INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE SODIUM-BICARBONATE COTRANSPORTERS (NBCs) are a family of integral membrane proteins that mediate electrogenic or electroneutral sodium-bicarbonate cotransport (1, 2, 4, 12, 29, 46, 47, 51). Members of the NBC protein family play an important role in intracellular pH regulation and transepithelial transport of sodium and bicarbonate in several tissues. In various cell types in the eye, electrogenic sodium-bicarbonate cotransport plays an important role in intracellular pH regulation and transepithelial ion flux. The importance of electrogenic sodium-bicarbonate cotransport to normal ocular function is highlighted by recent evidence that patients with homozygous mutations in the NBC1 (SLC4A4) gene have an ocular phenotype characterized by blindness, glaucoma, cataracts, and band keratopathy (28). Extraocular manifestations include severe proximal renal tubular acidosis, mental retardation, short stature, an elevated serum amylase, and thyroid abnormalities. The cellular mechanism(s) responsible for the ocular abnormalities in these patients is presently not known. Characterization of the expression pattern of NBC1 proteins in the eye would improve our understanding of the molecular pathogenesis of the cause of blindness in this disorder.

There are several cell types in the eye where functional electrogenic sodium-bicarbonate cotransport has been reported. Electrogenic sodium-bicarbonate cotransport has been described in corneal endothelial cells (10, 30), pigmented ciliary body epithelium (64), lens epithelium (62), retinal astrocytes and Müller cells (42-45), and retinal pigment epithelium (26, 31, 35). These studies have utilized several different model systems including intact epithelial preparations, cultured cells, and vesicle preparations from various species. Depending on the specific cell type, electrogenic bicarbonate transport is thought to play an important role in mediating transepithelial sodium, bicarbonate, and fluid transport (e.g., corneal endothelial cells, ciliary body epithelium, retinal pigmented epithelium), intracellular pH regulation, and maintenance of extracellular pH (retinal astrocytes and Müller cells). Ion and fluid transport in these tissues are thought to play an important role in the control of corneal transparency, lens transparency, intraocular pressure, retinal function, and attachment.

In humans, electrogenic sodium-bicarbonate cotransport is mediated by kNBC1 and pNBC1 proteins, which are encoded by the NBC1 gene (2). kNBC1 and pNBC1 are highly homologous proteins that have different NH₂ termini, in that in kNBC1 41 amino acids replace the initial 85 amino acids of pNBC1 (1). pNBC1 is highly expressed in the pancreas, where it is thought to play an important role in mediating basolateral bicarbonate influx in ductal cells and resultant transepithelial bicarbonate secretion, whereas kNBC1 is thought to play an important role in mediating renal proximal tubular bicarbonate absorption (1, 12, 51, 52). The expression pattern of these transporter proteins in various cells in the eye has not been characterized. Knowledge of the cellular distribution of electrogenic sodium-bicarbonate cotransport proteins in ocular tissues would begin to provide a molecular basis for understanding the potential role of loss of function mutations in pNBC1 and kNBC1 in causing blindness (28). In a recent study, Sun et al. (56) reported the presence of NBC1 on the basolateral membrane of corneal endothelial cells. However, there are presently no studies that have utilized specific antibodies to determine the tissue expression pattern of pNBC1 and kNBC1 proteins. To gain further insight into the molecular mechanisms responsible for electrogenic sodium and bicarbonate transport in the eye, we generated rabbit polyclonal antibodies against the unique NH₂ terminus of pNBC1 and kNBC1 and determined the cellular localization of each protein in the rat eye. A comprehensive analysis of the expression of pNBC1 and kNBC1 in various ocular tissues is reported.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Generation and characterization of polyclonal antibodies to pNBC1 and kNBC1. A polyclonal antibody to pNBC1 was raised in rabbits against a synthetic peptide derived from the NH₂ terminus of the protein (amino acids 1-19, coupled to an NH₂-terminal cysteine). The affinity-purified polyclonal antibody to kNBC1 was raised against a synthetic peptide, corresponding to amino acids 11-24, coupled to an NH₂-terminal cysteine. Both antibodies were affinity purified using Sepharose 4B columns with covalently attached pNBC1 or kNBC1 peptides. To characterize the antibodies, HEK-293T cells grown on fibronectin-coated coverslips were transfected using the calcium phosphate precipitation method with plasmids (pcDNA3.1, Invitrogen, Carlsbad, CA) containing the coding region for pNBC1 or kNBC1. The cells were rinsed in methanol at 20°C for 5 min and washed in PBS several times. Each primary antibody (1:100 dilution) was applied for 1 h and, after several washes in PBS, goat anti-rabbit IgG conjugated with Alexa 594 (1:500 dilution, Molecular Probes, Eugene, OR) was applied for an additional 1 h. The slides were washed in PBS and mounted in Cytoseal 60 (Stephens Scientific, Riverdale, NJ). A liquid-cooled PXL charge-coupled device camera (model CH1, Photometrics, Osnabruck, Germany), coupled to a Nikon Microphot-FXA epifluorescence microscope, was used to capture and digitize the fluorescence images of the cells expressing pNBC1 or kNBC1. The images were transferred to a Silicon Graphics Indy 5000 computer using ISEE 4.0 (c) software (Inovision, Raleigh, NC).

