Differential renal distribution of NHERF isoforms and their colocalization with NHE3, ezrin, and ROMK

James B. Wade, Paul A. Welling, Mark Donowitz, Shirish Shenolikar, Edward J. Weinman

Abstract

Na+/H+ exchanger regulatory factor (NHERF) and NHERF2 are PDZ motif proteins that mediate the inhibitory effect of cAMP on Na+/H+ exchanger 3 (NHE3) by facilitating the formation of a multiprotein signaling complex. With the use of antibodies specific for NHERF and NHERF2, immunocytochemical analysis of rat kidney was undertaken to determine the nephron distribution of both proteins and their colocalization with other transporters and with ezrin. NHERF was most abundant in apical membrane of proximal tubule cells, where it colocalized with ezrin and NHE3. NHERF2 was detected in the glomerulus and in other renal vascular structures. In addition, NHERF2 was strongly expressed in collecting duct principal cells, where it colocalized with ROMK. These results indicate a striking difference in the nephron distribution of NHERF and NHERF2 and suggests NHERF is most likely to be the relevant biological regulator of NHE3 in the proximal tubule, while NHERF2 may interact with ROMK or other targets in the collecting duct. The finding that NHERF isoforms occur in different cell types suggests that NHERF and NHERF2 may subserve different functions in the kidney.

  • sodium hydrogen exchanger regulatory protein 2
  • sodium hydrogen exchanger 3 protein kinase A regulatory protein
  • signal complexes
  • PDZ proteins

the na+ /h+ exchanger regulatory factor (NHERF) and the Na+/H+ exchanger 3 protein kinase A regulatory protein (E3KARP, NHERF2) are two recently described PDZ (PSD-95/Dlg/ZO-1) domain-containing proteins (14, 16,19). In a model fibroblast cell line, both proteins appear to facilitate the assembly of a multiprotein complex that includes Na+/H+ exchanger 3 (NHE3), ezrin, and cAMP-dependent protein kinase II (PKA) (7, 18-20). This signaling complex links NHE3 to the cytoskeleton and is necessary for cAMP-dependent phosphorylation of NHE3 with subsequent inhibition of Na+/H+ exchange transport. Other studies have suggested a role for the NHERF family of proteins in the targeting and endocytic retrieval and sorting of membrane receptors, transporters, channels, and signaling proteins (5, 6, 9-11,13, 15). In addition, a role for this family of proteins in the function of membrane receptors such as the platelet-derived growth factor receptor has been advanced. Accordingly, NHERF and NHERF2 may play important roles in the regulation of kidney function.

NHERF and NHERF2 are similar proteins that share 57% overall amino acid identity with even higher degrees of identity within the two tandem PDZ motifs (16, 19). A major difference between the proteins is that NHERF, but not NHERF2, is a phosphoprotein in vivo (1, 7, 20). Although both proteins support cAMP-mediated inhibition of NHE3 when expressed in PS120 cell fibroblasts, the precise roles of NHERF and NHERF2 in the kidney are not known (7,19, 20). Moreover, it has not been shown that the model of the signal complex regulation of NHE3 developed in fibroblasts is applicable to the kidney. To understand more fully the role of the NHERF family of proteins in kidney, the present studies were undertaken to examine the nephron distribution of NHERF and NHERF2 and to determine their colocalization with transporters and with ezrin. The results indicate a markedly different nephron distribution of NHERF compared with NHERF2. The colocalization of NHERF with NHE3 in the proximal tubule and NHERF2 with ROMK in the cortical collecting duct suggests functional differences between these proteins in kidney. The colocalization of NHE3 with NHERF and ezrin provides the first experimental evidence in epithelial cells that is consistent with the signal complex model of regulation of NHE3 by cAMP developed in a model fibroblast cell line. Finally, the wide expression of NHERF2 in the renal vasculature suggests the potential for a currently unknown regulatory role for NHERF2.

