Clathrin-mediated endocytosis and recycling of the neuron-specific Na+/H+ exchanger NHE5 isoform. Regulation by phosphatidylinositol 3'-kinase and the actin cytoskeleton.

Mammalian Na+/H+ exchangers (NHEs) are a family of integral membrane proteins that play central roles in sodium, acid-base, and cell volume homeostasis. The recently cloned NHE5 isoform is expressed predominantly in brain, but its functional and cellular properties are poorly understood. To facilitate its characterization, an epitope-tagged construct of NHE5 was ectopically expressed in nonneuronal and neuronal cells. In NHE-deficient Chinese hamster ovary AP-1 cells, NHE5 localized at the plasmalemma, but a significant fraction accumulated intracellularly in vesicles that concentrated in a juxtanuclear region. Similarly, in nerve growth factor-differentiated neuroendocrine PC12 cells and primary hippocampal neurons, immunolabeling of NHE5 was detected in endomembrane vesicles in the perinuclear region of the cell body but also along the processes. More detailed characterization in AP-1 cells using organelle-specific markers showed that NHE5 co-localized with internalized transferrin, a marker of recycling endosomes. Transient transfection of a dominant negative mutant of dynamin-1, which inhibits clathrin-mediated endocytosis, blocked uptake of transferrin as well as internalization of NHE5. Likewise, wortmannin inhibition of phosphatidylinositol 3'-kinase, a lipid kinase implicated in endosomal traffic, induced coalescence of vesicles containing NHE5 and caused a pronounced inhibition of plasmalemmal Na+/H+ exchange. By contrast, disruption of the F-actin cytoskeleton with cytochalasin D increased cell surface NHE5 activity and abundance. These observations demonstrate that NHE5 is localized to the recycling endosomal pathway and is dynamically regulated by phosphatidylinositol 3'-kinase and by the state of F-actin assembly.

Subtle fluctuations in intra-or extracellular pH of neurons can significantly modulate membrane excitability and are thought to fulfill a regulatory role in central nervous system function (1)(2)(3)(4)(5). On the other hand, extreme disturbances in acid-base homeostasis are associated with the progression of certain neuropathies, such as acute ischemic stroke (6 -8), glial swelling (9), and cerebral edema (10). Hence, fine pH control of the neural milieu is a vital biological process. Maintenance of acid-base homeostasis in the central nervous system is complex and involves the coordinated activities of several distinct plasma membrane ion transporters, including Na ϩ -HCO 3 Ϫ cotransporters, Na ϩ -dependent and -independent Cl Ϫ /HCO 3 Ϫ exchangers, and Na ϩ /H ϩ exchangers (for an overview, see Ref. 5). This diversity is further enriched by the presence of multiple isoforms for some of these transporters (11)(12)(13)(14). However, their specific regulatory properties and contributions to neural physiology are not well understood.
In mammals, seven distinct Na ϩ /H ϩ exchangers (NHEs) 1 have been described to date; five (NHE1 to -5) are resident primarily in the plasma membrane, whereas two (NHE6 and -7) localize predominantly to intracellular organelles, and all are present in brain (13)(14)(15)(16)(17). NHE1 is ubiquitously expressed and is noted for its high sensitivity to inhibitory drugs such as amiloride. It is chiefly responsible for restoration of steadystate pH i following cytosolic acidification and for maintenance of cell volume. Recent findings indicate that it is also crucial for neural function and viability. Mice with null mutations of Nhe1 exhibit locomotor abnormalities, epileptic-like seizures, and considerable mortality (67%) prior to weaning (18,19). Moreover, hippocampal CA1 neurons isolated from these animals display enhanced membrane excitability and increased Na ϩ current density, implicating a functional association between NHE1 and voltage-sensitive Na ϩ channels that seemingly contributes to neural dysfunction when disrupted (20). However, the precise molecular nature of this relationship is unknown.
By comparison, NHE2 to -4 are predominantly expressed in epithelia of the kidney and gastrointestinal tract (15) but are also detected at low levels in discrete regions of the brain (13). The best studied of these is the amiloride-resistant isoform, NHE3. In epithelia, NHE3 resides along the apical membrane as well as in recycling endosomes and fulfills an important role in luminal Na ϩ and HCO 3 Ϫ (re)absorption; the latter effected by H ϩ extrusion (21)(22)(23)(24). Its activity is also subject to acute regulation by several protein and lipid kinases, including protein kinase A (25)(26)(27)(28)(29), protein kinase C (25,30,31), Rho-associated kinase (32), and phosphatidylinositol 3Ј-kinase (PI3-K) (33,34). These kinases act on NHE3 by direct phosphorylation or through intermediary effectors and alter one or more of its biological parameters including proton affinity, cell surface abundance, and/or association with the actin cytoskeleton. In brain, NHE3 is present primarily in cerebellar Purkinje and glial cells (13) but is also detected in chemosensitive neurons of the ventrolateral medulla oblongata, where it has been proposed to play a role in the regulation of breathing rhythm (35)(36)(37).
NHE5 is distinguished from the other isoforms by its almost exclusive expression in the central nervous system. This isoform is found in multiple regions of the brain including the dentate gyrus, cerebral cortex, hippocampus, amygdala, caudate nucleus, hypothalamus, subthalamic nucleus, and thalamus but not in glia-enriched structures such as the corpus callosum, suggesting that NHE5 might be neuron-specific (38,39). Structurally, it is most closely related to NHE3 (ϳ50% amino acid identity) and shares similar pharmacological (39,40) and regulatory (41) properties. These observations have led to the suggestion that NHE5 may be the amiloride-resistant Na ϩ /H ϩ exchanger reported in hippocampal neurons (42). Aside from these few reports, however, little else is known about the functional properties of NHE5. Information regarding its subcellular distribution and regulation are hampered by the absence of isoform-specific antibodies and by the difficulties of isolating, maintaining, and transfecting neurons in primary culture. Additionally, the presence of multiple NHE isoforms in neurons further complicates interpretation of functional measurements.
To gain further insight into the sorting and regulation of NHE5, we have stably expressed an epitope-tagged version of human NHE5 cDNA in a subline (AP-1) of Chinese hamster ovary (CHO) cells that lacks endogenous plasmalemmal NHE activity, thereby facilitating transport measurements. The usefulness of this cell line for studying membrane proteins normally resident in polarized cells is bolstered by studies showing that undifferentiated cells, such as CHO cells, contain both regulated and constitutive secretory as well as endocytic vesicles analogous to those found in more specialized cell types such as epithelia and neurons (43)(44)(45)(46)(47)(48)(49). In this report, we show that NHE5, like NHE3, is present both in the plasma membrane and in clathrin-associated recycling endosomes that are regulated by PI3-K when expressed in AP-1 cells. Pharmacological disruption of the actin cytoskeleton leads to increased rates of transport and cell surface-exposed NHE5, possibly by blocking internalization. Interestingly, this latter finding is in marked contrast to that observed for NHE3, which is inhibited under comparable conditions (50). Although characterized less extensively, a comparable subcellular distribution of NHE5 was also observed in transiently transfected neuroendocrine PC12 cells and hippocampal neurons in primary culture.
