Rhbg and Rhcg, the Putative Ammonia Transporters, Are Expressed in the Same Cells in the Distal Nephron

ARTICLES J Am Soc Nephrol 14: 545–554, 2003

RhBG and RhCG, the Putative Ammonia Transporters, Are Expressed in the Same Cells in the Distal Nephron

FABIENNE QUENTIN,* DOMINIQUE ELADARI,*† LYDIE CHEVAL,‡ CLAUDE LOPEZ,§ DOMINIQUE GOOSSENS,§ YVES COLIN,§ JEAN-PIERRE CARTRON,§ MICHEL PAILLARD,*† and RE´ GINE CHAMBREY* *Institut National de la Santeét de la Recherche Me´dicale Unite´356, Institut Fe´de´ratif de Recherche 58, Universite´Pierre et Marie Curie, Paris, France; †Hoˆpital Europeén Georges Pompidou, Assistance Publique- Hoˆpitaux de Paris, Paris, France; ‡Centre National de la Recherche Scientifique FRE 2468, Paris, France; and §Institut National de la Santeét de la Recherche Me´dicale Unite´76, Institut National de la Transfusion Sanguine, Paris, France.

Abstract. Two nonerythroid homologs of the blood group Rh RhBG expression in distal nephron segments within the cortical proteins, RhCG and RhBG, which share homologies with specific labyrinth, medullary rays, and outer and inner medulla. RhBG ammonia transporters in primitive organisms and plants, could expression was restricted to the basolateral membrane of epithelial represent members of a new family of proteins involved in am- cells. The same localization was observed in rat and mouse monia transport in the mammalian kidney. Consistent with this kidney. RT-PCR analysis on microdissected rat nephron segments hypothesis, the expression of RhCG was recently reported at the confirmed that RhBG mRNAs were chiefly expressed in CNT apical pole of all connecting tubule (CNT) cells as well as in and cortical and outer medullary CD. Double immunostaining intercalated cells of collecting duct (CD). To assess the localiza- with RhCG demonstrated that RhBG and RhCG were coex- tion along the nephron of RhBG, polyclonal antibodies against the pressed in the same cells, but with a basolateral and apical local- Rh type B glycoprotein were generated. In immunoblot experi- ization, respectively. In conclusion, RhBG and RhCG are present ments, a specific polypeptide of Mr approximately 50 kD was in a major site of ammonia secretion in the kidney, i.e., the CNT detected in rat kidney cortex and in outer and inner medulla and CD, in agreement with their putative role in ammonium membrane fractions. Immunocytochemical studies revealed transport.

Ammonium is the preferred nitrogen source of bacteria, fungi, the result of NH3 diffusion, indirectly driven by (coupled to) ϩ ϩ and plants; however, in mammals, NH4 is synthesized as an active H secretion (for review, see reference 7). NH3 is end product of nitrogen and has to be eliminated. Renal excre- believed to pass freely through lipid bilayers down its concen- ϩ ϩ tion of NH4 also allows the elimination of the daily acid load, tration gradient generated by the pH gradient, whereas NH4 is provided merely by amino acid metabolism (1). Many studies orders of magnitude less permeant. However, NH may not be ϩ 3 have focused on NH4 transport mechanisms in the mamma- so lipid soluble, as its partition coefficient between chloroform lian kidney but at variance with primitive organisms, in which and water is only 0.04 (8); accordingly, some particular cell ammonium transport is mediated by the Mep/Amt (for Methyl- membranes appear to be virtually impermeable to NH3 (9,10). ammonium Permease/Ammonium Transporters) superfamily In addition, recent pieces of evidence have been provided that members (2–4), no specific ammonium transport protein has ϩ transport across cell membranes of other “freely permeant” been described (1,5). NH4 transport through mammalian cell ϩ volatile solutes such as CO2 could be accelerated by specific membranes is rather believed to occur by substitution of NH4 channel proteins (for review, see reference 11). for Hϩ or Kϩ on diverse cation transporters such as the New insight in ammonium transport mechanisms has been Naϩ/Hϩ exchangers, the Naϩ/Kϩ/2ClϪ cotransporters, and the ϩ ϩ recently provided by studies focusing on the red blood cell Na /K -ATPase (6). It has also been suggested that it could be Rhesus (Rh) proteins, the major targets of hemolytic disease of the newborn. The Rh complex is defined as the association of Received October 1, 2002. Accepted November 8, 2002. membrane polypeptides that includes the non-glycosylated Correspondence to Dr. Re´gine Chambrey, Unite´ INSERM 356, Institut de RhD and RhCE proteins, which carry the Rh antigens, and the Recherche des Cordeliers, 15 rue de l’Ecole de Me´decine, 75270 Paris Cedex RhAG glycoprotein (Rhesus associated glycoprotein), which is 06, France. Phone: 33-1-44-41-37-10; Fax: 33-1-44-41-37-17 E-mail: [email protected] required for cell surface expression of the whole complex 1046-6673/1403-0545 (12,13). Their membrane topology and the sequence homology Journal of the American Society of Nephrology between RhAG and the Mep/Amt protein family, followed by Copyright © 2003 by the American Society of Nephrology the consecutive cloning of two nonerythroid homologs, RhBG DOI: 10.1097/01.ASN.0000050413.43662.55 and RhCG, found to be expressed highly, but not exclusively, 546 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 545–554, 2003

