Cell Physiology

Differential expression and targeting of endogenous Arf1 and Arf6 small GTPases in kidney epithelial cells in situ

Jaafar El Annan, Dennis Brown, Sylvie Breton, Sylvain Bourgoin, Dennis A. Ausiello, Vladimir Marshansky


ADP-ribosylation factors (Arfs) are small GTPases that regulate vesicular trafficking in exo- and endocytotic pathways. As a first step in understanding the role of Arfs in renal physiology, immunocytochemistry and Western blotting were performed to characterize the expression and targeting of Arf1 and Arf6 in epithelial cells in situ. Arf1 and Arf6 were associated with apical membranes and subapical vesicles in proximal tubules, where they colocalized with megalin. Arf1 was also apically expressed in the distal tubule, connecting segment, and collecting duct (CD). Arf1 was abundant in intercalated cells (IC) and colocalized with V-ATPase in A-IC (apical) and B-IC (apical and/or basolateral). In contrast, Arf6 was associated exclusively with basolateral membranes and vesicles in the CD. In the medulla, basolateral Arf6 was detectable mainly in A-IC. Expression in principal cells became weaker throughout the outer medulla, and Arf6 was not detectable in principal cells in the inner medulla. In some kidney epithelial cells Arf1 but not Arf6 was also targeted to a perinuclear patch, where it colocalized with TGN38, a marker of the trans-Golgi network. Quantitative Western blotting showed that expression of endogenous Arf1 was 26–180 times higher than Arf6. These data indicate that Arf GTPases are expressed and targeted in a cell- and membrane-specific pattern in kidney epithelial cells in situ. The results provide a framework on which to base and interpret future studies on the role of Arf GTPases in the multitude of cellular trafficking events that occur in renal tubular epithelial cells.

  • protein trafficking
  • immunofluorescence microscopy
  • Western blotting
  • endocytosis

in epithelial cells from different regions of the kidney, vesicular trafficking and plasma membrane receptor signaling play a crucial role in many aspects of tubule function (10, 37, 38). Recently, important progress has been made in our understanding of the mechanisms responsible for the regulation of both exo- and endocytosis pathways (12, 16, 18). In current models of the regulation of these events, key roles are played by a family of ADP-ribosylation factors (Arf) that are members of the Ras superfamily of small GTPases. On the basis of their sequence similarities, members of the Arf family are divided into three classes. Class I consists of Arf1, Arf2, and Arf3; class II consists of Arf4 and Arf5; and class III has only one member, Arf6. Little is known about class II Arfs (Arf4 and Arf5), whereas the two most divergent members, Arf1 and Arf6, have been the most widely studied (16, 17, 26).

Similar to the other Ras family small GTPases, Arfs function as molecular binary switches. The GDP-bound form of Arfs is inactive, whereas replacement of GDP with GTP activates Arfs and promotes their recruitment to target membranes. This activation is catalyzed by guanine nucleotide exchange factors (GEFs). Inactivation of Arfs (conversion into the Arf-GDP form) is catalyzed by GTPase-activating proteins (GAPs), which enhances the very low intrinsic GTPase activity of Arfs and triggers their release from target membranes. Thus both regulation of specific activities and targeting of Arf isoforms to different intracellular membranes are most likely achieved through selective sets of GEF and GAP proteins associated with target membranes.

As a consequence of activation and recruitment to their target membranes, Arf proteins have been shown to regulate various vesicular trafficking pathways (12, 16, 48, 51), phospholipid metabolism (11), and actin cytoskeleton assembly (23). This occurs via the activation and/or recruitment of a variety of downstream effectors. It is generally accepted that Arf1 is involved in the formation of Golgi-derived COP-coated vesicles, is an important regulator of the intra-Golgi or Golgi-endoplasmic reticulum vesicular trafficking and thus is involved in regulation of the exocytotic pathway (48, 51). In contrast, Arf6 has been implicated in coat formation on endosome-derived carrier vesicles (45, 54) and in actin cytoskeleton rearrangement (15). It appears to be a critical regulator of endocytotic pathways (16, 42).

Recently, a novel role of an Arf6 small GTPase in G protein-coupled receptor desensitization has been uncovered (29). Upon agonist stimulation, a direct interaction of β2-arrestin with Arf6 and with its cognate GEF ARNO (as in “ADP-ribosylation nucleotide site opener”) followed by their recruitment to plasma membrane has been demonstrated (14). Thus activation of Arf6 by ARNO on the plasma membrane promotes endocytosis and downregulation of β2-adrenergic (β2AR) receptors in HEK-293 (14) as well as luteinizing hormone-choriogonadotropin hormone (LH-CGR) receptors in ovarian follicular membranes (43, 49).

We have previously reported the expression and codistribution of Arf6 and ARNO in the receptor-mediated endocytosis pathway of kidney proximal tubule epithelial cells (36). These proteins were colocalized with the endocytotic receptor megalin and with V-ATPase in apical endosomal vesicles in situ as well as in purified early endosomes in vitro (38, 39). We also showed that V-ATPase-dependent intra-endosomal acidification stimulates the recruitment of ARNO and Arf6 from proximal tubule cytosol to endosomal membranes, implicating this process in endosomal function in situ (36, 41). On the basis of these data, we proposed a key role for Arf6/ARNO in the regulation of megalin/cubilin-mediated endocytosis in proximal tubule epithelial cells (37) and a possible role in the pathogenesis of Dent's disease and Fanconi syndrome (37, 40, 41).

