Urea inducibility of egr-1 in murine inner medullary collecting duct cells is mediated by the serum response element and adjacent Ets motifs.

The renal medullary solute urea increases transcription and protein expression of the zinc finger-containing transcription factor Egr-1 in a renal epithelial cell-specific fashion. Transient transfection of mIMCD3 cells with a luciferase reporter gene driven by 1.2 kilobases of the murine egr-1 5'-flanking sequence showed 4-fold increase in reporter gene activity with 200 mM urea treatment. The effect of impermeant solutes such as NaCl was much less pronounced, whereas the permeant solute glycerol had no effect. In addition, this same sequence, minus the egr-1 minimal promoter, conferred urea responsiveness to a heterologous (thymidine kinase) promoter. Whereas deletion of two putative AP-1 sites from the sequence had no effect upon urea inducibility, elimination of the five putative serum response elements (SREs) abolished the urea effect. Progressive deletion of the SREs caused a corresponding diminution in urea effect. Two key tandem SREs (SRE-3 and SRE-4), in conjunction with their two adjacent clusters of Ets motifs, were sufficient to confer urea responsiveness to a reporter gene. This response was markedly attenuated in the absence of either cluster of Ets motifs and was abolished if both clusters were deleted. By electrophoretic mobility shift assay, formation of the ternary complex was constitutive and was demonstrable in vitro despite the presence of 200 mosm urea or NaCl. Therefore urea-inducible egr-1 transcription in renal medullary cells is mediated through the SRE and adjacent Ets motifs; ternary complex formation is not inhibited even in the presence of physiological hyperosmolality.

The renal medullary solute urea increases transcription and protein expression of the zinc finger-containing transcription factor Egr-1 in a renal epithelial cellspecific fashion. Transient transfection of mIMCD3 cells with a luciferase reporter gene driven by 1.2 kilobases of the murine egr-1 5-flanking sequence showed 4-fold increase in reporter gene activity with 200 mM urea treatment. The effect of impermeant solutes such as NaCl was much less pronounced, whereas the permeant solute glycerol had no effect. In addition, this same sequence, minus the egr-1 minimal promoter, conferred urea responsiveness to a heterologous (thymidine kinase) promoter. Whereas deletion of two putative AP-1 sites from the sequence had no effect upon urea inducibility, elimination of the five putative serum response elements (SREs) abolished the urea effect. Progressive deletion of the SREs caused a corresponding diminution in urea effect. Two key tandem SREs (SRE-3 and SRE-4), in conjunction with their two adjacent clusters of Ets motifs, were sufficient to confer urea responsiveness to a reporter gene. This response was markedly attenuated in the absence of either cluster of Ets motifs and was abolished if both clusters were deleted. By electrophoretic mobility shift assay, formation of the ternary complex was constitutive and was demonstrable in vitro despite the presence of 200 mosm urea or NaCl. Therefore ureainducible egr-1 transcription in renal medullary cells is mediated through the SRE and adjacent Ets motifs; ternary complex formation is not inhibited even in the presence of physiological hyperosmolality.
Excretion of a concentrated urine is essential for water conservation in mammals; however, the renal concentrating mechanism subjects cells of the renal medulla to extraordinarily high concentrations of NaCl and urea. Urea concentration can exceed 1 M in the medullae of desert rodents adapted for maximal urine concentration to avoid dehydration (1,2). Numerous studies have examined the physiological and biochemical response to hyperosmotic NaCl (reviewed in Refs. 2 and 3). In contrast, hyperosmotic urea has received comparatively little attention. Unlike NaCl, urea is readily membrane-permeant and hence has traditionally been viewed as playing a relatively passive role in renal cell function. Recently, however, urea was shown to be a potent effector of gene expression. Urea, in concentrations physiologically relevant in vivo only to the renal medulla (i.e. 100 -400 mM range) promptly up-regulated expression of several immediate-early gene transcription factors at the mRNA and protein levels (4,5). This phenomenon occurred specifically in cells of renal epithelial origin (5). Moreover, one of these genes, egr-1, was transcriptionally up-regulated, thereby classifying it as the only eukaryotic gene described to date to be transcriptionally responsive to hyperosmotic urea (4). The trans-acting factors and cis-elements involved in this process are unknown.
