We have identified a urea transporter from the mucosa of the human colon that has characteristics consistent with a Kidd antigen/UT-B urea transporter. This intestinal urea transporter encodes a 389-amino acid peptide with a sequence identical to that previously reported for the UT-B urea transporter in erythrocytes. Expression of a UT-B 2-kb mRNA transcript and of ∼50- and ∼98-kDa UT-B proteins is detected in human colonic mucosa by Northern and Western blot analysis. The UT-B protein is localized in the cell membrane and cytoplasm of the superficial intestinal epithelium and in the epithelial cells in the crypts. A 2-kb UT-B mRNA transcript and the UT-B protein were also identified in the intestinal cell line Caco-2. The transepithelial flux of 14C urea was examined in Caco-2 cells growing on porous membrane support and was significantly inhibited by phloretin, 1,3-dimethylurea, and thiourea, suggesting that the transfer of urea across the Caco-2 monolayer could be mediated, at least in part, by the UT-B urea transporter. We conclude that the Kidd antigen/UT-B urea transporter is physiologically expressed in the human colon epithelium, where it could participate in the transport of urea across the colon mucosa.
- urea transport
urea is synthesized by the mammalian liver as a result of protein catabolism and is excreted by the kidney. Urea is present in high concentration in the kidney inner medulla and is moved across epithelial, endothelial, and red blood cells by urea transporter proteins. The UT-A urea transporter, which includes several isoforms (5, 9, 31), is encoded by the Slc14A2 gene (16). The UT-B urea transporter (20) is encoded by the Slc14A1 gene (11) and carries the Kidd blood antigen, which is expressed in erythrocytes (19). Expression of UT-A and UT-B urea transporters has also been documented in other tissues, such as testis and brain (3, 4). Liver expression of the UT-A transporter protein has been described (10). In the liver, however, aquaglyceroporin AQP9 is likely to mediate, at least in part, the transport of urea from the hepatocytes (2). Urea present in the blood can pass from the circulation of the host into the digestive tract, possibly by carrier-mediated transport, and is hydrolyzed by the intestinal bacterial flora into ammonia and carbon dioxide. A considerable fraction of the nitrogen derived from this process is recycled in metabolic conversion to nitrogen-containing molecules, which can be reabsorbed and used by the host. This may be a mechanism for nitrogen salvage in cases of inadequate protein intake or malnutrition (13, 15). Studies with [15N]urea directly placed in the lumen of the human colon have shown that urea can be absorbed intact from the colon (14), and intestinal permeability to urea has been described in studies using perfused rat intestinal segments (7). However, a specific urea transporter in the intestine has never been cloned, although evidence for expression of UT-A and UT-B transporters in the gut has been presented (18, 22, 27, 31).
In this study, we provide the first description of a cloned intestinal urea transporter in the human colon and its characterization as a UT-B urea transporter.
MATERIALS AND METHODS
Northern blot analysis.
Anonymous samples of normal human tissue were obtained from the Emory University Tissue Bank in accordance with institutionally and federally approved guidelines. Northern hybridization was performed with the use of a full-size murine UT-B cDNA probe 86% identical to the human cDNA sequence (kindly provided by Dr. Baoxue Yang, Cardiovascular Research Institute, Departments of Medicine and Physiology, University of California, San Francisco, CA) as previously described (21).
Western blot analysis.
Proteins (10 μg/lane) were separated on 10% SDS-polyacrylamide gels and then transferred to a polyvinylidene difluoride membrane (Gelman Scientific, Ann Arbor, MI). Blots were blocked with 5% nonfat dry milk in Tris-buffered saline (TBS) at room temperature for 30 min and then incubated with an antibody to the COOH terminus of the rat UT-B (27) (kindly provided by Dr. Jeff Sands, Division of Nephrology, Department of Medicine, Emory University School of Medicine, Atlanta, GA), an antibody that reacts with rat, mouse, and human UT-B, at a dilution of 1:1,000 overnight at 4°C. After three washes in TBS with 0.5% Tween (TBS-Tween), blots were incubated for 2 h with horseradish peroxidase-linked goat anti-rabbit IgG at a dilution of 1:5,000 (Amersham, Arlington Heights, IL) at room temperature. After two washes with TBS-Tween, the bound secondary antibody was visualized by chemiluminescence (ECL kit; Amersham).
