Identification of a Novel cis-Element Required for the Constitutive Activity and Osmotic Response of the Rat Aldose Reductase Promoter*

A new and essential cis-element AEE (aldose reductase enhancerelement), necessary for the constitutive activity and the osmotic stress response of rat aldose reductase transcription in a rat liver cell line, has been identified. In transient transfection assays, an increase in promoter activity, up to 3.8-fold, was observed with osmotic stress (600 mosm/kg H2O) using a luciferase reporter gene construct containing aldose reductase promoter sequence from −1,094 base pair (bp) to +23 bp. A deletion between −1,071 and −895 bp reduced the constitutive activity and abolished the osmotic response of the promoter. Exonuclease III mediated in vivoDNA footprinting and dimethyl sulfate in vivo footprinting revealed DNA protection of a 32-bp region and two guanosines (G) within this region protected from methylation, respectively. Electrophoretic gel mobility shift assays using whole liver cell extracts showed protein binding, under both normal and stressed conditions. Deletion of the sequence between the two guanosines protected by in vivo dimethyl sulfate DNA footprinting (GAAGAGTG) in a luciferase construct (−1,094 bp to +23 bp) abolished the constitutive promoter activity. One copy of AEE fused to the thymidine kinase promoter gave a maximum constitutive activity of 7.7-fold and a maximum osmotic response activity of 6.7-fold.

The elevation of renal medullary extracellular NaCl and urea during antidiuresis has been known to be adjusted intracellularly by accumulation of osmolytes such as sorbitol, myoinositol, betaine, and glycerophosphorylcholine. Aldose reductase (AR), 1 an enzyme which catalyzes the reduction of various sugars to alcohols, including glucose to sorbitol, is highly expressed in kidney inner medulla (1,2), and is believed to play a role in normal renal osmoregulation. A similar role for AR has been reported in plants that accumulate sorbitol during seed development (3). Contrary to the role of AR in osmoregulation, AR-mediated accumulation of sorbitol from excess glucose in diabetic patients has been suggested to be one of the triggering events in diabetic complications, including cataract, retinopathy, neuropathy, and nephropathy (4 -6). The inhibition of this enzyme in diabetic animal models, such as galactosefed rats has been successful in retarding or preventing such complications (7).
Elevation of AR enzymatic activity and protein synthesis in cultured rabbit renal inner medulla cells (PAP-HT25) exposed to hypertonic medium was first reported by Moriyama et al. (2) followed by additional findings that AR induction by hypertonic stress also occurs in other cell types and species by Kaneko et al. (8). Wang et al. (9) analyzed the human AR promoter up to Ϫ609 bp from the transcriptional start site using HepG2 cells (9). DNA protection in a CGGAA(A/G) motif (Ϫ186 to Ϫ146), and a GC-rich region (Ϫ87 to Ϫ31) was observed by in vitro DNase I footprinting, but no response to osmotic stress was observed with this promoter region in transfection studies. Ferraris et al. (10) observed a 40-fold increase in promoter activity with a Ϫ3.6 kilobase to ϩ34-bp rabbit AR construct under osmotic stress (10). Three recent reports describe the involvement of a tonE-like osmotic response element of the human AR promoter at position Ϫ1,235 bp (11), the rabbit AR promoter at Ϫ1,108 bp (12), and the mouse AR promoter at Ϫ1,053 bp (13). A signal transduction study revealed that the p38 or SAPK/JNK pathway are not necessary for the transcriptional regulation of the AR promoter through osmotic response element (14).
In this paper, we describe the identification of a novel ciselement necessary for the constitutive activity and the osmotic response activity of the rat AR promoter which is located in the proximity of the previously described tonE-like element. This novel AEE element is shown to be occupied in vivo in both transiently transfected templates and in the endogenous promoter, suggesting that a stable association with transcription factors at this site is required for effective rat AR promoter activity.