However, it is likely that Arf small GTPases are involved in regulating vesicle trafficking as well as other intracellular processes in other renal epithelial cell types. As a first step toward understanding their function in such events, we have carried out an extensive examination of the expression and localization of two small GTPases, Arf1 and Arf6, in cells lining the urinary tubule. The characteristic distribution of Arf1 and Arf6 in these renal epithelial cells in situ suggests cell- and membrane-specific regulatory properties that are related to diverse epithelial cell functions in this complex organ.


Antibodies. Goat anti-rabbit (GAR-Alexa488), goat anti-mouse (GAM-Alexa488), and goat anti-chicken (GAC-Alexa488) IgG conjugated to Alexa488 were obtained from Molecular Probes (Eugene, OR). Goat anti-rabbit (GAR-Cy3), goat anti-mouse (GAM-Cy3), and goat anti-chicken (GAC-Cy3) IgG conjugated to cyanine-3 (Cy3) were obtained from Jackson Immunoresearch (West Grove, PA). Mouse monoclonal anti-calbindin antibodies were purchased from Sigma. Mouse monoclonal anti-TGN38 antibodies were purchased from BD Transduction Laboratories. Mouse monoclonal anti-megalin antibodies were obtained from Dr. R. T. McCluskey, Department of Pathology, Massachusetts General Hospital (30). Production and purification of recombinant recArf1 and recArf6 proteins as well as production and characterization of the rabbit polyclonal anti-Arf1 and monoclonal anti-Arf6 antibodies have been previously described (36, 39). Rabbit polyclonal anti-V-ATPase (subunit E, 31 kDa) antibodies were raised against a peptide (amino acids 217–226, C-GANANRKFLD), corresponding to the last 10 COOH-terminal amino acids of human, bovine, and mouse V-type ATPase subunit E, coupled to keyhole limpet hemocyanin (KLH). This antibody specifically recognizes V-ATPase subunit E (31-kDa protein) in Western blots of purified rat and dog kidney endosomes as well as of purified rat liver Golgi. The chicken polyclonal anti-V-ATPase antibodies were also generated against a KLH-coupled COOH-terminal peptide and affinity purified using a SulfoLink kit (Pierce, Rockford, IL). These reagents have been characterized previously (1, 4).

Animals and kidney fixation. Adult Sprague-Dawley rats were maintained on a standard diet and had free access to water. For immunofluorescence analysis the tissues were prepared as follows. The animals were anesthetized by intraperitoneal injection of 40 mg/kg body wt of Nembutal (Abbott, North Chicago, IL) and perfused through the left ventricle first with phosphate-buffered saline (PBS) (0.09% NaCl in 10 mM phosphate buffer, pH 7.4) followed by PLP fixative (4% paraformaldehyde, 10 mM sodium periodate, 70 mM lysine, and 5% sucrose in 10 mM sodium phosphate) as previously described (1, 5). The kidneys were PLP perfusion fixed for 5 min in situ, and slices were further fixed by immersion overnight in PLP at 4°C, washed three times in PBS, and stored until use in the same buffer containing 0.02% sodium azide. Tissues were cryoprotected by immersion in 0.9 M (30%) sucrose in PBS for at least 1 h and mounted for cryosectioning in Tissue-Tek (Miles, Elkhart, IN), before freezing in liquid nitrogen and sectioning at 4 μm with a Reichert-Jung 2800 Frigocut cryostat (Spencer Scientific, Derry, NH). Sections were picked up on SuperFrost Plus charged glass slides (Fisher, Pittsburgh, PA).

Preparation of kidney membranes. For Western blot analysis the tissues were prepared as follows. Kidneys of anesthetized rats were perfused with cold PBS for 1–2 min to remove blood, and the cortex, outer medulla, inner medulla and papilla were separated under a dissecting microscope. Pieces of the corresponding kidney parts (0.5 g) were homogenized in 5 ml of homogenization buffer (HB; 0.25 M sucrose, 1 mM EDTA, 10 mM Tris·HCl, pH 7.4, with Complete protease inhibitor from Roche Molecular Biochemicals) using a Wheaton glass Potter homogenizer fitted with a Teflon pestle attached to an electric Glas-Col (Fisher) stirrer (2,000 rpm with 20 complete strokes). Postnuclear supernatants of the different kidney regions were prepared by centrifugation of the homogenates for 15 min at 1,000 g (3,000 rpm). Protein concentration was measured with the Pierce BCA protein assay kit using albumin as a standard.