Bicarbonate-free medium RPMI 1640 was buffered with 25 mM HEPES to pH 7.4 at 37°C. Phosphate-buffered saline consisted of 140 mM NaCl, 10 mM KCl, 8 mM sodium phosphate, 2 mM potassium phosphate, pH 7.4. The isotonic Na ϩ -rich medium used in the fluorimetric pH measurements contained 140 mM NaCl, 3 mM KCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 10 mM glucose, 20 mM HEPES, pH 7.4, at 37°C. Isotonic K ϩ -rich medium had the same composition as Na ϩ -rich medium, except that NaCl was replaced by KCl. All solutions were nominally bicarbonate-free and adjusted to 290 Ϯ 10 mosM with the major salt.
DNA Constructs-To allow for immunological detection of human NHE5, a triple influenza virus HA epitope YPYDVPDYA(G/A) was engineered into the first predicted exomembranous loop of NHE5. Briefly, a unique NotI restriction endonuclease site was inserted into the coding region of this loop, which altered amino acid residues at positions 36 and 37 from Leu-Phe to Arg-Gly ( 35 ELFR to 35 ERGR). A NotI-NotI DNA fragment encoding a triple HA epitope (residues encoded by NotI sites are underlined; rgrifYPYDVPDYAgYPYDVPDYAgsY-PYDVPDYAaqcgr) was then inserted into the NotI site (tag inserted between amino acids Glu 35 and Arg 38 ), and the modified cDNA-protein was called NHE5 36HA3 . All constructs were subcloned into the mammalian expression vector pCMV under the control of the enhancer/ promoter region of the immediate early gene of human cytomegalovirus. In control experiments, these modifications had no obvious effect on the functional properties of NHE5 when expressed in AP-1 cells. The cDNA was sequenced to confirm the presence of the mutations and to ensure that other random mutations were not introduced.
Mammalian expression vectors containing cDNA constructs encoding the mitochondrion-targeted cytochrome oxidase subunit XIII (COX8) presequence fused to the amino terminus of green fluorescent protein (mito-GFP) and the endoplasmic reticulum retention sequence (SEKDEL) fused to the carboxyl terminus of a bovine preprolactingreen fluorescent protein chimera (KDEL-GFP) were kindly provided by Dr. A. S. Verkman (University of California, San Francisco) (51). The lysosome-targeted Rab7-GFP cDNA was kindly provided by Dr. P. Stahl (Washington University, St. Louis, MO). The cDNA encoding the enhanced green fluorescent protein was obtained from BD Biosciences CLONTECH. A pcDNA3 vector containing a dominant-negative mutant of dynamin (DynK44A), tagged with an HA epitope at its amino terminus, was a generous gift of Dr. S. Schmid (Scripps Institute, La Jolla, CA).
Cell Transfection, Selection, and Culture-AP-1 cells are CHO cells devoid of endogenous plasmalemmal NHE activity (52). They were generated by chemical mutagenesis followed by selection using the H ϩ suicide technique (53). The pNHE5 36HA3 plasmid was stably transfected into AP-1 cells using the calcium phosphate-DNA co-precipitation method (54) and acid-selection (55,56). Several clonal cell lines were screened for NHE5 36HA3 expression by measuring amiloride-inhibitable H ϩ -activated 22 Na ϩ influx. The cells were maintained in complete ␣-minimal essential medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 25 mM NaHCO 3 , pH 7.4, and incubated in an humidified atmosphere of 95% air, 5% CO 2 at 37°C. For transient transfections, DNA was introduced into cells plated on coverslips by transfection using FuGENE6 as recommended by the manufacturer using 1 g of the mito-GFP, KDEL-GFP, or Rab7-GFP constructs. For cotransfection of DynK44A and enhanced green fluorescent protein, a 5:1 ratio was used (1.5 g of DynK44A plus 0.3 g of enhanced green fluorescent protein cDNA/coverslip). Cells were analyzed 24 -48 h following transfection.
The pheochromocytoma PC12 cell line was cultured in 35-mm dishes on laminin/poly-D-lysine-coated glass coverslips in RPMI medium containing 5% fetal bovine serum. For transient expression of NHE5 36HA3 , DNA was introduced into subconfluent, undifferentiated cells by transfection using LipofectAMINE™ 2000 as recommended by the manufacturer (Invitrogen). Following transfection, cells were induced to differentiate by inclusion of 50 ng/ml nerve growth factor (NGF) in the culture medium for the indicated times.
Primary cultures of rat hippocampal neurons were prepared as previously described (57). Briefly, hippocampi from 3-4-day-old rats were dissociated using papain (Worthington) and triturated using a glass pipette (58). After centrifugation at 400 ϫ g for 2 min, the cells were plated at a density of 2 ϫ 10 5 cells/35-mm dish containing laminin/poly-D-lysine-coated glass coverslips in Neurobasal™ medium supplemented with B-27, penicillin, streptomycin, and 0.5 mM L-glutamine (Invitrogen). The cells were kept at 37°C in 5% CO 2 . Half the medium was replaced every 3 days. L-Glutamic acid (25 M) was included for the first 4 days and cytosine arabinoside (5 M) from days 2-5 to arrest nonneuronal cell proliferation. On day 4 after plating, cells were transfected using the LipofectAMINE™ 2000 reagent according to the manufacturer's instructions for primary neuron cultures. Following transfection, the cells were incubated for 24 h in culture medium supplemented with 20 mM KCl. Transient incubation in high K ϩ has been reported to enhance expression from vectors using the human cytomegalovirus promoter (57).
Immunoblotting-NHE5 36HA3 stable transfectants were grown to confluence in 10-cm dishes and lysed with 1% Triton X-100. Total cellular protein extracts were resolved by 6% SDS-polyacrylamide gel electrophoresis and transferred onto polyvinylidene difluoride membranes (Amersham Biosciences). The blots were briefly rinsed with PBS, blocked with 5% nonfat milk in phosphate-buffered saline with 0.1% Tween 20, and then incubated with a mouse monoclonal anti-HA antibody (dilution 1:5000). After extensive washes with phosphatebuffered saline containing 0.1% Tween 20, the blots were incubated with goat anti-mouse IgG secondary antibody conjugated with horseradish peroxidase (dilution 1:50,000). Immunoreactive bands were visualized by enhanced chemiluminescence (Amersham Biosciences) recorded on x-ray film.
Immunolocalization of HA-tagged NHE5-For immunolabeling experiments, NHE5 36HA3 transfectants were plated and grown on glass coverslips to ϳ70% confluence and then incubated in serum-free RPMI medium for 3 h prior to the experiment. The cells were then untreated or treated as indicated prior to immunostaining.
