Isolation and Immunolocalization of a Rat Renal Cortical Membrane Urate Transporter"

Two modalities of urate transport have been reported in rat kidney, a uratdanion exchanger and a potential sensitive, uricase-like uniporter. As an initial attempt to isolate and charac4erize the responsible transport pro- tein(&, rat renal cortical membranes were harvested, solubilized, and subjected to affinity chromatography with urate or xanthine as the affinity ligand. Pig liver peroxisomal uricase was purified with the same system, and the enzymatically active protein was used to generate polyclonal antibodies in rabbit. Silver stain of SDS- polyacrylamide gel electrophoresis gels of the eluted fraction containing the affinity-purified renal mem- brane protei&) demonstrated bands at 25,32,36, and 41 ma. On Western blot, two of these bands (32 and 36 kDa) were immunoreactive to the polyclonal antibody to pig liver uricase. In 6 of 10 studies, the affinity-purified re- nal membrane protein(s) also oxidized urate. Anti-pig liver uricase produced a selective and dose-dependent inhibition of the uricase-like urate uniporter in renal membrane vesicles, but did not affect the uratelanion exchanger or the sodium-dependent glucose transporter. Immunocytochemical studies of rat renal cortex with the same antibody indicated that the immunoreactivity was localized to proximal tubules. These studies demonstrate that the renal cortical plasma membranes contain urate-binding proteins, which have some func- tional and immunological homology to the hepatic peroxisomal core protein, uricase. Within the renal cortex, these proteins are localized to proximal tubules, the site of urate transport. Since the antibody that reacts with the affinity-purified urate-binding proteins on Western blot selectively inhibits urate transport in intact mem- brane vesicles, it is concluded that at least one of the affinity-purified urate-binding proteins is a uricase-like urate transporter. membrane vesicles the renal the transport mechanisms that are responsible for urate reab-sorption and secretion. Two modalities of urate transport have been reported in rat renal proximal

isolate and charac4erize the responsible transport protein(&, rat renal cortical membranes were harvested, solubilized, and subjected to affinity chromatography with urate or xanthine as the affinity ligand. Pig liver peroxisomal uricase was purified with the same system, and the enzymatically active protein was used to generate polyclonal antibodies in rabbit. Silver stain of SDSpolyacrylamide gel electrophoresis gels of the eluted fraction containing the affinity-purified renal membrane protei&) demonstrated bands at 25,32,36, and 41 m a . On Western blot, two of these bands (32 and 36 kDa) were immunoreactive to the polyclonal antibody to pig liver uricase. In 6 of 10 studies, the affinity-purified renal membrane protein(s) also oxidized urate. Anti-pig liver uricase produced a selective and dose-dependent inhibition of the uricase-like urate uniporter in renal membrane vesicles, but did not affect the uratelanion exchanger or the sodium-dependent glucose transporter. Immunocytochemical studies of rat renal cortex with the same antibody indicated that the immunoreactivity was localized to proximal tubules. These studies demonstrate that the renal cortical plasma membranes contain urate-binding proteins, which have some functional and immunological homology to the hepatic peroxisomal core protein, uricase. Within the renal cortex, these proteins are localized to proximal tubules, the site of urate transport. Since the antibody that reacts with the affinity-purified urate-binding proteins on Western blot selectively inhibits urate transport in intact membrane vesicles, it is concluded that at least one of the affinity-purified urate-binding proteins is a uricase-like urate transporter.