METHODS

The specificity of the antibodies was tested by using PS120/NHE3V fibroblasts stably transfected with rabbit NHE3 tagged at its COOH terminus with an epitope (YTDIEMNRLGK) derived from vesicular stomatitis virus glycoprotein as previously described (19). Mouse NHERF or human NHERF2 cDNAs were cloned into pcDNA3.1/Hygro+ vectors and transfected into the PS120/NHE3V fibroblasts by using Lipofectin (GIBCO BRL). Cells resistant to 600 U/ml of hygromycin were selected through eight passages before the study. The cells were scraped and suspended in a solution that contained 10 mM NaPO4 (pH 7.4), 100 mM NaCl, 5 mM EDTA, 1 mM Na3VO4, 50 mM NaF, and protease inhibitors (0.1 mM phenylmethylsulfonyl fluoride, 1 mM phenanthroline, and 5 mg/ml each of aprotinin, leupeptin, pepstatin, and trypsin inhibitor). Cells were collected by centrifugation for 10 min at 12,000g, resuspended in 1 ml of the above solution containing 1% Triton X-100, lysed by being drawn several times through a 27-gauge needle, boiled in Laemmli buffer, and resolved on 10% SDS-PAGE. The proteins were transferred to nitrocellulose and immunoblotted using a polyclonal antibody to full-length recombinant rabbit NHERF or an antibody to the terminal 103 amino acids of NHERF2. Both antibodies have previously been described (7, 18-20). Immune complexes were detected by enhanced chemiluminescence (Amersham, Arlington Heights, IL).

Antibodies to aquaporin-2 (LC-54) and ROMK (LC-35) were raised in chicken to COOH-terminal peptides of these proteins. Antibody to ezrin produced in goat (SC-6407) was obtained from Santa Cruz Biotechnology, Santa Cruz, CA. For localization of NHE3 with respect to NHERF, monoclonal antibody 2B9 was kindly provided by Dr. D. Biemesderfer (2).

Male Sprague-Dawley rat kidneys (180–250 g) were fixed by perfusion via the abdominal aorta. Perfusion was for 2 min in PBS to clear the kidneys of blood and 5 min in 2% paraformaldehyde. Preliminary localizations done on such briefly fixed kidneys showed uneven and inconsistent localizations of the studied proteins unless the kidneys underwent additional fixation as sections or as sliced tissue for 1 h at 4°C. Other procedures followed those previously described in detail (12). Sections were treated with either 1% SDS or 6 M guanidine for 10 min to unmask antigenic sites. Similar localizations were obtained with either treatment or with no unmasking treatment (with less labeling). Sections were incubated with primary antibody overnight at 4°C. After being washed, they were incubated with secondary antibody for 2 h at 4°C. Appropriate species-specific antibodies coupled to Alexa 488 or 568 dyes (Molecular Probes, Eugene, OR) were diluted 1:100. Sections were examined with a Zeiss LSM410 confocal microscope.

RESULTS

The specificity of the NHERF and NHERF2 antibodies was tested by using Western immunoblot analysis of lysates from PS120 cells expressing either rabbit NHERF or human NHERF2. As shown in Fig. 1 A, the anti-NHERF antibody recognized a protein of appropriate size in PS120 cells expressing NHERF but not in PS120 cells transfected with NHERF2 or in control nontransfected PS120 cells. As shown in Fig. 1 B, the anti-NHERF2 antibody recognized a protein of appropriate size in PS120 cells expressing NHERF2 but not in PS120 cells transfected with NHERF or control cells. This latter result is identical to prior studies in which the same antibody was used (18). Collectively, these findings establish that the anti-NHERF antibody was specific for NHERF with little or no cross reactivity to NHERF2 and that the antibody to NHERF2 is specific for NHERF2 with little cross reactivity to NHERF. These antibodies were also found to recognize single bands the size of NHERF and NHERF2 in rat kidney cortex, as shown in Fig.2.

Fig. 1.

Western immunoblot of cell lysates from control PS120 cell fibroblasts (PS120), PS120 cells transfected with mouse Na+/H+ exchanger regulatory factor (NHERF; PS120 NHERF), or transfected with human NHERF2 (PS120 NHERF2). An antibody to recombinant full-length rabbit NHERF was used inA, and an antibody to a polypeptide representing the terminal 103 amino acids of NHERF2 was used in B.