Various conditions were used to stain different subcellular pools of NHE5 36HA3 . To label plasmalemmal exchangers, cells were chilled and incubated at 4°C with anti-HA antibody (1:1000) in RPMI for 1 h in the absence or presence of the indicated drug. The cells were then washed three times with PBS and fixed with 4% paraformaldehyde at room temperature for 20 min. Paraformaldehyde was quenched for 10 min with 100 mM glycine in PBS, followed by blocking with 5% nonfat milk in PBS for 1 h. The coverslips were then incubated with Cy3-conjugated anti-mouse second antibody for 1 h in PBS, washed extensively with PBS, and finally mounted on slides using DAKO mounting medium (DAKO Corp., Carpinteria, CA).
To visualize the total cellular distribution of NHE5 36HA3 , cells were fixed for 15 min using 4% paraformaldehyde in PBS, incubated with 100 mM glycine, and then permeabilized using 0.1% Triton X-100 in PBS for 30 min, all at room temperature. To reduce nonspecific binding, the cells were blocked with 5% nonfat milk in PBS for 1 h and then incubated with a mouse monoclonal anti-HA antibody for 1 h. The coverslips were next washed four or five times with PBS and incubated with Cy3-conjugated anti-mouse IgG antibody for 1 h. After incubation with the secondary antibody, the cells were washed 3-5 times over a 15-min period with PBS and then mounted onto glass slides with DAKO mounting medium. Where specified, F-actin was labeled by incubating fixed and permeabilized cells with rhodamine-labeled phalloidin (1:500) for 1 h at room temperature.
To assess internalization of the antiporter, cells were incubated with anti-HA antibody (1:1000) at 37°C for 1 h in the presence or absence of the indicated drugs. Unbound antibody was removed by washing with cold PBS. The cells were then fixed, permeabilized, blocked, and labeled with Cy3-conjugated secondary antibody as above.
Cells were visualized using the 100ϫ objective of a Leica DM1RB fluorescence microscope (Heidelberg, Germany) equipped with a Micromax cooled CCD camera (Princeton Instruments, Trenton, NJ), operated from a Dell computer using Winview® software (Princeton Instruments). Where indicated, cells were also visualized using a Zeiss LSM 510 confocal microscope. Serial optical slices (0.5 m thick) were acquired using the LSM 510 software. Digitized images were processed using Adobe Photoshop (Adobe Systems, Inc.). All images are representative of at least three separate experiments.
Quantitation of Surface NHE5-A radioisotopic method was used to quantify surface expression of external epitope-tagged NHE5. AP-1 cells stably expressing NHE5 36HA3 were plated onto 12-well plates and grown to ϳ70% confluence. Cells were incubated with anti-HA antibody (1:1000 dilution) for 1 h at 4°C to prevent endocytosis in HEPESbuffered RPMI medium supplemented with 10% goat serum. After washing the cells six times with cold medium to remove excess unbound antibody, the cells were incubated with 125 I-labeled goat anti-mouse IgG (0.4 Ci/sample) in HEPES-buffered RPMI with 10% goat serum for 45 min at 4°C. At the end of the incubation, they were washed three times with HEPES-buffered RPMI to remove unbound radiolabel and detached from the plates using PBS containing 1% Triton X-100. Radioactivity was counted with a 1282 Compugamma LKB counter. The radioactivity bound to cells exposed only to 125 I-IgG without prior incubation with anti-HA antibody was also determined and subtracted from all determinations. Data are expressed as percentage of surface labeling of untreated control cells (100%) and are the means Ϯ S.E. of the number of experiments indicated, each performed in duplicate or triplicate.
Immunostaining of Various Intracellular Organelles for Colocalization with NHE5-To label the plasma membrane, cells were washed with PBS and then incubated with a mixture of FITC-labeled lectins (peanut, wheat germ, and pea; 2 mg/ml each) for 45 min at 4°C, followed by fixing with 4% paraformaldehyde. Recycling endosomes were labeled by cellular uptake of r-Tfn. Briefly, serum-deprived cells were incubated with 20 mg/ml r-Tfn for 45-60 min at 37°C. Unbound r-Tfn was washed extensively with PBS, and the cells were fixed for 45 min with 4% paraformaldehyde and mounted as described above. Total or recycling NHE5 was co-stained as described above.
The trans-Golgi network was visualized using an anti-mannosidase II antibody (59). Cells were fixed, permeabilized, and blocked as above. This was followed by co-incubation with a rabbit polyclonal antimannosidase II (1:500) and anti-HA antibody. After washes, cells were incubated with Cy3-conjugated anti-mouse and FITC-labeled antirabbit antibodies to visualize HA and mannosidase antibodies, respectively.
Measurement of Na ϩ /H ϩ Exchanger Activity-NHE activity was assessed by both (i) fluorimetric (rate of Na ϩ -mediated recovery of cytoplasmic pH (pH i ) from an acid load imposed by prepulsing with NH 4 Cl) and The pH i of small groups of cells was determined by microphotometry of the fluorescence emission of the pH-sensitive dye BCECF using dual wavelength excitation. Briefly, cells were grown to ϳ70% confluence on 25-mm glass coverslips and then serum-depleted by incubation in HEPES-buffered RPMI medium for 3 h at 37°C, followed by treatment in the same medium with or without 100 nM wortmannin or 10 M cytochalasin D, as specified. In the last 10 min of treatment, 2 g/ml BCECF acetoxymethyl ester and 50 mM NH 4 Cl were added to the medium. After the incubation, cells were washed with isotonic Na ϩ -free medium three times to remove extracellular NH 4 Cl and dye. The coverslips were then mounted in a Leiden coverslip holder (Medical System Corp., Greenvale, NY) and placed into a thermostatted holding chamber heated to 37°C (Medical Systems Corp., Greenvale, NY) attached to the stage of a Nikon Diaphot TMD inverted microscope (Nikon Canada, Toronto, Ontario, Canada). Cells were visualized using a Nikon Fluor 3 40/1.3 numerical aperture oil immersion objective and a Hoffman modulation contrast video system with an angled condenser (Modulation Optics) through a CCD-72 video camera and control unit (Dage-MTI, Michigan City, IN) connected to a Panasonic monitor. Fluorescence of small groups of cells (6 -12 cells) was followed using an M Series dual wavelength illumination system from Photon Technologies, Inc. (South Brunswick, NJ) in a dual excitation/single emission configuration, with excitation wavelengths of 440 and 490 nm and emission wavelength of 510 nm. Na ϩ /H ϩ exchange was initiated by reintroduction of extracellular Na ϩ , and activity was estimated from the rate of pH i recovery. The fluorescence was analyzed with the Felix software (Photon Technologies Inc., South Brunswick, NJ). Calibration of the fluorescence intensity in terms of pH i was performed in the presence of 5 M nigericin in high potassium medium (140 mM KCl, 20 mM HEPES, 1 mM MgCl 2 , 1 mM CaCl 2 , and 5 mM glucose) as detailed previously (60). Buffering power in the intracellular pH range used was determined by measuring pH changes caused by small pulses of NH 4 Cl, as described previously (61).