SDS-PAGE and Western blot analysis. Kidney cell proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions according to Laemmli (32). Prestained, broad-range Precision protein standards (Bio-Rad, Hercules, CA) were used as molecular weight markers. Kidney tissue postnuclear supernatant samples were diluted 3:1 in 4× reducing sample buffer, boiled for 5 min, loaded at 10 μg protein/lane onto 4–20% gradient Tris-glycine Novex minigels and electrophoretically separated using a Novex-XCell electrophoresis unit (Invitrogen, Carlsbad, CA). Proteins were transferred to Immobilon-P PVDF membranes (Millipore, Bedford, MA) using a Hoefer SemiPhor transfer unit (Amersham-Pharmacia, Piscataway, NJ). Nonspecific binding sites were blocked by exposing the membrane to 5% nonfat dry milk for 1 h at 20°C. Membranes were incubated for 1 h with primary antibody against Arf6 (1:1,000) or Arf1 (1:1,000) diluted in 3% albumin in TBS-Tween buffer. After being washed four times in TBS-Tween buffer, the membranes were exposed, respectively, to donkey anti-mouse (1:50,000) or donkey anti-rabbit (1:50,000) horseradish peroxidase-conjugated antibodies. Membranes were then washed four times in TBS-Tween buffer, and detection of antibodies binding was performed with the enhanced chemiluminescence method using Fuji Super RX films. Films were scanned and quantitative densitometric analysis was performed using NIH Image 1.62 software. Levels of expression of endogenous Arf6 and Arf1 in kidney tissue were quantified using calibration curves of highly purified recombinant recArf1 and recArf6 running simultaneously.

Conventional and confocal immunofluorescence microscopy. Immunofluorescence staining was performed on cryostat sections of rat kidney tissue after antigen retrieval. Kidney tissue slices were prepared as described above. Permeabilization and antigen retrieval was performed by treatment of cryostat sections with 1% SDS in PBS for 4 min (9). After being washed in PBS buffer (three times for 5 min each time) sections were blocked with 1% BSA in PBS for 30 min. Primary polyclonal rabbit anti-Arf1 (1:1,000), anti-mannosidase II (1:500), and chicken anti-V-ATPase (1:500), as well as monoclonal anti-Arf6 (1:10), anti-megalin (1:500), anti-calbindin (1:500), and anti-TGN38 (1:100), were diluted in DAKO antibody diluent and incubated overnight at 4°C. Secondary GAR/GAM/GAC-Alexa488 (1:100) as well as GAR/GAM/GAC-Cy3 (1:800) were also diluted in DAKO antibody diluent and incubated for 60 min at room temperature. After being washed in PBS, slides were also stained with 2 μg/ml 4′,6-diamidino-2-phenylindole (DAPI) for 15 min. Sections were mounted in a 2:1 mixture of Vectashield mounting medium (Vector Laboratories, Burlingame, CA) in 1.5 M Tris solution (pH 8.9). Epifluorescence analysis was performed on a Nikon Eclipse E800 epifluorescence microscope connected to a Macintosh G4 computer. Images were captured using an Orca C4742-95 (Hamamatsu Photonics, Bridgewater, NJ) digital camera and IPLab Spectrum image processing software (version 3.1a; Scanalytics, Fairfax, VA). Confocal microscopy analysis was performed on a Bio-Rad Radiance 2000 confocal microscope using LaserSharp 2000 software. All images were transferred into Photoshop 7.0, paginated using Illustrator 9.0 (Adobe, Seattle, WA), and printed on a Stylus Photo 750 (Epson, Long Beach, CA) color printer.


Identification and localization of endogenous Arf isoforms in kidney proximal tubules. Figure 1, A and D, shows an apical localization of endogenous Arf6 below the brush-border membrane (BBM) in cortical proximal tubules (S1 and S2 segments). Arf6 is not detectable in glomeruli. Anti-Arf1 antibodies also heavily labeled the apical pole of S1 and S2 segments of proximal tubules (Fig. 1B) immediately adjacent to the glomeruli (Fig. 1C). Figure 1C shows a differential-interference contrast (DIC) microscopy image of the kidney corresponding to the immunofluorescence image depicted in Fig. 1B. Arf1 was also concentrated in the apical submicrovillar region just below the BBM (Fig. 1E), where it colocalized with Arf6 (Fig. 1F, merge). In contrast to Arf6, Arf1 was also detected in a perinuclear location in some kidney epithelial cells, possibly in the trans-Golgi network (TGN)/Golgi complex (see below). Confocal microscopy analysis also reveals that in the apical pole of proximal tubules, Arf1 is colocalized with megalin (Fig. 2, AC, arrowheads), the LDL family member protein scavenger receptor that is involved in the reabsorption and trafficking of filtered proteins. Megalin is not expressed and thus is not colocalized with Arf1 in cells of glomeruli (Fig. 2 A and C, arrows) and collecting duct (Fig. 2, A and C, asterisks).

Fig. 1.

Targeting and distribution of ADP-ribosylation factors Arf6 and Arf1 in kidney proximal tubule epithelial cells in situ. Immunofluorescence microscopy shows the localization of Arf6 (A) and Arf1 (B) in the apical pole of S1 and S2 segments of proximal tubule epithelial cells. C: a differential-interference contrast (DIC) image of the kidney corresponding to the immunofluorescence image depicted in B, illustrating the position of glomeruli (GL) as well as S1 and S2 segments of proximal tubule. At higher magnification, both Arf6 (D) and Arf1 (E) are shown to partially colocalize (F) at the apical pole of proximal tubules and in apical vesicles below the brush-border membrane (BBM). Nuclei of some tissue sections were stained with 4′,6-diamidino-2-phenylindole (DAPI) (B and DF). Proximal tubule lumen (PT lumen) and BBM are indicated in DF. Bars = 25 μm (AC) or 10 μm (DF).