Western blotting. The eyes from normal Long-Evans rats were divided into cornea, ciliary body, conjunctiva, and retina. These tissues were disrupted at 0°C in a glass homogenizer with 20 mM Tris · HCl, pH 7.5, containing 140 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 1 mM EDTA, 1 µg/ml pepstatin, 1 µg/ml leupeptin, and 1 µg/ml aprotinin. The homogenate was centrifuged at 300 g for 10 min, and the supernatant was analyzed by SDS-PAGE. Membranes from rat pancreas and kidney were isolated using differential centrifugation. The proteins were separated by SDS-PAGE and electrotransferred onto a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA). Nonspecific binding was blocked by incubation of the membrane in 20 mM Tris · HCl, pH 7.5, 140 mM NaCl, and 0.1% Tween 20 (TBST) containing 5% dry milk. Primary pNBC1- or kNBC1-specific antibodies were used at a dilution of 1:1,000. Secondary horseradish peroxidase-conjugated mouse anti-rabbit antibody (Jackson ImmunoResearch, West Grove, PA) was used at a dilution of 1:10,000. The bands were visualized using an enhanced chemiluminescence (ECL) kit and ECL Hyperfilm (Amersham Pharmacia Biotech, Keene, PA).

Immunocytochemistry. For immunofluorescence experiments, rat eyes were removed and immediately frozen in dry ice. The pNBC1 and kNBC1 primary antibodies (1:100 dilution) were applied to 5-µm cryostat sections for 45 min and, after several washes in PBS, goat anti-rabbit IgG conjugated with Alexa 594 (1:500 dilution, Molecular Probes) was applied for 40 min. In control experiments, the primary antibodies were preabsorbed with the specific immunizing peptide (10 µg/ml). The slides were washed in PBS and mounted in Cytoseal 60 (Stephens Scientific). The confocal images were captured with a Leica TCS SP inverted confocal microscope (Leica) using a krypton laser (model 643, Melles Griot, Irvine, CA). In some experiments, rat eyes were dissected and immersed immediately in PBS with 4% paraformaldehyde. After fixation for 24 h at 4°C, the samples were washed in PBS, dehydrated through a graded series of ethanol washes, cleared in xylene, and embedded in paraffin. Sections were cut at a thickness of 7 µm, mounted onto glass slides, and stored at room temperature until use. The slides were deparaffinized in graded degrees of ethanol and treated with xylene. The sections were immersed in freshly made 3% hydrogen peroxide prepared in water for 30 min and then rinsed in water. The pNBC1 and kNBC1 primary antibodies (1:100 dilution) were applied for 20 min and, after a 5-s wash, were incubated in a biotinylated anti-rabbit secondary antibody (STAT-Q, Innovex Biosciences, Richmond, CA) for 10 min. The sections were then incubated in peroxidase-labeled strepavidin for 10 min, immersed in a diaminobenzidine substrate solution for 5 min, and subsequently counterstained with hematoxylin and mounted.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

To establish the presence and examine the cellular distribution of pNBC1 and kNBC1 in the rat eye, rabbit affinity-purified polyclonal antibodies were raised against the specific NH₂ terminus of each protein. The polyclonal antibodies against pNBC1 and kNBC1 were initially characterized by transfection experiments in HEK-293T cells. As shown in Fig. 1, immunofluorescence microscopy revealed that pNBC1 and kNBC1 are readily detected in cells expressing each protein. No labeling was observed in untransfected cells. Labeling of the pancreas with the anti-pNBC1 antibody and of the kidney with the anti-kNBC1 antibody are also shown.

View larger version (49K): [in this window] [in a new window]   Fig. 1.   A-D: immunocytochemical analysis of the sodium-bicarbonate cotransporter expression of pNBC1 and kNBC1 in HEK-293T cells. A: transfected cells labeled with the anti-pNBC1 antibody, showing a predominantly plasma membrane localization of pNBC1. These cells did not label with the anti-kNBC1 antibody. B: untransfected cells labeled with the anti-pNBC1 antibody. C: transfected cells labeled with the anti-kNBC1 antibody, showing a predominantly plasma membrane localization of kNBC1. These cells did not label with the anti-pNBC1 antibody. D: untransfected cells labeled with the anti-kNBC1 antibody. E: overlapping Nomarski and immunofluorescence image showing labeling of pancreatic ducts with the anti-pNBC1 antibody. F: immunofluorescence staining of proximal tubules in the kidney with the anti-kNBC1 antibody.