in the kidney, have raised the possibility that RhAG and its measurement. RNAs were extracted according to the technique pre- nonerythroid homologs may represent mammalian equivalents viously described (22) and briefly summarized below. Pools of tu- of Mep/Amt proteins (14–16), an hypothesis strengthened by bules were transferred with 10 ␮l of microdissection solution in 400 ␮ two functional studies strongly suggesting that RhAG and l of denaturing solution (4 M guanidium thiocyanate, 25 mM sodium ␤ RhCG could be involved in ammonium transport (17,18). citrate [pH 7.0], 0.1 M -mercaptoethanol, and 0.5% sarcosyl) containing 20 ␮g of yeast RNA used as carrier. After phenol-chloroform Although the recent study by Westhoff et al. (18) using Xe- extraction and isopropyl alcohol precipitation, the RNA pellets were nopus oocyte expression system concludes that RhAG protein ϩ ϩ ϩ dried under vacuum and dissolved in RNA dilution buffer (10 mM could mediate NH4 /H exchange, whether NH4 is preferen- Tris [pH 7.6], 1 mM ethylenediamine tetra-acetic acid, 2 mM dithio- tially transported by the renal Rh proteins over NH3 remains threitol, and 40 U/ml Rnasin; Promega, Madison, WI) at a dilution undetermined. We recently reported the expression of RhCG in corresponding to 1-mm tubular length or 1 glomerulus/3 ␮l of dilution the apical membrane of connecting tubule and collecting duct buffer. cells in the rat and demonstrated that RhCG was coexpressed with the Hϩ-ATPase in intercalated cells of the collecting duct (19). This localization favors the hypothesis that Rh proteins RT-PCR Analyses of mRNAs Encoding RhBG ϩ The presence of mRNAs encoding RhBG along the nephron was may mediate net NH3 diffusion through membranes, as NH4 secretion by the collecting ducts is believed to result from determined by RT-PCR. Primers were designed from the rat EST coupled Hϩ active secretion by the Hϩ-ATPase, along with (AW920339) showing homologies with the 5' coding region of human RhBG mRNA (NM_020407). NH “passive diffusion” (20). However, definitive demonstra- 3 The sense (5'-TTCACACATTTGGGGCCTACTT-3') and anti- tion of NH3 transport by the Rh proteins remains difficult sense (5'-CTGTCGGAGCGGAGTTGAAA-3') primers correspond to because of the absence of models of Rh protein inactivation bases 205 to 226 and 352 to 375 of the rat EST, respectively. and the difficulties of expressing these proteins at sufficient RT was carried out for 45 min at 45°C in a final volume of 50 ␮l level in heterologous systems of expression. in the presence of mRNA corresponding to 0.5 mm of nephron No study has rendered a complete description of the distri- segment or to 0.5 glomerulus, antisense primers (6.25 pmol) and bution along the nephron of the second renal Rh homolog, Moloney Murine Leukemia virus reverse transcriptase (200 U). RT RhBG. We therefore investigated RhBG expression from glo- was also carried out in the absence of reverse transcriptase as negative meruli down to inner medullary collecting duct cells by spe- control. ␮ cific nephron segments RT-PCR and immunohistochemically PCR was subsequently carried out in a final volume of 100 l after ␣32 ␮ by using polyclonal antibodies raised against peptides of the addition of sense primer (6.25 pmol), ( P)dCTP (5 Ci/sample; 6000 Ci/mmol), and Taq Polymerase (1.25 U). Samples were submit- Rh type B glycoprotein. ted to 31 cycles of three temperature steps: 95°C for 30 s, 56°C for 30 s, and 72°C for 1 min. During the final cycle, the elongation step Materials and Methods lasted 5 min. Animals Aliquots of 16 ␮l of each RT-PCR sample were electrophoresed Experiments were performed on male Sprague-Dawley rats weigh- through 2% agarose slab gels. The gels were fixed in 10% acetic acid, ing 200 to 300 g or on male mice (Iffa-Credo, L’Arbresle, France). All dried under vacuum at 70°C, and submitted to PhosphorImager (Mo- animals had access to a standard laboratory diet and water ad libitum. lecular Dynamics) to determine the intensity of the signal. For each Animals were anesthetized with intraperitoneal injection of sodium animal and each experiment, the signal intensity (arbitrary units) of pentobarbital, 5 mg/100 g body wt. the different samples was expressed as percent of the signal detected in the CCD. Results are presented as mean Ϯ SEM from different animals. Isolation of Nephron Segments and RNA Extraction The left kidney was perfused with microdissection solution (see composition below) containing 0.16% (wt/vol) collagenase (Serva, Antibodies Heidelberg, Germany). The microdissected solution was prepared Two rabbit polyclonal antibodies were raised against the Rh type B from sterile Hank’s solution (Eurobio, France) supplemented with 1 glycoprotein. Two peptides were custom-synthesized (Neosystem, mM acetate, 1 mM lactate, 1 mM glutamine, 1 mM pyruvate, 20 mM Strasbourg, France): Y-KAQRPLRVEEADTQA (human RhBG 4-(2-hydroxyethyl)-1-piperazineethane-sulfonic acid (HEPES), and amino acids 444 to 458) and Y-ETQRPLRGGESDTRA (mouse Rhbg RNAse-free bovine serum albumin (BSA, 1 mg/ml), at pH 7.4. Thin amino acids 441 to 455), corresponding to the 15 terminal residues of pyramids were cut from the kidney, incubated for 20 min at 30°Cin the cytoplasmic tail predicted from the human RhBG and mouse Rhbg 0.12% collagenase solution and then thoroughly rinsed in ice-cold sequences, respectively (15). An initial tyrosine was included to allow solution. coupling of the peptide to KLH. The immunization schedule was as The glomeruli and the successive nephron segments (proximal follows (Bioatlantic, Nantes, France): New Zealand rabbits were convoluted and straight tubules [PCT and PST]; medullary and cor- immunized by six injections of 200 ␮g of immunogen (KLH-peptide), tical thick ascending limb of Henle’s loop [MTAL and CTAL]; distal the first injection with Freund’s complete adjuvant, the following convoluted tubules and connecting tubules [DCT and CNT]; and injections with Freund’s incomplete adjuvant. Sera were screened by cortical and outer medullary collecting ducts [CCD and OMCD]) ELISA on the unconjugated peptide. were isolated under stereomicroscopic observation in microdissection Specificity of the anti-human RhBG was confirmed by indirect solution at 0 to 4°C as described previously (21). immunofluorescence labeling of stable HEK293-RhBG transfectants Pools of 20 to 50 microdissected tubules were transferred on a (Figure 1B), but not of parental HEK293 cells (Figure 1A), with the microscopic slide and photographed for subsequent tubular length rabbit anti-RhBG antibody. J Am Soc Nephrol 14: 545–554, 2003 RhBG Localization in the Kidney 547