Red blood cell membranes were prepared from human blood collected in heparinized tubes and centrifuged at 2,000 rpm for 10 min, the serum and buffy coat were removed, and cells were washed three times in 10 mM MgCl2. Cells were then washed with 5 mM Na2HPO4, centrifuged three times at 16,000 rpm for 10 min each, and solubilized in isolation buffer (10 mM triethanolamine, 250 mM sucrose, 1 μg/ml leupeptin, and 0.1 mg/ml phenylmethylsulfonyl fluoride, pH 7.6).
Glycosylation of the UT-B protein in the human colon, the rat outer medulla, and human red blood cell membranes was examined with N-glycosidase F, followed by separation on a 12.5% SDS-polyacrylamide gel, with the use of reagents and protocol provided in the PNGase F kit (New England BioLabs, Beverly, MA).
For immunohistochemistry, 5-μm sections of formalin-fixed, paraffin-embedded tissue were stained with the UT-B antibody at 1:2,000 dilution, followed by goat, horseradish peroxidase-conjugated anti-rabbit secondary antibody (Dako, Carpinteria, CA), as previously described (27). The primary antibody was omitted for negative control. For immunofluorescence, Caco-2 cells were grown onto eight-chamber sterile glass slides (Lab-Tek; Nalge Nunc, Naperville, IL), fixed with 3.7% formaldehyde, permeabilized with 3% Triton X-100, and stained at 37°C with the UT-B antibody (1:100), followed by FITC-conjugated goat anti-rabbit antibody at 1:40 dilution (Sigma, St. Louis, MO). Cells were also stained with a previously characterized primary rabbit polyclonal antibody to the COOH terminus of the rat UT-A1 urea transporter (an antibody that reacts with rat, mouse, and human UT-A1; generous gift of Dr. Jeff Sands) to test specificity (1:100 dilution). Madin-Darby canine kidney (MDCK) cells transiently transfected with the coding region of rat UT-A4, which has the same COOH terminus of UT-A1 and UT-A2 (9), subcloned into the pCDNA3 vector, were used as positive control for the UT-A antibody. The culture conditions and transfection procedures for the MDCK cells were as previously described (17).
Measurement of urea flux in Caco-2 cells.
Caco-2 cells (kindly provided by Dr. Asma Nusrat, Department of Pathology and Lab Medicine, Emory University School of Medicine, Atlanta, GA) were grown on plastic dishes or permeable membrane inserts (Corning, Marietta, GA) in 80% Dulbecco's modified Eagle's medium (Mediatech, Herndon, VA) with 2 mM l-glutamine, 0.1 mM nonessential amino acids, 15 mM HEPES, pH 7.4, 10% FBS, 100 IU/ml penicillin, 100 IU/ml streptomycin. Cells were grown to confluence on 1-cm2 permeable membrane support (Cluster 12 Transwell; Corning). Testing of transepithelial urea flux was usually conducted when cells reached transepithelial resistance close to 280 Ω·cm2 (277 ± 16 Ω·cm2) (8). Assay of urea transport was conducted at 37°C in basic DMEM with 1 mM cold urea added. At the start of the experiment, the normal growth medium was replaced by medium containing 2 μl/ml [14C]urea (54 mCi/mmol; Amersham, Piscataway, NJ), with 0.5 ml on the apical side for apical-to basal flux direction and 1.5 ml on the basal side for basal-to-apical flux direction. Assay medium without tracer was placed in the opposite side of each membrane. Urea transfer in both directions was tested with or without phloretin. Samples of the medium in the basal (100 μl) or apical (33 μl) side of each well were collected after 1, 2, 5, 15, 30, 60 min, and, after addition of 3 ml scintillation fluid (National Diagnostics, Atlanta, GA), radioactivity was measured in a scintillation counter (6000 LS; Beckman, Fullerton, CA). The appearance of [14C]urea in either side of the well, calculated as nanomoles per square centimeter per minute (mean ± SD in 3 or 4 wells), was used to determine the transport of urea across the cell layer. Urea transport was measured in the presence of urea transporter inhibitors phloretin (0.6 mM), 1,3-dimethylurea (150 mM), and thiourea (150 mM) (Sigma), and the effect of individual inhibitors compared with control was analyzed by performing Student's t-test, with P < 0.05 indicative of statistically significant difference. Experiments were repeated two or three times.
