Urea movement across mouse colonic plasma membranes is mediated by UT-A urea transporters

GASTROENTEROLOGY 2004;126:765–773

BASIC–ALIMENTARY TRACT Urea Movement Across Mouse Colonic Plasma Membranes Is Mediated by UT-A Urea Transporters GAVIN S. STEWART,* ROBERT A. FENTON,‡ FRANK THE´VENOD,§ and CRAIG P. SMITH* *School of Biological Sciences, University of Manchester, Manchester, England; ‡Laboratory of Kidney and Electrolyte Metabolism, Bethesda, Maryland; §Department of Physiology, University of Witten-Herdecke, Witten, Germany

Background & Aims: Urea is a major nitrogen source for commensal bacteria that inhabit the large intestine. UT-A urea transporters mediate urea movement across plasma membranes. The aim of this study was to determine whether UT-A proteins are expressed in the mouse colon and, if so, whether they have a functional role in transcellular urea transport. Methods: Mouse colonic UT-A transporters were investigated with Northern blot analysis, immunoblotting, immunolocalization, and refractive light flux experiments. Results: Northern blot analysis showed that 4 UT-A transcripts were present in mouse colon. Two peptide-targeted polyclonal antibodies showed the presence of UT-A immunoreactive proteins in mouse colon. Antiserum ML446 targeted to the N-terminus of mouse UT-A1 detected proteins of 34 and 48 kilodaltons. Antiserum ML194 targeted to the Cterminus of mouse UT-A1 detected proteins of 48, 75, and 100 kilodaltons. Immunolocalization studies using ML446 showed the presence of UT-A proteins in cells throughout the colonic crypts. ML194 specifically stained cells located in the proliferative and stem regions of the lower portion of colonic crypts. Differential centrifugation and immunoblotting of colonic epithelia showed that UT-A proteins were present in plasma membrane– enriched fractions. Refractive light flux experiments using colonic plasma membrane vesicles showed a significant urea flux, which was completely inhibited by the UT-A inhibitor phloretin. Conclusions: Functional UT-A transporters are expressed in the plasma membranes of mouse colon, indicating that these proteins may play a key role in host/bacterial interaction.

rea is the major end product of mammalian protein catabolism. The regulated movement of urea across plasma membranes is mediated via specialized transporters that are derived from the UT-A and UT-B genes.1,2 To date, 5 UT-A isoforms have been characterized (Figure 1). UT-A1, UT-A2, UT-A3, and UT-A4 are expressed in the kidney and play a central role in the urinary-concentrating mechanism.3,4 In contrast, UT-A5 is expressed in testis,5 where its function remains unclear.

U

The UT-B gene gives rise to a single isoform, UT-B1, which is primarily expressed in erythrocytes but is also expressed in the endothelial cells of the renal vasa recta6 and brain.7 UT-A and UT-B proteins have been detected in other tissues, including liver, heart, and colon, of monogastric animals including humans, rats, and mice.8,9 Functionally, UT-A and UT-B proteins are facilitative transporters that are characteristically blocked by phloretin.4 In contrast, mercury compounds inhibit UT-B proteins10,11 but do not affect UT-A proteins.10 Urea is a major source of nitrogen for the commensal bacteria that inhabit the colon. It passes from the circulation of the host into the digestive tract and is broken down by the resident bacterial enzyme urease into ammonia and carbon dioxide. Ammonia is then used to synthesize the amino acids and nucleotides required for bacterial growth.12 The host can absorb some of the molecules synthesized by the microbiota, thus completing “salvaging” of urea nitrogen.12 Indeed, there is compelling evidence that this process is important in humans during situations of low nitrogen uptake.13–16 Experiments using the oral administration of nitrogen-15– labeled urea showed that urea nitrogen retrieval from the gut was greater in subjects fed a low-protein diet than in those on a normal diet.17 It is now believed that urea nitrogen salvaging is also important in situations of high protein demand, e.g., growth in children.16,17 In addition to their role in urea nitrogen recycling, evidence suggests that colonic microflora are in other ways beneficial to the health of the host. For example, it has been proposed that they may guard against the invasion of pathogenic bacteria and colonic neoplasia.18,19 Abbreviations used in this paper: BSA, bovine serum albumin; HRP, horseradish peroxidase; kb, kilobase; MUTB, novel antiserum raised in rabbits to amino acids 1 to 19 of mouse UT-B; PBS, phosphate buffer solution; SDS, sodium dodecyl sulfate. © 2004 by the American Gastroenterological Association 0016-5085/04/$30.00 doi:10.1053/j.gastro.2003.11.045

766

STEWART ET AL.

Figure 1. The UT-A urea transporter family and epitopes for the ML446 and ML194 antisera. Schematic representation of mouse UT-A proteins: mUT-A1 (Genbank accession no. AF366052), mUT-A2 (Genbank accession no. AF367359), mUT-A3 (Genbank accession no. AF258602), mUT-A4 (GenBank accession no. AY221737), and mUT-A5 (Genbank accession no. AF258601). Black lines represent proteins and are shown relative to UT-A1 (top). Hatched vertical lines highlight ML446 and ML194 epitopes that are common between UT-A family members. It is shown that UT-A1, UT-A3, and UT-A4 express the ML446 epitope, whereas the ML194 epitope is expressed in UT-A1, UT-A2, and UT-A4.