reducing buffer (Dako Corp., Copenhagen, Denmark) to block nonspecific staining. The rabbit polyclonal antibody to human RhBG peptide diluted 1:400 in background-reducing buffer was applied for 1 h at room temperature. After three washes, sections were incubated sequentially with a 1:200 dilution (in background-reducing buffer) of goat biotinylated anti-rabbit IgG, (Vector Laboratories, Burlingame, CA), Vectastain Elite ABC reagent (Vector Laboratories, Inc.), each for 30 min at room temperature, with three TBS washes in between. Peroxidase activity was revealed with 3-amino-9-ethylcarbazol (AEC; Dako Corp). Sections were mounted and examined by light photomi- croscopy. RhBG labeling was also performed by using a three-layer immunofluorescence labeling procedure; sections were incubated sequentially with rabbit antibody against the human RhBG, goat-biotinylated anti-rabbit IgG, and Cy2-conjugated streptavidin. After wash- ing, sections were mounted and observed using a LeicaTCS SP confocal laser microscope equipped with an Ar-Kr laser (excitation at 488 nm; detection at 502 to 601 nm). In control experiments, the anti-human RhBG antibody was preadsorbed with the immunizing peptide (100 ␮g/ml), and the resulting medium was used as a negative control. Control using preimmune serum was also negative. For mouse Rhbg, sections were first placed in a plastic tank filled with 1 mM EDTA, pH 8.0, and heated in a microwave oven for two periods of 5 min. (500 W). This step unmasked antigen and allowed immunostaining on paraformaldehyde-fixed paraffin sections, as determined in preliminary experiments. Rhbg labeling was then performed by using a three-layer immunoperoxidase labeling procedure as described above. The rabbit polyclonal anti-Rhbg antibody was applied used at a dilution of 1:200 for1hatroom temperature. Some rat kidney sections were double-immunolabeled using the rabbit polyclonal anti-human RhCG peptide antibody that has been previously characterized (17,19). Because this antibody was also raised in rabbit, an amplification procedure was used to allow staining of sections with two primary antibodies raised in the same species as described previously (23). The first primary antibody, anti-RhBG, was applied at a dilution of 1:5000, a concentration that is too low to be detected using the three-layer immunofluorescence labeling proce- Figure 1. Indirect immunofluorescence staining of parental HEK293 dure. The dilute RhBG antibody was detected using a tyramide cells and HEK293-RhBG transfectants with the rabbit anti-RhBG amplification kit (NEN Life Science Products) with tyramide-CY5 as antibody. 4 ϫ 104 stable HEK293-RhBG transfectants and parental a fluorescence reagent, according to the manufacturer’s instructions. HEK293 cells washed in phosphate-buffered saline (PBS), were se- As heating greatly increased RhCG staining (19), microwave oven– quentially fixed with 4% paraformaldehyde for 20 min, permeablized heated sections (two periods of 5 min at 500 W in 1 mM EDTA with 0.5% Triton X-100 and 1% SDS as described (13,19), and buffer, pH 8.0) were then stained for the RhCG using a three-layer incubated with the rabbit polyclonal anti-Cter human RhBG (1/400) immunofluorescence labeling procedure; sections were incubated se- for1hat22°C. After three washes in PBS containing BSA, the cells quentially with rabbit antibody against the human RhCG peptide at a were stained with Alexa Fluor 488-conjugated anti-rabbit IgG. After dilution of 1:100, goat-biotinylated anti-rabbit IgG (diluted 1:200), washes in PBS, parental HEK293 cells (panel A) and stable HEK293- and Cy2-conjugated streptavidin (Amersham Pharmacia Biotech) di- RhBG transfectants (panel B) were mounted in Prolong (Molecular- luted 1:500. No cross-reactivity between the two sets of reagents was Probes; Interchim, Asnie`res, France), and membrane fluorescence was detectable under these conditions. Sections were observed using a observed with a Nikon TE 300 equipped with a mercury lamp and a Leica TCS SP confocal laser microscope. Cy2 was excited at 488 nm 60XA/1.4 oil objective lens. and detected at 498 to 550 nm, and then, on exactly the same field, Cy5 was excited at 647 nm and detected at 663 to 758 nm. Immunolabeling After clamping the abdominal aorta above and below the renal arteries and cutting open the inferior vena cava, the rat or mouse Crude Membrane Preparation kidneys were perfused with cold 4% paraformaldehyde freshly pre- Membranes were prepared from kidneys from Sprague-Dawley rats pared in Dulbecco modified Eagle’s-Ham’s F-12 medium. After re- as follows. The inner stripe of the outer medulla was excised under a moving the kidneys, coronal sections were cut, post-fixed for4hat dissecting microscope and placed into ice-cold isolation buffer (250 4°C in 4% paraformaldehyde, and embedded in paraffin. Subse- mM sucrose, 20 mM Tris-Hepes, pH 7.4) containing protease inhib- quently, 4-␮m sections of the paraffin block were deparaffinized with itors: 4 ␮g/ml aprotinin, 4 ␮g/ml leupeptin, 1.5 ␮g/ml pepstatin A, toluene, washed in ethanol, and rehydrated in water. and 28 ␮g/ml 4-(2-aminoethyl)-benzenesulfonyl fluoride (AEBSF). For RhBG, rat sections were rinsed in Tris-buffered saline (TBS), Minced tissues were homogenized in a Dounce homogenizer (pestle A pH 7.6, for 10 min and preincubated for 20 min with background ϫ 5) followed by five passes through a Teflon-glass homogenizer 548 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 545–554, 2003

