点击显示 收起
【摘要】 The amphibian urea transporter (fUT) shares many properties with the mammalian urea transporters (UT) derived from UT-A and UT-B genes. The transport of urea by fUT is inhibited by the mercurial agent p -chloromercuribenzenesulfonic acid (pCMBS). We found that in oocytes expressing cRNA encoding fUT, a 5-min preincubation in 0.5 mM mercury chloride (HgCl 2 ) also significantly reduced urea uptake. The transport of urea by fUT was rendered mercury (Hg 2+ ) insensitive by mutating either of the residues C185 or H187, both of which lie within the M-I region (close to the hypothetical UT pore). In oocytes expressing a mixture of the C185 and H187 mutants, Hg 2+ sensitivity was reestablished. The transport of urea by the mouse UTs mUT-A2 and mUT-A3 was not sensitive to Hg 2+. Introducing cysteine residues analogous to that mutated in fUT into mUT-A2 or mUT-A3 did not induce Hg 2+ sensitivity. Additionally, introducing the double cysteine, histidine mutations into mUT-A2 or mUT-A3 still did not induce Hg 2+ sensitivity, indicating that a region outside of the M-I region also contributes to the Hg 2+ -induced block of fUT. Using a series of chimeras formed between UT-A3 and fUT, we found that as well as C185 and H187, residues within the COOH terminal of fUT determine Hg 2+ sensitivity, and we propose that differences in the folding of this region between fUT and mUT-A2/mUT-A3 allow access of Hg 2+ to the fUT channel pore.
【关键词】 mercury chloride Xenopus laevis oocyte
THE ABILITY OF THE MAMMALIAN kidney to maximally concentrate urine requires the inner medullary interstitium to be hypertonic compared with plasma, with the small solute urea contributing 50% to inner medullary hypertonicity. The transport of urea across membranes is facilitated by the urea transporter (UT) proteins. In mammals, these transporters are derived from two genes, UT-A and UT-B (reviewed in Ref. 18 ). The major UT-A isoforms 1-3 are expressed in the nephron. UT-A1 and UT-A3 are present in the inner medullary collecting duct (IMCD) ( 3, 4, 7, 19 ), and UT-A2 is expressed in the thin descending limb of the loop of Henle ( 19 ).
The frog urinary bladder expresses a urea transporter (fUT) that has a high identity (66% at the amino acid level) to UT-A2 and UT-B (61%). In terms of its transport properties, fUT appears to be a hybrid transporter, sharing properties of both the UT-A and UT-B proteins. Like all UTs identified to date, fUT produces a phloretin-sensitive increase in membrane urea permeability (reviewed in Ref. 16 ). As is the case for the UT-A family, fUT is impermeable to the urea analog thiourea, whereas UT-B is permeable to thiourea ( 2, 9 ). Thiourea reduces urea transport in oocytes expressing either UT-A or UT-B ( 9 ). Similarly, the transport of urea by fUT is sensitive to thiourea ( 2 ). Finally, like the UT-B proteins, fUT is inhibited by the organic mercurial (pCMBS) ( 12 ), whereas the transport of urea by the UT-A isoforms when expressed in Xenopus laevis oocytes is insensitive to pCMBS ( 9 ).
Very little is known about the membrane topology of the urea transporters and how they interact to produce a functional urea pore. Initial analysis of the sequence of UTs using the Kyte-Doolittle algorithm predicted proteins with intracellular NH 2 and COOH termini, with 10 transmembrane spans, split by a large extracellular loop containing a N-linked glycosylation site ( 21 ). A more detailed analysis by Sands et al. ( 15 ) proposed a model with six membrane-spanning domains, split into two groups separated by a large extracellular loop (see Fig. 1 ). This model also predicts a short-integral membrane domain that dips into the membrane from the external face. This theoretical short-integral membrane domain, which has been designated the M-I region ( 15 ) ( Fig. 1 A ), resembles membrane loops described for many channel proteins, such as the water channel aquaporin-1 (AQP1) ( 6 ). In the case of AQP1, the second asparagine-proline-alanine (NPA)-containing domain, which is crucial to the formation of the channel pore, would be analogous to the M-I region in the UTs. It seems feasible that the M-I loop of the UTs might fulfill a similar role and could be linked to the formation of a channel pore. The M-I region is highly conserved among UTs. The similarity between UT-A2, UT-A3, and fUT in the M-I region is on the order of 90%, and the similarity between fUT and mUT-B in this domain is 70% ( Fig. 1 ).
