Despite a wealth of knowledge regarding the early steps of epithelial differentiation, little is known about the mechanisms responsible for terminal nephron differentiation. The bradykinin B2 receptor (B2R) regulates renal function and integrity, and its expression is induced during terminal nephron differentiation. This study investigates the transcriptional regulation of the B2R during kidney development. The rat B2R 5′-flanking region has a highly conserved cis-acting enhancer in the proximal promoter consisting of contiguous binding sites for the transcription factors cAMP response element binding protein (CREB), p53, and Krüppel-like factor (KLF-4). The B2R enhancer drives reporter gene expression in inner medullary collecting duct-3 cells but is considerably weaker in other cell types. Site-directed mutagenesis and expression of dominant negative mutants demonstrated the requirement of CREB DNA binding and Ser-133 phosphorylation for optimal enhancer function. Moreover, helical phasing experiments showed that disruption of the spatial organization of the enhancer inhibits B2R promoter activity. Several lines of evidence indicate that cooperative interactions among the three transcription factors occur in vivo during terminal nephron differentiation: 1) CREB, p53, and KLF-4 are coexpressed in B2R-positive differentiating cells; 2) the maturational expression of B2R correlates with CREB/p53/KLF-4 DNA-binding activity; 3) assembly of CREB, p53, and KLF-4 on chromatin at the endogenous B2R promoter is developmentally regulated and is accompanied by CBP recruitment and histone hyperacetylation; and 4) CREB and p53 occupancy of the B2R enhancer is cooperative. These results demonstrate that combinatorial interactions among the transcription factors, CREB, p53, and KLF-4, and the coactivator CBP, may be critical for the regulation of B2R gene expression during terminal nephron differentiation.
- gene transcription
- histone acetylation
- kidney development
- terminal differentiation
the metanephric kidney is derived from the intermediate mesoderm at embryonic day 10.5 in mice and nephrogenesis continues until postnatal day 10 (25). The initial inductive events are regulated by a vast array of signaling and transcription factors that are expressed in specific spatial and temporal patterns. The nephric duct-derived ureteric bud cells give rise to the collecting duct system, whereas the mesenchyme-derived epithelia form the glomeruli and the remaining tubular structures of the nephron. The mature nephron is subdivided into segments that are specialized for specific tasks. This specialization, which is acquired at terminal differentiation, is reflected in protein expression profiles of various segments that complement function. In several developmental systems, overlapping and differential expression of transcription factors generate combinations of interacting proteins that regulate cell-specific gene expression as well as developmental fate (42). With respect to the developing kidney, such interactions might also explain how segment identity and functional status are acquired.
The bradykinin B2 receptor (B2R) is a G protein-coupled receptor that mediates the actions of bradykinin on renal function causing renal vasodilation and natriuresis (7). B2R function is required for normal homeostasis, as B2R-null mice have life-long salt-sensitive hypertension that begins early in life (1, 11, 21). B2R is also critical for the maintenance of epithelial cell integrity during nephrogenesis as indicated by the development of renal dysgenesis in embryonic salt-stressed B2R-null mice (14, 16). Ontogeny studies have demonstrated that B2R gene expression is developmentally regulated in the kidney. Specifically, the onset of tubular differentiation is marked by a surge in B2R gene expression. B2R mRNA levels are maintained at high levels throughout nephrogenesis, followed by a gradual decline by adulthood (14–16). Investigations of the molecular mechanisms governing the developmental regulation of the B2R gene and its spatial restriction to differentiating renal epithelial cells provide an opportunity to understand the transcriptional control of terminal nephron differentiation, an important yet understudied area in renal development.
In previous studies, we identified the tumor suppressor protein, p53, as an upstream sequence-specific transcriptional regulator of several terminal differentiation genes in the kidney including the B2R gene (32). B2R is expressed in p53-enriched differentiating nephron segments, and kidneys of p53-null pups exhibit ectopic spatial expression of B2R (32) and epithelial cell dedifferentiation (34). These findings suggested that p53 is one of the critical factors involved in the spatial specification of B2R expression during the process of terminal nephron differentiation. We have since extended our analysis of the rat B2R promoter (31) and found that the positive p53 response element, located at nucleotide positions −51 to −70 in the B2R promoter, is flanked by binding sites for CREB at nucleotide positions −44 to −50 and KLF-4 at positions −70 to −82, respectively. Although p53, CREB, and KLF-4 play important roles in transcriptional regulation of developmental genes, their combinatorial role in gene regulation has not been explored. Of additional interest is that in vitro studies have shown that p53 can interact physically with both CREB and KLF-4 and that KLF-4 is an intermediate transcriptional target for p53 in the pathway leading to p21 gene activation and growth arrest.