RT-PCR and 3′ rapid amplification of cDNA ends.
Total RNA was purified with TriPure isolation reagent (Roche, Indianapolis, IN), and synthesis of cDNA was performed with 1 μg of total RNA as previously described (1). PCR amplification of the first-strand cDNA was performed with sense primer 5′-AGATAGCCATGGAGGACA-3′ and antisense primer 5′-GTTCTCACAAAGGGCTTT-3′, designed to amplify the coding sequence of UT-B (based on GenBank accession no. Y19039).
Antisense primers 5′-ATACTCGTGAAAAACAGCAG-3′, 5′-ACATTCCAGTCTTAGTGCCA-3′, 5′-TTAGTTGGTTTATGGGGTTT-3′, 5′-TGCCTAAGCTGAGTATTTCA-3′, 5′-CAATCTGACCTTTGCCTTCA-3′, 5′-AAAGTGCTGGGATTACAGGC-3′, and 5′-GATTTCACTCTTGTCGCCCA-3′, designed from the 3′ end sequence of the UT-B/Kidd mRNA sequence (GenBank accession no. NM_015865), in combination with sense primer 5′-GTGCATTCCAGGTGATTTAT-3′, were used to amplify the 3′ UTR sequence of the intestinal UT-B transcript and, after sequence analysis of the PCR products, to map the 3′ end of the transcript.
The PCR products were gel purified and sequenced with an automatic DNA sequencing system (3100 Genetic Analyzer; Applied Biosystems, Foster City, CA). Sequence analysis was performed with the use of DNAStar/Lasergene software (DNAStar, Madison, WI).
Expression of UT-B in human intestinal epithelium.
Expression of UT-B mRNA was detected in the mucosa of the normal human colon by Northern blot analysis (Fig. 1A). A predominant 2.0-kb UT-B mRNA was identified, with only faint evidence of a larger mRNA of ∼4.0 kb. These transcripts are similar in size to those originally described by Olivès et al. (20) in erythroid precursor cells, although in a different study by Olivès et al. (18), a predominant band was detected at 3.6 kb for UT-B in the small intestine and the colon. By performing Western blot analysis with a previously described antibody to the COOH terminus of UT-B (27), we identified proteins of ∼50 kDa (a weaker, ∼40-kDa band was also seen, although with variable intensity) and ∼98 kDa (Fig. 1B). The 50-kDa protein matches the size previously described for the kidney UT-B transporter. The renal protein usually appears as a broader band in Western blots (Fig. 1B), possibly reflecting different patterns of glycosylation in the two tissues. The limited effect of treatment with N-glycosidase F on the size of the UT-B protein in the colon compared with UT-B in the kidney and red blood cell membranes supports this possibility (Fig. 1C). Immunolocalization studies showed staining of UT-B in the superficial cells of the colon mucosa, which is somewhat weaker in the deeper aspects of the colonic crypts (Fig. 2, A and C). The staining is present in the cytoplasm of the cells and appears most intense in the apical cell membrane. Figure 2D shows negative staining of the endothelium but faint staining of erythrocyte membranes in intestinal small vessels. Strong staining for UT-B in the endothelial cells of the vasa recta in the human kidney inner medulla is shown as a positive control (Fig. 2E).
Expression of UT-B urea transporter in Caco-2 cells.
Expression of the UT-B urea transporter was identified in Caco-2 cells, a cell line derived from the human neoplastic colon epithelium. Northern blot analysis of these cells shows a 2.0-kb UT-B transcript identical in size to the one seen in the colon mucosa (Fig. 3A), without clear evidence of larger transcripts. Caco-2 cells express the UT-B protein, and the staining pattern apparent from immunohistochemical studies suggests prominent localization to the cell membrane (Fig. 3B).