Fig. 2.

Colocalization of Arf1 with megalin in the apical pole of proximal tubules. Confocal microscopy shows distribution of Arf1 (A) and megalin (B) in apical region of kidney proximal tubules. Arf1 is partially colocalized with megalin in S1 and S2 segments of the proximal tubule (AC, arrowheads). Megalin is not expressed and thus is not colocalized with Arf1 in cells of glomeruli (A and C, arrows) and collecting duct (A and C, asterisks). The localization of glomeruli (GL) and collecting duct (CD), as well as S1 and S2 segments of proximal tubule (PT), are indicated. Bars = 25 μm.

Localization of Arf1 in distal tubule and connecting segment. Strong expression of Arf1 was also observed in distal tubule and connecting segment epithelial cells. Similar to proximal tubules, Arf1 was predominantly localized in the apical pole of these tubules (Fig. 3A). Double-staining experiments identified distal tubules as calbindin negative (Fig. 3, B and C) and connecting segments as calbindin positive (Fig. 3, B and C), respectively. In addition to staining the apical pole of distal tubule and connecting segment cells, perinuclear Arf1 staining was also detectable, possibly representing TGN/Golgi staining (Fig. 3, D and E). In connecting segments, the apical pole of calbindin-negative cells was stained more strongly than calbindin-positive cells (Fig. 3C). We have previously shown that calbindin-negative cells in this segment are intercalated cells (47).

Fig. 3.

Expression and targeting of Arf1 in distal tubule and connecting segment. In distal tubules and connecting segments, Arf1 is predominantly localized at the apical pole (AC). In distal tubules, in addition to being targeted to the apical pole, Arf1 is also associated with juxtanuclear structures, possibly the trans-Golgi network (TGN)/Golgi complex (D and E). F: a DIC image of the distal tubule corresponding to the immunofluorescence image depicted in E. To distinguish between distal tubules and connecting segments, double staining with calbindin was performed. Distal tubules are poorly stained, whereas connecting segments are positive. Polyclonal anti-Arf1 antibodies were detected with goat anti-rabbit IgG conjugated to Alexa488 (GAR-Alexa488; A, green), and monoclonal anti-calbindin antibodies were detected with goat anti-mouse IgG conjugated to cyanine-3 (GAM-Cy3; B, red). The merge is shown in C. The strongest apical Arf1 staining is seen in calbindin-negative cells in the connecting segment: these are intercalated cells. DAPI staining was applied to stain nuclei of epithelial cells. DT, distal tubule; CS, connecting segment. Bars = 10 μm (AC, E and F) or 25 μm (D).

Distribution of Arf1 in collecting duct and thick ascending limb epithelial cells. In cortical collecting duct epithelial cells, Arf1 is found in both the apical and basolateral poles (Fig. 4A, green) of some cells. Note that in these epithelial cells, Arf1 is not colocalized with TGN38 (Fig. 4A, red), a marker of the TGN/Golgi compartment. Figure 4B shows a DIC microscopy image of the collecting duct corresponding to the immunofluorescence image depicted in Fig. 4A. Double labeling with the 31-kDa subunit E of the V-ATPase was used to identify various cell types in the cortical collecting duct (Fig. 4, C and D). Cells with exclusively apical localization of V-ATPase (Fig. 4C2) are type A intercalated cells (A-IC). Cells with basolateral and/or with basolateral/apical localization (Fig. 4D2) are type B intercalated cells (B-IC). There is predominant apical targeting of Arf1 in A-IC cells (Fig. 4C3, merge), whereas both apical and basolateral Arf1 staining was detectable in B-IC, in which it was colocalized with the V-ATPase (Fig. 4D3, merge). Principal cells also expressed apical (and cytoplasmic) Arf1, but the intensity of staining was lower than in the intercalated cells (Fig. 4A).

Fig. 4.

Localization of Arf1 in cortical CD. In the cortical CD Arf1 (A, green) is targeted to and associated with both apical and basolateral membrane domains of different epithelial cells, as well as being present in the cytosol. In the cortical CD, Arf1 (A, green) is not colocalized with TGN38 (A, red), a marker of the TGN/Golgi complex. B: a DIC image of the CD corresponding to the immunofluorescence image depicted in A. Double staining with Arf1 and V-ATPase demonstrates predominant apical targeting of Arf1 in type A intercalated cells (A-IC) (C) and additional (sometimes predominant) basolateral targeting in type B intercalated cells (B-IC) (D). In both cell types, Arf1 and V-ATPase distribution overlapped. Nuclei were stained with DAPI. Bars = 10 μm (A and B) or 5 μm (C and D).