The present study presents evidence that CREB, p53, and KLF-4 form a higher-order complex, in conjunction with the coactivator CBP, on the B2R enhancer in vivo. Assembly of this complex is subject to developmental regulation. Remarkably, KLF-4, p53, and CREB expression is restricted to the differentiating epithelial cells in an overlapping manner with B2R. Interference with enhancer spatial organization disrupts B2R promoter function. The combinatorial regulation of the B2R promoter by CREB/p53/KLF-4 provides a physiological mechanism to specify expression of B2R in terminally differentiating renal epithelial cells.
MATERIALS AND METHODS
Rat B2R promoter constructs have been described (27, 33). Site-directed mutagenesis of the B2R promoter at the P1 p53-binding site and insertions of spacers with either a half-helical turn (5 bp) or full-helical turn (10 bp) between the p53 and CREB binding sites were generated by PCR mutagenesis (Stratagene, Quick-Change). The sequences of the spacer DNA were generously provided on request by A. Goldfeld (Harvard Medical School, MA); these were previously tested and shown not to have transcriptional activity (39). The CREB expression plasmids were kind gifts of R. Goodman, M. Greenberg, and M. Montminy; KLF-4 expression plasmid was from V. Yang; expression plasmids for wild-type p53 and wild-type and mutant E1A were from G. Morris; CBP expression plasmid was from R. Kwok; and CBPKIX triple mutant expression plasmid was from P. Brindle.
Tissue culture and transfections.
HeLa and inner medullary collecting duct (IMCD)-3 cells were maintained in DMEM supplemented with 10% FBS, penicillin (100 U/ml), and streptomycin (100 μg/ml; Invitrogen) at 37°C, 5% CO2 in a humidified incubator. Isogenic HCT116 p53+/+ and −/− cells (gift of B. Vogelstein) were maintained in McCoy's medium supplemented with 10% FBS. Cells were plated in six-well plates at 4 × 105 cells/ml 1 day before transfection and transfected with 1.2 μg reporter plasmid along with various concentrations of expression plasmids. A β-galactosidase expression plasmid pSV-lacZ (0.4 μg/well; Promega) was cotransfected to correct for transfection efficiency. Transfections were performed using Lipofectamine Plus reagent (Invitrogen), and CAT assays were performed after normalization for β-galactosidase activity as described (33).
Kidney nuclear extracts were prepared as described (32). In addition, nuclear extracts from embryonic SVT-2 fibroblasts were used as a source of p53. Double-stranded oligodeoxynucleotides were 5′ end-labeled with [γ-32P]dATP using T4 kinase for use as probes. DNA-protein complexes were resolved on a 5% polyacrylamide gel with 0.25× TBE buffer at 200 V for 1–2 h, and the gel was transferred to and dried on Whatman paper and exposed to X-ray film. Sequences of oligonucleotides (double-stranded) used for EMSA are shown in Table 1.
Tissue chromatin immunoprecipitation (ChIP) was performed using reagents and protocols from an Upstate Biotechnology ChIP kit, with modifications. Freshly isolated or snap-frozen kidneys were rapidly minced into fine pieces and immediately immersed in a 1% formaldehyde solution in PBS for cross-linking, for 15 min at room temperature with rotation. Reaction was quenched by addition of 0.125 M glycine. Tissue was rinsed two times in ice-cold 1× PBS, homogenized with a dounce A homogenizer (∼10 strokes), and the homogenate was lysed in SDS-lysis buffer. DNA was sheared by sonication to produce an average DNA fragment size between 500 and 1,000 bp and diluted 10-fold in ChIP dilution buffer. Immunoprecipitation was performed with antibodies to p53 (Santa Cruz Biotechnology, sc-FL393; 2 μg), CREB1 (sc-58; 2 μg), KLF-4 (sc-20691x; 5–20 μg), CBP (sc-369+sc-370; 20 μg), acetyl H4 (Upstate Cell Signaling, 06–866; 1:150 dilution), or control normal immunoglobulin (IgG) antibodies overnight at 4°C. DNA-protein-antibody complexes were captured on protein A/G-conjugated agarose beads. After being washed and elution of the complexes from the beads, DNA-protein cross-links were reversed at 65°C overnight. Immunoprecipitated DNA was ethanol-precipitated after proteins were removed by phenol-chloroform-isoamyl alcohol extraction following proteinase K treatment and used for PCR. The number of PCR cycles was determined based on linear amplification of the amplicons (25–35 cycles). Sequences of the primers used for PCR of the rat or mouse B2R gene flanking the KLF4-p53-CRE enhancer are: forward primer, 5′-GACCTTAGTCTGCACCAATGGAG-3′, and reverse primer, 5′-GGTTCTGTGTTGTAGGGAGTCC-3′. Sequences of the primers used for PCR of human B2R gene flanking the KLF4-p53-CRE enhancer are forward primer, 5′-GCAGAGCGGAGAGCGAAGG, and reverse primer, 5′-GCCTGATGTCCCCACCGTC.