Fig. 2.

Western immunoblot of a cortical homogenate from rat kidneys. The antibody to rabbit NHERF was used in the leftlane (NHERF) and the antibody to NHERF2 was used in theright lane (NHERF2). These antibodies recognize single bands in the kidney that correspond to the proteins detected in transfected cells in Fig. 1.

Figure 3 A shows the presence of NHERF in the proximal tubule. It occurs strongly in all proximal regions. The greatest abundance of the protein is detected in or at the brush-border membrane, but protein can also be weakly detected in the cytosol and in the basolateral membrane. The specificity of the antibody signal is confirmed in Fig. 3 B, which demonstrates that the NHERF signal can be blocked by preincubation of the antibody with its antigen, recombinant full-length rabbit NHERF. Despite the use of a variety of fixative and epitope-uncovering techniques (seemethods), NHERF was not regularly detected in any other nephron structures, including the cells of thick ascending limb or collecting ducts. In some sections, weak staining was seen in the glomerulus and in the inner medullary collecting duct, but this staining was not consistent and was not blocked by preabsorption of the antibody with its antigen.

Fig. 3.

Immunolocalization of NHERF in the renal cortex. A: antibody labels the apical brush-border domains of renal proximal tubules but not adjacent distal (D) tubules. Weaker staining of the basolateral membrane is also seen. B: sections incubated with the NHERF antibody and an excess of the recombinant protein antigen have only background labeling. Scale bar, 100 μm.

Figure 4 A shows the detection of NHERF2 in the glomerulus and the elimination of this signal when the antibody was preincubated with its peptide antigen (B). Examination of the glomerular staining by NHERF2 indicated intense staining of capillary loops in a pattern consistent with labeling of endothelial cells. It is also possible that NHERF2 occurs in glomerular mesangial and epithelial cells, but additional studies are needed to establish with certainty which glomerular cells express NHERF2. Figure4 C shows that NHERF2 is absent in proximal tubules, but antibody to it strongly labels certain cells of the cortical collecting duct. Comparison of Fig. 4 C with the same section stained with an antibody to aquaporin-2 (D) shows that NHERF2 staining is confined to the principal cells of the cortical collecting duct. Figure 4 E shows NHERF2 labeling in vascular structures in the inner medulla. Label was not detected in the descending thin limbs of the loop of Henle (not shown). The cells of the thick ascending limb (marked by asterisks in Fig. 4, E andF) are not labeled for NHERF2, but these structures are strongly labeled in the same section by antibody to ROMK (F). NHERF2 was also detectable in principal cells of outer medullary collecting duct but not in the inner medullary collecting duct (data not shown).

Fig. 4.

Immunolocalization of NHERF2. A: antibody to NHERF2 strongly labels the glomerulus. B: incubation of sections with the NHERF2 antibody and an excess of the recombinant protein antigen blocks glomerular labeling completely. Scale bar, 100 μm.C: NHERF2 is present in a subset of collecting duct cells (arrows) in the endothelial cells of peritubular capillaries (PC) but not in adjacent proximal tubules (PT). D: immunolocalization of aquaporin-2 in the same section shown in C, indicating that the collecting duct cells labeled by NHERF2 (arrows) are principal cells. Scale bar, 25 μm. E: immunolocalization of NHERF2 in the outer medulla shows very bright labeling of vessels in the vascular bundle (VB) and peritubular capillaries. *Thick ascending limbs of the loop of Henle are unlabeled. F: labeling using an antibody to ROMK in the same region shown in E indicates the position of strongly labeled thick ascending limbs of the loop of Henle (*). Scale bar, 25 μm.