Because host-derived urea is a major source of nitrogen for colonic bacteria, urea transporters may, by regulating the flux of urea from host to bacteria, therefore act as key regulators of microfloral growth. As a first step to understanding the role urea transporters may play in the colon, we have determined the molecular characteristics and distribution of UT-A transporters in mouse colon and have shown that a phloretin-inhibitable urea pathway, characteristic of facilitative urea transporters, is present in colonic plasma membranes.

Materials and Methods Northern Analysis To investigate the distribution of UT-A urea transporter transcripts in mouse colon, poly(A⫹) RNA was isolated from MF1 mouse colon and from kidney and testis as positive controls. Poly(A⫹) RNA was obtained by using an oligo(dT)cellulose batch method5 (Amersham Pharmacia Biotech, Little Chalfont, Buckinghamshire, UK). Poly(A⫹) RNA (3 ␮g per lane) was separated in a 1% agarose gel in the presence of 2.2 mol/L formaldehyde and transferred to Hybond-N filters (Amersham Pharmacia Biotech). Filters were probed with a phosphorus-32–labeled full-length mouse UT-A1 complementary DNA (cDNA) probe (GenBank accession number AF366052). Hybridization was for 16 hours at 42°C (50% formamide) and washing was at 55°C in 0.1⫻ standard sodium citrate and 0.1% sodium dodecyl sulfate (SDS). Autoradiographs were produced with Biomax MS Film (Kodak, Rochester, NY).

Antisera To study the distribution of UT-A proteins in mouse colon, 2 previously characterized UT-A antisera20 were used:

GASTROENTEROLOGY Vol. 126, No. 3

antiserum ML446, raised against amino acids 57 to 75 (H2NEEKDLRSSDEDSHIVKIEK-CONH2) of mUT-A1, and antiserum ML194, raised against amino acids 912 to 930 (H2NQEKNRRASTITKYQAYDVS-COOH) at the carboxyl terminus of mUT-A1. In addition, a novel antiserum raised in rabbits to amino acids 1 to 19 (H2N-MEDSPTMVKVDRGENQILS-CONH2) of mouse UT-B (GenBank accession number AJ420967.1), termed MUTB, was also used (Eurogentec, Seraing, Belgium). This antiserum was affinitypurified by using an Affigel support column (Bio-Rad, Hemel Hempstead, UK) that contained immobilized immunizing peptide, as previously described.21 As an apical membrane marker, an antibody targeted to the apical Na⫹/H⫹ exchanger (NHE3)22 was used (gift of Dr. Mark Knepper, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD). A mouse monoclonal antibody targeted against the Na⫹/K⫹-adenosine triphosphatase (ATPase) ␣1 subunit (no. 05-369; Upstate Biotechnology, Lake Placid, NY) was used as a basolateral membrane marker.

Immunoblotting Tissues were obtained from dead male adult MF1 mice and homogenized in ice-cold buffer with a handheld Dounce homogenizer. The homogenization buffer (pH 7.6) contained HEPES 12 mmol/L, mannitol 300 mmol/L, and several peptidase inhibitors that were added immediately before use— pepstatin 1 ␮g/mL, leupeptin 2 ␮g/mL, and phenylmethylsulfonyl fluoride 1 ␮g/mL (Sigma, St. Louis, MO). Homogenates were initially centrifuged at 2500g for 15 minutes at 4°C. The resulting supernatant was centrifuged at 200,000g for 30 minutes at 4°C. These plasma membrane– enriched pellets were retained and resuspended in homogenization buffer. For subcellular fractionation experiments, homogenates underwent a series of differential centrifugations at 4°C: (1) 1000g for 10 minutes (pellet contains unhomogenized tissue); (2) 4000g for 20 minutes (pellet contains nuclei, mitochondria, and some plasma membranes); (3) 17,000g for 20 minutes (pellet contains mostly plasma membranes); and (4) 200,000g for 1 hour (pellet contains mainly intracellular vesicles).23 Total protein concentrations were determined by using a Protein Assay Reagent Kit (Bio-Rad). Reducing Laemmli sample buffer 5⫻ (5% SDS, 25% glycerol, 0.32 mol/L Tris(hydroxymethyl)aminomethane [pH 6.8], bromophenol blue, and 5% ␤-mercaptoethanol) was added to protein samples in a ratio of 1:4 and then heated at 60°C for 15 minutes. SDS-polyacrylamide gel electrophoresis was performed on minigels of 8% polyacrylamide by loading 10 ␮g per lane of protein. Proteins were then transferred electrophoretically to nitrocellulose membranes (Gelman Sciences, Ann Arbor, MI). After blocking with 5% nonfat dry milk in washing buffer (15 mmol/L Tris HCl [pH 8.0], 150 mmol/L NaCl, and 0.01% Tween-20) for 1 hour, membranes were probed with affinity-purified antisera (ML446, ML194, MUTB, or NHE3) for 16 hours at 4°C. Membranes were rinsed in washing buffer for 3 ⫻ 10 minutes and then probed with goat anti-rabbit horseradish peroxidase (HRP)-linked

March 2004

COLONIC MOUSE UT–A TRANSPORTERS

767

secondary antiserum (Dako, Bucks, UK) at 1:5000 dilution in 5% nonfat milk in washing buffer for 1 hour. After another 3 ⫻ 10 minute rinse in washing buffer, detection of protein was performed by using enhanced chemiluminescence Western blotting detection reagents and enhanced chemiluminescence film (Amersham Pharmacia Biotech). The exception to this protocol was immunoblotting with the anti–Na⫹/K⫹-ATPase primary antibody, which was used in conjunction with a 1:10,000 dilution of anti-mouse HRP-linked secondary antibody (no. NA931V; Amersham Pharmacia Biotech).