Western blot analysis.
Protein extraction, SDS-PAGE, and Western blotting were performed as described (32).
Formalin-fixed, paraffin-embedded kidneys were sectioned (5-μm thick) and immunostained using the immunoperoxidase technique (ABC Elite Vectastain kit, Vector Laboratories), as described (15). In negative controls, primary antibody was either omitted or replaced by nonimmune serum. Colocalization studies were performed on serial sections subjected to antigen retrieval (microwave heat treatment for 20 min in citrate buffer, pH 8.0).
Proximal enhancer element in the B2R promoter is sufficient to drive reporter transcription in IMCD3 cells.
We analyzed the rat B2R gene promoter (31) for potential regulatory elements responsible for conferring developmental specificity. We previously reported the presence of p53 response elements at nucleotide positions −70 (P1 site) and −707 (P2 site; with respect to the transcription start site) of the rat B2R promoter (27, 33). Sequence analysis of the rat B2R promoter using the TRANSFAC program (www.gene-regulation.com) revealed the presence of cannonical KLF-4 and CREB binding sites flanking the P1 p53 site, at positions −70 to −82 and −44 to 51, respectively (Fig. 1A). The KLF4-p53-CRE sequence (nucleotide −44 to −82 of rat B2R sequence) is 100% identical between rat and mouse and 85% between rat and human B2R genes. To test the functional relevance of this module, we performed promoter deletion analysis. Mouse IMCD-3 were transfected with CAT reporter constructs driven by various segments of the B2R promoter. As shown in Fig. 1B, cells transfected with eight different B2R promoter-CAT constructs, B2R (−1184/+55), (−1087/+55), (−827/+55), (−635/+55), (−563/+55), (−384/+55), (−200/+55), and (−94/+55), showed relatively high baseline CAT activity compared with the promoter-less construct pCAT3-Basic. Importantly, the −38/+55-CAT constructs, lacking the KLF4-p53-CREB binding sites, exhibited no significant reporter activity. Thus the KLF4-p53-CRE module is sufficient to drive reporter expression in renal epithelial cells. In comparison, baseline CAT activity using the −94/+55 B2R promoter construct is undetectable in HeLa cervical carcinoma cells (33) and only minimally active in HCT116 colon carcinoma cells (data not shown).
Coordinate developmental expression of B2R, p53, CREB, and KLF4.
To implicate KLF-4, p53, and CREB in the regulation of B2R gene expression in vivo, it is important to demonstrate evidence of spatiotemporal coexpression. Previously, we demonstrated that newborn rat kidneys express ∼30-fold higher levels of B2R mRNA than adult kidneys (15). Immunoblotting of nuclear extracts from newborn or adult rat kidneys with antibodies against phospho-Ser20 p53, phosphoSer133-CREB, CREB, and ATF-1 revealed that these transcription factors are subject to developmental regulation, with higher levels in newborn than adult kidneys (Fig. 2, B–E). In a previous study, we demonstrated that phosphor-Ser15 p53 levels are also higher in newborn than adult kidney extracts (27). This developmental profile fits well with that of B2R (14, 15) (Fig. 2A). EMSA results indicated that KLF-4 DNA binding activity is higher in newborn than adult kidneys (Fig. 2F). Complex formation was efficiently competed in the presence of unlabeled oligoduplexes containing either a KLF-4 binding site from the p21 promoter (43) or a GC-rich Sp1 consensus sequence. A similar temporal profile was obtained for binding of kidney nuclear extracts to the B2R p53-binding site and CRE (Fig. 2, G and H). Addition of anti-acetylated-p53 antibody to binding reactions produced a supershift, indicating the presence of acetylated p53 in kidney nuclear extracts (Fig. 2G, arrow). This finding is consistent with our previous results, indicating that newborn nuclear extracts contain transcriptionally active p53 (27).