The pNBC1 antibody recognized a ~145-kDa band on immunoblots of rat cornea, ciliary body, retina, and conjunctiva (Fig. 2). The specificity of the labeling was confirmed using the pNBC1 antibody preabsorbed with the immunizing peptide. In addition, data from rat pancreatic membranes that strongly express pNBC1 are shown for comparison. The size of the recognized protein in all tissues is larger than the predicted size of ~121 kDa, suggesting that pNBC1 is glycosylated or posttranslationally modified in other ways.

View larger version (86K): [in this window] [in a new window]   Fig. 2.   Immunoblots from various rat ocular tissues. A: lanes 1 and 6: cornea; lanes 2 and 7: ciliary body; lanes 3 and 8: retina; lanes 4 and 9: conjunctiva; and lanes 5 and 10: pancreas. Loading: lanes 1-4 and 6-9: 80 µg; lanes 5 and 10: 20 µg. Lanes 1-5 were probed with the anti-pNBC1 antibody; lanes 6-10 were probed with the anti-pNBC1 antibody preincubated with 10 µg/ml of specific immunizing peptide. B: lanes 1 and 6: cornea; lanes 2 and 7: ciliary body; lanes 3 and 8: retina; lanes 4 and 9: conjunctiva; lanes 5 and 10: kidney. Loading: lanes 1-4 and 6-9: 80 µg; lanes 5 and 10: 40 µg. Lanes 1-5 were probed with the anti-kNBC1 antibody; lanes 6-10 were probed with the anti-kNBC1 antibody preincubated with 10 µg/ml of specific immunizing peptide.

Unlike pNBC1, kNBC1 was found in conjunctiva only. As shown in Fig. 2, the kNBC1 antibody recognized a ~130-kDa band in conjunctiva only. Furthermore, kNBC1 was not detectable in cornea, ciliary body, or retina. These findings suggest that kNBC1 plays a more specialized role in mediating electrogenic sodium-bicarbonate cotransport in the conjunctiva. Given a predicted size of ~116 kDa, the results indicate that the kNBC1 is posttranslationally modified. In an immunoblot of rat kidney membranes known to strongly express kNBC1, a band of ~130 kDa was also detected and is illustrated in Fig. 2 for comparison.

pNBC1 immunostaining was observed in many ocular tissues, and in a cell-specific pattern. Figure 3 depicts the immunolocalization of pNBC1 in the cornea. In the cornea, epithelial cells in the basal and intermediate cell layers were stained most intensely, with only occasional cells staining in the surface layer (Fig. 3, A-D). Cells in the basal and intermediate layers in the peripheral cornea (Fig. 3A) were more uniformly stained than those of the central cornea (Fig. 3, B and D). Within the central cornea, some of the cells in the basal layer were not labeled (Fig. 3, B and D). There was no clearly defined preferential staining of apical and basolateral membrane domains in the basal layer or elsewhere. As shown in Fig. 3, A, B, and D, keratocytes of the corneal stroma were also stained for pNBC1. Unlike pNBC1, kNBC1 was not detected in corneal epithelial cells (Fig. 3, E and F). In the corneal endothelium (Fig. 4), pNBC1 was predominantly localized to the basolateral cell membrane of endothelial cells, as shown in Fig. 4C. Corneal endothelial cells also failed to stain with the anti-kNBC1 antibody (Fig. 4, E and F).

View larger version (52K): [in this window] [in a new window]   Fig. 3.   A-F: immunolocalization of pNBC1 in rat corneal epithelium and stroma. A and B: immunoperoxidase labeling of paraffin-embedded rat cornea. A: peripheral cornea showing the plasma membrane of epithelial cells in the basal layer labeled with the anti-pNBC1 antibody (arrows). Bar, 46 µm. B: the central cornea showing cells in the basal layer staining for pNBC1. Note that epithelial cells in this region of the cornea were less frequently labeled (arrows). Bar, 45 µm. In addition to epithelial cells, stromal keratocytes were also labeled (arrowheads, A and B). C: Nomarski image of cornea showing the corneal epithelium and stroma. Bar, 35 µm. D: immunofluorescence staining of cryostat cornea sections showing epithelial cells in the basal, intermediate, and surface layer labeled with the anti-pNBC1 antibody (arrows). Labeling of stromal keratocytes with the anti-pNBC1 antibody is also shown (arrowheads). Bar, 35 µm. E: Nomarski image of cornea depicting the epithelial layer and stroma. Bar, 35 µm. F: specific peptide blocking of pNBC1 staining. Bar, 35 µm.