Immunolocalization For colonic tissue, MF1 mice were killed by cervical dislocation; the tissue was removed and then immediately placed in 4% paraformaldehyde in a phosphate buffer solution (PBS) for up to 4 hours before being embedded overnight in paraffin blocks. For kidney tissue, the kidneys of thiobutabarbital-anesthetized MF1 mice were perfused with 4% paraformaldehyde solution plus sucrose (approximately 1200 mOsm) through the ascending aorta, removed, and then treated as the colonic tissue. Sections of tissue (5 ␮m) were sliced with a Leica RM2135 microtome (Leica Instruments, Wetzlar, Germany), placed on Superfrost Plus slides (VWR International, West Chester, PA), and then left to dry overnight at 37°C. Sections were placed in xylene overnight and were then rehydrated by using decreasing concentrations of ethanol. To facilitate epitope retrieval, sections were boiled in TEG buffer (10 mmol/L Tris and 0.4 mmol/L ethylene glycol-bis[␤-aminoethyl ether]-N,N,N⬘,N⬘-tetraacetic acid; Sigma) for 3–7 minutes. After cooling to room temperature, sections were washed for 30 minutes in NH4Cl 50 mmol/L in PBS and then for 3 ⫻ 10 minutes in 1% bovine serum albumin (BSA) buffer (1% BSA, 0.2% gelatin, and 0.05% saponin in PBS; Sigma). Sections were incubated at 4°C overnight with primary antiserum (ML446, ML194, or MUTB), washed for 3 ⫻ 10 minutes in 0.1% BSA buffer (0.1% BSA, 0.2% gelatin, and 0.05% saponin in PBS), and incubated in a 1:200 dilution of secondary antiserum (goat anti-rabbit HRP-linked; Dako) for 1 hour at room temperature. After another 3 ⫻ 10 minute washes in 0.1% BSA buffer, sections were counterstained with hematoxylin (BDH, Poole, Dorset, UK), and immunoperoxidase signals were detected by using diaminobenzidine (BDH). Coverslips were mounted with Eukitt (O. Kinder GmBH, Freiburg, Germany), and sections were viewed with an Axiophot microscope coupled to a color charge-coupled device camera (Zeiss).

Urea Flux Experiments Purified plasma membrane vesicles were obtained from colonic scrapings and dissected superficial kidney cortex by using the thiocyanate method.24 A total of 10 ␮L of purified plasma membrane solution containing 10 ␮g of protein was added to 90 ␮L of 1 of the following buffers: (1) control (mannitol 100 mmol/L and HEPES 1 mmol/L; pH 7.4), (2) control plus phloretin 100 ␮mol/L, (3) control plus HgCl2 500 ␮mol/L, (4) urea 2 mmol/L (control buffer plus urea 2 mmol/

Figure 2. Northern analysis of mouse kidney, testes, and colon poly(A⫹) RNA (3 ␮g per lane). K, kidney; T, testis; PC, proximal colon; DC, distal colon probed with full-length mUT-A1. Final washes were in 0.1⫻ standard sodium citrate at 55°C. Exposures were as follows: kidney, 1 hour; testis, 3 hours; and colon, 3 days. The kidney shows strong signals corresponding to mUT-A1 (4.1 kb), mUT-A2 (3.0 kb), and mUT-A3 (2.1 kb). In testis, the mUT-A5 signal at 1.5 kb is clearly visible. Less intense signals at 2.7 and 3.3 kb were present in testis and in the proximal and distal colon. Bands at 1.5 and 1.4 kb were also present in the proximal colon.

L), (5) urea 2 mmol/L plus phloretin 100 ␮mol/L, (6) urea 2 mmol/L plus HgCl2 500 ␮mol/L, (6) glutamate 2 mmol/L (control buffer plus glutamate 2 mmol/L), or (7) glutamate 2 mmol/L plus phloretin 100 ␮mol/L. Inhibitors were dissolved in ethanol to produce a 500 mmol/L stock solution and then diluted in the experimental buffer. The addition of ethanol alone to the mannitol control buffer at the appropriate dilution produced no effect on urea flux (data not shown). Flux into plasma membrane vesicles was assessed by measuring the time-dependent decrease in light scattering due to vesicle swelling25 by using a DU530 spectrophotometer (Beckman, High Wycombe, Bucks, UK) set to measure absorbance at 470 nm.

Statistical Analysis The mean absorbance after 3 minutes was compared for each group by using 1-way analysis of variance (ANOVA). If the ANOVA indicated a difference, treatment comparison between groups was performed with the Student–Newman– Keuls method. Values are shown as mean ⫾ SEM; statistical significance was assumed at the 5% level.

Results Northern Analysis of Mouse Tissue Northern analysis using a full-length mUT-A1 probe showed several signals corresponding to UT-A messenger RNAs (mRNAs) in kidney, testis, and colon (Figure 2). As expected, strong signals were observed in

768

STEWART ET AL.