The SRE consensus sequence, CC(A/T) 6 GG, is bound by the serum response factor (SRF) protein (15). An additional family of nuclear proteins, termed ternary complex factors (TCFs), interact with the SRE⅐SRF complex but likely not the SRE alone (16). Ets motif sequences (core sequence GGA (17)) adjacent to many native SREs (e.g., in the c-fos promoter) bind TCFs and thereby permit TCF⅐SRF interaction. The aminoterminal Ets domain of the TCF proteins (e.g., SAP-1 (18) and Elk-1 (19,20)) confers Ets motif DNA-binding specificity; SRF-TCF interaction is mediated through an additional conserved structural motif (17). Post-translational modification of TCFs by members of the mitogen-activated protein kinase family regulate SRE/Ets-mediated transcription (21)(22)(23)(24)(25). Alternatively, lysophosphatidic acid, a constituent of serum, may act upon SRF via a small G-protein-dependent mechanism to mediate Ets-independent transcription from the SRE (26). In the present study, we investigated the 5Ј-flanking sequence of the murine egr-1 gene to identify the cis-elements responsible for conferring urea inducibility to this key regulatory gene in cultured murine renal inner medullary cells.

EXPERIMENTAL PROCEDURES
Cell Culture and Solute Treatment-mIMCD3 cells were maintained in Dulbecco's modified Eagle's medium/Ham's F-12 medium (Life Technologies, Inc.) supplemented with 10% fetal bovine serum (JRH) as described previously (4). Cells were growth-suppressed in Dulbecco's modified Eagle's medium/Ham's F-12 without serum for 24 h prior to treatment with medium supplemented with hyperosmotic solute to a final concentration of 200 mosm as described previously (27); final total osmolality, unless otherwise noted, was approximately 510 mosm.
Promoter Deletion Mutants-Full-length (1.2 kilobase pairs) egr-1 5Ј-flanking sequence was excised as a SalI fragment from construct #632 (pEgr-1p1.2; from V. P. Sukhatme, Beth Israel Hospital, Boston, MA) and ligated into the polylinker SalI site upstream of the luciferase reporter gene in the promoterless vector pXP2 (28) to create Egr-1(1.2)Luc (A in Fig. 5). This construct includes all sequence data listed under EMBL accession number X12617 (13). An additional polymerase chain reaction product including the sequence from nucleotides 88 and 1013 (inclusive) in EMBL sequence accession number X12617 was amplified and ligated into HindIII/BglII-digested pXP2 to control for the additional 3Ј sequences present in construct A but absent from other deletion mutants (see below); an amplification product of sequences from approximately 1 through 1000 could not be obtained despite multiple primer combinations. In all transfection experiments, the resultant construct (88 -1013) behaved identically with construct A. Truncation and deletion mutants were prepared using restriction site-tailed commercially synthesized oligodeoxynucleotides (Life Technologies, Inc.) in the polymerase chain reaction. Constructs B through G represent amplification products cloned into SalI/XhoI-digested pXP2 (B, C, and D) or HindIII/BglII-digested pXP2 (E, F, and G). Construct B includes sequences from 1 through 386 and from 886 through 1000 (amplified from construct #696 from V. P. Sukhatme); C includes sequences from 511 through 1000; D includes from 511 through 686 and from 886 through 1000 (amplified from construct #668 from V.P. Sukhatme); E includes from 811 through 1000; F includes from 845 through 1000; and G includes from 866 through 1000.
