motif that plays an essential role in protein-protein interaction. We can only speculate at this point as to the functional importance of the interaction of NBC3 and the vacuolar proton pump. NBC3 may modulate the activity of the vacuolar H+-ATPase by altering the local bicarbonate concentration in cells in which the two transporters colocalize and/or by protein-protein interactions through NHERF-1 and other PDZ protein/s. In this regard, it is interesting to note that studies of rat renal endocytotic vesicles have shown that bicarbonate, in addition to chloride, can stimulate proton pump activity (14, 32). Similar findings have been reported with the gastric H+-K+-ATPase and the mitochondrial H+-ATPase (3, 9); however, the mechanism of this stimulatory effect is unknown. In vesicles derived from Dictostelium bicarbonate stimulates vacuolar H+-ATPase activity and can shunt the electrical potential generated by electrogenic proton pumping (11). Further studies are needed to clarify the physiological importance of the interaction of NBC3 and the 56-kDa subunit of the vacuolar proton pump in intercalated cells.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-58563 and DK-07789, the Max Factor Family Foundation, the Richard and Hinda Rosenthal Foundation, and the Fredericka Taubitz Foundation. N. Abuladze is supported by a training grant from the National Kidney Foundation of Southern California J891002.
Address for reprint requests and other correspondence: I. Kurtz (E-mail:) or A. Pushkin (E-mail: ), UCLA Div. of Nephrology, 10833 Le Conte Ave., Rm. 7-155 Factor Bldg., Los Angeles, CA 90095-1689.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
First published November 20, 2002;10.1152/ajpcell.00225.2002
- Copyright © 2003 the American Physiological Society