It is now generally accepted that urate is bidirectionally transported in the renal proximal tubule (reviewed in Ref. 1). Since little, if any, net urate flux has been detected at nephron sites distal to the pars recta of the proximal tubule (2,3), it has been concluded that the vast majority of urate flux occurs within the renal cortex (1). As a consequence it has been possible to utilize brush border and basolateral membrane vesicles * This work was supported in part by a National Kidney Foundation Fellowship with matching funds from its affiliate, the National Kidney Foundation of New YorWNew Jersey (to B. A. K.), by National Institutes of Health Grants DK08419 (to B. A. K.), EY09244 (to S. K. M.), and DK37315 (to R. G. A,), and by a grant-in-aid from the American Heart Association, New York City Affiliate (to R. G. A,). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. prepared from the renal cortex to physiologically characterize the transport mechanisms that are responsible for urate reabsorption and secretion. Two modalities of urate transport have been reported in rat renal proximal tubule brush-border and basolateral membranes. One, a n anion exchanger, transports urate in exchange for a variety of organic and inorganic anions (4, 5). The second, described by our laboratory, is a voltagesensitive urate uniporter with a number of characteristics similar to those of the hepatic peroxisomal enzyme uricase (6, 7); both the transporter and uricase are Cu2+-dependent, have virtually the same affinity for urate, oxidize urate to allantoin, and are inhibited by oxonate, a specific inhibitor of the oxidative activity of uricase (8). Although uricase that resides in pig liver peroxisomes functions solely as an oxidative enzyme (9-111, this protein was found to be capable of functioning as a saturable urate transporter when inserted in a lipid bilayer (12). Oxonate, which inhibited urate transport in the renal membranes, also inhibited transport in uricase-containing proteoliposomes (12). While these studies in proteoliposomes reinforced the hypothesis that the renal urate uniporter has some homology to hepatic uricase, to date uricase has not been identified in the rat kidney; in rat, uricase has only been localized to liver peroxisomes (9,11,(13)(14)(15)(16)(17).
As a first approach to purifying and characterizing the proteins responsible for urate transport in renal cortical plasma membranes, an affinity chromatography system was used to isolate the urate transporteds). Evidence has been obtained indicating that rat renal cortical plasma membranes contain urate-binding protein(s), which have some homology to hepatic peroxisomal uricase. Functional homology was revealed by uricase-like enzymatic activity of the purified renal membrane protein. Immunologic homology was demonstrated by immunoreactivity of the purified renal membrane protein to an antibody to hepatic uricase. This antibody also inhibited urate transport in intact membrane vesicles, suggesting that the immunoreactive purified proteins represent components of the uricase-like urate transporter. On the basis of immunocytochemical labeling with the antibody to hepatic uricase, this putative urate transporter has been localized in the renal cortex to proximal tubules, the site of urate transport.

MATERIALS AND METHODS
Isolation of Plasma Membranes-Renal cortical membrane vesicles were prepared with minor modifications of methods previously described (6, 18). In brief, in each experiment 45-50 male rats (Charles River Breeding Laboratory, Wilmington, MA) were anesthetized with an intraperitoneal injection of pentobarbital (45 mgkg body weight). The kidneys were harvested, placed in ice-cold 250 m~ sucrose, 10 m~ Tris buffered to pH 7.4 with HCl (ST buffer), decapsulated, and slices of renal cortices were obtained. The cortical slices were weighed (79.2 f 3.7 g, wet weight), placed in fresh ST buffer (2 d g tissue) containing 0.2 m~ phenylmethylsulfonyl fluoride, finely minced with scissors, and ho-

6759
Isolation and Immunolocalization of a Urate Dansporter mogenized. The homogenate was then subjected to differential centrifugation using a Sorvall model RC-5B refrigerated centrifuge with a SS-34 rotor (6,19). The final membrane pellet was suspended in 280 mM sucrose, 8.5 mM Tris acetate buffered to pH 7.4 with NaOH in a volume estimated to yield a protein concentration of 8-10 mg/ml. Brush border and basolateral membranes were subsequently separated and isolated by free flow electrophoresis (model FF5, Garching Instruments, Bender and Hobein, Munich, Germany) using methods identical to those previously described (6,19). The electrophoretic fractions containing brush border and basolateral membranes were identified on the basis of the marker enzymes alkaline phosphatase (determined with a Sigma test kit) and potassium-activated para-nitrophenyl phosphatase (20), respectively. Thereafter, the purified brush border and basolateral membrane fractions were recombined, diluted in a n equal volume of 100 mM mannitol, 20 UM CuCI,, 10 mM HEPES buffered with Tris to pH 7.4, homogenized, and centrifuged at 32,000 x g for 30 min. The resulting pellet was solubilized in a volume of 0.1 M borate, 0.5% CHAPS,' 10% glycerol a t pH 9.0 estimated to yield a detergent:protein ratio of approximately 1:2 w/w. Following 30 min on ice the solubilized membranes were centrifuged at 32,000 x g for 30 min. The supernatant was aspirated and subsequently subjected to affinity chromatography.