View larger version (45K): [in this window] [in a new window]   Fig. 4.   Immunolocalization of pNBC1 (A-D) and kNBC1 (E and F) in corneal endothelium. A: Nomarski image of the cornea depicting the stroma and corneal endothelial layer. Bar, 35 micros B: immunofluorescence staining of the corneal endothelial cells labeled with the anti-pNBC1 antibody (arrow). Bar, 35 µm. C: high-power image of corneal endothelial cells. Overlapping Nomarski and immunofluorescence image showing labeling of endothelial cells using the anti-pNBC1 antibody (arrows). Bar, 9 µm. D: specific peptide blocking of pNBC1 staining. Bar, 35 µm. E: Nomarski image of cornea showing the corneal epithelium, stroma, and endothelium. Bar, 35 µm. F: immunofluorescence image showing the lack of any cell staining in the cornea with the anti-kNBC1 antibody. Bars, 35 µm.

In contrast to the cornea, pNBC1 was not detected in the iris (Fig. 5, A and B). The posterior pigmented epithelium, the dilator iridis muscle, the stroma, and the anterior border layer were not stained with anti-pNBC1 antibody. Similarly, kNBC1 was undetectable in the iris (Fig. 5, C and D). In addition, components of the outflow pathway for aqueous humor such as the trabecular meshwork and the canal of Schlemn were not stained with either antibody.

View larger version (46K): [in this window] [in a new window]   Fig. 5.   Nomarski (A and C) and corresponding immunofluorescence images (B and D) showing lack of staining of the iris with either the anti-pNBC1 antibody (B) or the anti-kNBC1 antibody (D). The canal of Schlemn (arrow) depicted in the Nomarski image (E) was not stained with the anti-pNBC1 antibody (F). Bars, 35 µm (A-D); 26 µm (E and F).

As shown in Fig. 6, the lens epithelium was stained uniformly from the anterior pole and extended posteriorly to the region of lens fiber cell differentiation. Apical and basolateral plasma membranes were labeled with the anti-pNBC1 antibody. kNBC1 was not detectable (Fig. 6, E and F).

View larger version (35K): [in this window] [in a new window]   Fig. 6.   The lens epithelium (arrows) was labeled with the anti-pNBC1antibody. Shown here is the anterior epithelium labeled with the anti-pNBC1 antibody, however, the entire epithelium extending to the equatorial zone of differentiation into lens fiber cells was also labeled. A: Nomarski image of the lens showing the capsule, lens fiber cells, and epithelium. Bar, 35 µm. B: immunofluorescence image of the lens stained with anti-pNBC1. The labeling was not polarized to a specific membrane domain. Bar, 35 µm. C: overlapping Nomarski and immunofluorescence image. Bar, 35 µm. D: specific peptide blocking of pNBC1 staining. Bar, 35 µm. E: Nomarski image. Bar, 35 µm. F: immunofluorescence image showing the lack of staining in the lens with the anti-kNBC1 antibody. Bar, 35 µm.

The conjunctival epithelium was labeled by the anti-pNBC1 antibody (Fig. 7) and by the anti-kNBC1 antibody (Fig. 8); however, the staining patterns were different. pNBC1 labeled surface and wing cells; however, kNBC1 was detected predominantly in basal cells. This is the first evidence of kNBC1 protein expression in an extrarenal location. No labeling was observed in the conjunctival stroma.

View larger version (74K): [in this window] [in a new window]   Fig. 7.   Immunolocalization of pNBC1 (A-F) in the conjunctiva. Nomarski image (A) and immunofluorescence image (B) showing surface cells (arrow) and wing cells (arrowhead) stained with the anti-pNBC1 antibody (A and B; bar, 35 µm). C and D: high-power images of the conjunctiva showing surface cells (arrow) and wing cells (arrowhead) stained with the anti-pNBC1 antibody. Nomarski image (C) and immunofluorescence image (D). C and D; bar, 9 µm. Overlapping low-power (E; bar, 35 µm) and high-power (F; bar, 9 µm) images showing the anti-pNBC1 antibody labeling of conjunctival epithelial cells.