Fig. 1. Predicted topological structure of frog urea transporter (fUT). Top : predicted topological map for fUT based on the Sands et al. ( 15 ) model. The M-I domain is indicated by the dotted line. Bottom : table showing a comparison of the residues in the M-I region for mUT-A2, mUT-A3, hUT-B, and fUT, where m indicates mouse and h, human. Residues shown on a gray background are identical to mUT-A2, and differences are indicated by residues on a white background.
Like fUT and UT-B, the water channel AQP1 is inhibited by mercurial agents such as mercury chloride (HgCl 2 ) and pCMBS ( 1, 13 ). The molecular basis of the mercurial sensitivity of AQP1 has been established and involves the cysteine residue at position 189 ( 14 ). Mutation of this cysteine to a serine residue (AQP1-C189S) renders AQP1 insensitive to mercurials. The C189 residue is close to the second NPA motif in AQP1, and in the hourglass model of AQP1 it was predicted that the C189 residue bordered the opening of the channel pore on the extracellular face of the protein ( 6 ). This prediction has been confirmed by analysis of the three-dimensional crystal structure of AQP1 ( 11 ). The mercurial-induced inhibition of human UT-B (hUT-B) has been investigated, but the basis of mercurial sensitivity was not established ( 8 ). Two individual cysteine residues were mutated (C151 and C236), but mercurial sensitivity remained intact ( 8 ). Besides these two cysteine residues, there are no immediate candidates in the region bordering the M-I loop that might account for the mercurial sensitivity of hUT-B.
We predict that residues in the M-I region confer Hg 2+ sensitivity to fUT, and we have used a combination of point mutants in fUT and chimeras between fUT and mUT-A3 to test this hypothesis. We identified two residues (a cysteine and histidine), which both lie in the M-I domain, that are required to produce Hg 2+ -dependent inhibition of urea transport by fUT.
METHODS
Solutions. The control amphibian Ringer solution (ND-96) contained (in mM) 96 NaCl, 2 KCl, 1.8 CaCl 2, 1 MgCl 2, and 5 HEPES. For the calcium-free ND-96 solution, CaCl 2 was replaced by NaCl. The OR3 solution contained 6.85 g/l of Leibovitz L-15 cell culture medium (Invitrogen, Paisley, UK); 10,000 U/ml penicillin G sodium; 10,000 µg/ml streptomycin sulfate (Invitrogen); and 5 mM HEPES. The urea uptake solution contained (in mM) 1 urea, 1 KCl, 1 MgCl 2, 1.8 CaCl 2, 175 mannitol, and 5 HEPES. The uptake solution was supplemented with 2.67 µCi [ 14 C]urea/ml. The stop solution contains (in mM) 10 urea, 1 KCl, 1.8 CaCl 2, 1 MgCl 2, 165 mannitol, and 5 HEPES. The pH of all solutions was adjusted to 7.5 using HCl, NaOH, or KOH as appropriate, and the osmolarity was adjusted to 195 ± 5 mosmol/kgH 2 O using NaCl, H 2 O, or mannitol. HgCl 2 was prepared freshly as a stock solution of 100 mM in H 2 O and added to ND-96 to give a final concentration of 0.5 mM.
Isolation and injection of X. laevis oocytes. Mature female X. laevis were killed using a procedure approved by the University of Sheffield Field Laboratories and in accordance with current UK legislation. The ovarian lobes were removed, and oocytes were isolated as described in depth previously ( 1 ). Briefly, oocytes were isolated from lobes and agitated in calcium-free ND-96 for 60 min. This was followed by two 20-min incubations in calcium-free ND-96 containing 2 mg/ml collagenase type I (Sigma, Poole, UK), separated by a 15-min wash in calcium-free ND-96. Following the second collagenase treatment, the oocytes were washed in calcium-free ND-96 for 30 min. The isolation was completed with a 30-min wash in standard ND-96. The oocytes were transferred to OR3 media and sorted by size and stage. Stage V and stage VI oocytes were selected for injection.