Pilot studies demonstrated that identical results were obtained when either solubilized, electrophoretically purified membrane fractions or less pure, combined cortical membrane suspensions (pre-electrophoresis) were applied to the affinity gel. Thus, in subsequent studies the membrane suspensions were not subjected to free-flow electrophoresis. In these studies CuCl, was added to the membrane suspension to yield approximately 70 nm Cu2+/mg protein. After 30 min on ice, the membranes were centrifuged at 32,000 x g for 30 min, the supernatant was discarded, and the pellet was solubilized and subsequently handled in a manner identical to that detailed above. In all studies the protein concentrations of the membrane suspensions and solubilized membranes were determined by the method of Lowry et al. (21) using bovine serum albumin as the standard.
Afinity Chromatography-An affinity gel, in which the ligand urate was linked to epoxy-activated Sepharose 6B (Pharmacia LKB Biotechnology Inc.), was prepared by a modification of the method of Batista-Viera et al. (22). The affinity gel was hydrated, suspended in 50 mM urate in 1% KOH, pH 12.0 (5 g of geVlO ml of potassium urate), rotated in the dark at 37 "C for 24 h to link urate to the spacer arms and then rinsed sequentially with 1 liter of 0.1 M borate, pH 8.0, distilled water, 0.1 M acetate, pH 4.0, and distilled water. To block free spacer arms, the gel was then suspended in 100 ml of 2.0 M glycine, rotated in the dark a t room temperature for 15 h, and suspended in 0.1 M borate pH 9.0 (23). The gel was stored at 4 "C until use. In additional studies xanthine-Sepharose (Sigma) was substituted for the urate-Sepharose. All other aspects of the affinity chromatography were identical to those detailed below.
Water-jacketed columns that were maintained at 4 "C were packed with gel and rinsed with 0.1 M borate, pH 9.0. Solubilized membranes (60-117 mg), which were also maintained at 4 "C, were then applied to the gel. Unbound proteins were washed out of the gel with the same borate buffer, which contained CHAPS and glycerol at 1/10,000 of the concentrations used to solubilize the membranes. Subsequently, bound protein was specifically eluted with 10 mM hypoxanthine (a uricase inhibitor) in the same buffer at 22 "C; 5-ml eluate fractions were collected on ice.
Protein Electrophoresis-Aliquots of protein samples that were applied to and eluted from the affinity gel were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on 8-25% gradient gels using the PhastSystem (Pharmacia) (24).
All protein samples were reduced with P-mercaptoethanol. Gels were subsequently silver stained using a Pharmacia kit. The molecular weights of the proteins detected on SDS-PAGE were estimated using an Ultrascan XL laser densitometer (PharmaciaLKB). Rabbit Anti-pig Liver Uricase-Commercial pig liver uricase (Sigma) was affinity-purified with the urate affinity gel described above. The specifically eluted protein, which was shown to have enzymatic activity, was used to generate polyclonal antibodies in rabbits (Pocono Rabbit Farm & Laboratory, Inc., Canadensis, PA). Assays for antibody specificity indicated that the antibody was monospecific to the antigen and that the purified antigen was a single protein; Ouchterlony double diffusion analysis revealed a single precipitin line, and immunoelectro-The abbreviations used are: CHAPS, 3-[(3-cholamidopropyl)dimethylammoniol-1-propanesulfonic acid; PAGE, polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; DIDS, 4,4'-diisothiocyanostilbene-2,2'-disulfonic acid.