Additional studies were performed to study the colocalization of NHERF, ezrin, and NHE3 in the proximal tubule and the colocalization of NHERF2 and ROMK. Figure 5, A–C, demonstrates the presence of NHERF (A) and ezrin (B) and the colocalization of the proteins (C) in the proximal tubule. Figure 5, D–F, demonstrates the presence of NHERF (D), NHE3 (E), and the colocalization of NHERF and NHE3 (F) in the proximal tubule. Figure 5, G–I, demonstrates the presence of NHE3 (G), ezrin (H), and the colocalization (yellow) of NHE3 (green) and ezrin (red) in the proximal tubule(I). While there is an extensive band of yellow colabel in the brush-border membranes, there are also regions in proximity to the areas of colocalization where ezrin and NHE3 are not associated with one another. Collectively, these findings indicate that NHERF, NHE3, and ezrin are in close proximity in the renal proximal tubule and suggest that they could function as a signal complex. Figure6, A–C, demonstrates the presence of NHERF2 (A), ROMK (B), and the colocalization of the two proteins in the cortical collecting duct (C). While NHERF2 and ROMK colocalized at or near the apical membrane, there was considerable cell-to-cell variation with both ROMK and NHERF2 occurring within the cytoplasm.

Fig. 5.

Colocalization of NHERF with ezrin and Na+/H+ exchanger 3 (NHE3) in the renal cortex.A–C: double labeling with antibodies to NHERF (A) and ezrin (B) are shown together inC with NHERF labeling (green) and ezrin labeling (red). The colocalization of these components in proximal tubules results in a yellow labeling of the proximal brush border in C. Scale bar, 100 μm. D–F: double labeling with antibodies to NHERF (D) and NHE3 (E) are shown together inF with NHERF labeling (green) and NHE3 labeling (red). The colocalization of these components in proximal tubules results in a yellow labeling of the proximal brush border in F. Scale bar for D–F, 25 μm. Of note, there is a zone of NHE3 label (red) at the base of the brush border that does not appear to be associated with significant NHERF labeling. G–I: double labeling with antibodies to NHE3 (G) and ezrin (F) are shown together in I with NHE3 labeling (red) and ezrin labeling (green). A region of colocalization (yellow) is seen in the brush-border membrane, but regions predominated by ezrin (green) and NHE3 (red) occur above and below the region of colocalization, respectively. Scale bar for panels G–I, 25 μm.

Fig. 6.

Double labeling with antibodies to NHERF2 (A) and ROMK (B) are shown together in C with NHERF2 labeling (green) and ROMK labeling (red). Scale bar, 100 μm.

DISCUSSION

The biological role of the NHERF family of proteins (NHERF and NHERF2) has increasingly been defined in recent studies from a number of laboratories (5–7, 9–11, 13–16, 18–20; reviewed in Ref. 8). NHERF was originally isolated from renal brush-border membranes and identified to be a necessary cofactor in cAMP regulation of NHE3 (14). Shortly after the cloning of NHERF, NHERF2 was isolated in a yeast 2 hybrid screen of a human lung library using the COOH terminus of NHE3 as bait (16, 19). NHERF and NHERF2 are similar proteins that share 57% overall identity with an even higher degree of identity within the two tandem PDZ protein-protein interactive motifs of each protein. Studies in a model PS120 fibroblast cell line have indicated that NHERF facilitates the assembly of a signal complex of proteins including NHE3, ezrin, and PKA to phosphorylate NHE3 and inhibit its activity (7, 18,20). In the same PS120 cells, NHERF2 also supports cAMP-mediated inhibition of NHE3 activity (19). The overlap in function regarding the effect of cAMP on NHE3 activity in PS120 cells has led to the assumption that NHERF and NHERF2 have redundant cellular effects. However, the distribution of the NHERF and NHERF2 proteins along the length of the nephron has not been determined. Moreover, there has been no evidence to date to indicate that the signal complex model of regulation of NHE3 by NHERF or NHERF2 developed in fibroblast cells is operative in an epithelial tissue. To understand more fully the potential role of the NHERF family of proteins in renal physiology and to attempt to distinguish NHERF from NHERF2, immunocytochemical studies were undertaken to determine the nephron distribution of the two proteins and to determine the colocalization of these proteins with other transporters and regulatory proteins.