View larger version (41K): [in this window] [in a new window]   Fig. 8.   Immunolocalization of kNBC1 (A-F) in the conjunctiva. A: immunofluorescence image showing basal epithelial cells (arrow) stained with the anti-kNBC1 antibody. Bar, 35 µm. B: specific peptide blocking of kNBC1 staining. Bar, 35 µm. C: immunofluorescence image showing basal epithelial cell kNBC1 staining (arrow), and overlapping Nomarski/immunofluorescence image (D) of the conjunctiva. Bar, 35 µm. High-power Nomarski image (E) and immunofluorescence image (F) of the conjunctiva showing kNBC1 basal epithelial cell membrane staining. (E and F; Bars, 5.5 µm).

The pigmented epithelium of the ciliary body processes was stained intensely for pNBC1 (Fig. 9). Labeling was present in the basolateral membrane of ciliary pigmented epithelial cells. In contrast to the strong labeling of pigmented epithelial cells, the anti-pNBC1 antibody did not label the nonpigmented epithelial cell layer. Furthermore, the capillaries, the postcapillary venules, and the ciliary stroma were unlabeled. As shown in Fig. 9, E and F, kNBC1 was not detected in either the pigmented or nonpigmented ciliary epithelium.

View larger version (64K): [in this window] [in a new window]   Fig. 9.   Immunolocalization of pNBC1 (A-F) in the ciliary body processes. A: Nomarski image showing a longitudinal section of a ciliary body process. Bar, 35 µm. B: immunofluorescence image showing staining of the basolateral membrane of pigmented epithelial cells. Bar, 35 µm. C: overlapping image showing unlabeled nonpigmented epithelial cells (arrows) and basolateral membrane labeling of pigmented epithelial cells (arrowheads). Bar, 35 µm. D: specific peptide blocking of pNBC1 staining. Bar, 35 µm. E: Nomarski image of a longitudinal section of a ciliary body process and the corresponding immunofluorescence image (F) showing lack of staining with the anti-kNBC1 antibody. Bar, 35 µm.

Within the retina, anti-pNBC1 labeling was detected in both the retinal pigment epithelium (RPE) and the neurosensory retina. As shown in Fig. 10, A and B, the apical membrane of RPE cells was labeled with the anti-pNBC1 antibody. Furthermore, within the choroid and sclera, occasionally labeled fibroblasts were detected (Fig. 10, C and D). As shown in Fig. 10, E and F, kNBC1 was not detected in either the RPE or the neurosensory retina. Within the neurosensory retina, Müller glial cells were labeled with the anti-pNBC1 antibody. As shown in Fig. 11, A-C, the apical microvilli of Müller cells labeled strongly. Figure 11E shows Müller cell membrane staining in the inner nuclear layer where the Müller cells pass among neurons. In the nerve fiber layer, pNBC1 staining was detected in Müller cell end feet (Fig. 11F). Therefore, pNBC1 staining of Müller cells extended from the apical microvilli in the outer limiting membrane and continued radially throughout the retina. The anti-pNBC1 antibody failed to label rod and cone photoreceptor cells and other neural elements.

View larger version (62K): [in this window] [in a new window]   Fig. 10.   Immunolocalization of pNBC1 in the retina, choroid, and sclera. A: immunofluorescence image of the retinal pigment epithelium labeled with the anti-pNBC1 antibody. Bar, 9 µm. B: overlapping Nomarski/immunofluorescence image of the retinal pigment epithelium and choroid showing localization of pNBC1 to the apical membrane (arrows). Some cells were unlabeled. Occasional fibroblasts in the choroid are also labeled (arrowheads). Bar, 9 µm. C: high-power Nomarski image and corresponding immunofluorescence image (D) showing fibroblasts in the choroid and sclera labeled with the anti-pNBC1 antibody. Bar (C and D), 9 µm. E: Nomarski image of the neurosensory retina and the corresponding immunofluorescence image (F) showing lack of staining with the anti-kNBC1 antibody. Bar (E and F), 35 µm.

View larger version (68K): [in this window] [in a new window]   Fig. 11.   Labeling of Müller cells with the anti-pNBC1 antibody. A: Nomarski image and corresponding immunofluorescence (B) and overlapping image (C) showing pNBC1 staining of the apical microvilli (arrows) and the processes (arrowheads) of Müller cells surrounding the nuclei of photoreceptors. Bar (A-C), 10 µm. D: specific peptide blocking of pNBC1 staining in the region shown in A-C. Bar, 10 µm. E: overlapping Nomarski/immunofluorescence image of Müller cell bodies and their processes in the inner nuclear layer. Bar, 10 µm. F: immunofluorescence image of Müller cell end feet at the vitreal surface of the retina. Bar, 35 µm.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study, we report the immunolocalization of pNBC1 and kNBC1 proteins in the rat eye using antibodies specific for each protein (summarized in Fig. 12). The results demonstrate that pNBC1 is widely distributed in the eye in various cell types including cornea, conjunctiva, lens epithelium, ciliary body, and retina. In contrast, kNBC1 was only expressed in the basal epithelial cells of the conjunctiva. The widespread localization of pNBC1 was documented in immunoblots of cornea, conjunctiva, ciliary body, and retina, which confirmed the presence of a ~145-kDa band corresponding to pNBC1 in these tissues. The results suggest that the electrogenic sodium-bicarbonate cotransporter pNBC1 plays an important physiological role in several cell types in the eye. In contrast, a ~130-kDa band corresponding to kNBC1 was only detected in conjunctiva. kNBC1 is localized in the basolateral membrane of the renal proximal tubule, where it is known to play an important role in mediating basolateral sodium-bicarbonate efflux. The finding that kNBC1 protein is also expressed in the basal epithelial cells of the conjunctiva is the first suggestion that this NBC1 variant plays a physiological role in extrarenal cells.