In the inner stripe of the outer medulla, Arf1 is expressed in thick ascending limb epithelial cells (Fig. 5A). In the thick ascending limb, Arf1 was predominantly subapical and perinuclear (Fig. 5A, inset). In the inner medulla and papilla, Arf1 was also abundantly expressed in collecting duct epithelial cells (Fig. 5B). Arf1 staining was detectable in the perinuclear region and was associated with multiple vesicles distributed throughout the cytoplasm of collecting duct epithelial cells (Fig. 5B, bottom inset). In contrast the distribution of Arf1 was perinuclear, showing a TGN/Golgi-like pattern in non-collecting duct cells in the interstitium of the papilla (Fig. 5B, top inset).

Fig. 5.

Distribution and targeting of Arf1 in epithelial cells in inner stripe and papilla. In the inner stripe (A) Arf1 was expressed in thick ascending limbs (TAL) with a cytosolic and perinuclear localization (A, inset). In the papilla (B) Arf1 is more abundantly expressed in CDs but is also detectable in surrounding, unidentified non-CD cells. In CDs (B, bottom inset) Arf1 was found in the perinuclear cytosol, as well as on vesicles throughout the cell. In non-CD cells (B, top inset) Arf1 is exclusively found in a perinuclear patch, probably the TGN/Golgi complex (see Fig. 6). Bars = 25 μm.

Targeting and localization of Arf1 in TGN/Golgi complex of kidney epithelial cells in situ. Analysis of the expression and localization of Arf small GTPases along the nephron revealed the selective targeting of Arf1 but not Arf6 to the perinuclear region of the some kidney cells in situ. This TGN/Golgi-like pattern of localization was observed in proximal tubule (Fig. 6A, arrows) and was especially pronounced in glomeruli (Fig. 6C, arrows) and in non-collecting duct cells in the papilla (Fig. 6E and Fig. 5B, top inset). Double staining with anti-TGN38 antibodies demonstrated a pattern of staining that overlapped with perinuclear Arf1 staining in proximal tubules (Fig. 6B, arrows), glomeruli (Fig. 6D, arrows), and non-collecting duct epithelial cells of the papilla (Fig. 6F). The cells in the glomeruli that show Arf1 perinuclear staining have the shape and distribution of podocytes.

Fig. 6.

Distribution and targeting of Arf1 to TGN/Golgi complex in kidney epithelial cells in situ. Double staining with Arf1 (green) and a marker of the TGN/Golgi complex TGN38 (red) shows targeting of Arf1 to the TGN/Golgi region in various types of kidney cells in situ. Arf1 (green) has a similar distribution to TGN38 (red) in proximal tubule cells (A and B, arrows), glomerular podocytes (C and D, arrows), and papillary non-CD cells (E and F). Nuclei were stained with DAPI. Bars = 5 μm (A and B), 25 μm (C and D), or 2 μm (E and F).

Localization and distribution of endogenous Arf6 in collecting ducts. The distribution and targeting of Arf6 was distinct from the distribution of Arf1 in collecting duct epithelial cells. Double staining for Arf6 and Arf1 is shown in Fig. 7A. In cortical collecting ducts, Arf6 is targeted predominantly to the basolateral pole where it is associated with vesicular structures (Fig. 7A, green, arrowheads), whereas Arf1 is located at the apical pole (Fig. 7A, red, arrows). Double labeling of sections for Arf6 and mannosidase II clearly demonstrated the targeting of Arf6 to basolateral vesicles (Fig. 7B1, green) which are distinct from the perinuclear TGN/Golgi complex (Fig. 7B1, red). Figure 7B2 shows a DIC image (with nuclei highlighted) of the collecting duct corresponding to the immunofluorescence image depicted in Fig. 7B1.

Fig. 7.

Distribution and targeting of Arf6 in cortical CD epithelial cells. Double staining with Arf isoforms shows predominant targeting of Arf6 to the basolateral pole of cortical CD epithelial cells where it is associated with vesicular structures (A; green, arrowheads). Arf1 is located in a narrow band at the apical pole of the same cells (A; red, arrows). Double staining with Arf6 (B1; green) and mannosidase II (B1; red), a Golgi marker, shows no targeting of Arf6 to TGN/Golgi complex and absence of its colocalization with mannosidase II in situ. B2: a DIC image of the CD corresponding to the immunofluorescence image depicted in B1. Nuclei have been identified in this DIC image by superimposing a gray-scale, DAPI-stained image. Double staining with Arf6 and V-ATPase (C) demonstrates targeting of Arf6 to the basolateral pole of all cell types in the cortical CD. Monoclonal anti-Arf6 antibodies were detected with GAM-Alexa488 (C2 and C3; green), whereas polyclonal anti-V-ATPase (subunit E) antibodies were detected with GAR-Cy3 (C1 and C3; red). C1C3: A-IC are indicated by asterisks, B-IC are indicated by arrowheads, and principal cells are indicated by arrows. Similar double staining for the V-ATPase (red) shows that Arf6 maintains its basolateral distribution in all cell types of CD in the outer stripe (D, arrows) and in A-ICs in the inner stripe (E, arrows) and inner medulla/papilla (F, arrows). Principal cell Arf6 staining is reduced in the inner stripe (E) and undetectable in the inner medulla/papilla (F). Arrowheads show apical V-ATPase staining and arrows show basolateral Arf6 staining in A-IC in these regions. In B, nuclei were stained with DAPI. Bars = 10 μm (A, C, F) or 5 μm in (B, D, E).