As validated in PS120 cells expressing either NHERF or NHERF2, the rabbit polyclonal antibody to recombinant full-length NHERF and a polyclonal antibody to a peptide representing the COOH terminus of NHERF2 did not cross-react with one another. Immunocytochemical studies showed that the distribution of NHERF in the kidney was significantly different from NHERF2, and, in fact, there were no areas of overlap in the distribution of the two proteins. NHERF was found in the renal proximal tubule with prominent staining located in the apical brush-border membrane, the cytosol, and, to a lesser degree, in the basolateral membrane. This finding is consistent with prior fractionation studies using brush-border membrane, basolateral membrane, and cytosolic fractions of renal tubular cells (14, 15). NHERF was not detected in any other nephron structures including the glomerulus, thick ascending limb, collecting ducts, or in renal vascular structures. By contrast, NHERF2 was strongly expressed in the glomerulus and other vascular elements of the kidney and in the cortical and medullary collecting ducts. In the glomerulus, bright staining was seen in a pattern consistent with labeling of endothelial cells. In preliminary studies, however, we found that the distribution of NHERF2 was not absolutely identical to that of von Willebrand factor, suggesting the presence of NHERF2 in other glomerular cells, including the mesangial and epithelial cells. Of particular interest, NHERF2 could not be detected in the proximal tubules, and if present in the proximal tubules, it is in amounts below the level of sensitivity of the NHERF2 antibody and much less abundant than at other renal sites. The finding that NHERF and NHERF2 have a different renal distribution suggests important differences in the physiological effects of these proteins. As studied in PS120 cell fibroblasts that express NHE3, coexpression of either NHERF or NHERF2 confers sensitivity of the exchanger to cAMP (19). Despite the apparent overlap in function of the two proteins when assayed in a model cell system, the finding that NHERF, but not NHERF2, is present in renal proximal tubule cells indicates that NHERF is the important biological regulator of Na+/H+ exchange activity in the renal proximal tubule.

Recent experiments in PS120 cells have established that NHERF functions as an adaptor protein to facilitate formation of a multiprotein signal complex including NHE3, ezrin, and PKA. NHERF coimmunoprecipitates with both ezrin and NHE3 and is suggested to bring PKA bound to ezrin into physical proximity to the COOH terminus of NHE3 (7, 18,20). Subsequent PKA-mediated phosphorylation of NHE3 results in a decrease in transport activity. Before the present observations, however, there was no evidence that this model is applicable in epithelial cells. The present results provide evidence that NHERF colocalizes with ezrin and with NHE3 in the proximal tubules of rat. In turn, NHE3 colocalizes with ezrin. These three proteins appear to be localized near or in the apical membrane, findings that provide the first morphological evidence supporting the signal complex model in renal tubule cells. Additional studies will be required to demonstrate that the signal complex is functional in this tissue.

The present experiments confirm previous studies and indicate the presence of NHE3 in the apical border of thick ascending limb cells (2). Moreover, it has been demonstrated that cAMP inhibits Na+/H+ exchange transport in this nephron segment (4). Using reverse transcription and PCR techniques, we have reported the presence of NHERF mRNA in rabbit thick ascending limb cells (16). In preliminary studies of SV40 transformed mouse medullary thick ascending limb cells, we have been able to detect NHERF protein. Nonetheless, in the present experiments we were unable to detect either NHERF or NHERF2 in the thick ascending limb cells. The failure to detect NHERF in the thick ascending limb of rat kidney with the use of immunocytochemistry, compared with its detection in a mouse thick ascending limb cell line with the use of Western immunoblot analysis, may be related to sensitivity of the antibodies when used in different assays. Nonetheless, these findings raise the interesting speculation that the relation between NHERF and NHE3 as elucidated in PS120 cells and by the present immunocytochemical analysis of proximal tubule cells may not be generalized to all nephron segments. It is distinctly possible, for example, that there are other unidentified NHERF isoforms or that other PDZ proteins function to modulate the effect of cAMP on NHE3 in the thick ascending limb.