Cell Culture, Transfection, and Osmotic Shock-A normal rat liver cell line (Clone 9, ATCC, Rockville, MD) which expresses AR and responds to osmotic stress (data not shown) was cultured in 35-mm * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
In Vivo Exonuclease III-mediated Footprinting-Exonuclease III (Ex-oIII)-mediated footprinting was carried out as described by Archer et al. with slight modification (17)(18)(19). Cells were cultured in 10-cm diameter dishes up to 60% confluency and were then transfected with 10 g of the luciferase construct 4, which contains 1,094 bp of upstream AR promoter sequence. After 6 h, the cells were washed with medium and grown an additional 48 h. Cells were scraped and washed with phosphate-buffered saline followed by suspension in 5 ml of homogenization buffer (10 mM Tris-HCl, pH 7.4, 15 mM NaCl, 60 mM KCl, 1 mM EDTA, 0.1 mM EGTA, 0.1% Nonidet P-40, 0.15 mM spermine, 0.5 mM spermi-dine, 5% sucrose). The suspension was homogenized in a Dounce homogenizer and 1 ml of sucrose solution (10 mM Tris-HCl, pH 7.4, 15 mM NaCl, 60 mM KCl, 0.15 mM spermine, 0.5 mM spermidine, 10% sucrose) was added. After centrifugation at 1,400 ϫ g for 20 min the pelleted nuclei were washed once with wash buffer (10 mM Tris-HCl, pH 7.4, 15 mM NaCl, 60 mM KCl, 0.15 mM spermine, 0.5 mM spermidine). Isolated nuclei (4.5 ϫ 10 5 ) were resuspended in 200 l of enzyme digestion buffer (10 mM Tris-HCl, pH 7.4, 15 mM NaCl, 60 mM KCl, 0.1 mM EDTA, 5 mM MgCl 2 , 5% glycerol, 1 mM dithiothreitol) and treated at 30°C for 15 min with 200 units of Asp718 or HindIII restriction enzyme (Boehringer Mannheim) and 650 units of ExoIII (Boehringer Mannheim). One ml of proteinase K buffer (10 mM Tris-HCl, pH 7.6, 10 mM EDTA, 0.5% SDS, 0.2 mg/ml proteinase K) was added and incubated at 37°C for 2 h to digest the nuclei. DNA was purified by phenol/chloroform extraction followed by ethanol precipitation. The DNA was resuspended and treated at 30°C for 30 min with 50 units of mung bean nuclease (Boehringer Mannheim) in mung bean buffer (50 mM NaOAc, pH 5, 30 mM NaCl, 1 mM ZnSO 4 ). Twenty micrograms of isolated DNA was used for linear PCR. Primers used for linear PCR were 5Ј-AAGCATGAC-CCAGCAGAAGGAGA-3Ј (Ϫ974 bp to Ϫ952 bp, primer 1) for the coding strand and 5Ј-AGTTGCCCCAAGAACAATGGCGGAA-3Ј (Ϫ877 bp to FIG. 1. Analysis of rat AR promoter constructs in transfected rat liver cells. A, scheme of rat AR promoter constructs used in the transfection experiments. B, the luciferase activity (relative luciferase activity) from rat liver cells transfected with constructs 1 to 9 and subsequently grown in isotonic medium (nonstress, open bar) or hypertonic medium (stress, filled bar) for 18 h. Error bars indicate standard deviation (S.D.). C, luciferase activity Ϯ S.D. under isotonic or hypertonic conditions. pX/p1 indicates the increase over p1 luciferase activity. Ratio H/I indicates the ratio between hypertonic and isotonic activity; n indicates the number of replicates.
Ϫ901 bp, primer 2) for the noncoding strand. For control experiments construct 4 was digested with HindIII and linear PCR was performed with primer 1 (Fig. 2, lane 1) or the construct 4 was digested with Asp718 and linear PCR was performed with primer 2 (Fig. 2, lane 3). Linear PCR product were resolved in a 6% sequencing gel (Sequagel-6, National Diagnostics, Atlanta, GA).