For radioisotopic measurement of Na ϩ /H ϩ exchanger activity, the cells were grown to confluence in 24-well plates and then serumdeprived overnight, followed by treatment with vehicle or drug as specified. Prior to 22 Na ϩ influx, the cells were either at rest or acidified using the NH 4 Cl prepulse technique, as indicated (40). The assays were initiated by incubating the cell monolayers in isotonic choline chloride solution (125 mM choline chloride, 1 mM MgCl 2 , 2 mM CaCl 2 , 5 mM glucose, 20 mM HEPES-Tris, pH 7.4) containing 1 mM ouabain and carrier-free 22 Na ϩ (1 Ci/ml) in the absence or presence of the NHE inhibitor amiloride (1 mM). The lack of K ϩ and the presence of ouabain minimized transport of Na ϩ catalyzed by the Na ϩ -K ϩ -2Cl Ϫ cotransporter and the Na ϩ ,K ϩ -ATPase. The influx of 22 Na ϩ was terminated by rapidly washing the cells three times with four volumes of ice-cold NaCl stop solution (130 mM NaCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 20 mM HEPES-NaOH, pH 7.4). To extract the radiolabel, the monolayers were solubilized with 0.25 ml of 0.5 N NaOH, and the wells were washed with 0.25 ml of 0.5 N HCl. Both the NaOH cell extract and the HCl wash solution were combined in 5 ml of scintillation fluid and transferred to scintillation vials. The radioactivity was assayed by liquid scintillation spectroscopy. Protein content was determined using the Bio-Rad DC protein assay kit according to the manufacturer's protocol. Under these conditions, the uptake of 22 Na ϩ at nominal Na ϩ concentrations was linear over a 10-min period at room temperature. Therefore, a time course of 5 min was chosen for the experiments. All experiments represent the average of three or four experiments, each performed in quadruplicate. The data are presented as the mean Ϯ S.E.

Subcellular Localization of NHE5-Previous
functional studies of wild-type human NHE5 ectopically expressed in CHO AP-1 cells indicated that it was resident in the plasma membrane, based on its ability to mediate the influx of extracellular 22 Na ϩ in a H ϩ i -dependent manner (40). Because of its close relatedness to NHE3, which is distributed to the plasmalemma but also to endomembrane compartments, it was of interest to assess whether NHE5 also accumulates at other locations. Such information may give insight into potential mechanisms of regulation of NHE5 as well as its specific physiological roles. To facilitate detection and discrimination of transporters located at the cell surface from other intracellular compartments, a triple immunogenic epitope derived from the influenza virus HA protein was inserted into the predicted first extracellular loop of NHE5, yielding NHE5 36HA3 . This site was chosen because prior analyses of its closest homologue, NHE3, showed that this region could accommodate peptide insertions without compromising transport function or subcellular distribution (34). Schematic representation of the construct is shown in Fig. 1A. Like untagged NHE5 (40), transfection of this modified construct in NHE-deficient AP-1 cells restored plasmalemmal Na ϩ /H ϩ exchange activity. Importantly, measurement of the kinetic and pharmacological properties of Na ϩ /H ϩ ex-change in stable isolates indicated that they were indistinguishable from the untagged version (data not shown). The abundance and intactness of the protein were assessed by immunoblotting with an anti-HA antibody. As shown in Fig.  1B, a single immunoreactive band that migrated close to the predicted molecular mass of NHE5 36HA3 (ϳ102 kDa) was readily detected in whole cell extracts, whereas untransfected cells failed to react with the antibody, implying that the immunoreactivity is specific for the HA epitope.
To determine the site(s) where NHE5 36HA3 accumulates in AP-1 cells, dual labeling studies were performed using recognized organelle-specific markers and fluorescence microscopy. The cell surface expression and exofacial orientation of the epitope tag of NHE5 36HA3 were verified by exposing intact cells to anti-HA antibodies. After a 1-h incubation at 4°C, the cells were washed to remove excess unbound antibody, fixed, and exposed to a secondary Cy3-conjugated anti-mouse antibody. As expected, examination of the cells using confocal microscopy showed that NHE5 36HA3 was present in the plasmalemma, as verified by colocalization with lectins used as surface markers (Fig. 2, A and B), whereas no signal was detected in untransfected cells (data not shown). By comparison, immunolabeling of fixed and permeabilized cells to visualize total cellular NHE5 36HA3 using a conventional fluorescence microscope revealed the existence of a large intracellular pool (Fig. 2C). The signal was partly dispersed throughout the cell in a reticular pattern and partly in a punctate compartment with a noticeable accumulation in a juxtanuclear location (indicated by arrowheads). The former staining showed marked colocalization with a KDEL-GFP construct (Fig. 2D) that accumulates in the endoplasmic reticulum of transiently transfected cells. The presence of NHE5 36HA3 in this compartment is probably a consequence of heterologous overexpression, which is often associated with accumulation and trapping in the endoplasmic reticulum. On the other hand, the vesicular juxtanuclear accumulation of intracellular NHE5 36HA3 was clearly distinct from the endoplasmic reticulum. The identity of the juxtanuclear organelle expressing NHE5 was examined, and representative data are presented in Figs. 3 and 4. To visualize mitochondria, an expression vector containing enhanced GFP linked at its amino terminus to a mitochondrion-targeting sequence (mito-GFP) was cotransfected into NHE5 36HA3 -expressing cells. In these cells, GFP-stained mitochondria appeared as punctate elongated structures that did not significantly overlap with NHE5 36HA3 (see arrowheads; Fig. 3, compare A and B), indicating that NHE5 is not a mitochondrial isoform. Immunolabeling of the medial and trans-Golgi cisternae using an antibody to ␣-mannosidase II (59) revealed that although the site of accumulation of NHE5 36HA3 was near the Golgi (Fig. 3, compare C and D), the location and morphology of the cisternae were clearly distinct from that of NHE5.