rotating at 1000 rpm. The homogenate was centrifuged at 1000 ϫ g for 10 min, and the supernatant was centrifuged at 100,000 ϫ g for1h at 4°C. The pellet was resuspended in isolation buffer.

SDS-PAGE and Western Blotting Proteins were solubilized in loading buffer, incubated at 20°C for 30 min. Proteins were then separated by 7.5% SDS-polyacrylamide gel electrophoresis and transferred to Hybond ECL nitrocellulose (Amersham Pharmacia Biotech) and analyzed by Western blotting. Nitrocellulose membranes were first incubated in 5% nonfat dry milk in phosphate-buffered saline (PBS), pH 7.4, for1hatroom temperature to block nonspecific binding of antibody, followed by overnight at 4°C with anti-human RhBG peptide antibody diluted 1:10000 in PBS containing 1% nonfat dry milk. After four 5-min washes in PBS containing 0.1% Tween-20, membranes were incubated with 1:3000 dilution of goat anti-rabbit IgG–conjugated to horseradish peroxidase (Bio-Rad, Hercules, CA) in PBS containing 5% nonfat dry milk for 2 h at room temperature. Blots were washed as above, and luminol- enhanced chemiluminescence (NEN Life Science Products) was used to visualize bound antibodies on Polaroid film. Photographs of im- munoblots were numerized with NIH image software. For peptide Figure 2. Expression of RhBG mRNAs along the rat nephron. mRNAs inhibition experiments, the anti-RhBG antibody was preincubated corresponding to 0.5 glomerulus or to 0.5-mm microdissected nephron with the specific immunizing peptide (100 ␮g/ml) before immuno- segment were reverse transcribed, and the cDNA were amplified by PCR blotting renal membrane fractions. using primers specific for RhBG. The DNA fragments were separated on 2% agarose gels and visualized using a PhosphoImager. A representative gel is shown below the histogram. In each experiment, values were Enzymatic Deglycosylation expressed as percent of the CCD value, and data are means Ϯ SEM from Membrane samples were denatured in loading buffer for 30 min at several animals (number in circles). G, glomerulus; PCT and PST, room temperature. Conditions for deglycosylation reaction were as proximal convoluted and straight tubule; MTAL and CTAL, medullary ␮ suggested by the manufacturer. Denatured membranes (35 gof and cortical thick ascending limbs of Henle’s loop; DCT, distal convo- protein) were incubated at 37°C for 1 h with 12 units of N-glycosidase luted tubule; CNT, connecting tubule; CCD and OMCD, cortical and F (Roche Diagnostics) in sodium phosphate, pH 7.2, and nonionic outer medullary collecting duct, respectively. detergent. Sample proteins (20 ␮g) were resolved by SDS-PAGE and transferred to nitrocellulose as described above. outer medulla and the inner medulla. Digestion with N-glyco- Results sidase F followed by Western blot analysis (Figure 4) resulted Expression of RhBG mRNAs along the Rat Nephron in products that had similar sizes, suggesting that renal cortical Figure 2 illustrates one representative RT-PCR experiment RhBG is more heavily glycosylated that renal medulla RhBG. performed to localize RhBG mRNAs in the glomerulus and different nephron segments of rats. Using dilutions of RNAs corresponding to 0.5 glomerulus or 0.5 mm of tubule, the presence of RhBG mRNAs was consistently observed in DCT, CNT, CCD, and OMCD. In contrast, RhBG mRNAs were barely or not detectable in glomerulus (G), PCT, PST, MTAL, and CTAL. In the distal nephron, the expression level of RhBG mRNAs was highest in the CNT, CCD, and OMCD and lowest in DCT.

Immunoblotting of Membrane Fractions Figure 3 shows the regional localization of RhBG in rat kidney as determined by immunoblotting of crude membrane preparations from cortex, outer medulla, and inner medulla. Figure 3. Immunoblot analysis of RhBG peptide in rat renal mem- ␮ Western blot analyses revealed that the anti-RhBG antibody brane preparations. Each lane was loaded with 30 g of membrane detected 50-kD and 70-kD polypeptides (Figure 3, left) in all protein. Left: Anti-human RhBG antibody detects 70-kD and 50-kD polypeptides in all three regions of the kidney namely cortex (C), three regions. However, only the 50-KDa band was undetect- outer medulla (OM), and inner medulla (IM). Right: When the anti- able when the antibody was preincubated with an excess of RhBG antibody was preincubated with an excess of the immunizing immunizing peptide before immunoblotting membrane pro- peptide (100 ␮g/ml) before application to nitrocellulose, only band teins (compare in Figure 3, left with right). In several such detected at 50-kD was competed by the immunizing peptide. * non- experiments, RhBG polypeptide ran slightly slower using specific reactivities not competed by immunizing peptide. Left mar- membranes from the cortex than with membranes from the gin: molecular mass expressed as 10Ϫ3 ϫ Mr. J Am Soc Nephrol 14: 545–554, 2003 RhBG Localization in the Kidney 549

double-stained using a rabbit polyclonal antibody against the human RhCG peptide. Based on double-labeling experiments, we have previously shown using this antibody that RhCG is present at the apical pole of connecting tubule cells as well as intercalated cells in both the connecting tubule and collecting duct. In these double-stained sections (Figure 7), tubular localization pattern of RhBG overlaps with that of the RhCG in all the kidney regions namely cortex (Figure 7, A through C), outer medulla (Figure 7, D through F), and inner medulla (Figure 7, G through I). All cells with apical RhCG staining showed basolateral staining for RhBG. The same pattern of staining was observed in mouse kidney (Figure 8) using a rabbit polyclonal antibody to mouse Rhbg peptide. Rhbg appears to be expressed in all cells within connecting tubules (Figure 8A). In cortical collecting ducts within the medullary ray (Figure 8A), cells exhibited two distinct patterns of labeling; the majority of cells had thin or no basolateral staining, whereas a subpopulation of cells had more Figure 4. Western blot analysis and N-glycosidase F treatment of rat intense basolateral staining. In some cortical distal tubules, one renal RhBG. Digestion with N-glycosidase F of samples from cortex or two cells were heavily stained while the remaining cells (C) and outer medulla (OM). Deglycosylated RhBG was of similar were negative (Figure 8B). These cells could correspond to size in both cortex and outer medulla. ϩ and Ϫ indicate the presence intercalated cells that make their first appearance in late parts and absence of N-glycosidase F, respectively. * nonspecific reactiv- of the distal convoluted tubule. In Figures 8C and 8D, showing ities not competed by immunizing peptide. Left margin: molecular inner medullary and outer medullary collecting ducts, respec- Ϫ3 ϫ mass expressed as 10 Mr. tively, only intercalated cells appeared to be stained.