Whole Cell Extract Preparation for Electrophoretic Mobility Shift Assay (EMSA)-Whole cell extracts were prepared as described previously (22). Rat liver cells were cultured in three 10-cm diameter dishes to 90% confluency. The cells were then grown in normal medium or hypertonic medium for 30 min or 3 h. After stress, cells were pelleted, washed with phosphate-buffered saline, and resuspended in 2 packed cell volumes of Buffer C (20 mM Hepes, pH 7.9, containing 0.42 M NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA, 0.5 mM dithiothreitol, 25% (v/v) glycerol, 2 mM proteinase inhibitor 4-(2-aminoethyl)-benzenesulfonyl fluoride hydrochloride). The cells were snap frozen in ethanol/dry ice. Frozen cells were quickly thawed for homogenization and centrifuged for 1 h at 34,000 ϫ g. The supernatant was used for EMSA without dialysis.

Identification of a Region Controlling the Constitutive Transcription and Osmotic Response of the Rat AR Promoter-Var-
ious luciferase reporter constructs of the rat AR promoter were made and transfected into rat liver cells to measure the promoter activity under normal and hypertonic conditions (Fig. 1). Rat AR constitutive promoter activity was constant for constructs 1 (Ϫ3,104 to ϩ23 bp) and 2 (Ϫ2,371 to ϩ23 bp), while further deletions increased the activity up to construct 5 (Ϫ1,052 to ϩ23 bp) which gave the highest constitutive activity. Deletion to construct 6 (Ϫ895 to ϩ23 bp) had a significant effect on the reduction of constitutive promoter activity.
Induction of promoter activity by osmotic response of construct 1 was 4.4-fold while deletion to construct 2 resulted in an increase of 11.8-fold (Fig. 1). Further deletion to construct 4 (1,094 to ϩ23 bp) gave the highest absolute promoter activity by osmotic response (100.00 under hypertonic conditions). Construct 5 (Ϫ1,052 to ϩ23 bp) which lacks the tonE element still  ity and the osmotic response capability. These results indicated the presence of several regions involved in modulating the AR promoter activity and suggest the presence of an element(s) crucial for the constitutive activity and osmotic response located between Ϫ1,071 and Ϫ895 bp.
In Vivo Factor Binding to the Region Ϫ926 to Ϫ895 bp-Based on the luciferase gene expression experiments with construct 4, we tested transcription factor binding between Ϫ1,094 and Ϫ649 bp by using the ExoIII-mediated in vivo DNA footprinting method. This technique has been used to detect constitutive and hormone-induced factor binding to the mouse mammary tumor virus long terminal repeat (19), and is based on the fact that transcription factors upon binding, protect DNA from digestion with ExoIII nuclease. Under normal osmotic conditions, transient templates (construct 4) in isolated nuclei were digested with ExoIII and Asp718 (to supply an entry point for ExoIII) in the sense-directed digestion or with ExoIII and HindIII (to supply an entry point for ExoIII) in the antisense-directed digestion. A strong block to ExoIII digestion was detected at Ϫ926 bp for the sense strand-directed digestion and at Ϫ895 bp for the antisense strand-directed digestion (Fig.  2).
Constitutive Occupancy of the Endogeneous AR Promoter in the Region Protected by ExoIII in Vivo DNA Footprinting-To observe whether the ExoIII protected region (Ϫ926 to Ϫ895 bp) was also protected in the endogeneous gene, dimethyl sulfate in vivo footprinting was performed (20). Liver cells were grown in normal or hypertonic medium for 30 min or 3 h. The cells were then treated for 90 s with 0.1% dimethyl sulfate. As a control, DNA extracted from cells and treated with 0.1% dimethyl sulfate in vitro for 90 s was also analyzed by ligation-mediated PCR (LM-PCR) (Fig. 3). When the intensity of the bands be- tween the in vitro lane and the in vivo lanes were compared, two guanosines, at positions Ϫ915 and Ϫ908 bp, were strongly and reproducibly protected from methylation in cells grown in both normal or hypertonic medium as shown in comparison between lanes 4 and lanes 1-3. The relative intensity of the two guanosine bands was similar for normal and osmotic stress. This result confirms that a factor(s) binding to the region Ϫ926 to Ϫ895 bp occurs in vivo, and shows that occupancy of this site is constitutive in the endogeneous rat AR promoter.