GASTROENTEROLOGY Vol. 126, No. 3

kidney, representing mUT-A1 (4.1 kilobases [kb]; GenBank accession no. AF366052), mUT-A2 (3.0 kb; GenBank accession no. AF367359), and mUT-A3 (2.1 kb; GenBank accession no. AF258602) and also in testis, representing mUT-A5 (1.5 kb; GenBank accession no. AF258601).5,9 In testis, 2 weaker bands at 3.3 and 2.7 kb were also observed. In proximal colon, signals were weaker and corresponded to transcripts with molecular weights of 3.3, 2.7, 1.5, and 1.4 kb. Similar signals were observed in distal colon, except that the 1.5-kb band was absent. This analysis indicates that UT-A transcripts detected in the colon are different from those present in kidney; that the 3.3-, 2.7-, and 1.5-kb transcripts detected in testis are also present in colon; and that the 1.4-kb colonic transcript is novel and not apparent in either kidney or testis. Immunoblotting On the basis of our current knowledge of mouse UT-A proteins, it is predicted that ML446 and ML194 should recognize several known UT-A isoforms (Figure 1). In kidney, as previously observed,20 ML446 detected bands at 89 and 119 kilodaltons, characteristic of UT-A1 in different glycosylation states,26 and a signal at 48 –53 kilodaltons (mUT-A3) in the kidney inner medulla (Figure 3A). ML194 also strongly detected UT-A1 (89- and 119-kilodalton bands), but no other bands, in kidney inner medulla (Figure 3C). In contrast to ML446, ML194 detected a diffuse band at 43–55 kilodaltons in the inner stripe of the outer medulla, corresponding to mUT-A2 in the thin descending limbs. ML446 detected numerous protein bands in the testis, most notably at 48, 60, and 77 kilodaltons, whereas ML194 detected bands at 48, 75, and 100 kilodaltons. In colon, ML446 detected a strong band at 34 kilodaltons and a weaker band at 48 kilodaltons. As observed in testis, ML194 detected bands at 48, 75, and 100 kilodaltons in colon. In each case, the signal was stronger in the testis. None of these bands was present after preincubation with the corresponding immunizing peptide (Figure 3B and D), but they were unaffected by incubation with a nonrelated peptide (data not shown). Together these data and those from Northern analysis show that UT-A proteins expressed in colon differ from those detected in kidney. However, the 48-, 75-, and 100kilodalton UT-A proteins detected in colon are also present in testis. It is interesting to note that the 34kilodalton protein detected in colon by ML446 is not present in either kidney or testis. Antiserum MUTB was targeted to the N-terminus of mouse UT-B. MUTB detected a signal at 45 kilodaltons

Figure 3. Western analysis of mouse MF1 kidney, testis, and colon (10 ␮g per lane). (A) ML446 detected bands at 89 and 119 kilodaltons (mUT-A1), along with a signal at 48 –53 kilodaltons (mUT-A3), in the kidney inner medulla. In testis, ML446 detected major bands at 48, 60, and 77 kilodaltons, whereas in the colon, ML446 detected a strong band at 34 kilodaltons and a weaker band at 48 kilodaltons. (B) All bands were ablated after preincubation with the ML446 immunizing peptide. (C) ML194 detected 89- and 119-kilodalton bands (mUT-A1) in kidney inner medulla and a diffuse band at 43–55 kilodaltons (mUT-A2) in the kidney inner stripe of the outer medulla. In testis and colon, ML194 detected bands at 48, 75, and 100 kilodaltons. (D) None of these bands was present after preincubation with the ML194 immunizing peptide. (E) MUTB detected bands at 45 kilodaltons in kidney medulla (KM) and at 65 kilodaltons in colon (Co). (F ) These bands were ablated after preincubation with the MUTB immunizing peptide.

in the mouse kidney medulla (Figure 3E), similar to the reported 41–54-kilodalton size for UT-B protein in rat kidney.27 In colon, MUTB detected a 65-kilodalton protein (Figure 3E). Both bands were ablated by preincubating the antibody with immunizing peptide (Figure 3F). In data not shown, MUTB specifically stained the descending vasa recta in the kidney medulla, thus confirming its specific affinity for UT-B protein.

March 2004

COLONIC MOUSE UT–A TRANSPORTERS

769

Immunolocalization of UT-A Proteins in Mouse Kidney and Colon In agreement with our previously published data,20 ML446 detected a strong signal in the middle and terminal inner medullary collecting ducts corresponding to mUT-A1 and mUT-A3 (Figure 4A). Antiserum ML194 also strongly stained the inner medullary collecting ducts (Figure 4B). In addition, ML194 strongly labeled short (type 1) and long (type 3) thin descending limbs in the inner stripe of the outer medulla corresponding to mUT-A2 (data not shown). In all cases, staining was absent in sections in which preincubation with immunizing peptide was performed (insets, Figure 4A and B). These data show that the antisera labeled the appropriate structures in the kidney and were specific for UT-A proteins. Using these reagents, we determined the distribution of labeling in colon. Antiserum ML446 strongly stained cells in the bottom half of colonic crypts (Figure 4C). Cells in the proliferative and stem cell regions were stained. Staining decreased toward the plateau region but was still evident in the absorptive columnar cells lining the luminal surface. The staining was predominantly cytoplasmic (Figure 4E and I) and may represent UT-A proteins present in intracellular vesicles that may shuttle to the membrane. In kidney, inner medullary collecting duct UT-A proteins have been detected in subcellular vesicles.23 Antiserum ML194 also stained proteins in cells in the bottom half of colonic crypts (Figure 4D). Again, proliferative cells and putative stem cells were stained; the staining sharply decreased toward the plateau region and was absent in mature goblet cells. ML194 stained lateral and apical membranes of the cells lining the lower crypt (Figure 4G and J), suggesting a role in transcellular urea movement. All staining was absent after preincubation with immunizing peptide (Figure 4F and H). Subcellular Fractionation Experiments Serial centrifugation and immunoblotting of colon homogenate samples allowed us to investigate the presence of UT-A and UT-B proteins in plasma membranes. Using an antibody to detect the Na⫹/K⫹-ATPase ␣1 subunit (Na⫹/K⫹-ATPase), which is expressed on the basolateral plasma membrane, we confirmed that the 17,000g sample was enriched for basolateral plasma membranes compared with the 4000g sample (Figure 5). In addition, we also confirmed that the 200,000g sample did not contain significant amounts of basolateral plasma membrane. Because NHE3 is an apical membrane marker in the mouse colon,28 we used an NHE3 anti-