The location of recycling endosomes was probed using r-Tfn. Transferrin binds to its surface receptor, becomes internalized (passing sequentially through early endosomes and perinuclear recycling endosomes), and then is returned to the cell surface (62). As shown in Fig. 3F, internalized r-Tfn showed a punctate distribution with accumulation in a compact pericentriolar complex. This intracellular distribution closely overlapped the vesicular juxtanuclear compartment of intracellular NHE5 when the total NHE5 pool is visualized (Fig. 3E). This suggests that NHE5 might be retrieved from the surface into recycling vesicles. To examine this possibility, intact NHE5 36HA3 cells were co-incubated with r-Tfn and the anti-HA antibody at 37°C for 1 h. Under these conditions, the antibody bound to surface exchangers is internalized and can be detected inside the cell. As shown in Fig. 4, A and B, the signals for NHE5 36HA3 and r-Tfn overlapped extensively, indicating that they are internalized into similar, if not identical, endomembrane vesicles. Importantly, the distribution of antibody-bound NHE5 36HA3 in recycling endosomes resembles that observed in untreated cells that were fixed and permeabilized prior to immunostaining (cf. Fig. 3, A, C, and E). Thus, binding of monoclonal antibody to the extracellular epitope did not appreciably alter the subcellular distribution of NHE5 36HA3 , implying normal trafficking within the cell. Last, the site of NHE5 36HA3 accumulation was also readily discernible from lysosomes labeled with transiently expressed Rab7-GFP, which in AP-1 cells appear as neighboring, but discrete, vesicles throughout the cell (cf. Fig. 4

, C and D). NHE5 Is Internalized by Clathrin-mediated Endocytosis-
The preceding experiments provided evidence that NHE5 is internalized from the cell surface in a manner analogous to that reported for the ligand-bound Tfn receptor and the NHE3 isoform. Endocytosis of these latter membrane proteins occurs through the clathrin-mediated pathway (22,63). Hence, it is conceivable that a parallel mechanism also underlies the internalization of NHE5. In order to evaluate the potential involvement of clathrin, we used a dominant-negative mutant of the neuronal isoform of dynamin, dynamin-1/K44A (DynK44A), that has been shown to block the fission of clathrin-coated pits from the plasma membrane (64). The inhibitory effectiveness of DynK44A was verified by confirming its ability to block endocytosis of ligand-bound Tfn receptors. To identify successfully transfected cells, DynK44A was cotransfected with enhanced GFP in a 5:1 ratio, which results in over 90% co-expression of the two proteins (data not shown). Representative results are presented in Fig. 5. Rhodamine-Tfn readily accumulated in typical perinuclear recycling endosomes in NHE5 36HA3 cells that were transiently transfected with GFP (indicated by the arrowheads in Fig. 5, A and B), a pattern that was indistinguishable from neighboring untransfected cells. In contrast, little r-Tfn was internalized by cells co-expressing DynK44A (indicated by an asterisk in Fig. 5, C and D). Note that sur-

FIG. 3. Accumulation of NHE5 in an intracellular compartment.
A and B, AP-1/NHE5 36HA3 cells were transiently transfected with a mitochondrion-targeted GFP construct (mito-GFP). Forty-eight h after transfection, the cells were fixed, permeabilized, and immunostained to visualize NHE5 36HA3 with anti-HA primary antibody and Cy3-conjugated secondary antibody. Total NHE5 36HA3 staining (A) and mito-GFP (B) of the same field are shown. The arrowheads indicate the juxtanuclear location of NHE5 36HA3 . C and D, cells were fixed, permeabilized, and co-stained to visualize NHE5 36HA3 as described above (C) and to label the Golgi cisternae with a polyclonal anti-mannosidase II primary antibody and FITC-labeled anti-rabbit secondary antibody (D). E and F, cells were preincubated in serum-free, HEPES-buffered RPMI medium for 3 h. r-Tfn was added to the medium for 1 h at 37°C. The cells were then washed, fixed, and permeabilized. After blocking for 1 h with 5% milk in PBS, total NHE5 36HA3 was labeled using an anti-HA primary antibody followed by incubation with an FITC-coupled antimouse secondary antibody. NHE5 36HA3 staining (E) and fluorescence of r-Tfn (F) of the same field are shown. Images were taken using a Leica fluorescence microscope.

FIG. 4. Plasmalemmal NHE5 is internalized and sorted to recycling vesicles.
A and B, AP-1/NHE5 36HA3 cells were cultured in serum-free RPMI medium for 3 h, followed by incubation with anti-HA antibody (1:1000) and r-Tfn for 1 h at 37°C. After extensive washes, cells were fixed, permeabilized, and blocked with 5% milk in PBS. NHE5 36HA3 staining was visualized using an FITC-coupled anti-mouse secondary antibody. The images in A and B represent a section through the middle of the same cells using a Zeiss confocal microscope and show recycling NHE5 36HA3 and r-Tfn, respectively. C and D, cells were transiently transfected with a mammalian expression vector containing Rab7-GFP, a lysosomal marker. Twenty-four h after transfection, cells were maintained in serum-free, HEPES-buffered RPMI medium for 3 h, followed by incubation with anti-HA antibody (1:1000) for 1 h at 37°C. After extensive washes, cells were fixed, permeabilized, and blocked with 5% milk in PBS. NHE5 36HA3 staining was visualized using a Cy3-coupled anti-mouse secondary antibody. The images in C and D represent a section through the middle of the same cells using a Zeiss confocal microscope and show recycling NHE5 36HA3 and Rab7-GFP, respectively. rounding cells lacking DynK44A plus GFP readily internalized r-Tfn. Thus, DynK44A effectively prevented clathrin-mediated endocytosis.
The effect of DynK44A on the recycling of surface NHE5 36HA3 was investigated next. NHE5 36HA3 internalization was not affected by expression of GFP alone (indicated by the arrowheads in Fig. 5, E and F). Importantly, dominant negative dynamin-1 impaired the internalization of NHE5 36HA3 (indicated by asterisks in Fig. 5, G and H) in a manner comparable with that observed for r-Tfn, indicative of a clathrin-dependent mechanism.
Trafficking of NHE5 Is Dependent on Phosphatidylinositol 3Ј-Kinase-Trafficking of vesicles along the recycling endocytic pathway is known to be regulated by PI3-K (65). Evidence supporting the involvement of this inositide kinase is derived mainly from experiments using wortmannin, a cell-permeant fungal toxin that specifically and irreversibly inhibits the enzyme (IC 50 of 5-10 nM) by alkylating its catalytic p110 subunit (66,67). Thus, treatment of CHO cells with wortmannin was found to decrease the number of plasma membrane Tfn receptors by increasing the rate of endocytosis while decreasing exocytosis (68). Moreover, wortmannin was found to induce a marked tubulation and expansion of endosomes containing Tfn receptors and fluid phase markers (65).