phoresis revealed a single band with the purified protein. Enzymelinked immunosorbent assay revealed antibody titers of at least 1/12,000 to the monomeric form of porcine uricase. The IgG fraction of A antibody purification kit (RepliGen, Cambridge, MA). The IgG frac-the polyclonal antibody to porcine uricase was purified using a protein tion was then concentrated with a Centricon 30 concentrator (Amicon, Beverly, M A ) to 10 mg/ml. Additionally, aliquots of preimmune and immune sera were purified against the 33-kDa monomer of porcine uricase as previously described (25,261. Briefly, reduced porcine uricase was electrophoresed on a preparative 10% SDS-PAGE gel. The gel was stained with Rapid Reversible Stain (Diversified Biotech, Newton Centre, MA). The 33-kDa monomer was cut out of the gel and subsequently electroeluted from the gel slice with the Little Blue Tank (Isco, Lincoln, NE). The electroeluted monomer was electrophoresed on a second 10% SDS-PAGE gel and transferred to nitrocellulose filters using the Multiphor I1 Nova Blot semidry transfer unit (Pharmacia). Undiluted preimmune or immune serum was incubated with the transferred antigen and subsequently eluted with 1 M potassium thiocyanate (26). The eluted antibody was desalted on a Sephadex G25 column equilibrated with phosphate-buffered saline (PBS) and stored with 2% albumin. buffer (0.125 M Tris-HC1, 4% SDS, 20% glycerol, 10% mercaptoethanol, Western Blot Analysis-Samples were suspended in Laemmli sample pH 6.81, resolved on 10% SDS-polyacrylamide gels (27) using a Mighty Small I1 gel electrophoresis unit (Hoefer, San Francisco, CA) and electrophoretically transferred to nitrocellulose filters. The blots were blocked with 2.0% milk in PBS, pH 7.4, for 1 h a t room temperature, incubated overnight at 4 "C in either a 1/250 dilution of serum, a 11500 dilution of the IgG fraction of anti-porcine uricase, or undiluted antigenpurified anti-uricase antibody in the milkPBS solution. The blots were washed at room temperature with milWBS followed by PBS, then incubated a t room temperature with a 1/250 dilution of peroxidase labeled goat anti-rabbit IgG in the PBS solution. After washing with PBS, the blots were developed with the 4 CN Peroxidase substrate system kit (Kirkegaard & Perry, Gaithersburg, MD). In some studies, solubilized membrane samples were examined with and without deglycosylation using a De-N-Glycosylation kit (Oxford GlycoSystems, Rosedale, N Y ) prior to Western blot analysis. Urate Oxidation-Since hypoxanthine is a uricase inhibitor, hypoxanthine was removed from the eluate fractions by desalting with detergent and glycerol-free 0.1 M borate buffer on Sephadex G25 columns (Pharmacia). Solubilized membranes were similarly treated. [2-l4C1-Urate (49.1 mCi/mM, Amersham Carp.) was added to a final concentration of 2 p~ to aliquots of the solubilized membranes and hypoxanthinefree eluate fractions and incubated for 60 min at room temperature. The fraction of [2-l4C1urate that was oxidized to [2-14Clallantoin was determined by separating urate and allantoin with a previously described column chromatographic technique (28).
'Dunsport Studies-Membrane vesicles were isolated and treated with CuCl, as detailed above. Paired vesicles were preincubated for 30 min with non-immune IgG or the IgG fraction of anti-porcine uricase (0.01-1 pg of IgG/pg of vesicle protein). Thereafter, the paired vesicles were incubated for 5 s in medium containing 50 PM [2-l4C1urate, 100 mM mannitol, 100 mM NaC1, 10 mM Tris-Hepes, at pH 7.5. In a subset of experiments the uptake of ~-[6-~Hlglucose (6.4 pCiinmo1; DuPont NEN) was examined; paired vesicles that were preincubated with non-immune or immune IgG were then incubated 10 and 30 s in media containing 50 p~ D-glucose, 100 mM mannitol, 100 mM NaC1, 10 mM Tris-Hepes a t pH 7.5. Uptake was terminated by the addition of 4 ml of ice-cold stop solution (154 mM NaCl, 10 m~ Tris-Hepes, pH 7.5) followed by filtration through 0.65-pm Millipore filters. The tubes and filters were subsequently rinsed with a n additional 12 ml of the same solution. In additional studies, brush-border membrane vesicles were prepared by magnesium aggregation (4). Urate uptake was examined under conditions identical to those previously described (4); paired vesicles were prepared in 201 mM mannitol, 49 mM K' , 80 mM HEPES, 10 m~ MgSO,, pH 7.5, and then preincubated 30 min with non-immune or immune

I&
(1 pg of IgG/pg of vesicle protein) that was diluted in the same buffer.  Immunocytochemistry-Rat kidneys were perfused in situ with normal saline followed by 4% paraformaldehyde. Slices (2-3 mm thick) of kidney were then dehydrated, cleared, and embedded in paraffin. Sections (3 pm) were cut and placed on polylysine-coated slides. f i r deparaffinization, immunoperoxidase staining was performed using a Super Sensitive Multilink Immunodetection System (BioGenex, San Ramon, CA) in which the antibody is localized via a biotin-streptavidin horseradish peroxidase complex using diaminobenzidine as substrate. Thereafter the tissue was counterstained with Hams hematoxylin. The primary antibody was a l/6400 dilution of the IgG fraction of antiporcine uricase or undiluted antigen-purified antibody. A 116400 dilution of non-immune IgG or undiluted antigen-purified preimmune serum served as the controls. The tissue was examined with a Zeiss Axiomat using light and differential interference contrast microscopy.