Figure 4. Immunolocalization of UT-A transporters in kidney and colon. Immunoperoxidase labeling was performed on paraffin-embedded mouse kidney sections (5 ␮m) stained with (A) ML446 or (B) ML194. Both antisera stained apical (small arrows) and basolateral (arrowheads) membranes in the inner medullary collecting ducts (IMCD; 100⫻ magnification). *Tubular lumen. IMCD staining was not present after preincubation with immunizing peptide (inset, A and B, 40⫻). Immunoperoxidase labeling was also performed on paraffin-embedded mouse colon sections (5 ␮m). (C) ML446 strongly stained mainly the cytoplasm of cells in the lower half of the colonic crypts and some cytoplasm of absorptive columnar cells lining the lumen (small arrows; 40⫻). (D) ML194 stained plasma membranes of cells in the bottom half of colonic crypts (40⫻). (E) Longitudinal view of a crypt showing ML446 staining (100⫻). (F ) ML446 staining was not present in colon after peptide preincubation (100⫻). (G) Longitudinal view of a crypt showing ML194 staining (100⫻): apical (small arrows) and lateral (arrowheads) membranes. (H ) ML194 staining was absent in colon after peptide preincubation (100⫻). (I ) Transverse view showing ML446 cytoplasmic staining of crypt cells (100⫻). ( J) Transverse view showing ML194 staining of crypt cell plasma membranes. Lateral membranes (arrowheads) and apical membranes (arrows) are labeled (100⫻).

serum22 to confirm similar results for apical plasma membranes. ML446 strongly detected the 34-kilodalton protein in the plasma membrane– enriched 17,000g sample (Figure

770

STEWART ET AL.

GASTROENTEROLOGY Vol. 126, No. 3

Urea Flux Experiments

Figure 5. Semiquantitative immunoblotting of colonic subcellular fractionation samples, plus purified plasma membrane vesicle samples. Each sample contains different subcellular fractions: 4000g, nuclei, mitochondria, and some plasma membranes; 17,000g, plasma membrane enriched; 200,000g, intracellular vesicles; vesicles; prepared colonic vesicles—protein from the purified plasma membrane vesicle preparations used in urea flux experiments (flux expt prep). Blots were probed with antibodies for Na⫹/K⫹-ATPase, NHE3, ML446, ML194, or MUTB.

5). This 34-kilodalton protein was also present in the 4000g and 200,000g samples. ML194 recognized both 75- and 100-kilodalton signals in the 4000g, 17,000g, and 200,000g samples. MUTB detected a 65-kilodalton signal only in the 200,000g sample. The pattern of distribution of mUT-B was therefore completely different from that of Na⫹/K⫹-ATPase and NHE3. These results indicate that the localization of UT-B proteins is different from that of UT-A proteins and that UT-A proteins are present in the plasma membranes. Immunoblotting was also performed on the purified colonic plasma membrane vesicles that were used in urea flux experiments (see below). Bands representing Na⫹/K⫹-ATPase and NHE3 were observed in the colonic vesicle samples. This confirmed that the colonic vesicles contained both apical and basolateral plasma membranes. By using the UT-A–selective antisera, weak signals were observed in these vesicles that corresponded to the 34- and 100-kilodalton UT-A proteins. In contrast, the 65-kilodalton mUT-B protein was not detected. From these results we conclude that UT-A proteins, but not UT-B proteins, are present in the vesicle preparation of colonic plasma membranes used in the urea flux experiments.