Although B2R expression is detectable in the renal tubular epithelium by embryonic day 16 of rat gestation (14, 16), we focused our studies on the neonatal period as in this species the bulk of nephrogenesis and differentiation occurs in the neonate. Histologically, the developing kidney can be divided into two zones: a nephrogenic zone (NZ), which contains proliferating nephron precursors (PCNA-positive; Fig. 3A), and a differentiation zone (DZ), which contains nephrons entering terminal differentiation to express renal function genes, such as B2R (Fig. 3B) (32). Similar to B2R, epithelial expression of p53, CREB, and KLF-4 is predominantly localized in the DZ (Fig. 3, C–E). Ser133 phospho-CREB is also expressed in the DZ (data not shown). As expected, CBP, a key cofactor for multiple transcription factors, is ubiquitously expressed in epithelial cell nuclei of both NZ and DZ (Fig. 3F). These findings demonstrate a remarkable congruency in the spatial expression of CREB, p53, and KLF4 with that of B2R.
To further assess the spatial distribution of p53, CREB, and KLF-4 during nephrogenesis, we carried out immunohistochemical studies on serial sections in newborn (day 5) rat kidneys. As depicted in Fig. 4, A–C, p53, KLF-4, and B2R are expressed in a spatially congruent manner in the collecting ducts and other tubular structures (arrowheads). Figure 4, D and E, show high-power views of p53 and CREB immunostaining. There is an overlap in p53 and CREB expression in some tubules (e.g., tubules no. 6–8). Other nephron segments, having the morphological appearance of distal tubules and collecting ducts, express higher levels of p53 than CREB (e.g., tubules no. 1–5; Fig. 4, D and E). These results demonstrate that p53, CREB, KLF-4, and B2R share an overlapping distribution during nephron differentiation.
Cooperative binding of p53 and CREB to the B2R enhancer in vitro.
The presence of contiguous binding sites for p53 and CREB in the B2R promoter prompted us to examine whether these two transcription factors bind DNA in a cooperative manner. To this end, EMSA was performed using the individual p53 and CRE sites or a combined p53-CRE sequence from the B2R promoter as probes. Newborn kidney nuclear extracts, a rich source of p53 and CREB, were used as a source of proteins. The results demonstrated formation of a p53-containing complex with p53-binding site (Fig. 5A, lane 1, black arrow), a CREB-containing complex with CRE (Fig. 5A, lane 2, open arrow), and three distinct molecular complexes with p53-CRE sequence containing p53, CREB, or p53+CREB (Fig. 5A, lane 3). The lower intensity of the CREB band in lane 3 in the presence of the p53-CRE probe compared with the CRE probe alone (lane 2) may be explained by incorporation of CREB into the p53 complex (lane 3), resulting in more intense higher molecular mass bands, presumably containing both p53 and CREB. The high-molecular-mass complexes were attenuated or did not form when either unlabeled p53-binding site or CRE was used as competitors (Fig. 5A, lanes 4 and 5) or when either the p53-binding site or CRE binding site was mutated (Fig. 5A, lanes 6-11). The mutation in the p53-binding site eliminated p53 binding but allowed weak binding of CREB (lanes 9 and 10); the latter was eliminated by unlabeled CRE sequence (lane 11). In addition, enhanced efficiency of competition by a p53 consensus sequence (17) in the presence of CRE mutation in the p53-CRE probe suggests that less p53 is able to bind if CREB binding is compromised (Fig. 5A, lane 7). Together, these results demonstrate that p53 and CREB bind cooperatively to the B2R enhancer in vitro.
Cooperative occupancy of B2R promoter by CREB and p53 in vivo.
To demonstrate the cooperation in p53 and CREB binding suggested by the EMSA results in vivo, we compared the presence of CREB at the endogenous B2R promoter in isogenic p53 −/− or p53 +/+ HCT116 cells by ChIP. PCR primers were designed to amplify the region flanking the conserved enhancer in the human B2R gene. The results revealed a modest increase in the amount of PCR product in p53 +/+ compared with p53 −/− cells (range 1.3–3.9-fold, n = 3; Fig. 5B). More relevant were results from two independent experiments using kidneys from newborn p53 +/+ and p53 −/− mice showing an average of 4.5-fold higher CREB enrichment at the B2R promoter in p53 +/+ than −/− kidneys (Fig. 5C). These findings along with the EMSA results strongly support that p53 and CREB bind cooperatively to the B2R promoter in vivo.