Mutation of Nucleotides between the Two Guanosines Protected by in Vivo Footprinting Inhibits Binding of Transcription
Factor-DNA probes were used to further evaluate putative transcription factor binding at Ϫ926 and Ϫ895 bp in the rat AR promoter. Probe A, which encompasses the 32-bp of sequence protected by ExoIII-mediated footprinting (Fig. 2), gave several complexes in EMSA experiments (Fig. 4, lanes 2-4). These complexes were similar in normal and hypertonically-treated cells. In competition experiments, cold probe A caused inhibition of band complex formation (Fig. 4, lanes 5-7). Probe A8, which contains eight mutations, including the two guanosine nucleotides protected from methylation by dimethyl sulfate in vivo footprinting, did not show significant binding of the same factors (Fig. 4, lanes 8 -10). This result was confirmed by a competition assay with cold mutated probe A8 which did not compete with wild type probe A (Fig. 4, lanes 11-13).
The AEE Element Confers Osmotic Inducibility and Enhances the Constitutive Activity of Thymidine Kinase Promoter-The effect on constitutive activity and osmotic response of the thymidine kinase (TK) promoter (Ϫ105 to ϩ52 bp) was studied by fusing a portion of 35-bp ExoIII-protected region to the 5Ј-end of the TK promoter (Fig. 5). Construct pAEE35-TK, which contains the full 35-bp sequence of AEE, gave a 4-fold increase in basal activity as well as a 5.3-fold increase with osmotic stress. Construct pAEE28-TK (28 bp) which lacks 7 nucleotides of AEE gave the highest basal activity (7.7-fold) and the highest absolute activity under osmotic stress (100.00). Construct pAEE24-TK, which lacks 11 bp, was found to be the minimum required sequence for the functional induction of basal promoter activity and for the osmotic response. The constructs, pAEE20-TK and pTK, gave similar basal activity and no osmotic response. DISCUSSION The transfection studies of the AR promoter region between Ϫ1,071 and Ϫ895 indicate important elements necessary for the increase of AR constitutive promoter activity and osmotic response (Fig. 6). Within this region a tonE-like element (Ϫ1,071 to Ϫ1,059 bp), originally found as an osmotic response enhancer of the dog betaine transporter gene (Table I), was found (23). The sequence and the location of the element matches other tonE-like elements recently reported in the rabbit AR promoter between Ϫ1,108 and Ϫ1,092 bp (12), and in the mouse AR promoter between Ϫ1,053 and Ϫ1,040 bp (13). The human tonE-like element has been reported at two positions, between Ϫ1,230 to Ϫ1,220 bp and Ϫ1,157 to Ϫ1,148 bp (11).
Surprisingly in the rat promoter, the deletion of the tonElike element decreased, but did not abolish the osmotic response. Also, the lack of the tonE-like element had no effect on the high constitutive promoter activity. As the 5Ј-end of the promoter was deleted from construct 1 to construct 5, constitutive activity increased 7.0-fold suggesting that several negative regulatory elements may lie within this region. These results indicate that the most proximal cis-element required for the osmotic response and the constitutive promoter activity lie downstream of the tonE-like element.