Polymerase chain reactions (MJ Research) were performed in a 25-l volume with 100 ng of each primer/reaction, 100 ng of template, 0.  Fig. 6A). Fragments were subcloned into BamHI/HindIIIdigested construct G (containing enhancerless egr-1 minimal promoter, as described above) or upstream of the thymidine kinase promoter in BamHI/HindIII-digested PT109 luciferase reporter vector (28). Sequence from nucleotides 1-905 was amplified and cloned into HindIIIdigested PT109 to create Egr-1TKLuc (see Fig. 4). (N.B. to convert from accession sequence position number to nucleotide position relative to the murine egr-1 transcriptional start site, subtract 936 bp.) Transient Transfection and Reporter Gene Analysis-For transfection via electroporation, mIMCD3 cells were grown to 80 -90% confluence, trypsinized, and resuspended in warmed complete medium. They were then pelleted at 1000 ϫ g for 5 min, washed with ice-cold Dulbecco's modified Eagle's medium/Ham's F-12, repelleted, and resuspended in ice-cold Dulbecco's modified Eagle's medium/Ham's F-12 at a working concentration of approximately 5 ϫ 10 6 cells/ml. Cell suspension (0.5 ml) was added to 20 g of luciferase reporter plasmid and 5 g of the CMV-Gal vector (for normalization) in ice-cold electroporation cuvettes (Invitrogen), incubated on ice for 10 min, electroporated at 1000 microfarad and 300 V (GenePulser, Bio-Rad), incubated on ice for 10 min, diluted 1:20 with warmed complete medium, and plated. After 24 h, cells were taken out of serum; at 48 h, cells were treated for 6 h with the desired condition prior to harvest for determination of reporter gene activity. ␤-Galactosidase activity was determined using standard methods (29). To measure luciferase activity, individual wells of six-well plates were washed with ice-cold phosphate-buffered saline, and lysed with 150 l of luciferase lysis buffer (125 mM Tris, pH 7.6, 0.5% Triton X-100). Lysate (100 l) was incubated with 200 l of 5 mM ATP in luciferase buffer (25 mM glycylglycine, 15 mM MgSO 4 , pH 7.8) and 100 l of luciferin (60 g/ml; Analytical Luminescence) in luciferase buffer in an automated luminometer (Berthold), counted for 30 s, and normalized to ␤-galactosidase activity. Data are expressed as the means Ϯ S.E., except where noted; statistical comparison was achieved with ANOVA (StatView), wherein significance was ascribed to a p Ͻ 0.05 via Scheffe F-test.
EMSA-EMSA was performed as described previously (4). Oligonucleotides encoding the murine c-fos SRE (top strand: GATCCAGGAT-GTCCATATTAGGACATCTA) were commercially synthesized, annealed, and gel-purified prior to end-labeling with T4 polynucleotide kinase in the presence of [␥-32 P]ATP. Whole cell lysates were prepared as described previously (4). Competition experiments were performed using excess cold double-stranded oligonucleotide encoding the native c-fos SRE (SRE n ) or excess cold oligonucleotide encoding an SRE mutant (SRE m ) with which SRF interacts poorly (same sequence as above, absent the bold T (14,30,31)).

RESULTS
A 1.2-kilobase fragment of the murine egr-1 5Ј-flanking sequence conferred urea responsiveness to a luciferase reporter gene when transiently transfected into cells of the renal medullary mIMCD3 line, as evidenced by urea-inducible luciferase activity at 6 h of treatment (Fig. 1). In addition, the potent egr-1 inducers, fetal bovine serum (20%) and O-tetradecanoylphorbol 13-acetate (1 M), also up-regulated reporter gene activity, albeit to a lesser extent than hyperosmotic urea. The magnitude of the induction was consistent with that observed at the transcriptional, mRNA, and protein levels in response to urea treatment in these and other renal epithelial cells (4,5). There was a dose-dependent response to urea until 800 mM urea (Fig.  2), at which point luciferase activity was markedly suppressed. Cells remained viable in the presence of 800 mM urea for 6 h, as evidenced by trypan blue exclusion and persistent adherence to the plate; there was no gross morphological change detected. Not shown, urea had no effect upon expression of a luciferase reporter gene driven by tandem repeats of the previously identified NaCl-responsive element from the betaine transporter gene (32) subcloned upstream of the thymidine kinase promoter.
To assess the urea specificity of solute-inducible egr-1 expression, a panel of osmotically active and inactive solutes (added in equiosmolar concentration) was investigated (Fig. 3). The only other highly permeant solute examined, glycerol, failed to up-regulate luciferase activity. In contrast, all of the poorly permeant organic solutes (mannitol, sorbitol, sucrose, and glucose), as well as the functionally impermeant inorganic solute NaCl, increased egr-1 transcription less than 2-fold. Structural analogs of urea were also investigated. Thiourea (200 mM) was extremely toxic to cells and resulted in loss of adherence and viability. Double methyl-substituted urea ana- logs (e.g., 1,1-and 1,3-dimethylurea) exhibited less activity than the mono-substituted N-methyl urea; none was as potent as urea.