Our immunolocalization and immunoblotting studies indicate that UT-A proteins in colon are expressed in the plasma membranes of colonic epithelial cells. To test whether these proteins may function as urea transporters in these cells, we measured urea-induced swelling of colonic plasma membrane vesicles, using the time-dependent decrease in light scattering that occurs during cell swelling.29 In the absence of urea, vesicles did not swell markedly. On the addition of urea 2 mmol/L, vesicles swelled over 3 minutes, as indicated by the significant decrease in absorbance of light at a wavelength of 470 nm (A470) (Figure 6A; P ⬍ 0.05; ANOVA). Although inclusion of the UT-A inhibitor phloretin (100 ␮mol/L) had no effect on control values, it completely inhibited the urea-induced swelling (Figure 6B; P ⬍ 0.01; ANOVA). In contrast, HgCl2—a known inhibitor of UT-B but not UT-A transporters— had no effect on the urea-induced swelling (Figure 6C), suggesting that UT-B proteins are not involved in the process. Kidney cortex is known to express a plethora of transporters—for example, sodium-coupled transporters30—some of which have been reported to transport urea at low levels. It does not, however, express UT-A transporters.4 Although swelling was observed in vesicles prepared from mouse kidney cortex, it was neither increased by the presence of urea nor inhibited by phloretin (data not shown). Other transporters, such as the Naglutamate transporter,31 have been reported to transport urea passively and are also expressed in colon. To investigate the involvement of these proteins, we tested the specificity of the phloretin inhibition on urea-induced swelling. Colonic vesicles were placed in buffer containing glutamate 2 mmol/L instead of urea. In this solution, vesicles were observed to swell in a similar fashion to that observed with urea, but in contrast to the urea-induced swelling, the addition of phloretin had no effect (Figure 6D), indicating that urea was not translocated via glutamate transporters. For both colonic and kidney cortex flux experiments, the inclusion of NaCl in the buffer had no effect on urea-induced swelling (data not shown). These results strongly suggest that UT-A urea transporters facilitate urea transport across colonic plasma membranes.

Discussion The primary aims of this study were to characterize the expression and regional distribution of UT-A urea transporters in mouse colon and to establish whether these transporters mediate urea movement across colonic plasma membranes. To achieve this, we performed

March 2004

COLONIC MOUSE UT–A TRANSPORTERS

771

Figure 6. Refractive light experiments to measure colonic urea flux. For each experiment, 10 ␮g of colonic vesicles was suspended in 100 ␮L of test solution. (A) Time course of typical vesicular swelling experiments. (B) Summary of data for the experiment shown in (A) (mean changes in A470 absorbance after 3 minutes ⫾ SEM; numbers below bars indicate n). Urea 2 mmol/L caused a significant increase in swelling (P ⬍ 0.05; ANOVA), which was inhibited by phloretin 100 ␮mol/L (P ⬍ 0.01; ANOVA). (C) HgCl2 500 ␮mol/L had no effect on urea-induced swelling (P ⬎ 0.05; ANOVA). (D) Glutamate 2 mmol/L produced an increase in cell swelling (P ⬍ 0.05; ANOVA) that was not inhibited by phloretin 100 ␮mol/L (P ⬎ 0.05; ANOVA). *Significantly different from control; P ⬍ 0.05 (ANOVA).

Northern analysis and Western analysis to determine the molecular weight of expressed proteins and performed immunolocalization studies to determine patterns of expression. In addition, we isolated colonic plasma membrane vesicles and tested for the presence of functional, phloretin-inhibitable urea transporters. Previously we have isolated and characterized cDNAs from mouse kidney that encode UT-A1 (4.1 kb), UT-A2 (3.0 kb), and UT-A3 (2.1 kb).9 By conventional libraryscreening methods, we have been unable to isolate the mouse homolog of UT-A4, although reverse-transcription polymerase chain reaction analysis of renal mRNA indicated that the UT-A4 transcript is present in kidney medulla at low levels.32 We have also isolated a cDNA from mouse testis encoding UT-A5 (1.5 kb).5 Northern analysis showed that 4 distinct mRNA transcripts (3.3, 2.7, 1.5, and 1.4 kb) were present in the colon. All 4 colonic transcripts differ in size from the mouse renal UT-A isoforms, suggesting that these transcripts represent nonrenal UT-A isoforms. Three of the transcripts

detected in colon (3.3, 2.7, and 1.5 kb) were also present in testis, indicating a tissue-specific pattern of expression shared by testis and colon that is distinct from that of renal UT-A expression. On the basis of molecular weight, the 1.5-kb transcript may represent UT-A5. The 1.4-kb transcript detected in colon was detected in neither kidney nor testis and may represent a unique colonic transcript. Western analysis and immunocytochemistry of mouse kidney showed that both ML446 and ML194 UT-A antisera, as previously observed,20 were specific for UT-A proteins. Antiserum ML446 detected UT-A1 and UTA3, whereas ML194 detected UT-A1 and UT-A2. The mUT-B antiserum, MUTB, detected a 45-kilodalton protein in mouse kidney medulla and stained the descending vasa recta, confirming its specificity for UT-B.27 In colon, UT-A antisera detected proteins that differed in size from those expressed in the kidney. This supported our Northern analysis, which showed that the molecular weight of colonic mRNAs differed from those

772

STEWART ET AL.