View larger version (29K): [in this window] [in a new window]   Fig. 12.   Localization of pNBC1 and kNBC1 proteins in the eye.

In pancreatic ductal cells, pNBC1 is expressed on the basolateral membrane, where it mediates electrogenic sodium-bicarbonate influx with a 2 HCO:1 Na⁺ stoichiometry (18). In the proximal tubule, kNBC1 has a stoichiometry of 3 HCO:1 Na⁺ and mediates cellular bicarbonate efflux because the reversal potential of the transporter is less negative than the resting membrane potential (69). We have recently demonstrated that both pNBC1 and kNBC1 can function in either a 2 HCO:1 Na⁺ or 3 HCO:1 Na⁺ mode depending on the cell type in which each transporter is expressed, indicating that an unknown cellular factor(s) can alter the stoichiometry of both transporters (19). These findings suggest the intriguing possibility that the various cell types in the eye expressing pNBC1 may exhibit differences in HCO:Na⁺ stoichiometry. In keeping with this possibility, RPE and retinal Müller cells each express pNBC1, and yet the HCO:Na⁺ stoichiometry appears to be 2:1 in the RPE (26) and 3:1 in retinal Müller cells (43). Therefore kNBC1 may not necessarily exhibit a 3 HCO:1 Na⁺ stoichiometry in conjunctival basal epithelial cells. Whether a difference in the function or regulation of pNBC1 and kNBC1 necessitates the unique expression pattern of each transporter in conjunctival cells requires further study.

Previous studies have demonstrated that the corneal endothelium transports bicarbonate ions from the stroma to the surrounding aqueous humor (22, 25). Sodium and bicarbonate play an important role in maintaining corneal hydration (34, 49, 50) and in endothelial cell intracellular pH regulation (56, 59). Although electrogenic sodium-bicarbonate cotransport has been implicated in this process, these studies have been based on cultured cells, which may express proteins that are not necessarily present in corneal endothelial cells in vivo. In contrast to experiments utilizing cultured endothelial cells, studies of vesicles derived from native corneal endothelial cells have failed to demonstrate electrogenic sodium-bicarbonate cotransport (36-38). Indeed, it has been postulated that a novel chloride-dependent electrogenic sodium- and bicarbonate-dependent transporter may mediate corneal endothelial bicarbonate efflux (38). Furthermore, these authors have recently reported their failure to detect pNBC1 and kNBC1 transcripts in freshly isolated corneal endothelial cells (38). The latter results are in contrast to a recent report showing pNBC1 and kNBC1 transcripts in cultured human corneal endothelial cells (59), and a separate study, which demonstrated that NBC1 protein is expressed in freshly dissected and cultured bovine corneal endothelial cells (56). However, the antibody used in the latter study was unable to distinguish which NBC1 variant, pNBC1 or kNBC1, is present. With the use of antibodies that are specific for each NBC1 variant, the results in the present study definitively demonstrate that pNBC1 protein is localized to the basolateral membrane of corneal endothelial cells in vivo, where it likely plays an important role in mediating endothelial cell basolateral sodium-bicarbonate cotransport. Furthermore, kNBC1 protein is not expressed in any cells in the cornea.

Of interest, in addition to endothelial cells, corneal epithelial cells also expressed pNBC1, with cells in the basal and intermediate layers expressing this protein most frequently. Unlike corneal endothelial cells, which expressed pNBC1 in a polarized manner, the entire plasma membrane of corneal epithelial cells was stained. There is presently limited information on the transport properties of corneal epithelial cells, despite their importance in functioning as a barrier against the environment and in forming, along with the tear film, a refractive surface that is essential for visual acuity. Recent studies have identified a nonselective cation current, an outwardly rectifying K⁺ current, a voltage-gated Na⁺ current, and inwardly rectifying K⁺ currents in cultured human corneal epithelial cells (9). However, there is presently no information on the acid-base transport properties of corneal epithelial cells, nor information regarding transport mechanisms responsible for pH regulation. Interestingly, the PCO₂ in corneal epithelial cells varies from ~5 Torr in the open eye to 55 Torr in the closed eye (16). This variation in the carbonic acid load requires potent acid-base transporters to maintain intracellular pH in the physiological range. The finding that pNBC1 is expressed in corneal epithelial cells suggests that it may play an important role in regulating intracellular pH in this cell type, which potentially contributes to the maintenance of corneal transparency.