Discussion Immunolocalization of RhBG in the Rat and Mouse In this study, we determined the regional, cellular, and Kidney subcellular distribution of RhBG protein in rat and mouse The tubular and plasma membrane location of RhBG in rat kidneys. To this end, we have raised two polyclonal antibodies kidney was then investigated by indirect immunoperoxidase on to RhBG and Rhbg proteins by immunizing rabbits with syn- paraformaldehyde-fixed paraffin kidney sections (Figure 5). thetic peptides corresponding to the C-terminal 15 amino acids Staining for RhBG was restricted to some tubular structures of the human or the mouse sequence, respectively. The immu- within the cortex (Figure 5, A and B), outer medulla (Figure nizing peptides were selected to avoid epitopes similar to those 5C), and upper portion of the inner medulla (Figure 5D). In the of the other Rh proteins. The resulting antibodies are likely to cortical labyrinth, anti-RhBG antibody labeled the basolateral be specific for these proteins. Indeed, the immunohistochemi- pole of CNT (Figure 5A). The cortical collecting duct con- cal localization of RhBG in the kidney differed from that seen tained cells with strong basolateral labeling, but in this segment previously for RhCG (19). The sequence corresponding to the the remaining cells stained weakly or not at all (Figure 5B). rat RhBG peptide exhibits significant overlap with the human RhBG staining was restricted to the basolateral pole of a RhBG; as expected, the antibody to human RhBG peptide also subpopulation of cells within the OMCD (Figure 5C). The detects rat RhBG protein. Importantly, both anti-Rh type B number of stained cells fell sharply in the inner medulla (Fig- glycoprotein antibodies show the same pattern of labeling in ure 5D). In the more distal portions of the IMCD, the number the kidney. of stained cells gradually decreased; none was found in deeper Immunoblot analysis of rat renal membrane proteins using regions of the inner medulla (data not shown). There was no rabbit polyclonal antibody to human RhBG peptide showed a evidence of RhBG staining in the proximal tubule or in the thin broad band at 50 kD. Specificity of this 50-kD polypeptide was or the thick limbs of Henle’s loop. Immunofluorescence-based demonstrated by the ability of the immunizing peptide to block labeling gave the same general pattern of staining as immuno- the reaction. On the basis of prior observations (15), the 50-kD peroxidase (Figure 6). Specificity of the labeling was demon- polypeptide is likely to be the glycosylated form of the RhBG strated by the absence of signal when anti-RhBG antibodies protein. Indeed, digestion with N-glycosidase F resulted in a were incubated before application to sections with 100 ␮g/ml product that had an apparent size of 40 kD corresponding to the of the peptide that was used as an immunogen (compare in predicted molecular mass for the unglycosylated form (15). In Figure 6, E with F and G with H) and when the primary samples from cortex, the band consistently ran at slightly antibody was replaced by the preimmune serum (data not higher molecular weight than in samples from outer and inner shown). medulla. Deglycosylated products from all samples had a sim- Because the tubular localization of RhBG was reminiscent ilar apparent molecular mass of 40 kD, suggesting differences of that of its homologue RhCG, rat kidney sections were in posttranslational glycosylation of RhBG. 550 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 545–554, 2003

Figure 5. Immunoperoxidase staining of RhBG in rat kidney sections. (A) View of the cortex, showing basolateral staining of connecting tubules. Proximal tubules and glomerulus showed no detectable staining with the RhBG antibody. (B) Detail of cortex showing strong basolateral staining of a subpopulation of cells within a cortical collecting duct, whereas the remaining cells were not stained. (C) Basolateral RhBG staining appeared in a subpopulation of cells within collecting ducts in the inner stripe of the outer medulla. No labeling was observed in the thin and thick limbs of the loop of Henle. (D) View of the upper part of the inner medulla showing basolateral staining of few cells within inner medullary collecting ducts. G, glomerulus; PT, proximal tubule; CNT, connecting tubule; TAL, thick ascending limb; CD, collecting duct. Scale bar ϭ 50 ␮m for A; 25 ␮m for B, C, and D. J Am Soc Nephrol 14: 545–554, 2003 RhBG Localization in the Kidney 551