To evaluate the role of PI3-K on the distribution and function of NHE5 36HA3 , cells were pretreated with 100 nM wortmannin for 75 min at 37°C. Inhibition of PI3-K enhanced the accumulation of NHE5 36HA3 in the juxtanuclear region and the formation of larger tubulovesicular structures, while concurrently diminishing peripheral staining (Fig. 6, A and B). This was accompanied by a marked reduction (50 -60%) in plasmalemmal Na ϩ /H ϩ exchange over a broad pH i range, as assessed fluorimetrically by measuring the rates of Na ϩ o -dependent pH recovery following an imposed acid load (Fig. 6C). As a complementary approach, NHE5 activity was also measured by radioisotopic means. Similar to the fluorimetric assay, the rate of amiloride-inhibitable, H ϩ i -activated 22 Na ϩ uptake was significantly decreased by ϳ30% following a 30-min preincubation with 100 nM wortmannin (Fig. 6D). The quantitative difference between the two measures of Na ϩ /H ϩ exchange activity probably reflects variations in the assay conditions, but qualitatively the results are comparable. Similar effects on transport were also observed with LY294002, another inhibitor of PI3-K that is chemically unrelated to wortmannin and impairs the enzyme by a different mechanism (data not shown) (69). Taken  6. Exocytosis of NHE5 is regulated by phosphatidylinositol 3-kinase. A and B, cellular distribution of NHE5 36HA3 following inhibition of PI3-K. AP-1/NHE5 36HA3 cells were grown on 25-mm glass coverslips to ϳ70% confluence and then incubated in serum-free RPMI medium for 3 h at 37°C, followed by treatment in the same medium with vehicle (Con) or 100 nM wortmannin (Wtm), a highly selective antagonist of PI3-K, for 30 min. The cells were then incubated with anti-HA antibody at 37°C in the absence or presence of the inhibitor for 45 min, followed by washes, fixation, and permeabilization. Visualization of the signals for NHE5 36HA3 was performed as described in Fig. 3. The images were taken using a confocal microscope. C, measurement of NHE5 36HA3 activity assayed as the rate of Na ϩ -dependent H ϩ efflux following an NH 4 ϩ -induced acid load. AP-1/NHE5 36HA3 cells were cultured and then pretreated with vehicle (squares) or 100 nM wortmannin (circles) as described above. In the last 10 min of the incubation, the cells were loaded with 2 M BCECF-acetoxymethyl ester in bicarbonate-free isotonic medium containing 50 mM NH 4 Cl. At the end of the incubation, the cells were washed with Na ϩ -free medium, and their pH i was monitored fluorimetrically, as described under "Experimental Procedures." Recording was initiated upon reintroduction of extracellular Na ϩ to induce Na ϩ /H ϩ exchange. The results are means Ϯ S.E. of six determinations from three separate experiments. D, measurement of NHE5 36HA3 activity assayed as the rate of amiloride-inhibitable, acidactivated 22 Na ϩ influx. The cells were grown to confluence in 24-well plates, serum-deprived, and then pretreated with vehicle or 100 nM wortmannin for 30 min. Prior to radioisotope uptake, the cells were acidified using the NH 4 Cl prepulse technique. Rates of transport activity were measured as described under "Experimental Procedures." Significant difference from control values was determined by a paired Student's t test and is indicated by an asterisk (p Ͻ 0.05).
together, these data suggest that PI-3K regulates plasmalemmal NHE5 activity by modulating the number of exchangers at the cell surface, although changes in the intrinsic activity of NHE5 cannot be discounted.
Disruption of the Actin Cytoskeleton Increases NHE5 Activity-We have previously shown that optimal function of NHE3 requires an intact cytoskeleton (50). When the cytoskeleton was disrupted, the activity of this isoform was largely inhibited. Since both the function and subcellular distribution of NHE5 resemble that of NHE3, we postulated that NHE5 might be similarly regulated by the state of F-actin assembly. To examine this possibility, NHE5 36HA3 -transfected cells were treated with cytochalasin D, a membrane-permeant drug that induces gradual depolymerization of actin filaments by binding to their fast-growing barbed ends (70). As illustrated in Fig. 7, A and B, the well defined stress fibers and actin accumulation near focal adhesions seen in control cells were no longer apparent after a 30-min treatment with 10 M cytochalasin D. Instead, small punctate and large amorphous accumulations of F-actin were present throughout the cells. However, in contrast to the inhibitory effects observed for NHE3, these changes were associated with a marked activation of NHE5 36HA3 , as measured fluorimetrically in acid-loaded cells (Fig. 7G). This altered transport activity does not reflect changes in the cellular buffering capacity, since it was unchanged following exposure to cytochalasin D (the buffering power was 15.31 Ϯ 0.99 mM/pH unit/liter and 15.8 Ϯ 1.75 mM/pH unit/liter in untreated and cytochalasin-treated cells, respectively; mean Ϯ S.E., n ϭ 4). Elevation of NHE5 36HA3 activity was also detected in nonacidified cells using the radioisotopic flux assay (Fig. 7H), which is a more sensitive measure of transport activity at resting pH i .
The preceding data suggested that NHE5 may interact physically with the actin cytoskeleton. If so, redistribution of the exchangers might be expected upon disruption of the actin filament network. To selectively visualize the plasmalemmal exchangers before and after treatment with cytochalasin D, NHE5 36HA3 was immunolabeled without cell permeabilization. Typical results are illustrated in Fig. 7, C and D. The distribution of NHE5 36HA3 at the surface of control cells is rather homogeneous, with faint punctation over a diffuse background (Fig. 7C). This pattern was similar to that observed in cells treated with cytochalasin D (Fig. 7D), although there was a tendency for these cells to display increased surface immunostaining. However, visual estimation of plasmalemmal exchangers is limited by possible differences in focal plain and by heterogeneity in the population, which introduces considerable variance when comparing small numbers of cells by microscopy. To quantify the effects of actin depolymerization on the surface density of NHE5, the number of exchangers was measured by binding anti-HA antibody to the external epitope tag, followed by a secondary 125 I-coupled antibody (see "Experimental Procedures"). As shown in Fig. 7I, cytochalasin treatment   FIG. 7. NHE5 activity is stimulated by the actin depolymerizing agent cytochalasin D. A-F, AP-1/NHE5 36HA3 cells were grown on 25-mm glass coverslips to ϳ70% confluence and then incubated in serum-free, HEPES-buffered RPMI medium for 3 h at 37°C. In B, D, and F, the cells were treated with 10 M cytochalasin D for 30 min. Following drug treatment, in A and B, the cells were fixed and permeabilized, and F-actin was labeled by incubating with rhodamine-labeled phalloidin (1:500) for 1 h at room temperature. In C and D, cells were incubated with anti-HA antibody at 4°C in the absence or presence of cytochalasin D, respectively, for 45 min to label surface NHE5 36HA3 . After washing, the cells were fixed and blocked with 5% milk in PBS for 1 h, and the anti-HA antibody-NHE5 36HA3 complex was visualized using a Cy3-labeled secondary antibody. The images shown were taken using a conventional Leica fluorescence microscope. In E and F, cells were incubated with anti-HA antibody at 37°C in the absence or presence of cytochalasin D for 45 min, followed by fixing, permeabilization, and labeling with a secondary antibody. G, NHE5 36HA3 activity assayed as cellular pH i recovery following an NH 4 ϩ -induced acid load. Cells expressing NHE5 36HA3 were pretreated with vehicle (circles) or 10 M cytochalasin D (squares) for 30 min. The cells were loaded with BCECF and acidified, and pH recovery was measured. The results are means Ϯ S.E. of six determinations from three separate experiments. H, NHE5 36HA3 activity assayed as the rate of amilorideinhibitable 22 Na ϩ influx in resting cells. The cells were grown to confluence in 24-well plates and then pretreated with vehicle (hatched) or 10 M cytochalasin D (solid) for 30 min. Rates of transport activity were measured as described under "Experimental Procedures." I, cells were untreated or treated with 10 M cytochalasin D for 30 min and then incubated with anti-HA antibody (1:1000 dilution) in RPMI supplemented with 5% goat serum for 1 h at 4°C to prevent endocytosis. Following washing, the cells were incubated with 125 I-labeled goat anti-mouse IgG (0.4 Ci/sample) in RPMI with 10% goat serum for 45 min at 4°C. At the end of the incubation period, cells were washed three times with HEPES-buffered RPMI to remove unbound radiolabel and then detached from the plates using PBS containing 1% Triton-X-100. Radioactivity was counted with a 1282 Compugamma LKB counter. The radioactivity bound to cells exposed only to 125 I-IgG without prior incubation with anti-HA antibody was subtracted from all determinations. Data are expressed as the percentage of surface labeling of untreated cells (100%), and the values represent the mean Ϯ S.E. of seven experiments, each performed in duplicate or triplicate. Significant difference from control values was determined by a paired Student's t test and is indicated by an asterisk (p Ͻ 0.05). resulted in an approximate 70% increase in the number of transporters at the cell surface, an amount proportionate to the increase in NHE5 activity.