To confirm that urea responsiveness was a consequence of enhancer activity and not an intrinsic property of the egr-1 promoter, the 5Ј-flanking sequence, devoid of the egr-1 minimal promoter (13), was cloned upstream of the heterologous thymidine kinase promoter in the PT109 luciferase vector (28). In addition to increasing basal reporter gene activity, transfection of this construct into mIMCD3 cells resulted in marked ureainducible reporter gene activity (Fig. 4), suggesting that urea inducibility was a function of upstream enhancer elements and not a general urea effect operating at the egr-1 minimal promoter.
To "map" the urea-responsive site(s) in the egr-1 5Ј-flanking sequence, deletion and truncation analyses were performed (Fig. 5). The initial series of deletion/truncation mutants (constructs A, B, and C) was designed to permit comparison of 1.2 kilobases of the native murine egr-1 5Ј-flanking sequence (containing all enhancer sequences (A)) with mutant sequences devoid of putative SREs (B) or AP-1 sites (C). Urea inducibility was well preserved in the absence of AP-1 sites (C) but in the absence of SREs (B) was essentially at the level of induction seen with only the minimal egr-1 promoter (G). Further deletional analysis of the 5-SRE-containing proximal 5Ј-flanking sequence (constructs D, E, F, and G) revealed a diminution in urea inducibility with progressive loss of SREs. Of note, all deletion constructs contained the minimal egr-1 promoter, as present in construct G.
GGA (Ets) motifs adjacent to SREs have been implicated in mediating a component of their enhancer activity (17). Two of the murine egr-1 SREs (SRE-3 and SRE-4) were suspected of conferring much of the inducible activity in the present study; this was observed by McMahon and Monroe (33) in the context of egr-1 transcription in response to antigen receptor crosslinking in B cells. Therefore, the contribution of Ets motifs to the urea responsiveness of the murine egr-1 promoter in mIMCD3 cells was investigated utilizing native sequences in the vicinity of SRE-3 and SRE-4. A cluster of three GGA motifs resides within 30 bp upstream of SRE-4, and a cluster of two GGA motifs is found within 40 bp downstream of SRE-3. Sequences including neither, either, or both of these clusters of Ets motifs, in conjunction with the intervening SREs (Fig. 6A), were examined for urea responsiveness upstream of the egr-1 minimal promoter (Fig. 6B) and the heterologous thymidine kinase promoter (Fig. 6C). The pair of SREs alone (E min 2-3 ( Fig. 6B) and TK 2-3 (Fig. 6C)) conferred no urea responsiveness, whereas the SREs in conjunction with both clusters of Ets motifs (E min 1-4 ( Fig. 6B) and TK 1-4 (Fig. 6C)) were modestly responsive. The presence of only a single cluster of Ets motifs, either 5Ј (E min 1-3 (Fig. 6B) and TK 1-3 (Fig. 6C)) or 3Ј (E min 2-4 ( Fig. 6B) and TK 2-4 (Fig. 6C)) to the pair of SREs, conferred intermediate urea responsiveness. For either empty pXP2 or PT109 vector alone, urea exerted no effect (not shown).
Urea, at high concentrations, is a potent denaturant of protein and nucleic acid and is a broadly acting competitive inhibitor of enzyme function; both effects can be noted in some circumstances in the 100 mM range (1). For these reasons, the ability of an SRE/Ets motif to support ternary complex formation despite the presence of several hundred millimolar urea was essential to establish. The human c-fos SRE in conjunction with its adjacent Ets motif comprises the best studied model of TCF formation. Therefore, this sequence was commercially synthesized and end-labeled for use in the EMSA. Whole cell lysates prepared from mIMCD3 cells treated with control medium or solute-supplemented (ϩ200 mosm) medium were incubated with the labeled oligonucleotide and resolved via nondenaturing polyacrylamide gel electrophoresis. A reproducible complex was noted in the presence but not the absence of lysate (Fig. 7). With urea treatment, formation of the putative ternary complex was neither up-nor down-regulated, suggesting a constitutive interaction. Specificity was demonstrated by the   FIG. 2. Effect of medium supplemented with increasing concentrations of urea upon egr-1 transcription. mIMCD3 cells were transfected as in Fig. 1 and treated with 0, 100, 200, 400, or 800 mM urea for 6 h. ability of excess of cold (unlabeled) native c-fos SRE/Ets to strongly compete for binding, whereas excess mutant SRE/Ets oligonucleotide (with which SRF interacts poorly (30,31,34)) exhibited much weaker competitive activity (Fig. 8). DISCUSSION These data indicate that the urea inducibility of egr-1 expression that occurs in a renal epithelial cell-specific fashion (5) is likely mediated at the transcriptional level by multiple SREs and their adjacent Ets motifs contained within the egr-1 promoter. Consistent with this observation, ternary complex factor formation occurs in vitro despite the presence of markedly elevated (e.g., 200 mosm) solute concentration. This is the first example of transcriptional activation of a renal solute-responsive gene to be defined at the molecular level. Thus far, egr-1 is the only eukaryotic gene known to be transcriptionally regulated by hyperosmotic urea, and the mammalian renal medulla is the only mammalian tissue physiologically exposed to an elevated urea concentration.