Oocytes were injected with 1 ng of cRNA (50 nl of a 0.02 µg/µl cRNA solution) or an equal volume of water using a Drummond microinjector (Drummond Instruments, Broomall, PA). The oocytes were incubated at 18°C in OR3, and experiments were performed on day 3 or 4 following injection.
Urea uptake. Uptake experiments were performed in 12-well plates using a protocol based on those described previously ( 17 ). Before uptake was measured, oocytes were placed in ND-96. The ND-96 solution was removed, and the uptake solution was added. At the end of the uptake period (90 s), ice-cold stop solution was added to the halt the uptake. This was followed by three rapid washes in ice-cold stop solution. The oocytes were transferred individually to scintillation tubes and dissolved in 10% SDS. The amount of 14 C in each oocyte was measured by scintillation counting. Where necessary, the oocytes were preincubated in ND-96 containing 0.5 mM HgCl 2 for 5 min before the uptake phase. We found that addition of HgCl 2 to the uptake solution caused a precipitate to form; therefore, the uptake measurements were carried out in the absence of HgCl 2.
Preparation of cRNA. In all cases, capped cRNA was synthesized using an Ambion mMessage mMachine kit (Ambion, Huntingdon, UK). The cDNA encoding for fUT was a gift from Dr. G. Rousselet (Institut National de la Santé et de la Recherche Médicale). The gene encoding fUT had previously been subcloned into the pT7TS Xenopus expression vector ( 2 ). The plasmid was linearized using Bam H1, and cRNA was prepared using T7 polymerase. The mUT-A2 and mUT-A3 clones were inserted in the pT7TS Xenopus expression vector. The plasmid was linearized with Not 1, and cRNA was prepared using T7 polymerase.
Production of single- and double-point mutations. Point mutations were introduced into fUT, mUT-A2, and mUT-A3 using the QuikChange protocol (Stratagene). In fUT, the C185S mutation was created by changing TGC to AGC, and the H187Y mutation was created by changing CAC to TAC. In mUT-A2, the T184C mutation was created by changing ACC to TGC, and in the double CH mutation Y186 was mutated from TAC to CAC. In mUT-A3, T254 was mutated by changing ACA to TGC, and in the double CH mutation Y256 was altered from TAC to CAT. The mutations were confirmed by nucleotide sequencing.
Production of fUT/mUT-A3 chimeras. To facilitate the production of chimeras, between fUT and mUT-A3 restriction endonuclease sites were silently engineered into fUT and mUT-A3 using site-directed mutagenesis. The design of the restriction sites was facilitated by the SILMUT software package ( 11 ). A Stu I site was produced at residues 129-131 of fUT and residues 198-200 of mUT-A3. A Nhe I site was engineered in fUT between residues 268 and 270. mUT-A3 already contains an Nhe 1 site between residues 337 and 339. Three chimeras were formed: mUT-A3/fUTM, mUT-A3/fUTF, and mUT-A3/fUTB (see Fig. 5 A for a diagrammatic representation of the composition of each chimera).
Fig. 5. Effect of Hg 2+ on the urea permeability of mUT-A3/fUT chimeras. A : diagrammatic representation of the components used to form each of the 3 chimeras used in this study. B and C : effect of a 5-min incubation in 0.5 mM Hg 2+ on the uptake of urea in oocytes expressing the chimeras mUT-A3/fUTM ( B ) and mUT-A3/fUTF and mUT-A3/fUTB ( C ), respectively. Values are means ± SE, with the no. of observations in parentheses. *Significantly different from the paired control group as judged by ANOVA and subsequent treatment comparison, P < 0.05.
Statistics. Unless otherwise stated, results are presented as means ± SE. Analysis was performed using paired t -tests or ANOVA as appropriate, and significance is assumed at the 5% level. If ANOVA indicated a difference between groups, further comparison was performed using the Student-Newman-Keuls method with InStat (GraphPad Software, San Diego, CA).