Binding of CREB, p53, and KLF4 to the endogenous B2R promoter is developmentally regulated.
Our results thus far indicate that CREB and p53 bind to the CRE-p53 sequence in the B2R promoter in vitro and in vivo. To determine whether assembly of these transcription factors, along with KLF-4 and CBP, on the endogenous B2R promoter is subject to developmental regulation in a manner similar to that of endogenous B2R gene expression (i.e., newborn > adult), ChIP assays were performed on newborn and adult rat kidneys. Positions of primer pairs used are shown in Fig. 6A. The results revealed that p53, CREB, KLF-4, and CBP assemble on the endogenous B2R promoter and that their assembly is developmentally regulated (Fig. 6, B–D).
In addition to acetylating some transcription factors, including p53 (3, 13, 19) and CREB (26), histone acetylase activity of CBP acetylates histones H3 and H4 rendering chromatin accessible to transcription factors (2, 6, 10). Therefore, we tested whether the differential CBP enrichment in the newborn kidney is associated with the presence of hyperacetylated H4 at the B2R promoter. ChIP analysis using anti-acetylated H4 revealed slightly higher acetylated H4 levels associated with chromatin at the B2R enhancer in newborn than adult kidneys (range 1.5–2.3 fold, n = 3; Fig. 6D).
Combinatorial activation of the B2R promoter by CREB, p53, and KLF4.
We next evaluated the functional relevance of the binding of CREB, p53, and KLF-4 to contiguous sites in the B2R enhancer using transient transfection assays in HeLa cells. These cells were chosen because they lack a functional p53 protein. To be able to measure additive or synergistic responses, concentrations of individual expression plasmids were adjusted to elicit a detectable submaximal transcriptional output (reporter activity). Pilot studies indicated that these concentrations corresponded to 5 ng/well of p53 expression plasmid, 10 ng/well of CREB, and 50 ng/well of KLF-4 expression plasmid. At these concentrations, the increase in reporter gene activity driven by the −94/+55 B2R promoter amounted to nearly 15-, 8-, and 2-fold in response to p53, CREB, or KLF-4, respectively (Fig. 7). Combinations of any two transcription factors on cotransfection elicited more than additive activation: CREB+p53 (36-fold increase), KLF-4+p53 (67-fold increase), and CREB+KLF-4 (22-fold increase). Also, the combination of p53+KLF-4 produced a stronger transcriptional output than CREB+KLF-4, suggesting that the relative position of the transcription factors on the enhancer is an important determinant of the final response. Cotransfection of all three expression plasmids together increased transcriptional output by 80-fold (Fig. 7).
To further delineate the role of the three transcription factors in B2R promoter activation, we introduced point mutations in the p53 and CRE-binding sites, which converted these high- to low-affinity binding sites. Because these binding sites are present in the proximal region of the promoter, deletions were not introduced to avoid interfering with basal promoter function. CRE mutations significantly decreased promoter output in response to p53, KLF-4, or CREB individually or in combination (Fig. 7). Point mutations in the p53-binding site, which decreased p53 binding by 50% (EMSA, data not shown), caused a modest decrease in promoter activity in response to the transcription factors (Fig. 7). The effect of p53-binding site mutagenesis became more apparent when this mutation was combined with mutation of CRE, resulting in marked inhibition of B2R promoter activation by various combinations of the transcription factors (Fig. 7). We did not test the transcriptional effect of mutagenesis of KLF-4-binding site.
To test the importance of the spatial organization and relative spacing of the individual transcription factor binding sites, we introduced a 5-bp spacer DNA between the CRE and p53 sites. Such a B2R construct with “out-of-phase” p53 and CRE has nearly 50% lower promoter activity than the wild-type promoter (Fig. 7). The decrease in promoter activity was observed with all combinations of transcription factors. Restoring phasing with a 10-bp spacer returned B2R promoter activity toward wild-type levels in response to the single transcription factors but was less effective in increasing transcriptional output in response to combinations of transcription factors. This finding suggests that proximity of transcription factors binding sites is important for the functional cooperation of the multiprotein complex.
CREB binding to DNA and CBP is required for B2R enhancer activity.