We tested whether we could detect flux of urea across the Caco-2 epithelial layer. For these studies, cells were grown on permeable supports, and the transfer of [14C]urea through the cells was measured with and without phloretin. The transepithelial flux of urea was detectable from the apical to the basal side, as well as from the basal to the apical side (Fig. 4A), and was inhibited by 0.6 mM phloretin in both cases (Fig. 4B). Flux in either direction ranged from 5 ± 1 to 7.6 ± 1.7 nmol·min−1·cm2. Significant inhibition of urea transport was observed in association with phloretin (50–60%), 1,3-dimethylurea (39%), and thiourea (31%) (Fig. 4, C and D). These findings indicate that movement of urea in Caco-2 cells may be mediated, at least in part, by the UT-B transporter and suggest that transepithelial flux of urea in Caco-2 cells may occur across both sides of the epithelial layer.
Identification of human intestinal urea transporter UT-B cDNA.
We identified a sequence for the intestinal urea transporter UT-B by analyzing a cDNA segment obtained by RT-PCR. The cDNA of human intestinal UT-B has an open reading frame of 1,179 bp, which is predicted to encode a peptide of 389 amino acids and is identical to the previously identified sequence Y19039 for the Kidd blood antigen/urea transporter (25), referred to as hUT-B1 in the recently proposed nomenclature of urea transporters (24) (Fig. 5). By PCR, we could not detect a sequence expressed in the transcript previously identified as RACH1 that has a longer 5′ UTR (GenBank accession no. U35735) than Y19039, indicating that this intestinal UT-B transcript has the same 5′ end of Y19039. The colon UT-B cDNA (2,127 bp) includes 884 nt beyond the known 3′ end of Y19039, which matches the 3 UTR sequence of the UT-B/Kidd blood group (GenBank accession no. NM_015865). The size of our UT-B cDNA clone corresponds to the main UT-B mRNA detected in the human colon mucosa by Northern blot analysis. The 3′ UTR of intestinal UT-B is transcribed from the last exon of the Slc14A1 gene and seems to involve the use of an alternative polyadenylation signal distal from the 3′ end of the Y19039 transcript.
In this study, we describe a human urea transporter expressed in the colon, with features consistent with a UT-B urea transporter/Kidd antigen encoded by the Slc14A1 gene. This is the first identification of a specific urea transporter in the gut and the first time that expression of a UT-B transporter has been detected in epithelial cells. By Northern blot hybridization, a 2-kb UT-B mRNA was clearly the predominant transcript in the colon mucosa, with marginal evidence of a larger 4-kb transcript, that must be expressed in low abundance. The significance of the longer transcript has not been elucidated fully, but Lucien et al. (11) showed that the 4-kb transcript is likely to arise from alternative use of polyadenylation signals at the 3′ end of the last exon of the gene encoding UT-B. The 2-kb mRNA UT-B transcript also was detected in Caco-2 cells. We identified a sequence of the UT-B transporter in the human intestine, which has the same open reading frame of the human erythrocyte urea transporter described by Sidoux-Walter et al. (25) as HUT11A, Y19039, but it includes a longer 3′ UTR. Sidoux-Walter et al. analyzed the Kidd/UT-B sequence using the DNA from several individuals and determined that the Y19039 sequence differs from the one originally cloned by Olivès et al. (20) and represents the normal transcript for the Kidd blood group/UT-B urea transporter in erythrocytes.
Evidence of mRNA expression was associated with demonstration of UT-B protein expression by Western blot hybridization and immunolocalization in human colonic epithelial cells. Of the proteins identified by the UT-B antibody, the 50-kDa protein is the closest to the expected molecular weight of the peptide encoded by the intestinal UT-B transcript. The 98-kDa protein has been noted before in the kidney, the brain, and the testis, and it is not clear whether it represents a complex of UT-B bound to a different protein (27). As Timmer et al. (27) previously noted, the pattern of glycosylation among the UT-B proteins expressed in the kidney or in other tissues may differ, and our findings in the colon seem to support these observations.