RESULTS
Effect of HgCl 2 on urea uptake by oocytes expressing fUT. In initial experiments, we discovered that a 5-min preincubation in 0.5 mM HgCl 2 inhibits urea transport in oocytes expressing fUT ( Fig. 2 A ). The rat and rabbit UT-A isoforms are insensitive to mercurials ( 9 ), but this property has not been investigated for the mUT-A isoforms. A 5-min preincubation in HgCl 2 had no effect on the uptake of urea in oocytes expressing mUT-A2 ( Fig. 2 B ) or mUT-A3 ( Fig. 2 C ).
Fig. 2. Effect of cysteine mutations on urea uptake by fUT, mUT-A2, and mUT-A3. A : uptake of urea by oocytes injected with H 2 O or expressing cRNA encoding for fUT or its C185S mutant in the absence (open bars) or after a 5-min preincubation in 0.5 mM HgCl 2 (filled bars). Values are means ± SE, with the no. of observations in parentheses. *Significantly different from the paired control group as judged by ANOVA and subsequent treatment comparison, P < 0.001. B and C : results of the same protocol performed on oocytes expressing mUT-A2 and its T184C mutant or mUT-A3 and its T254C mutant, respectively.
Effect of HgCl 2 on uptake of urea by oocytes expressing fUT-C185S, mUT-A2-T184C, or mUT-A3-T254C mutants. The alignment in the M-I region between fUT, mUT-A2, and mUT-A3 highlights a cysteine residue at position 185 in fUT, which is absent in mUT-A2 and mUT-A3 ( Fig. 1 ). This C185 residue in fUT was mutated to a serine. The fUT-C185S mutant transported urea ( Fig. 2 A ), but following preincubation in HgCl 2 for 5 min there was no significant reduction in urea uptake.
Cysteine residues were introduced into the equivalent positions in mUT-A2 (T184C) and mUT-A3 (T254C). Introduction of the cysteine residues into UT-A2 and UT-A3 did not effect urea transport ( Fig. 2, B and C ). Preincubation in HgCl 2 for 5 min had no effect on urea uptake compared with paired control experiments.
Role of H187 in the sensitivity of fUT to HgCl 2. The lack of effect of introducing cysteine residues into mUT-A2 and mUT-A3 suggests that there are either significant differences in structure between fUT and mUT-A2/mUT-A3 and/or other residues are involved in the inhibition of fUT by HgCl 2. In fUT, close to C185 there is a histidine residue (H187) that is absent in mUT-A2 and mUT-A3 ( Fig. 1 ). As well as being able to interact with cysteine residues, Hg 2+ can also interact with histidine residues ( 20 ). The fUT-H187Y mutation was produced (tyrosine occupies the equivalent position in mUT-A2 and mUT-A3; see Fig. 1 A ). When expressed in oocytes, the fUT-H187Y mutant transported urea effectively ( Fig. 3 A ), but there was no reduction in urea transport following a 5-min preincubation in HgCl 2 ( Fig. 3 A ).
Fig. 3. Effect of histidine or double cysteine-histidine mutations on Hg 2+ -sensitive urea uptake. A - C : effect of a 5-min preincubation in 0.5 mM Hg 2+ on uptake of urea by oocytes injected with H 2 O or expressing the fUT-H187Y ( A ), the mUT-A2 double T184C-Y186H mutant ( B ), or the mUT-A3 double T254C-Y256H mutant ( C ). Values are means ± SE, with the no. of observations in parentheses. *Significantly different from the paired control group as judged by ANOVA and subsequent treatment comparison, P < 0.05.
Effect of HgCl 2 on urea transport by oocytes expressing the double T184C-Y186H mUT-A2 mutant and the T254C-Y256H mUT-A3 mutant. The results indicate that both Cys185 and His187 are required for Hg 2+ to exert an effect on fUT. As introducing the C184 and C254 residues in mUT-A2 and mUT-A3, respectively, by themselves were not sufficient to introduce Hg 2+ sensitivity, the double mUT-A2-T184C-Y186H and mUT-A3-T254C-Y256H mutants were produced. These mutants were functional ( Fig. 3, B and C ); however, after a 5-min preincubation in Hg 2+, there was no reduction in urea permeability compared with control.