After gross promoter deletion analysis, we examined in detail two native adjacent murine egr-1 SREs (SRE-3 and SRE-4) FIG. 6. Adjacent Ets motifs are essential for urea-inducible SRE-mediated transcription. A, primers 1-4 (depicted diagrammatically) were used in polymerase chain reaction to amplify the native sequence in the region of the murine egr-1 SRE-3 and SRE-4 (1-4) and to construct deletion mutants devoid of one (1-3 and 2-4) or both (2-3) of the 5Ј and 3Ј clusters of Ets motifs (filled ovals); all were subcloned upstream of the minimal egr-1 promoter (plasmid G in Fig. 5) in PXP2 or the thymidine kinase promoter in vector PT109 (28). B and C, luciferase activity (relative to ␤-galactosidase activity) of mIMCD3 cells transfected with the indicated Ets motif deletion mutants upstream of the egr-1 minimal promoter (B) or the thymidine kinase promoter (C) in the presence (filled bars) and the absence (open bars) of 200 mM urea. The data are the means Ϯ S.E. of at least three wells.
with strong transcriptional activity in other contexts (33) to assess their ability to support urea-inducible transcription. This pair of elements, in isolation, did not confer urea inducibility to either the egr-1 minimal promoter or the heterologous TK promoter (Fig. 6). MacMahon and Monroe attributed much of the transcriptional activity of these two elements to the adjacent Ets motifs situated immediately 5Ј of SRE-4 and 3Ј of SRE-3 (Fig. 6A) (33). Accordingly, inclusion of both the upstream cluster of three Ets motifs and the downstream cluster of two Ets motifs conferred a modest degree of urea responsiveness to the tandem SRE-driven reporter gene construct. It should be emphasized that the magnitude of this response is less than that seen with the entire 5Ј-flanking sequence, which contains all five SREs. The absence of both clusters of Ets motifs eliminated urea responsiveness, whereas inclusion of either group of Ets motifs alone conferred intermediate urea responsiveness. These data, in conjunction with the mapping data depicted in Fig. 5, strongly suggested that SREs and their adjacent Ets motifs mediated urea-inducible egr-1 transcription. Importantly, although ternary complex formation was not inhibited by hyperosmotic solute, it was also not up-regulated, as might have been expected. This pattern of constitutive binding was consistent with most (35)(36)(37)(38) but not all (38, 39) models of SRE-mediated transcription.
The only other documented renal example of SRE-mediated egr-1 induction occurs in response to the mitogen plateletderived growth factor in glomerular mesangial cells (40). Other examples of mitogen-related egr-1 expression that have been shown to be mediated through the SRE include induction by serum and peptide growth factors (12,14) and by v-src (41) and v-raf (42) in 3T3 cells. Interestingly, urea-treated renal epithelial Madin-Darby canine kidney cells exhibit several hallmarks of mitogenesis (e.g., increased [ 3 H]thymidine incorporation and total DNA content); actual cell proliferation, however, does not occur (43). Other nonmitogenic stimuli, such as x-irradiation and exposure to reactive oxygen intermediates, also mediate egr-1 induction via the SRE (44,45). In contrast, other stimuli mediate all or part of their effect upon egr-1 transcription through DNA consensus elements other than the SRE, including the AP-1 site (46) and the cyclic AMP response element (47).