RESULTS
Purification of Urate-binding Proteins-The extent of purification of renal urate-binding proteins that was achieved with the above described affinity chromatography system was evaluated by SDS-PAGE (Fig. 1). As anticipated, the number of protein bands in both the sample of solubilized membranes that was applied to the affinity gel and in the fraction containing unbound protein were too numerous to count. In contrast, silver stain of the protein fraction that was specifically eluted with hypoxanthine revealed two to four protein bands. In virtually all studies bands were seen at 25 and 36 kDa; less frequently bands were detected at 32 and 41 kDa. Protein bands of comparable size were obtained when xanthine was substituted for urate as the affinity ligand. SDS-PAGE of hepatic uricase that was purified with the same affinity system yielded a protein band that approximated 33 kDa. Despite the difference in size in the purified renal proteins and hepatic uricase, these proteins bound to the same ligands and eluted with the same substrate. Additional studies were therefore performed to assess whether these proteins displayed any immunologic andlor functional homology.
Immunoreactivity of Urate-binding Proteins-Western blot analysis with rabbit anti-porcine hepatic uricase revealed that the solubilized renal membranes and the affinity-purified protein were highly reactive to the antibody (Fig. 2), but nonreactive to preimmune serum and non-immune rabbit IgG (not depicted). Identical immunoreactive bands were detected independent of whether immune serum, the purified IgG fraction of immune serum, or a n antigen-purified antibody was used. Although innumerable protein bands were evident on the silver stain of solubilized renal membranes (Fig. l), only three bands at 27, 32, and 36 kDa were immunoreactive (the 36-kDa band may represent a doublet) (Fig. 2). The sizes of these immunoreactive bands were indistinguishable before and after treatment with peptide-N-glycosidase:'F, suggesting that the differ- 36 kDa) that were present in solubilized membranes were detected in the purified protein fraction (Fig. 2); the 27-kDa immunoreactive band that was seen in solubilized membranes was not detected in the purified protein fraction. Since the smallest of the immunoreactive membrane proteins was not detected in the purified protein fraction, it may have been degraded or denatured after solubilization, obviating purification via the affinity gels employed.
Functional Assessment of Urate-binding Proteins-The uricase-like activity of the purified proteins and solubilized membranes was assessed by measuring the ability of these proteins to oxidize urate. In 6 of 10 experiments, the fraction containing the purified protein oxidized 8.0 2 2.6% of 2 [2-14C]urate; activity was not detected in the remaining 4 studies. In the same six experiments in which the purified proteins oxidized urate, the solubilized membranes oxidized a similar percent (9.25 2 2.9%). It is of note that the assays with solubilized membranes contained 0.74 0.18 mg of protein, yielding a specific activity of 161 * 64 pmol of urate oxidizedlmg of proteinh, whereas assays with the purified protein contained an amount of protein that was below the lower limit (1.0 pg) of detection with the Pierce Micro BSA protein assay. Although the specific activity could not be determined, the finding that a comparable amount of urate was oxidized by such disparate amounts of purified protein and solubilized membrane protein indicates that the specific activity of the purified urate-binding protein was markedly enriched (at least 740-fold) relative to that in the membrane from which it derived. Since uricase functions as an enzyme that oxidizes urate, the finding that the purified renal urate-binding protein(s) also oxidized urate indicates that this protein has some functional homology to hepatic peroxisomal uricase.
Since a very limited amount of purified urate-binding protein was recovered, it was not possible to assess transport function by the technique of reconstituting the purified proteins in liposomes. The strategy that was used to assess the role of these proteins in urate transport was based on the assumption that transport might be inhibited by a n interaction between antiuricase and proteins in the intact plasma membrane if one or more of the immunoreactive membrane protein(s) (Fig. 2) is involved in urate transport. However, two modalities of urate transport have been detected in rat renal brush border membranes; the uricase-like uniporter has been observed in membranes exposed to trace amounts of Cu2+ (6,7), while the anion exchanger has been evident in membranes prepared with Mg2+ (4, 5).