The lens possesses an epithelial cell layer that underlies the anterior capsule. Ion transport by these cells has been suggested to play an important role in controlling the hydration and transparency of the lens (7, 8, 13, 41). Measurement of unidirectional fluxes indicates that 35% of the translenticular short-circuit current is mediated by active sodium transport (13). The toad lens epithelium has been shown to have an electrogenic sodium-bicarbonate cotransporter that appears to function with a 2:1 stoichiometry (62). Although there are no studies that have characterized this transporter in the mammalian lens, our results suggest that pNBC1 mediates electrogenic sodium-bicarbonate cotransport in this part of the eye. Interestingly, a recent study indicates that -crystallin binds lens fiber membranes through hydrophobic interactions that are strikingly pH and ionic strength dependent (14). The amount of -crystallin that binds to the membrane increases under acidic pH conditions (14). Patients with loss-of-function pNBC1 mutations may have cataracts because of an alteration in the pH or ionic strength near lens fiber cells.

The conjunctival epithelium has been shown to have significant ion-transporting capabilities and is likely important, along with conjunctival goblet cells, in the formation and/or modification of the tear film (24, 54, 55, 60). Sodium, chloride, and glucose have been shown to be actively transported across the conjunctival epithelium (23, 24, 54). The conjunctival epithelium mediates active chloride and water secretion subject to modulation by cAMP, Ca²⁺, protein kinase C, and purine and pyrimidine nucleotides (3, 23, 24, 33, 58). The present model of electrolyte transport in conjunctival cells includes an apical membrane sodium-glucose cotransporter and a cAMP-stimulated chloride channel (23, 33). The Na⁺-K⁺-ATPase and a bumetinide-sensitive Na⁺-K⁺-2Cl cotransport process are localized to the basolateral membrane (32). Previous studies have demonstrated that aquaporin-3 is also expressed uniquely in the conjunctival epithelium (21). The relative proportion of tear volume attributable to fluid secreted by the conjunctiva is not known. The results of the present study suggest that H⁺/base transport may play a heretofore unrecognized role in tear film formation/modification. pNBC1 protein was localized to surface and wing cells in the conjunctiva. No labeling was observed in the conjunctival stroma. The finding that kNBC1 was localized uniquely to basal cells in the conjunctiva suggests an important role for this protein in transporting sodium and bicarbonate in this cell type. It has been postulated that conjunctival fluid secretion plays an important role in hydrating mucus secreted by goblet cells (55). Impaired NBC1-mediated sodium transport could secondarily affect passive chloride uptake via the basolateral Na⁺-K⁺-2Cl cotransporter and subsequent apical chloride secretion, thereby altering transepithelial fluid secretion. Whether conjunctival fluid secretion is abnormal in patients with mutations in the NBC1 (SLC4A4) gene requires further investigation.