in the kidney. Currently, we do not know the molecular structure of the colonic urea transporters, but some indication of their makeup can be gleaned from our results. For example, because the 34-kilodalton protein was detected by ML446 only in the colon, this protein may represent a novel colonic UT-A isoform (Figure 1). Although a cDNA encoding this isoform has not yet been isolated from mouse, it may represent the 1.4-kb transcript that was specific to the colon. We have recently discovered a novel 1.7-kb cDNA in the database that encodes a UT-A isoform isolated from human colon (GenBank accession no. AK074236). The open reading frame for this cDNA encodes a protein of approximately 30 kilodaltons, which hypothetically would be detected only by our ML446 antiserum. When heterologously expressed in Xenopus oocytes, this protein functions as a phloretin-inhibitable urea transporter (Smith et al., unpublished observations, 2003). Immunostaining of colon by using ML194 or ML446 showed an intriguing pattern of UT-A expression in that staining was mainly restricted to the lower half of the colonic crypts. Little or no staining was present in the secretory goblet cells that contain mucus or mature absorptive columnar cells at the crypt surface. Staining was observed on the plasma membranes of thin, tightly packed cells in the proliferative and stem cell regions. Subcellular fractionation experiments indicated that the ML446-sensitive 34-kilodalton protein was present in both plasma membranes and intracellular vesicles. Similarly, these experiments suggested that 75- and 100kilodalton proteins detected by ML194 were present throughout the cells. The finding that UT-A proteins were expressed in the plasma membranes of colonic epithelial cells indicated that these transporters may be involved in mediating transcellular urea flux. By comparison, UT-B was not detected in plasma membrane samples, indicating that this protein is likely to be expressed within the cell rather than in the plasma membrane. Refractive light experiments with colonic plasma membrane vesicles showed that urea caused vesicular swelling that was completely abolished by the addition of phloretin, but, importantly, not by HgCl2. Phloretin is known to inhibit both UT-A and UT-B urea transporters,4,5 whereas mercury compounds inhibit only UT-B transporters.10,11 In contrast, no urea-induced, phloretin-sensitive swelling was observed in vesicles prepared from kidney cortex, a tissue that expresses a plethora of sodium-coupled transporters (e.g., sodium-glucose cotransporter SGLT1),30 which have been proposed to transport urea with low affinity.33 It is important to note

GASTROENTEROLOGY Vol. 126, No. 3

that glutamate-induced swelling of colonic vesicles was completely insensitive to phloretin 100 ␮mol/L. Because urea transport by glutamate transporters is inhibited by phloretin with low affinity (inhibition constant ⬎1 mmol/L),33 it is unlikely that urea transport by these proteins contributes significantly to the urea transport we observed. Both 34- and 100-kilodalton UT-A proteins were detected in the colonic plasma membrane vesicles used for functional experiments. These data therefore strongly suggest that UT-A urea transporters are present in the membranes of the colonic vesicles. This finding, combined with the results of our immunocytochemical localization study, implicates UT-A urea transporters in urea flux across colonic cell membranes. The high urease activity imparted by the microorganisms that inhabit the colon suggests that a urea gradient exists in a basolateral to apical direction. We suggest that because of their location, UT-A proteins transport urea destined for microbial metabolism into the crypt. These transporters therefore possibly play a key role in the host/microbial relationship. In summary, we have identified mouse UT-A proteins expressed in colonic plasma membranes and have established that these membranes possess a phloretin-inhibitable, mercury-insensitive urea pathway that is characteristic of UT-A proteins. We suggest that these proteins mediate urea movement destined for microbial breakdown into the colonic lumen. These findings therefore have important implications for the host/microbial relationship and associated host gastrointestinal health.

References 1. You G, Smith CP, Kanai Y, Lee W, Stelzner M, Hediger M. Cloning and characterization of the vasopressin-regulated urea transporter. Nature 1993;365:844 – 847. 2. Olives B, Neau P, Bailly P, Hediger MA, Rousselet G, Cartron JP, Ripoche P. Cloning and functional expression of a urea transporter from human bone marrow cells. J Biol Chem 1994;269: 31649 –31652. 3. Smith CP, Lee WS, Martial S, Knepper MA, You G, Sands JM, Hediger MA. Cloning and regulation of expression of the rat kidney urea transporter (rUT2). J Clin Invest 1995;96:1556 – 1563. 4. Smith CP, Rousselet G. Urea transporters. J Membr Biol 2001; 183:1–14. 5. Fenton RA, Howorth A, Cooper GJ, Meccariello R, Morris ID, Smith CP. Molecular characterization of a novel UT-A urea transporter isoform (UT-A5) in testis. Am J Physiol Cell Physiol 2000;279: C1425–C1431. 6. Yang B, Bankir L, Gillespie A, Epstein CJ, Verkman AS. Urea selective concentrating defect in transgenic mice lacking urea transporter UT-B. J Biol Chem 2002;277:10633–10637. 7. Trinh-Trang-Tan MM, Lasbennes F, Gane P, Roudier N, Ripoche P, Cartron J-P, Bailly P. UT-B1 proteins in rat: tissue distribution and regulation by antidiuretic hormone in kidney. Am J Physiol Renal Physiol 2002;283:F912–F922. 8. Ritzhaupt A, Wood IS, Jackson AA, Moran BJ, Shirazi-Beechey

March 2004

9.

10.

11.

12. 13.

14. 15.

16. 17.

18.

19.

20.

21.

22.