Three clearly distinct cell types were identified by staining (Fig. 7C1) with antibodies against V-ATPase: 1) A-IC with exclusively apical localization of V-ATPase (asterisk); 2) B-IC with basolateral and/or with basolateral/apical localization (arrowheads); and 3) principal cells, without V-ATPase staining (arrows). Double staining for Arf6 (Fig. 7C2) and V-ATPase (Fig. 7C1) shows that Arf6 was basolaterally located in all of these cell types (Fig. 7C3, merge). Arf6 was not, therefore, always codistributed with the V-ATPase, as was found for Arf1 (see Fig. 4). In the outer stripe of the outer medulla, Arf6 maintained its basolateral localization in all cell types including principal cells and A-IC (Fig. 7D, arrows). However, in the inner stripe (Fig. 7E) and papilla (Fig. 7F) the level of expression of Arf6 was greatly reduced in principal cells of the collecting duct. Double staining for Arf6 and V-ATPase (Figs. 7, DF, arrowheads) demonstrates the basolateral targeting of Arf6 exclusively in A-IC cells of the inner stripe (Fig. 7E) and papillary (Fig. 7F) collecting ducts. The level of expression of Arf6 in thick ascending limbs was below the level of detection by immunofluorescence microscopy.

Expression and quantification of endogenous Arf1 and Arf6 in kidney epithelial cells in situ. Figure 8A shows expression levels of endogenous Arf1 and Arf6 in different parts of the kidney in situ. The level of expression of Arf1 was similar in different parts of kidney, whereas Arf6 was more abundantly expressed in outer medulla, inner medulla, and papilla than in the cortex. Absolute levels of expression of endogenous Arf1 and Arf6 isoforms in kidney were estimated by running highly purified recombinant recArf1 and recArf6 proteins in the same gel (Fig. 8A). Note that the calibration curve for recArf1 ranged from 2.5 to 20 ng/lane, whereas the calibration curve for recArf6 ranged from 0.25 to 2.5 ng/lane. The densitometry analysis and quantification results are presented in Fig. 8B and revealed that the basal level of expression of Arf1 is significantly higher (∼700–1,000 ng/mg of total protein) than Arf6 (∼4–40 ng/mg of total protein).

Fig. 8.

Quantification of expression of endogenous Arf1 and Arf6 in kidney epithelial cells in situ. Expression of endogenous Arf1 and Arf6 in different parts of rat kidney (A) was performed by Western blotting. Recombinant recArf1 and recArf6 were used as standards (A) to quantify the levels of expression of endogenous Arf1 and Arf6. Quantified results are expressed as nanograms of Arf1/Arf6 per milligram of total protein (B). Molecular sizes (kDa) of protein markers are indicated to left in A.


Arf family proteins are crucial regulators of various cellular processes including exo- and endocytotic vesicular trafficking and intracellular signaling. Defining the subcellular localization and targeting of Arf family isoforms is a necessary step to facilitate our understanding of their potential roles in these key cellular events. To date, our current knowledge of the distribution of these proteins has been derived mainly from cell cultures in which recombinant Arf isoforms are overexpressed (16, 27). Depending on the cell type examined and the level of expression achieved, the localization patterns reported vary considerably (13, 16, 27). In contrast, the present study reports the levels of expression, localization, and targeting of endogenous Arf1 and Arf6 small GTPases in native kidney epithelial cells in situ. This was made possible by the use of an antigen unmasking procedure (9), SDS pretreatment of tissue sections, that allowed lower levels of protein to be visualized by indirect immunofluorescence. Moreover, to compare the levels of expression of endogenous Arf1 and Arf6 in different parts of kidney, a quantitative Western blot analysis of Arf small GTPases was applied. This method was previously developed in our laboratory and successfully used to study recruitment of these proteins to endosomal membranes (39).

Arf1 is the best-studied member of the Arf family of small GTPases. In CHO cells, as well as other cultured cells, Arf1 was associated with the Golgi complex (19, 48), where it plays a crucial role in coatomer recruitment (48), phospholipid metabolism (28), and spectrin cytoskeleton rearrangement (23) during vesicular trafficking. A similar targeting of Arf1 and TGN38 to the perinuclear TGN/Golgi complex could also be clearly seen in some kidney epithelial cells in situ. However, in contrast to reports in other cell types in vitro, the localization of Arf1 is not restricted exclusively to the TGN/Golgi in epithelial cell in situ. Indeed, in some kidney cells, Arf1 was not detectable in the TGN/Golgi complex, indicating that other members of the Arf subfamily (i.e., Arf3) could be involved in the regulation of the secretory pathway in these cells. Indeed, the association of both Arf1 and Arf3 with the TGN/Golgi complex and their interaction with GGA proteins (Golgi-localizing, gamma-adaptin ear homology domain, Arf-binding protein) have been reported (2, 53). In the present study, we demonstrated that Arf1 was also localized at the apical plasma membrane of proximal tubules, as well as at the apical pole of distal tubules and connecting segments. Moreover, Arf1 was targeted both to the apical and basolateral poles of collecting duct intercalated cells and closely followed the distribution of the V-ATPase in these cells. Thus endogenous Arf1 can target 1) simultaneously both to the TGN/Golgi complex and to the plasma membrane or 2) exclusively to the plasma membrane of some kidney epithelial cells in situ.