Effect of HgCl 2 on urea transport by oocytes coexpressing the fUT-C185S and fUT-H187Y mutants. C185 and H187 are required for Hg 2+ to have an effect on fUT. In oocytes injected with cRNA encoding C185S or H187Y individually, Hg 2+ had no effect on urea permeability ( Figs. 2 A and 3 A ). However, when oocytes were injected with a mix of cRNAs encoding for the C185S and H187Y mutants (in the ratio 1:1), Hg 2+ sensitivity was reinstated and a 5-min preincubation in HgCl 2 significantly reduced urea uptake ( Fig. 4 ).
Fig. 4. Effect of Hg 2+ on urea uptake in oocytes injected with H 2 O or expressing cRNA encoding for a mixture of the fUT-C185S and fUT-H187Y mutants after a 5-min preincubation in 0.5 mM Hg 2+. Values are means ± SE, with the no. of observations in parentheses. *Significantly different from the paired control group as judged by ANOVA and subsequent treatment comparison, P < 0.001.
Effect of HgCl 2 on fUT/mUTA3 chimeras. The results from double-point mutations indicate that there is a region outside the M-I domain that contributes to the Hg 2+ sensitivity of fUT. To pinpoint this region, we constructed a series of fUT/mUT-A3 chimeras (see Fig. 5 A ). Replacing the central region of mUT-A3 with the equivalent portion of fUT (mUT-A3/fUTM) produced a clone insensitive to Hg 2+ ( Fig. 5 B ), indicating that a region in either the NH 2 or COOH terminal of the protein is required for Hg 2+ sensitivity. Replacing the NH 2 terminal of mUT-A3 in the mUT-A3/fUTM chimera with the equivalent sequence from fUT (mUT-A3/fUTF) did not impart Hg 2+ sensitivity, indicating that the NH 2 terminal of fUT is not involved ( Fig. 5 C ). However, when we replaced the COOH terminal of mUT-A3 in mUT-A3/fUTM with the COOH-terminal region of fUT (mUT-A3/fUTB), we produced a chimera sensitive to Hg 2+ ( Fig. 5 C ). This result suggested that the additional region accounting for Hg 2+ sensitivity lies in the final 121 amino acids of fUT.
DISCUSSION
The goal of the current study was to identify the basis of mercurial-induced inhibition of urea transport in the amphibian urea transporter fUT. Couriaud et al. ( 2 ) demonstrated that fUT is inhibited by the organic mercurial pCMBS. In the current study, we show that fUT is sensitive to HgCl 2 and also that mUT-A2 and mUT-A3, when expressed in X. laevis oocytes, are insensitive to Hg 2+.
The overall sequence similarity among fUT, UT-A2, and UT-A3 is 65%, but in the M-I region the homology is much higher, 90%. Based on our assumption that the M-I region is linked to the formation of the channel pore, and the similarities between this region to pore-forming structures in aquaporins, the M-I region formed the starting point in our search for potential Hg 2+ binding sites. Additionally, when the hydropathy profiles of mUT-A2 and fUT (obtained using the Kyte-Doolittle algorithm) were compared (data not shown), we found that the plots closely matched each other and it is fair to assume that the overall membrane topology of these two transporters is similar. Although there was no clear reason that Hg 2+ could be interacting with this region of fUT, one difference among mUT-A2, mUT-A3, and fUT in the M-I region is the cysteine residue at position 185 in fUT. The fUT-C185S mutant is insensitive to Hg 2+, suggesting that Hg 2+ in some way interacts with the C185 residue and that this residue may lie close to the channel pore (as for the C189 residue in AQP1). However, introducing cysteine residues into the equivalent positions in mUT-A2 (T184C) or mUT-A3 (T254C) did not impart Hg 2+ sensitivity. There are two possibilities for the lack of Hg 2+ sensitivity in the mUT-A2 and mUT-A3 mutants: either more than one residue is involved in the Hg 2+ binding, and/or there are differences in the protein folding that prevent Hg 2+ from accessing the Hg 2+ -sensitive site. The inability to recreate Hg 2+ sensitivity in proteins is not unusual. In the case of the mercurial-insensitive water channel AQP4, introduction of a cysteine residue into a position equivalent to the C189 residue in AQP1 did not reproduce Hg 2+ sensitivity in AQP4 ( 5 ).