Welling et al. (17) provided intriguing evidence for an association between NHERF and ROMK. With the use of a pull-down assay, the COOH terminus of ROMK bound to a kidney protein identified by immunoblot analysis to react with an anti-NHERF antibody. Further studies in which recombinant proteins and expression systems were used have suggested that NHERF may function to link ROMK to cystic fibrosis transmembrane conductance regulator (CFTR). The ability of NHERF2, on the other hand, to support the association between CFTR and ROMK has not yet been explicitly studied. This consideration becomes relevant given our observation that NHERF2 rather than NHERF colocalized with ROMK in rat kidney. As discussed above, given the similarities between NHERF and NHERF2, these proteins may overlap function in expression systems where only one of the proteins is expressed. On the basis of findings of the present experiments, NHERF2 is more likely to be the important biological regulator of ROMK in the cortical collecting duct. On the other hand, ROMK but not NHERF2 was detected in the thick ascending limb, suggesting that the association of ROMK with regulatory proteins in the thick ascending limb may differ from that in the cortical collecting duct.

A recent study by Breton et al. (3) also provides evidence for a cell type-specific functional expression of NHERF isoforms. Using a different NHERF antibody, these investigators detected NHERF in the apical membrane of the renal proximal tubule, as shown here, and also in B-type intercalated cells of collecting ducts where it interacts with the H+-ATPase of these cells. Together with our findings, it is of interest that two adjacent cell types in the collecting duct express different members of the NHERF PDZ family of proteins, NHERF2 in the principal cells and NHERF in B-intercalated cells.

In summary, the present experiments establish that the nephron distributions of NHERF and NHERF2 are markedly different and suggest that NHERF and NHERF2 may subserve different functions in kidney. NHERF is expressed in the renal proximal convoluted tubule, where it colocalizes near the apical membrane of the proximal tubule in association with ezrin and NHE3. These findings extend studies from fibroblasts to the renal proximal tubule and are consistent with the proposed model of PKA regulation of NHE3 involving NHERF and the formation of a signal complex. The absence of NHERF2 in the renal proximal tubule indicates that NHERF, rather than NHERF2, is more likely to be the important biological regulator of NHE3 in the proximal tubule. NHERF2, on the other hand, is expressed in the glomerulus and in other renal vascular structures and in the collecting duct. In the cortical collecting duct, NHERF2 and ROMK colocalize at or near the apical membrane of principal cells, suggesting a biological interaction between these proteins. On the other hand, NHE3 and ROMK, but not NHERF or NHERF2, were detected in the thick ascending limb cells. It is possible that the regulation of NHE3 and ROMK in this nephron segment involves mechanisms distinct from an interaction with either NHERF or NHERF2. Finally, an unanticipated finding of the present studies was the detection of NHERF2 throughout the renal vascular structures. Although the physiological role of NHERF2 in these structures is unknown at the present, the association suggests the potential for a new regulatory role of NHERF2 in renal function.

Acknowledgments

We acknowledge the expert technical assistance of Deborah Steplock and Jie Liu. The NHERF2 polyclonal antibody and the NHE3 monoclonal antibody were kindly provided by Dr. C.-H. Yun, Gastroenterology Div., Dept. of Medicine, Johns Hopkins Univ. School of Medicine, Baltimore, MD, and Dr. D. Biemesderfer, Depts. of Internal Medicine and Cellular and Molecular Physiology, Yale Univ. School of Medicine, New Haven, CT, respectively.

Footnotes

  • This study was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-32839 (to J. B. Wade), DK-54231 (to P. Welling), DK-26523 and DK-44484 (to M. Donowitz), DK-55881 (to E. J. Weinman and S. Shenolikar), and a Department of Veterans Affairs Research Service grant (to E. J. Weinman).

  • Dr. Welling is the recipient of an Established Investigator Award from the American Heart Association.

  • Address for reprint requests and other correspondence: J. B. Wade, Dept. of Physiology, 655 W. Baltimore St., Univ. of Maryland, Baltimore, MD 21201 (E-mail: jwade{at}umaryland.edu).

  • 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.

REFERENCES

View Abstract