SP. Isolation of a RT-PCR fragment from human colon and sheep rumen RNA with nucleotide sequence similarity to human and rat urea transporter isoforms. Biochem Soc Trans 1998;26:S122. Fenton RA, Stewart GS, Carpenter B, Howorth A, Potter E, Cooper GJ, Smith CP. Characterization of the mouse urea transporters UT-A1 and UT-A2. Am J Physiol Renal Physiol 2002;283:F817– F825. Martial S, Olives B, Abrami L, Couriaud C, Bailly P, You G, Hediger M, Cartron J-P, Ripoche P, Rousselet G. Functional differentiation of the human red blood cell and kidney urea transporters. Am J Physiol Renal Physiol 1996;271:F1264 –F1268. Tsukaguchi H, Shayakul C, Berger UV, Tokui T, Brown D, Hediger MA. Cloning and characterization of the urea transporter UT3. J Clin Invest 1997;99:1506 –1515. Fuller MF, Reeds PJ. Nitrogen recycling in the gut. Ann Rev Nutr 1998;18:385– 411. Meakins TS, Jackson AA. Salvage of exogenous urea nitrogen enhances nitrogen balance in normal men consuming marginally inadequate protein diets. Clin Sci 1996;90:215–225. Jackson AA. Salvage of urea-nitrogen and protein requirements. Proc Nutr Soc 1995;54:535–547. Metges CC, Petzke KJ, El-Khoury AE, Henneman L, Grant I, Bedri S, Regan MM, Fuller MF, Young VR. Incorporation of urea and ammonia nitrogen into ileal and fecal microbial proteins and plasma free amino acids in normal men and ileostomates. Am J Clin Nutr 1999;70:1046 –1058. Waterlow JC. The mysteries of nitrogen balance. Nutr Res Rev 1999;12:25–54. Tanaka N, Kubo K, Shiraki K, Koishi H, Yoshimura H. A pilot study on protein metabolism in the Papua New Guinea highlanders. J Nutr Sci Vitaminol 1980;26:247–259. Gibson GR, Mountzouris KC, McCartney AL, Gibson GR. Intestinal microflora of human infants and current trends for its nutritional modulation. Br J Nutr 2002;87:405– 420. Isolauri E, Kirjavainen PV, Salminen S. Probiotics: a role in the treatment of intestinal infection and inflammation? Gut 2002; 50(Suppl 3):54 –59. Stewart GS, Fenton RA, Wang W, Kwon T-H, White SJ, Collins VM, Cooper G, Nielsen S, Smith CP. The basolateral expression of mUT-A3 in the mouse kidney. Am J Physiol Renal Physiol (Epub 10.1152/ajprenal.00334.2003) Ferguson CJ, Wareing M, Ward DT, Green R, Smith CP, Riccardi D. Cellular localization of divalent metal transporter DMT-1 in rat kidney. Am J Physiol Renal Physiol 2001;280:F803–F814. Kim GH, Ecelbarger C, Knepper MA, Packer RK. Regulation of thick ascending limb ion transporter abundance in response to altered acid/base intake. J Am Soc Nephrol 1999;10:935– 942.

COLONIC MOUSE UT–A TRANSPORTERS

773

23. Nielsen S, Terris JM, Smith CP, Hediger MA, Ecelbarger CA, Knepper MA. Cellular and subcellular localization of the vasopressin-regulated urea transporter in rat kidney. Proc Natl Acad Sci U S A 1996;93:5495–5500. 24. Hopfer U, Crowe TD, Tandler B. Purification of brush border membrane by thiocyanate treatment. Anal Biochem 1983;131: 447– 452. 25. Verkman AS. Optical methods to measure membrane transport processes. J Membr Biol 1995;148:99 –110. 26. Bradford AD, Terris JM, Ecelbarger CA, Klein JD, Sands JM, Chou C, Knepper MA. 97- and 117-kDa forms of collecting duct urea transporter UT-A1 are due to different states of glycosylation. Am J Physiol Renal Physiol 2001;281:F133–F141. 27. Timmer RT, Klein JD, Bagnasco SM, Doran JJ, Verlander JW, Gunn RB, Sands JM. Localization of the urea transporter UT-B protein in human and rat erythrocytes and tissues. Am J Physiol Cell Physiol 2001;281:C1318 –C1325. 28. Chu J, Chu S, Montrose MH. Apical Na⫹/H⫹ exchange near the base of mouse colonic crypts. Am J Physiol Cell Physiol 2002; 283:C358 –C372. 29. Benharouga M, Lipecka J, Fanen P, Baudoin-Legros M, Banting G, Fritsch J, Edelman A. Properties of a Cl(⫺)-conductive pathway(s) in microsomes from rat kidney inner medulla. Involvement of cystic fibrosis transmembrane regulator protein. Eur J Biochem 1996;240:268 –273. 30. Cramer SC, Pardridge WM, Hirayama BA, Wright EM. Colocalization of GLUT2 glucose transporter, sodium/glucose cotransporter, gamma glutamyl transpeptidase in rat kidney with double-peroxidase immunocytochemistry. Diabetes 1992;41: 766 –770. 31. MacAulay N, Gether U, Klaerke DA, Zeuthen T. Passive water and urea permeability of a human Na-glutamate cotransporter expressed in Xenopus oocytes. J Physiol 2002;542:817– 828. 32. Fenton RA, Cottingham CA, Stewart GS, Howorth A, Hewitt JA, Smith CP. Structure and characterization of the mouse UT-A gene (Slc14a2). Am J Physiol Renal Physiol 2002;282:F630 –F638. 33. Leung DW, Loo DD, Hirayama BA, Zeuthen T, Wright EM. Urea transport by cotransporters. J Physiol 2000;528:251–257.

Received July 11, 2002. Accepted November 13, 2003. Address requests for reprints to: Craig P. Smith, M.D., School of Biological Sciences, University of Manchester, G.38, Stopford Building, Oxford Road, Manchester M13 9PT, England. e-mail: cpsmith@man. ac.uk; fax: (01) 61-275-5600. Supported by the Biotechnology and Biological Sciences Research Council (Grant 34/D10935) and the Royal Society (C.P.S.). We are very grateful to Dr. Catherine Booth, Dr. Greg Tudor, Huw Morgan, and Elizabeth Potter for their expertise and advice.