To address the possibility that additional residues are involved in the Hg 2+ inhibition of fUT, we reexamined the sequences of fUT, mUT-A2, and mUT-A3 in the M-I region. Another difference between these three sequences is a histidine residue at position 187 in fUT. Hg 2+, as well as binding to cysteine residues, is able to coordinate with histidine residues ( 20 ). In fUT, H187 was also found to be critical for the Hg 2+ -dependent inhibition of urea transport. However, introducing both the cysteine and histidine residues into mUT-A2 or mUT-A3 did not impart Hg 2+ sensitivity. The lack of effect of Hg 2+ in these double CH mutants of mUT-A2 and mUT-A3 strongly suggests that either residues lying outside of the M-I region are involved in sensitivity of urea transport to Hg 2+ or that differences in the folding of mUT-A2/mUT-A3 prevent access of Hg 2+ to the M-I region.
To determine whether regions of fUT outside the M-I domain contribute to Hg 2+ sensitivity, we constructed a series of chimeras between mUT-A3 and fUT. The results indicate that residues in the final 121 amino acids of fUT contribute to Hg 2+ sensitivity. However, a comparison of the COOH-terminal regions of mUT-A2 and mUT-A3 to fUT revealed no residues predicted to interact with mercurials, which were not present in mUT-A2 or mUT-A3. This would suggest that subtle differences in the folding of the COOH terminals of mUT-A2, mUT-A3, and fUT determine the accessibility of crucial residues within the M-I region to mercurials.
The amphibian urea transporter fUT and the mammalian isoform UT-B are sensitive to mercuricals. We have identified two residues which confer Hg 2+ sensitivity within fUT. These two residues, C185 and H187, which lie in the putative integral membrane loop M-I, are absent from the UT-B family and suggest that the basis of sensitivity to Hg 2+ varies between these two transporters. Interestingly, the cysteine and histidine residues responsible for Hg 2+ sensitivity in fUT are preserved in the UTs from other amphibians ( Bufo marinus, GenBank accession no. AB212932 ) and teleost fish ( Takifugu rubripes, AB181946 ; Anguilla japonica, AB049726 ; Alcolapia grahami, AF278537 ; Danio rerio, AY788989 ; Opsanus beta, AF165893 ) but not elasmobranchs ( Dasyatis sayi, AY277796 and AY277793 ; Triakis scyllium, AB094993 ; Squalus acanthias, AF257331 ). We predict that the UTs from amphibians and teleosts would also exhibit Hg 2+ sensitivity, although this remains to be tested.
One surprising result was the rescue of Hg 2+ sensitivity in fUT on the coexpression of the C185S and H187Y mutants. To date, there is no understanding of how the functional unit of urea transport forms. Two possibilities exist: each individual unit acts as a stand-alone urea channel, or, as has been proposed in a recent study, the functional urea channel is a dimer ( 10 ). While our results support this dimer hypothesis, without further studies it is not possible to exclude alternative explanations.
In conclusion, we have identified two residues in the M-I region that are required to confer Hg 2+ sensitivity on the frog urea transporter fUT. Introducing these residues into mUT-A2 or mUT-A3 did not induce Hg 2+ sensitivity in these proteins. Chimera studies indicated that addition residue(s) in the COOH terminal of fUT also contributes to Hg 2+ sensitivity, although the identity of these residues still needs to be determined.
GRANTS
The financial support of the Royal Society and Kidney Research UK is gratefully acknowledged.
ACKNOWLEDGMENTS
We are grateful to Prof. A. Surprenant for helpful comments during the preparation of this manuscript. The fUT clone was a generous gift from Dr. Germain Rousselet (Institut National de la Santé et de la Recherche Médicale, France).
【参考文献】
Cooper GJ and Boron WF. Effect of pCMBS on the CO 2 permeability of Xenopus oocytes expressing aquaporin 1 or its C189S mutant. Am J Physiol Cell Physiol 275: C1481-C1486, 1998.
Couriaud C, Leroy C, Simon M, Silberstein C, Bailly P, Ripoche P, and Rousselet G. Molecular and functional characterization of an amphibian urea transporter. Biochim Biophys Acta 1421: 347-352, 1999.