The pattern of cellular distribution and targeting of overexpressed recombinant Arf6 is different from Arf1 and is a controversial issue (13, 16, 22, 55). We previously characterized the expression and localization of endogenous Arf6 in kidney proximal tubule endosomes (35, 39), and a crucial role in regulation of the megalin/cubilin-mediated endocytosis pathway was proposed (36, 37). Although Arf1 and Arf6 have quite distinct patterns of distribution in many kidney epithelial cells, the functional significance of the targeting of two small GTPases (Arf6 and Arf1) to apical endosomes of proximal tubules, and their partial colocalization with megalin as well as with the V-ATPase (36) is unclear. Previously, we demonstrated the V-ATPase-driven and acidification-dependent recruitment of Arf6 but not Arf1 to proximal tubule endosomes (36). However, the GTP/GDP-cycle-dependent recruitment of both Arf6 and Arf1 to the same endosomal membranes clearly indicates a possible functional role of Arf1, in addition to Arf6, in the regulation of this endocytotic pathway. For example, Arf1 (recruited to membranes in a GTP-dependent manner) could influence the earliest steps in clathrin-mediated endocytosis while Arf6 (also recruited in an acidification-dependent manner to endosomes) could be active in steps distal to the initial clathrin-mediated event, after vesicle acidification has occurred. Recently, a novel role of Arf6 in G-protein-coupled receptor desensitization and function has been uncovered (29). In particular, the direct interaction of Arf6 with β-arrestin promotes the downregulation of the β2AR (14) as well as the LH-CGR (43). Furthermore, a direct and specific interaction of Arf1 but not Arf6 with 5-hydroxytryptamine 2A receptor (46) has been demonstrated. Thus the colocalization of Arf1 and Arf6 with megalin, described in this study, may imply a possible role in regulation of this receptor as well. However, megalin is only one of many receptors, transporters, and enzymes that are expressed apically in the proximal tubule and that are regulated by endo- and exocytosis (10). Thus it is possible that Arf1 and Arf6 have a more generic and wide-spread function in the regulation of apical endocytosis in the proximal tubule. Finally, another possibility is that Arf1 may have a role in early endosome trafficking (25), while Arf6 could be involved in the function of recycling endosomes (45). Both populations of early and recycling endosomes are present in apical pole of proximal tubules and cannot be distinguished by immunofluorescence analysis (36).

In contrast to proximal tubules, marked differences in the localization and targeting of Arf1 and Arf6 were seen in cortical collecting duct epithelial cell in situ. Double staining clearly demonstrated a basolateral/vesicular localization of Arf6 and an apical and/or basolateral localization of Arf1. Basolateral Arf6 was found in all cell types (principal and intercalated) of the cortical collecting duct. Interestingly, in collecting ducts of the inner stripe and papilla Arf6 gradually disappeared from the basolateral membrane of principal cells but remained strongly expressed in A-IC cells. As discussed earlier for the proximal tubule, a multitude of trafficking events involving membrane proteins occur in collecting duct epithelial cells, and the membrane localization of Arf small GTPases could reflect a role in some or all of these processes. However, several points can be raised that are related to major functional properties of principal cells and intercalated cells. First, the basolateral localization of Arf6 indicates that this small GTPase may not have an obligatory function in one of the major trafficking processes that occurs in this cell type, namely, aquaporin-2 (AQP2) water channel trafficking. Although basolateral AQP2 insertion can be seen in some collecting duct regions, vasopressin exposure results in a predominant apical insertion of AQP2 in most principal cells (7). The vasopressin receptor (V2R) is, however, an important basolateral G protein-coupled receptor in principal cells. As discussed above, Arf6 regulates β-adrenergic receptor endocytosis via β-arrestin binding (14). The vasopressin receptor also requires β-arrestin for its internalization after ligand binding, and basolateral Arf6 could, therefore, be involved in this process. Second, in all intercalated cells, Arf1 showed a striking colocalization with the V-ATPase, at both the apical and basolateral plasma membranes. This observation suggests that Arf1 may be involved in the targeting and trafficking of the V-ATPase in these cells, which are involved in acid-base transport across the collecting duct epithelium (8). Proton pumps (V-ATPase) are inserted into and removed from the cell surface of intercalated cells in a clathrin- and caveolin-independent manner (3, 8). Recent work has identified some accessory and regulatory proteins that are involved in V-ATPase trafficking, including members of the SNARE family of proteins (4, 33) and soluble adenylyl cyclase (sAC) (44). By analogy with the role of Arf1 in COP coat recruitment (34, 48), it is possible that Arf1 may be involved in the recruitment of specialized coat components to vesicles involved in V-ATPase recycling in intercalated cells, including some peripheral, cytoplasmic subunits of the V-ATPase itself.