Figure 6. Immunofluorescence staining of RhBG in rat kidney sections and absorption test for staining of RhBG. The localization of RhBG in the cortical labyrinth (A), a medullary ray (C), the inner stripe of the outer medulla (E), and the upper portion of the inner medulla (G) is shown. Differential interference contrast images corresponding to views shown in A and C are shown in B and D, respectively. (E through H) controls of specificity of RhBG staining. Under normal incubation conditions (E and G), anti-RhBG stains intercalated cells in the OMCD and the initial IMCD. This staining was abolished by preincubation of the antibody with the immunizing RhBG peptide (F and H). Scale bar ϭ 25 ␮m. 552 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 545–554, 2003

Figure 7. Sections of rat kidney double stained for RhBG (red) and RhCG (green). The localization of RhBG relative to RhCG in the cortex (A through C), the outer stripe of the outer medulla (D through F), and the inner medulla (G through I) is shown. When comparing the merged images, cells that stained for RhCG have strong basolateral RhBG staining in all three regions of the rat kidney. Scale bar ϭ 25 ␮m.

These experiments also demonstrated that RhBG is present or only weak staining. The incidence of labeled cells varied in all three regions of the kidney. This regional localization widely along collecting tubules consistently with the occur- corresponds well with immunohistochemical findings. Immu- rence of intercalated cells in the different segments of the nohistochemistry revealed that RhBG is present throughout collecting duct. Above all, identification of labeled cells was most of the collecting duct system, beginning in the CNT and also based on double-labeling experiments with previously extending to the inner medullary collecting duct. In most characterized anti-RhCG antibodies. Indeed, we recently dem- segments of the collecting duct, a subpopulation of cells that onstrated that RhCG expression in medullary collecting ducts were identified as intercalated cells showed an intense basalot- is restricted to intercalated cells (19). This heterogeneous pat- eral staining, whereas other cells (principal cells) exhibited no tern of labeling was not observed in the connecting tubule in J Am Soc Nephrol 14: 545–554, 2003 RhBG Localization in the Kidney 553