pNBC1 was immunolocalized to the basolateral membrane of pigmented ciliary epithelial cells. The ciliary body is comprised of two cell layers (pigmented and nonpigmented), whose apical cell membranes are juxtaposed. Secretion of aqueous humor is thought be mediated by the transfer of ions across the basolateral membrane of the pigmented epithelial cell layer, crossing through gap junctions linking the pigmented and nonpigmented cells (42). Water and ions then enter the aqueous humor across the basolateral membrane of the nonpigmented epithelium, although paracellular transport has also been suggested (48). Several transport mechanisms have been implicated in mediating the secretion of aqueous humor, including Na⁺-K⁺-2Cl cotransport, Na⁺/H⁺ exchange, Cl/HCO exchange, vacuolar H⁺-ATPase, and sodium-bicarbonate cotransport (42, 61, 63, 64, 66, 67). However, it is not completely understood how these transport mechanisms are integrated into a cell model of pigmented and nonpigmented epithelial cell transport to account for the formation of aqueous humor. Furthermore, the complexity of the dual-layered ciliary body epithelium imposes serious difficulties in determining the molecular mechanisms responsible for regulation of aqueous humor formation. The finding that pNBC1 protein is localized to the basolateral membrane of pigmented epithelial cells suggests that electrogenic sodium-bicarbonate cotransport plays an essential role in mediating transepithelial transport/intracellular pH (pH_i) regulation in these cells and in the formation of aqueous humor. This result is in keeping with the studies of Wolosin et al. (64), who functionally identified a sodium- and bicarbonate-dependent, apparently electrogenic, transport process in the freshly dissected rabbit ciliary body-pigmented epithelium. In contrast, in nonpigmented ciliary epithelial cells, bicarbonate transport was mediated by a chloride-bicarbonate exchange process (64). An understanding of the transport processes involved in aqueous humor production is clearly clinically important, given that inhibition of aqueous formation with subsequent lowering of intraocular pressure is a common approach for treating open-angle glaucoma (39, 65). Although carbonic anhydrase inhibitors are among the most potent inhibitors of aqueous humor formation, their cellular site of action is unclear. It has been established that carbonic anhydrase inhibitors reduce not only the rate of aqueous humor formation but also the net rate of sodium and bicarbonate transport from blood to aqueous humor (40). Carbonic anhydrase inhibitors can alter bicarbonate absorption in transporting epithelia by decreasing the rate of hydration of CO₂ and subsequent generation of bicarbonate (39, 40, 65). Our results suggest that inhibition of pNBC1-mediated bicarbonate uptake by pigmented ciliary body epithelial cells is a potential mechanism for inhibition of aqueous humor production by these drugs, which requires further investigation.

Previous studies have documented that retinal activation by light results in an increase in extracellular pH in the inner plexiform (synaptic) layer and in the photoreceptor layer (11, 68). This increase in extracellular pH is mediated by increased synaptic activity, in addition to a reduction in the metabolic activity of the photoreceptors. A change in intracellular or extracellular pH can result in marked changes in neuronal function, including gap junction conductance, N-methyl-D-aspartate receptor behavior, and voltage-gated ion channel transport (5, 6, 17, 57). For example, previous studies have shown that an increase in extracellular pH of only 0.05 pH units can increase synaptic transmission between photoreceptors and bipolar cells by ~25% (6). Newman (45) has hypothesized that bicarbonate uptake mediated by an electrogenic sodium-bicarbonate cotransporter in retinal Müller cells may prevent the elevation of extracellular pH in the inner plexiform (synaptic) layer and in the photoreceptor layer. Our finding that pNBC1 protein is localized to retinal Müller cells is compatible with Newman's hypothesis and suggests that pNBC1-mediated Müller cell bicarbonate uptake plays an important role in preventing the extracellular pH from increasing during retinal light stimulation.

The RPE mediates the transport of fluid, ions, and metabolites between the subretinal space and the choroidal blood supply and plays an essential role in maintaining the viability of the neural retina (27). In native RPE preparations, the cell transport model includes an apical Na⁺-K⁺-ATPase, Na⁺/H⁺exchange, electrogenic sodium-bicarbonate cotransport, H⁺-lactate cotransport, and basolateral Cl/HCO exchange (26, 27, 35). There is also recent evidence for a possible basolateral membrane sodium-bicarbonate cotransport process (31). These cells are normally exposed to lactic acid produced by photoreceptors under aerobic conditions (20). pNBC1 likely plays an important role in maintaining RPE intracellular pH constant, despite H⁺ influx mediated by the apical H⁺-lactate cotransporter. pNBC1 may also play a role in mediating apical-to- basolateral bicarbonate transport in the bullfrog RPE (26), although the relative importance of this transport process in mammals appears to be species dependent (15). In addition, pNBC1-mediated alterations in intracellular pH in response to apical potassium concentration changes may be an important signaling event after dark-light transition (31).

The potential cause of blindness in patients with mutations in the NBC1 gene reported by Igarashi et al. (28) is clearly very complex, given the various regions in the eye where NBC1 proteins are localized. The patients reported by Igarashi et al. had mutations in the coding region common to pNBC1 and kNBC1. Whether abnormal pNBC1 and/or kNBC1 function is responsible for the ocular phenotype in these patients requires further study. However, it is likely, given our findings of the widespread distribution of pNBC1 in the eye (cornea, conjunctiva, lens, ciliary body, and retina) and the fact that kNBC1 is only expressed in conjunctiva, that the phenotype reported in these patients is most likely due to abnormal pNBC1-mediated sodium-bicarbonate cotransport. Thus far, patients with mutations in the unique NH₂ terminus of pNBC1 or kNBC1 have not been reported. It would be predicted that those patients with mutations in the unique kNBC1 NH₂ terminus would present with predominantly a renal phenotype (proximal renal tubular acidosis), with minimal ocular abnormalities involving conjuntival tear film formation/modification. Knockout mice with specific NBC1 mutations will help address these issues.