It is generally accepted that Arf1 is targeted to the TGN/Golgi and shuttles between cytosol and Golgi membranes during its GDP/GTP cycle (17, 34). Both cytosolic and membrane-bound forms of Arf1 have been found in a variety of cell types (17, 34). In contrast, the pattern of cellular distribution and function of Arf6 during its GTP/GDP cycle is a controversial issue. Until recently, Arf6 had been considered as an “unconventional” member of the Arf family. In CHO cells, it was reported to be an exclusively membrane-bound protein, which, during its GDP/GTP cycle, shuttles between endosomal and plasma membranes without being released into cytosol (13). In contrast, we previously reported (38, 39) that both endogenous Arf6 and Arf1 are cytosolic as well as membrane-bound proteins in kidney proximal tubule epithelial cells. Moreover, our recent results on proximal tubule endosomes demonstrated (36) that both endogenous Arf6 and Arf1 are effectively recruited from cytosol to endosomal target membrane during their GTP/GDP cycle. Thus our data established that, at least in kidney proximal tubules in situ, Arf6 is both a cytosolic and a membrane-bound protein and that it cycles from cytosol to endosomal membranes during the GTP/GDP cycle in a similar manner to Arf1. This is supported by recent data demonstrating that release of membrane-bound Arf6 into the cytosol depends upon the presence of physiological concentrations of magnesium and that, under appropriate conditions, Arf6 cytosol-to-membrane shuttling also occurs during the GTP/GDP cycle in a variety of cultured cells (22).

Regulation of the GTP/GDP cycle and selective recruitment of specific Arf isoforms (Arf1 and Arf6) to selected intracellular target membranes is achieved through the concerted interaction of GEFs and GAPs. Various families of Arf-GEFs have been recently identified, and their specificity for Arf1 and Arf6 was determined. It was demonstrated that ARNO is four times more effective with Arf6 than Arf1 both in vitro (21) and in vivo (50). In in vitro experiments another Arf6-specific GEF, EFA6 (20), was ∼6 times more selective for Arf6 than for Arf1, while Arf-GEP100 (52) was ∼10 times more selective for Arf6 than for Arf1. Thus the selectivity for Arf6-specific GEFs is only between 4- and 10-fold greater than for Arf1 (20, 21, 50, 52), whereas the levels of expression of endogenous Arf6 reported here in situ are 26–180 times lower than Arf1. Under steady-state conditions, endogenous Arf6 (4–40 ng/mg protein) is found in an ocean of Arf1 (700–1,000 ng/mg protein), in close agreement with a quantitative analysis of Arf isoforms expression in yeast (31). These quantitative considerations raise an important question regarding how Arf6 is selectively recruited to its target membranes. One possibility to resolve this apparent paradox could be a highly cooperative function of Arf6-specific GEFs with Arf1-specific GAPs at the target membranes. The expression, targeting, and localization of the increasing number of Arf-GEF(s) and Arf-GAP(s) in kidney epithelial cells are largely unknown and will require further study to answer these questions.

Various Arf family members may be recruited to and function on the same membrane domains (TGN/Golgi, endosome, phagosome, plasma membrane) in many cell types. Indeed, different members of the Arf family (i.e., Arf1 and Arf6) could have a “mosaic” distribution and differential function on the same organelle/membrane (24). Our present finding that both Arf1 and Arf6 are colocalized at the apical pole of proximal tubule cells, but are in opposite domains of collecting duct epithelial cells, also suggests that they can have distinct and/or cooperative functions in trafficking pathways in these distinct cell types. Whether β-COP is involved in the downstream function of Arf1 at the plasma membrane, as it is on Golgi membranes, remains to be determined.

It is very likely that Arf-regulated intracellular events (vesicular trafficking, receptor downregulation, phospholipid metabolism, and cytoskeleton rearrangement) are related to the localization of specific Arf proteins in diverse kidney epithelial cells in situ. Our present data suggest that Arf1 and Arf6 small GTPases have distinct cell- and membrane-specific roles in the variety of epithelial cell types in which they are located, including regulation of the apical megalin/cubulin-mediated endocytosis pathway in proximal tubules, V-ATPase trafficking in intercalated cells, and vasopressin receptor downregulation in principal cells. Although Arf6-and Arf1-specific GEFs and GAPs have been recently identified and characterized in cultured cells, their expression, localization, and functional role in kidney epithelial cells in situ are completely unknown and will be the subject of future investigation.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 7.
  7. 8.
  8. 9.
  9. 10.
  10. 11.
  11. 12.
  12. 13.
  13. 14.
  14. 15.
  15. 16.
  16. 17.
  17. 18.
  18. 19.
  19. 20.
  20. 21.
  21. 22.
  22. 23.
  23. 24.
  24. 25.
  25. 26.
  26. 27.
  27. 28.
  28. 29.
  29. 30.
  30. 31.
  31. 32.
  32. 33.
  33. 34.
  34. 35.
  35. 36.
  36. 37.
  37. 38.
  38. 39.
  39. 40.
  40. 41.
  41. 42.
  42. 43.
  43. 44.
  44. 45.
  45. 46.
  46. 47.
  47. 48.
  48. 49.
  49. 50.
  50. 51.
  51. 52.
  52. 53.
  53. 54.
  54. 55.
View Abstract