One possible mechanism to account for this apparent increase in NHE5 is reduced endocytosis from the surface. Indeed, recent studies have shown that disruption of the cytoskeleton using cytochalasin D decreases the rate of endocytosis of transferrin in adherent CHO cells (71). To investigate this possibility, intact control and cytochalasin D-treated NHE5 36HA3 cells were incubated with the anti-HA antibody at 37°C for 1 h to selectively label surface transporters. After extensive washes, the cells were fixed and permeabilized, and the immunolabeled NHE5 36HA3 was visualized using an FITCcoupled anti-mouse secondary antibody. As shown in Fig. 7, E and F, the redistribution of surface-labeled NHE5 36HA3 into a juxtanuclear vesicular cluster was noticeably diminished in cytochalasin D-treated cells relative to control cells. These results are consistent with the notion that internalization of NHE5 36HA3 is attenuated when the actin cytoskeleton is disrupted, leading to increased functional exchangers at the cell surface.
Localization of NHE5 in Transfected Neuronal Cells-In an effort to gain some indication of the distribution of NHE5 in cells that more closely resemble its native environment, the epitope-tagged transporter was transiently overexpressed in NGF-differentiated PC12 cells and primary cultures of rat hippocampal neurons. In differentiated PC12 cells at 48 h (Fig. 8,  A and B) and 72 h (Fig. 8, C and D) post-transfection, immunostaining of NHE5 36HA3 localized to punctate vesicles of varying size in the perinuclear region of the cell body and along the neurite processes. Similar to the pattern in PC12 cells, NHE5 36HA3 immunoreactivity in hippocampal neurons (48 h post-transfection) accumulated in the perinuclear region of the somata (Fig. 8, E and G), although the staining was diffuse, but also in more punctate and tubular structures that were readily visualized at dendrite extremities (Fig. 8, E and F) and along the axons (Fig. 8, G and H). Endomembrane compartments containing NHE5 36HA3 within the somatodendritic region were difficult to resolve due to the intensity of the immunostaining in the limited number of neurons transfected (Ͻ0.1% transfection efficiency).

Subcellular Distribution of NHE5 to the Plasmalemma and
Recycling Endosomes-The Na ϩ /H ϩ exchanger NHE5 isoform is unique among NHE family members by its highly restricted expression in neural tissue (38,39). Whereas NHE5 mRNAs have been detected in human spleen and testis, analyses of these transcripts showed that they were either incompletely processed or aberrantly spliced and therefore unlikely to yield functional transporters (38). However, determination of the cellular and functional properties of NHE5 in neurons is complicated by the coexistence of other NHE isoforms and the current absence of specific molecular probes. To circumvent these difficulties, we transfected NHE5 into CHO AP-1 cells. These cells are deficient in plasmalemma NHE activity, thereby facilitating functional measurements. Additionally, despite being undifferentiated, they retain a variety of exocytic and endocytic sorting mechanisms that are analogous to the more specialized membrane trafficking pathways found in polarized cells such as epithelia and neurons (43)(44)(45)(46)(47)(48)(49). They are particularly advantageous for microscopic localization studies, since they are broader than neuronal cells and easy to transfect, and markers for their intracellular compartments are well defined. Using this approach, we recently determined (40) that human NHE5 has pharmacological properties closely resembling an amiloride-resistant NHE described in hippocampal neurons (42), consistent with the high abundance of NHE5 transcripts in this region of the brain (38,39). In the present report, we extend these studies by showing that a functionally competent, epitope-tagged form of NHE5 is present not only in the plasma membrane but also in endomembrane structures, a significant fraction of which is concentrated in a juxtanuclear location.
Several lines of evidence suggest that these intracellular vesicles correspond to recycling endosomes. First, NHE5 was retrieved from the plasmalemma into endomembrane vesicles that overlap extensively with internalized Tfn-bound receptors, a recognized marker for recycling endosomes (62). Moreover, NHE5 did not colocalize with markers of other intracellular organelles, such as mitochondria, lysosomes, and Golgi cisternae. Whereas a diffuse fraction of NHE5 was detected in the endoplasmic reticulum, it probably reflects biosynthetic accumulation due to ectopic overexpression rather than organellespecific retention. At present, there is little evidence to support the existence of resident Na ϩ /H ϩ exchange in the endoplasmic reticulum. Recent studies in fibroblasts using an endoplasmic reticulum-targeted pH-sensitive fluorophore showed that this compartment is highly permeable to H ϩ (or H ϩ equivalents) such that the prevailing luminal pH is equivalent to cytoplasmic pH and subject to indirect regulation by plasmalemmal acid-base transporters (72). Second, inhibition of PI3-K with wortmannin induced the coalescence of NHE5-containing vesicles in a manner that mirrored the effects reported for Tfn receptor-enriched endosomes (65,68). Treatment of CHO cells with this drug was found to reduce the surface density of Tfn receptors by moderately increasing the rate of internalization concomitantly with a marked inhibition in exocytosis (68). Importantly, the effect of wortmannin is seemingly independent of any effect on fluid phase endocytosis (73) and therefore selective for receptor-mediated endocytosis. The wortmannin-mediated concentration of NHE5 within internal structures shown in Fig. 6 is also consistent with the parallel reduction in NHE5 activity. Third, experiments using dominant-negative dynamin-1 suggest that endocytosis of NHE5 takes place by a clathrin-dependent mechanism. It is also worth mentioning that endosomal accumulation is unlikely to result from overexpression of NHE5 per se, since previous studies showed that the ubiquitous NHE1 isoform, when expressed ectopically in the same cell type at comparable levels, was restricted to the plasmalemma (49), thereby implicating the existence of an intrinsic endomembrane targeting mechanism. Taken together, these observations indicate that NHE5 cycles dynamically between the plasmalemma and perinuclear recycling endosomes and provides a means to regulate the density and therefore the activity of exchangers at the surface membrane.