To assess the effect of anti-uricase on each mode of urate transport, vesicles were prepared and urate uptake was examined under conditions in which either the uniporter or ex-Urate Dansporter changer would be operative. As demonstrated in Fig. 3, in three paired studies in Cu2+-exposed vesicles the purified IgG fraction of rabbit anti-porcine uricase produced a dose-dependent inhibition of urate uptake in renal cortical membrane vesicles. In contrast, in these same vesicles Na+-dependent glucose uptake (at 10 and 30 s) was unaffected by immune IgG (not depicted). Non-immune rabbit IgG did not alter urate (Fig. 3) or glucose uptake. In vesicles prepared with Mg2+, urate uptake was 5-fold greater at 10 and 30 s in the presence of a n outward OHgradient (pHin 7.5 > pH,,, 6.0) than in its absence (pHin 7.5 = pHout 7.5); anti-uricase did not inhibit basal urate uptake or the enhanced uptake that occurred on the urate/anion exchanger. These findings indicate that the effect of anti-pig liver uricase on urate transport is quite specific and suggest that at least one or more of the affinity-purified immunoreactive uratebinding proteins is a component of the uricase-like urate uniporter, but not the uratelanion exchanger.
Immunocytochemistry-Immunoperoxidase staining was absent from cortical sections incubated with non-immune IgG (Fig. 4A). In the presence of either the IgG fraction of antiuricase or antigen-purified antibody, immunoperoxidase labeling was confined to proximal tubules within the renal cortex (Fig. 4, B-F). Blood vessels, glomeruli, and cortical segments of the distal nephron were non-reactive to anti-uricase. Three patterns of immunolabeling of proximal tubules were observed.
In the very early portions of the proximal tubule (SI), immunolabeling appeared to be localized to the brush borders (Fig. 4,  B and C ) . In somewhat more distal segments of the proximal tubule, immunoreactivity was primarily observed in subapical regions (Fig. 4, B and D ) . Within proximal tubule segments near the corticomedullary junction (S3), the cytoplasm appeared to be diffusely stained, but the brush borders were unstained (Fig. 4, E and F ) . DISCUSSION The present studies have demonstrated that urate-binding proteins that were purified from brush border and basolateral membrane vesicles of rat renal cortex display some homology to the hepatic peroxisomal protein uricase. Immunologic homology was revealed by the immunoreactivity of two of the purified proteins with anti-porcine uricase on Western blot (Fig. 2). Functional homology was evidenced by uricase-like enzymatic activity. In addition to demonstrating shared properties of the purified renal proteins and hepatic uricase, the present studies also suggest that one or more of the purified renal cortical membrane urate-binding proteins is a component of a urate transporter. This conclusion is based on the finding that antiporcine uricase not only reacts with immobilized purified protein on Western blot (Fig. 2), it also reacts with intact membrane proteins to act as a potent and specific inhibitor of urate transport under conditions in which the uricase-like trans-porter is functional (Fig. 3). Finally, the immunoreactive uratebinding protein(s), the putative urate uniporter, has been immunolocalized within the renal cortex to proximal convoluted and straight tubules (Fig. 4, B-F).