Figure 8. Immunoperoxidase staining of Rhbg in mouse kidney sections. (A) view of the cortex, showing strong basolateral staining of connecting tubules. (B) Detail of cortex showing basolateral staining of distal tubules. Note the presence of one or two heavily stained intercalated cells in some segments that could correspond to later parts of the distal convoluted tubule. (C) View of the upper part of the inner medulla, showing basolateral staining of intercalated cells in medullary collecting ducts. (D) view of the inner stripe of the outer medulla, showing basolateral staining of intercalated cells in medullary collecting ducts. G, glomerulus; PT, proximal tubule; DCT, distal convoluted tubule; CNT, connecting tubule; TAL, thick ascending limb; CCD, cortical collecting duct; OMCD, outer medullary collecting duct; IMCD, inner medullary collecting duct. Scale bar ϭ 50 ␮m for A; 25 ␮m for B; 40 ␮m for C; and 25 ␮m for D. 554 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 545–554, 2003 which both the connecting cells and the intercalated cells had 10. Waisbren SJ, Geibel JP, Modlin IM, Boron WF: Unusual per- strong basolateral RhBG staining. The distribution of RhBG meability properties of gastric gland cells. Nature 368: 332–335, protein is consistent with that of RhBG mRNAs as determined 1994 by RT-PCR on rat microdissected tubules. The observed ex- 11. Cooper GJ, Zhou Y, Bouyer P, Grichtchenko II, Boron WF: pression of RhBG mRNAs in DCT could have been accounted Transport of volatile solutes through AQP1. J Physiol 542: for by the presence of connecting and/or intercalated cells in 17–29, 2002 the rat DCT segment. 12. Cartron JP: RH blood group system and molecular basis of Previous in situ hybridization studies did not provide precise Rh-deficiency. Baillie`re’s Clin Haematol 12: 655–689, 1999 data on the sites of Rhbg expression in mouse kidney (15). To 13. Mouro-Chanteloup I, D’Ambrosio AM, Gane P, Le Van Kim C, Raynal V, Dhermy D, Cartron JP, Colin Y: Cell-surface expres- further define the renal tubular and cellular distribution of sion of RhD blood group polypeptide is posttranscriptionally Rhbg in mouse, we used our rabbit polyclonal antibody against regulated by the RhAG glycoprotein. Blood 100: 1038–1047, the mouse Rhbg peptide. Our data support the presence of 2002 Rhbg in basolateral membrane of collecting duct intercalated 14. Marini AM, Urrestarazu A, Beauwens R, Andre B: The Rh cells and connecting tubule cells in the mouse kidney, which ϩ (rhesus) blood group polypeptides are related to NH4 transport- concurs with our data obtained in the rat. ers. Trends Biochem Sci 22: 460–461, 1997 The major finding of this study is that the renal homologs of 15. Liu Z, Peng J, Mo R, Hui C, Huang CH: Rh type B glycoprotein the Rh-associated glycoprotein RhAG, RhBG, and RhCG pro- is a new member of the Rh superfamily and a putative ammonia teins, are found in intercalated cells and connecting tubule cells transporter in mammals. J Biol Chem 276: 1424–1433, 2001 with RhCG at their apical pole and RhBG at their basolateral 16. Liu Z, Chen Y, Mo R, Hui C, Cheng JF, Mohandas N, Huang pole in rodent kidneys. This conclusion was based on double- CH: Characterization