Cell Surface NHE5 Activity Is Dependent on the State of F-actin Assembly-The subcellular distribution of NHE5 closely resembles that observed for its closest homologue, NHE3, which localizes to both the plasmalemma and recycling endosomes of transfected CHO cells (49) and to the brush border membrane and apical endosomes of native epithelia (21,22,29). Despite similarities in membrane trafficking of NHE3 and NHE5, we observed a striking difference in their regulation by the actin cytoskeleton. NHE3 requires an intact cytoskeleton for optimal function. Disruption of actin assembly with drugs such as cytochalasin B and D or latrunculin B (50) or by overexpression of dominant negative mutants of RhoA-GTPase and Rho-associated kinase (32) markedly decreases NHE3 activity in AP-1 cells. Recent preliminary results indicate that NHE3 responds similarly in renal epithelial OK cells. 2 In AP-1 cells, this inhibitory effect reflected an alteration in its intrinsic activity, since the density of cell surface exchangers was unchanged (32,50). In contrast, in the same cells and under similar conditions, cytochalasin-mediated disruption of the actin cytoskeleton elevated NHE5 activity. As shown by cell surface immunolabeling experiments (see Fig. 7), this elevation was attributable, at least in part, to an increase in the number of plasmalemmal exchangers concomitant with a reduction in internalization of NHE5. An association between actin and clathrin-mediated endocytosis is not without precedent in mammalian cells. Several studies have shown that disruption of the actin cytoskeleton inhibits receptor-mediated endocytosis (71,74,75), although this phenomenon is not universally observed (71,76) and seems dependent on cell type and the state of actin assembly and organization. At present, the molecular connections between clathrin-mediated endocytosis and the actin cytoskeleton are ill defined. Thus, whereas the activities of NHE3 and NHE5 are subject to dynamic cycles of actin polymerization and depolymerization, the molecular events underlying this phenomenon for the latter are clearly distinct from those influencing NHE3 function.
Possible Functional Role of NHE5 in Neurons-Neurons contain at least two distinct types of secretory endomembrane vesicles: large dense core vesicles that store and secrete peptide neurotransmitters and amines and synaptic vesicles involved in the rapid storage and release of nonpeptide hormones (77). At present, the subcellular distribution of native NHE5 in neurons is unknown and awaits the development of isoformspecific antibodies. However, our initial immunolabeling experiments of NHE5 36HA3 in transiently transfected NGFdifferentiated PC12 cells and primary cultures of rat hippocampal neurons show accumulation in somatodendritic vesicles but also in varicosities along the axonal processes (i.e. regions that are enriched in synaptic vesicles or, in the case of neuroendocrine cells, synaptic-like microvesicles). More precise identification of the nature of these NHE5-containing vesicles using neuronal organelle-specific markers is currently ongoing.
Despite this uncertainty, tentative morphological localization of NHE5 to synaptic vesicles or synaptic-like microvesicles, which are functionally analogous to vesicles of the receptormediated recycling pathway (78), suggests that the sorting behavior of NHE5 in CHO cells may also be conserved, at least to some extent, in neuronal cells. In this regard, it is interesting to note parallels with the subcellular distribution of the synaptic vesicle protein synaptophysin in nonpolarized and polarized cell types (43,79). When ectopically expressed in CHO cells, synaptophysin colocalizes with Tfn receptor-associated recycling endosomes. By comparison, in cultured hippocampal neurons, the distribution of synaptophysin partially coincides with Tfn receptors in the somatodendritic region but is selectively enriched in axonal synaptic vesicles that are devoid of Tfn receptors (43,79).
In view of the above, what might be the physiological role(s) of NHE5 in neuronal function? Although speculative, one intriguing possibility is that NHE5 resident in synaptic vesicles might influence synaptic transmission by modulating the luminal acidity of that compartment. Acidification of synaptic vesicles, generated in part by the vacuolar H ϩ -ATPase, provides the electrochemical driving force for the uptake of neurotransmitters, such as glutamate, GABA, acetylcholine, and catecholamines (80). By inference, the degree of vesicular acidification should serve as an important determinant of neurotransmitter concentration and ultimately of synaptic transmission. The extent of luminal acidity may also modulate the rate of dissociation of internalized ligands from their receptors, thereby facilitating the recycling of the receptors back to the cell surface. The ability of NHEs to regulate endosomal acidification is not without precedent. An amiloride-insensitive NHE was shown to modulate the acidification of rat liver endocytic vesicles immediately after their formation (81). Likewise, the epithelial NHE3 isoform was found to increase the acidification of recycling endosomes when expressed in CHO cells (49). Additionally, the passive release of luminal H ϩ (or H ϩ equivalents) into the synaptic cleft upon vesicle fusion with the plasma membrane might further influence synaptic transmission by modulating the activities of nearby proton-sensitive ligand receptor-mediated currents (82,83) and voltage-gated cation channels (84,85). Alternatively, or in conjunction, protons may also be actively transported across the plasmalemma into the synaptic cleft in a transient manner following recruitment of vesicular Na ϩ /H ϩ exchangers (i.e. NHE5) and H ϩ -ATPase to the cell surface. Evidence for synaptically released acid equivalents comes from studies by Krishtal et al. (1), who recorded transient millisecond fluctuations in extracellular pH of rat hippocampal slices following electrical stimulation of presynaptic pathways, an effect that was not observed under conditions (i.e. inhibition of Ca 2ϩ influx) that prevented synaptic transmission. Recently, in an elegant series of experiments, DeVries (86) showed that fusion of glutamate-enriched synaptic vesicles of cone photoreceptor cells elicits a transient inhi-bition of neighboring presynaptic inward Ca 2ϩ channels required for vesicle fusion, an effect that was suppressed by pH buffers and antagonists of the H ϩ -ATPase but independent of glutamate release. These data were interpreted as evidence of a proton-dependent feedback mechanism to inhibit ion channel function and neurotransmitter release, thereby suppressing synaptic transmission. Whether Na ϩ /H ϩ exchangers such as NHE5 contribute to pH-dependent regulation of synaptic transmission is unknown, but it remains an attractive hypothesis that warrants further investigation.