By immunohistochemistry, UT-B was detected in the superficial colonic epithelium, with a staining pattern indicating localization in the cytoplasm and the apical membrane of epithelial cells. Whether UT-B is expressed in the small intestine remains to be established. UT-B was detected in Caco-2 cells with a predominant localization to the cell membrane and a distribution similar to that described in cultured bovine aortic endothelial cells with the same antibody (27). The intensity of the staining in the colonic epithelium was weaker than it was in the human medullary vasa recta but definitely greater than that of red blood cells, whose membrane staining is barely detectable in the same tissue sections. These differences may reflect variations in the abundance of UT-B protein in different tissues, which may be determined by tissue-specific regulatory factors. We could not detect evidence of intestinal expression for mRNA or protein of the UT-A transporter, although its expression has been described in mouse colonic crypts (26).
The functional characterization of cloned UT-B erythrocyte urea transporter protein has been described previously and has been performed using the Xenopus oocyte expression system (12, 20, 30), in which the transport of urea mediated by UT-B was found to be sensitive to inhibition by phloretin, thiourea, 1,3-dimethylthiourea, and para-chloromercuribenzene sulfonate (pCMBS). In the present study, we assayed directly the transport of urea in cultured intestinal epithelial cells by using the Caco-2 cell line, which offers the possibility to study transepithelial movement of solutes when cells are grown on permeable support. Phloretin-inhibitable flux of urea was detected in the apical to basal and basal to apical directions, suggesting that passage of urea from the circulation into the intestinal lumen may be mediated by a carrier but also allowing for the possibility that transport of urea might occur in the opposite direction, as some studies have suggested (7, 14). The transport rate for urea across the Caco-2 cell monolayer appears to be significantly less than that estimated for erythrocytes (24). It is possible that more time might be required for urea to be transported across the polarized cells of the intestinal epithelium than to move through a single membrane enveloping the red blood cells, or that differences in membrane abundance and/or activity of UT-B in erythrocytes and intestinal epithelial cells might result in different rates of urea transfer.
Transport of urea in Caco-2 cells was significantly decreased in the presence of phloretin, thiourea, and 1,3-dimethylthiourea compared with control (commercial source for pCMBS was not found), and preliminary tests of pharmacological inhibition of urea transporter activity in Caco-2 cells are similar to those obtained in oocytes into which UT-B mRNA is injected. In our assay, the inhibitory effect of phloretin on urea transport was somewhat less than that reported previously in other studies. This may be accounted for in part by differences in the expression system used (e.g., measuring uptake in single oocytes or isolated cells vs. transepithelial transport). However, we cannot rule out the possibility that other carriers, such as aquaglyceroporins expressed in the colon, may contribute in part to urea transport in intestinal epithelial cells (6, 28). The substrate specificity of the UT-B intestinal urea transporter needs to be established more clearly, because studies in the oocyte expression system have indicated that in addition to urea, water and small solutes may be transported by the UT-B protein (12, 25, 30), and it is possible that, similarly to aquaporins, intestinal UT-B might also be involved in water absorption from the gut. Further studies are required to clarify the properties and physiological role of UT-B in the intestine.
The study of urea transporters has focused mostly on their involvement in the urine concentration mechanism. Aside from mild deficit in the ability to concentrate urine, no significant abnormalities have otherwise been reported in individuals who lack a functional UT-B transporter/Kidd antigen due to mutation in the Slc14a1 gene (11, 23), and a similar mild change in renal function has been described in UT-B knockout mice (29). Finding expression of urea transporters outside the kidney has raised new questions about the significance of urea in physiological processes in other tissues. It is tempting to hypothesize that regulation of urea transport in the intestine may occur in defined settings. For example, if urea moves physiologically from the circulation to the intestinal lumen, this could become a pathway through which to eliminate excess urea from the body when renal function is significantly impaired. It is not known whether long-term expression and abundance of urea transporters in the intestine may be upregulated in uremia or whether they may be affected by high or low content of proteins in the diet.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant R01 DK-53917.
Part of this work was presented in abstract form at the 36th Meeting of the American Society of Nephrology, November 14–17, 2003, San Diego, CA.
↵* H. Inoue and S. D. Jackson contributed equally to this work.
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.
- Copyright © 2004 the American Physiological Society