Attribution of urea inducibility to the SRE and associated Ets sequences provides important clues to the signaling pathway utilized by urea in the renal medulla. Multiple signaling pathways impinge upon the ternary complex. The MAP kinases, extracellular signal-regulated kinases 1 and 2, phosphorylate the TCF, Elk-1, in response to mitogen treatment as well as other ras-dependent and -independent stimuli (21)(22)(23)(24). Similarly, the stress-responsive mitogen-activated protein kinase, stress-activated protein kinase/jun kinase, can also phosphorylate Elk-1 (25). In addition, transcriptional competence of SRF (independent of the TCF⅐Ets interaction) is regulated by the serum constituent, lysophosphatidic acid, through a small G-protein-dependent pathway (26). Whereas the latter two pathways are completely unexplored in the context of solute signaling and the kidney, an extracellular signal-regulated kinase-dependent pathway has previously been implicated. Itoh et al. (48) and Terada et al. (49) showed that functionally impermeant solutes could activate mitogen-activated protein kinases of the appropriate molecular mass for extracellular signal-regulated kinases in Madin-Darby canine kidney cells, but these events were not linked to regulation of downstream genes. Each of these pathways could potentially mediate urea action through the SRE⅐Ets complex.
The effect of urea structural analogs upon egr-1 transcription is provocative. Although thiourea is an effective competitor for urea interaction with the urea transporter in vitro (50), it was extremely toxic to renal epithelial cells at the concentration and duration required for urea to be an effective reporter gene inducer. Of the other analogs examined, the dimethyl-substituted analogs (e.g., 1,1-and 1,3-dimethylurea) were less potent agonists than the mono-substituted N-methylurea. Whether this is a consequence of steric interference of the methyl groups with a cytoplasmic or cell surface urea-sensing molecule or whether this is a consequence of diminished entry into the cell has not been established.
Transcriptional regulation by the other principal renal medullary solute, NaCl, is also under active investigation. Transcription of multiple genes involved in the synthesis and transport of organic osmolytes, including the genes encoding the enzyme aldose reductase (51), as well as the Na ϩ /betaine (32) and Na ϩ /myo-inositol (52) cotransporters, is up-regulated in response to treatment with impermeant solutes such as NaCl or raffinose. In the case of the betaine transporter gene, solute inducibility has been attributed to a unique oligonucleotide enhancer sequence (32) that failed to confer urea responsiveness to a reporter gene and heterologous promoter in the present study. Thus far, the SRE or Ets motif has not been implicated in impermeant solute-mediated transcription. The data presented here suggest a small effect of impermeant solutes upon egr-1 (and SRE/Ets-mediated) transcription, but the effect was considerably less than that seen with hyperosmotic urea. Nonetheless, even a small effect upon reporter gene expression is somewhat surprising, because impermeant solutes in the concentrations used generally inhibit protein synthesis in renal epithelial cells (27).
That distinct signaling pathways likely exist for hyperosmotic urea and impermeant solutes should not be surprising, because the physiological responses to these stressors are distinct as well. Relatively membrane-impermeant solutes such as NaCl and raffinose induce an acute change in cell volume, whereas the membrane-permeant solute, urea, does not (2). In FIG. 8. Competition experiments to confirm specificity of putative ternary complex formation in Fig. 7. EMSA performed with whole cell lysates prepared from control mIMCD3 cells and radiolabeled double-stranded oligonucleotide encoding human c-fos SRE and adjacent Ets motif. The filled arrowhead represents the specific complex in the presence of 10ϫ or 100ϫ excess cold (unlabeled) native c-fos SRE/ Ets (SRE n ) or 10ϫ or 100ϫ excess cold mutant c-fos SRE/Ets (SRE m ), with which SRF interacts poorly (see "Experimental Procedures"). addition, in contrast to NaCl, treatment with hyperosmotic urea results in only a partial osmolyte response (2). Interestingly, although intramedullary concentrations of NaCl and urea tend to move in concert, some physiological and pathophysiological states result in preferential accumulation of renal medullary urea. Examples include dehydration in the normal rat and anti-diuretic hormone replacement in the congenitally anti-diuretic hormone-deficient Brattleboro rat (53). Under such circumstances, hyperosmotic urea-inducible signaling and gene regulation may play a particularly prominent role.
In conclusion, the renal epithelial cell-specific induction of egr-1 transcription by hyperosmotic urea is a consequence of the multiple SREs and their adjacent Ets motifs in the murine egr-1 promoter. How urea and urea structural analogs signal to these promoter elements and whether a cell surface receptor or cytoplasmic urea sensor is involved remain to be explored. The present series of observations affords an ideal model system for the investigation of signaling to urea-inducible transcription and provides a framework, through promoter analysis, for a rational approach to the identification of other candidate urearesponsive genes.