The ExoIII-mediated DNA footprinting method was used to scan the region between Ϫ1,094 and Ϫ649 bp for any DNA FIG. 6. Schematic summary of promoter sequence necessary for constitutive and osmotic response of rat AR gene. Scheme and sequence of the deleted region (Ϫ1,071 to Ϫ895 bp) in the luciferase promoter construct 9 (Fig. 1). Box A indicates the deleted sequence to create construct 9 (Fig. 1). Box B indicates homologous sequence to the tonE element. Box C indicates ExoIII protected AEE sequence. Box D indicates the functional minimum AEE sequence. The two protected guanosines from dimethyl sulfate in vivo DNA footprinting (Fig. 3) are marked by asterisks. The deleted 8 bp of nucleotide sequence in construct 8 ( Fig. 1) are underlined. TGGAAAATCA protection under normal osmolarity. ExoIII digestion was halted at positions Ϫ926 bp in the coding direction and at Ϫ895 bp in the noncoding direction, indicating that a 32-bp region was protected under normal osmotic conditions. This putative transcription factor binding sequence, indicated by the ExoIII method, was further studied by dimethyl sulfate in vivo DNA footprinting to determine whether the factor was bound to the endogeneous AR promoter and how it would be affected by osmotic stress. Two guanosines located at Ϫ915 and Ϫ908 bp within the 32-bp protected region were found to be protected from dimethyl sulfate methylation under both normal and hypertonic conditions. No change in the intensity of the DNA protection was observed under normal culture conditions, 30 min osmotic stress, or 3 h osmotic stress. No other sequence surrounding this ExoIII-protected region was observed protected from methylation by these experiments. Further confirmation of transcription factor binding and delineation of the critical nucleotides involved in binding were determined by EMSA. A probe containing the 35-bp protected sequence gave several intense complexes under normal and osmotic stress, while a probe containing eight nucleotide mutations, between and including the two guanosines protected in the dimethyl sulfate in vivo experiment, significantly reduced the transcription factor binding. Competition assays also confirmed that the sequences between the two guanosines are crucial for the transcription factor(s) to bind. No match was found to this sequence in the transcription factor data base. This novel cis-element was named aldose reductase enhancer element (AEE).
The regulatory properties of this novel cis-element were further demonstrated when various lengths of AEE were fused to the 5Ј-end of the thymidine kinase promoter (Ϫ105 to ϩ52 bp). Surprisingly, construct pAEE28-TK gave a significant basal activity increase of 7.7-fold and an osmotic response of 5.7-fold, which is a higher increase than construct pAEE35-TK (Fig.  5). The minimum functional sequence for AEE was determined as 5Ј-GGGTGTTGGAAGAGTGCCAAATTT-3Ј by construct pAEE24-TK and pAEE20-TK.
Finally, luciferase assays were performed with constructs 8 and 9 to determine the effect of inhibition of AEE binding or inhibition of AEE plus tonE binding on promoter activity. Construct 8 was designed to only inhibit AEE binding by deleting 8 bp (5Ј-GAAGAGTG-3Ј) of the critical sequence determined by the EMSA experiment (Fig. 4). Both constructs 8 and 9 completely abolished the constitutive activity of the rat AR promoter, while construct 8 which includes tonE retained an osmotic response (2.5-fold). This unexpected finding suggests to us that the AEE element is critical for maintenance of constitutive activity. Deletion of AEE, which resulted in continued osmotic response, indicates the presence of other positive or negative elements.
In conclusion, we have identified a new cis-element AEE, located downstream of the tonE-like element, which has a strong effect on the constitutive activity and osmotic response of the rat aldose reductase promoter. This enhancer is capable of inducing the basal activity of a heterologous promoter by 7.7-fold and osmotic response up to 6.7-fold. From our data on AEE and considering the fact that tonE-like elements have been reported in different locations in the AR promoter (11), it is likely that several elements might be involved in the control of the osmotic stress response. Identification of the relative role of these respective elements and their connection to the signal transduction pathway will be important in understanding the process of cellular osmoregulation. The in vivo occupancy of the AEE in the endogeneous AR promoter strongly suggests that this element plays a physiological role in the regulation of AR expression.