Fenton RA, Chou CL, Stewart GS, Smith CP, and Knepper MA. Urinary concentrating defect in mice with selective deletion of phloretin-sensitive urea transporters in the renal collecting duct. Proc Natl Acad Sci USA 101: 7469-7474, 2004.
Fenton RA, Flynn A, Shodeinde A, Smith CP, Schnermann J, and Knepper MA. Renal phenotype of UT-A urea transporter knockout mice. J Am Soc Nephrol 16: 1583-1592, 2005.
Jung JS, Bhat RV, Preston GM, Guggino WB, Baraban JM, and Agre P. Molecular characterization of an aquaporin cDNA from brain: candidate osmoreceptor and regulator of water balance. Proc Natl Acad Sci USA 91: 13052-13056, 1994.
Jung JS, Preston GM, Smith BL, Guggino WB, and Agre P. Molecular structure of the water channel through aquaporin CHIP. The hourglass model. J Biol Chem 269: 14648-14654, 1994.
Karakashian A, Timmer RT, Klein JD, Gunn RB, Sands JM, and Bagnasco SM. Cloning and characterization of two new isoforms of the rat kidney urea transporter: UT-A3 and UT-A4. J Am Soc Nephrol 10: 230-237, 1999.
Lucien N, Sidoux-Walter F, Roudier N, Ripoche P, Huet M, Trinh-Trang-Tan MM, Cartron JP, and Bailly P. Antigenic and functional properties of the human red blood cell urea transporter hUT-B1. J Biol Chem 277: 34101-34108, 2002.
Martial S, Olives B, Abrami L, Couriaud C, Bailly P, You G, Hediger MA, Cartron JP, Ripoche P, and Rousselet G. Functional differentiation of the human red blood cell and kidney urea transporters. Am J Physiol Renal Fluid Electrolyte Physiol 271: F1264-F1268, 1996.
Minocha R, Studley K, and Saier MH Jr. The urea transporter (UT) family: bioinformatic analyses leading to structural, functional, and evolutionary predictions. Receptors Channels 9: 345-352, 2003.
Murata K, Mitsuoka K, Hirai T, Walz T, Agre P, Heymann JB, Engel A, and Fujiyoshi Y. Structural determinants of water permeation through aquaporin-1. Nature 407: 599-605, 2000.
Olives B, Neau P, Bailly P, Hediger MA, Rousselet G, Cartron JP, and Ripoche P. Cloning and functional expression of a urea transporter from human bone marrow cells. J Biol Chem 269: 31649-31652, 1994.
Preston GM, Carroll TP, Guggino WB, and Agre P. Appearance of water channels in Xenopus oocytes expressing red cell CHIP28 protein. Science 256: 385-387, 1992.
Preston GM, Jung JS, Guggino WB, and Agre P. The mercury-sensitive residue at cysteine 189 in the CHIP28 water channel. J Biol Chem 268: 17-20, 1993.
Sands JM, Timmer RT, and Gunn RB. Urea transporters in kidney and erythrocytes. Am J Physiol Renal Physiol 273: F321-F339, 1997.
Shayakul C and Hediger MA. The SLC14 gene family of urea transporters. Pflügers Arch 447: 603-609, 2004.
Smith CP, Lee WS, Martial S, Knepper MA, You G, Sands JM, and Hediger MA. Cloning and regulation of expression of the rat kidney urea transporter (rUT2). J Clin Invest 96: 1556-1563, 1995.
Smith CP and Rousselet G. Facilitative urea transporters. J Membr Biol 183: 1-14, 2001.
Stewart GS, Fenton RA, Wang W, Kwon TH, White SJ, Collins VM, Cooper G, Nielsen S, and Smith CP. The basolateral expression of mUT-A3 in the mouse kidney. Am J Physiol Renal Physiol 286: F979-F987, 2004.
Vallee BL and Ulmer DD. Biochemical effects of mercury, cadmium, and lead. Annu Rev Biochem 41: 91-128, 1972.
You G, Smith CP, Kanai Y, Lee WS, Stelzner M, and Hediger MA. Cloning and characterization of the vasopressin-regulated urea transporter. Nature 365: 844-847, 1993.
作者单位：2 Department of Biomedical Science, University of Sheffield, Sheffield; and 1 Faculty of Life Sciences, University of Manchester, Manchester, United Kingdom