of human RhCG and mouse Rhcg as novel labeling with an anti-RhCG peptide antibody that has been nonerythroid Rh glycoprotein homologues predominantly ex- previously characterized and our recent immunolocalization pressed in kidney and testis. J Biol Chem 275: 25641–25651, study of RhCG in rat kidney. Consistent with their putative 2000 ammonium transport function, these proteins were found in a 17. Marini AM, Matassi G, Raynal V, Andre B, Cartron JP, Cherif- major site of ammonia secretion in the kidney, i.e., the con- Zahar B: The human Rhesus-associated RhAG protein and a necting tubule and collecting duct. kidney homologue promote ammonium transport in yeast. Nat Genet 26: 341–344, 2000 References 18. Westhoff CM, Ferreri-Jacobia M, Mak DO, Foskett JK: Identi- 1. Knepper MA, Packer R, Good DW: Ammonium transport in the fication of the erythrocyte Rh blood group glycoprotein as a kidney. Physiol Rev 69: 179–249, 1989 mammalian ammonium transporter. J Biol Chem 277: 12499– ϩ 2. Kleiner D: The transport of NH3 and NH4 across biological 12502, 2002 membranes. Biochim Biophys Acta 639: 41–52, 1981 19. Eladari D, Cheval L, Quentin F, Bertrand O, Mouro I, Cherif- 3. Howitt SM, Udvardi MK: Structure, function and regulation of Zahar B, Cartron JP, Paillard M, Doucet A, Chambrey R: Ex- ammonium transporters in plants. Biochim Biophys Acta 1465: ϩ pression of RhCG, a New Putative NH3/NH4 Transporter, 152–170, 2000 along the Rat Nephron. J Am Soc Nephrol 13: 1999–2008, 2002 4. Marini AM, Soussi-Boudekou S, Vissers S, Andre B: A family of 20. Knepper MA, Good DW, Burg MB: Mechanism of ammonia ammonium transporters in Saccharomyces cerevisiae. Mol Cell secretion by cortical collecting ducts of rabbits. Am J Physiol Biol 17: 4282–4293, 1997 Renal Fluid Electrolyte Physiol 247: F729–F738, 1984 5. Good DW, Knepper MA: Ammonia transport in the mammalian 21. Imbert M, Chabardes D, Montegut M, Clique A, Morel F: kidney. Am J Physiol Renal Fluid Electrolyte Physiol 248: Adenylate cyclase activity along the rabbit nephron as measured F459–F471, 1985 in single isolated segments. Pflugers Arch 354: 213–228, 1975 6. Good DW: Ammonium transport by the thick ascending limb of Henle’s loop. Annu Rev Physiol 56: 623–647, 1994 22. Elalouf JM, Buhler JM, Tessiot C, Bellanger AC, Dublineau I, de 7. Alpern RJ, Rector FC: Renal acidification mechanisms. In: The Rouffignac C: Predominant expression of beta 1-adrenergic re- Kidney,5th ed., edited by Brenner BM, Philadelphia, WB Saun- ceptor in the thick ascending limb of rat kidney. Absolute mRNA ders Company, 1996, pp 408–471 quantitation by reverse transcription and polymerase chain reac- 8. Bell JM, Field AL: The distribution of ammonia between water tion. J Clin Invest 91: 264–272, 1993 and chloroform. J Am Chem Soc 33: 940–943, 1911 23. Breton S, Wiederhold T, Marshansky V, Nsumu NN, Ramesh V, ϩ 9. Singh SK, Binder HJ, Geibel JP, Boron WF: An apical perme- Brown D: The B1 subunit of the H -ATPase is a PDZ domain- ϩ ability barrier to NH3/NH4 in isolated, perfused colonic crypts. binding protein. Colocalization with NHE-RF in renal B-inter- Proc Natl Acad Sci USA 92: 11573–11577, 1995 calated cells. J Biol Chem 275: 18219–18224, 2000