In addition to the two immunoreactive proteins that were purified from the renal membranes, two other proteins (25 and 41 kDa) were purified that were non-reactive to anti-uricase. As all of the assays were performed under reduced conditions, it is possible that the four protein bands represent subunits of a single urate-binding protein in which some subunits contain, but others lack, epitopes recognized by anti-uricase. Alternatively, as two distinct modalities of urate transport have been described in rat kidney, a n anion exchanger (4, 5) and a potential sensitive uricase-like uniporter ( 6 , 7), the non-reactive and immunoreactive proteins may represent different urate-binding proteins with each responsible for one mode of urate transport. Based on the immunologic and functional homology between the purified renal proteins and porcine liver uricase, as well as the fact that anti-uricase inhibited urate uptake in intact membranes under conditions in which only the urate uniporter is detected (Fig. 31, it is concluded that the purified immunoreactive proteins are components of the uricase-like transporter. The failure for anti-uricase to inhibit transport under conditions in which only the urate/anion exchanger is observed implies that uratelanion exchange occurs on a protein that is different from the uricase-like urate uniporter. If such is the case, then the purified urate-binding proteins that do not react to anti-uricase are likely to be components of the uratel anion exchanger. However, the present studies do not totally exclude the possibility that the two modes of transport occur on a single protein whose structural configuration is altered when experimental conditions are varied; a change in the configuration of a single urate transport protein that interferes with binding of the antibody could obviate an effect of the antibody on urate transport via the exchanger. A definitive conclusion regarding this issue must await comparison of the protein sequences of the exchanger and uniporter. It is of interest that prior studies failed to detect immunoreactivity to an antibody to uricase in the rat kidney (13,15). The marked immunoreactivity that was demonstrated in the current studies may, in part, be consequent to the fortuitous generation of a high titer, high affinity antibody. It may also be consequent to the fact that the protein that was used to generate antibody was not denatured as demonstrated by its ability to oxidize urate. However, the relevance of using an intact protein is uncertain since the biological activity of the antigens used to generate antibodies in previous studies (13,15) was not commented upon. Perhaps more significantly, polyclonal or monoclonal antibodies were raised to rat uricase in prior studies (13, 15) whereas the present studies employed a polyclonal antibody that was raised to porcine uricase. Porcine uricase was selected for two reasons. First, studies in proteoliposomes demonstrated that porcine uricase, like the renal membrane protein, is capable of transporting urate when inserted into a lipid bilayer (12). Second, when antibodies to uricase have been raised in rabbit, porcine uricase has been shown to be more antigenic than dog, cow, horse, house musk shrew, and guinea pig uricase (29). Thus, in choosing the source of antigen the possibility was considered that porcine protein may also be more antigenic than the rat preparation. Regardless of the reason(s) for the difference in immunoreactivity of the antibody used in this and prior studies, the finding that anti-porcine uricase produced a profound inhibition of urate uptake in rat renal cortical membrane vesicles (Fig. 3) provides strong support for the use of this antibody as a marker of a urate transport protein.
Since the antibody that was employed was a selective and potent inhibitor of urate transport, immunocytochemistry was used to localize the nephron sites of the urate transporter within the renal cortex. In previous studies in rat, it has been suggested that the majority of urate transport occurs within the proximal tubule (1-3). The finding that immunoperoxidase labeling was confined to proximal tubules (Fig. 4, B-F) is thus consistent with the physiologic data. At the cellular level, the demonstration of immunolabeling of the brush border (Fig. 4, B and C) is in accord with the fact that urate is transported across this membrane. Since this luminal membrane transporter must be delivered to the brush border from intracellular organelles, the subapical staining pattern (Fig. 4, B and D ) may represent protein in vesicles involved in membrane trafficking or protein free within the cytoplasm. However, the precise intracellular localization that is represented by subapical labeling cannot be determined at the light microscopic level. Similarly, light microscopy is inadequate to determine the intracellular constituents that are labeled in the diffusely immunostained cells in S3 segments of the proximal tubule (Fig.   4, E and F). Since urate is also transported across the basolat-era1 membrane, it would not be surprising if some of the diffuse staining was localized within this membrane, but the present study does not reveal enhanced staining of the basolateral membrane relative to the remainder of the cell. The failure to detect specific labeling of the basolateral membrane may indicate that reactive antigenic sites in the basolateral membrane to the uricase-like urate transporter are masked or urate is transported across this membrane by a different protein. Resolution of these issues will require further analysis with immunoelectron microscopy.
Although the current studies indicate that the urate uniporter has homology to hepatic peroxisomal uricase, it seems clear that the transporter and uricase are not identical proteins. First, pig hepatic uricase and the immunoreactive renal membrane urate-binding proteins have slightly different molecular weights. Second, the uricase-like enzyme activity of the purified membrane proteins appears to be quite weak relative to that of hepatic uricase. Finally, as Northern blot analysis of rat kidney RNA probed with the cDNA for rat hepatic uricase failed to detect transcripts (15,301, it can be concluded that the mRNA for uricase, per se, is absent or present in very low copy number. However, some epitopes must be conserved between hepatic uricase and the renal membrane transport protein; anti-porcine uricase not only recognizes urate-binding proteins purified from renal membranes, but specifically inhibits the urate uniporter without affecting other transporters (ie. Na+-dependent glucose transporter and the uratelanion exchanger). A determination of the actual extent of homology between these proteins must await cloning of the renal urate uniporter and subsequent comparison of the nucleotide sequences of the renal and hepatic proteins. The availability of a high titer polyclonal antibody to hepatic uricase may facilitate the cloning of the renal transporter.