We previously showed the dependence of B2R gene transcription on p53 binding, as deletion of its binding site or coexpression of a dominant-negative p53 abolished B2R promoter-driven reporter activation by endogenous p53 (27, 33). To determine the role of endogenous CREB binding to the CRE in p53-mediated transcriptional activation, we cotransfected wild-type CREB, dominant-negative derivatives of CREB (M1 and A-CREB), or constitutively active CREB (CREB-DIEDML) with the B2R −94/+55-CAT construct and a p53 expression plasmid (5 ng; Fig. 8A). CREB dose dependently activated the B2R promoter. CREB phosphorylation at Ser133 is required for its association with CBP, resulting in enhanced transcriptional activity (9, 44). A nonphosphorytable mutant of CREB (Ser-to-Ala133, M1-CREB), which cannot bind CBP but retains the ability to bind DNA, increased B2R promoter activity modestly compared with the wild-type CREB (Fig. 8A). The role of CBP-CREB interactions in B2R promoter activation was assessed further by introducing CREBDIEDML, an activating mutant of CREB that exhibits high-affinity, phosphorylation-independent binding to CBP (9). Overexpression of CREBDIEDML markedly enhanced B2R promoter activation by p53 nearly eightfold (Fig. 8A). On the other hand, A-CREB, a mutant of CREB which cannot bind DNA, strongly suppressed p53-mediated activation in a dose-dependent manner (Fig. 8A). These results indicate that binding of CREB to its response element in the B2R enhancer is crucial for p53-stimulated transcription and that CREB phosphorylation on Serine 133 (and association with CBP) is required for full transcriptional activation.
To further examine the role of CBP-CREB interactions in B2R promoter activation, we transfected into HCT116 p53 −/− and +/+ cells with CBP expression plasmids containing either wild-type CBP or CBP with triple mutations in the KIX domain (designated CBPKIX) at amino acids Tyr650Ala, Ala654Gln, and Tyr658Ala. These mutations render CBP incapable of interacting with phospho-Ser133-CREB (23). It is important to note that this CBP mutant retains its ability to interact with the other transcription factors, as p53 and KLF-4 interact with CBP outside the KIX domain. Expression of the CBPKIX mutant decreased B2R promoter activity equally in p53 +/+ and −/− cells by ∼20% (Fig. 8B). These results suggest that the level of CBP at the B2R promoter is not limiting, possibly due to the CBP-recruiting ability of p53 (20). To test the latter hypothesis, we used an E1A expression plasmid to disrupt p53-mediated recruitment of CBP. E1A interacts with and sequesters CBP/p300 away from p53 (12). E1A repressed B2R promoter activity dose dependently in p53 +/+ cells (Fig. 8C). In comparison, E1A was significantly less effective in p53 −/− cells (Fig. 8C). As a control, cotransfection of mutant E1A (29), unable to interact with CBP/p300, had no significant effect on p53-mediated activation. These data demonstrate that p53-CBP and CREB-CBP interactions are critical for optimal B2R promoter activity.
Terminal differentiation is a crucial step in organogenesis as it signals the onset of functional identity. Experimental evidence indicates that cell-cycle arrest is required but not sufficient to complete terminal differentiation (8) and that transcriptional mechanisms leading to induction of functional genes must also be activated coordinately. Despite a wealth of knowledge regarding the early steps of mesenchymal-epithelial induction and nephron formation (4, 25, 40, 41), little is known about the molecular mechanisms responsible for the onset and maintenance of terminal renal epithelial cell differentiation. In the developing kidney, mature and functioning nephrons are located in the inner cortex, whereas the outer nephrogenic zone contains proliferating mesenchymal and epithelial nephron precursors. Cessation of epithelial cell proliferation is accompanied by functional differentiation, as evidenced by expression of renal function genes. However, the transcriptional programs governing and coordinating these two processes are largely unknown. To begin to understand the molecular basis for terminal differentiation in the kidney, we chose to study the regulation of the bradykinin B2R gene, a developmentally regulated renal function gene whose expression is restricted to differentiating tubular epithelial cells (15, 16). To our knowledge, the present study is the first to describe a differentiation gene in the kidney whose expression is dependent on the combinatorial interactions of three developmentally regulated transcription factors.
We previously reported that p53 is highly enriched in terminally differentiated renal epithelial cells and identified several renal function genes (bradykinin B2R, angiotensin type 1 receptor, Na-K-ATPase α1, and aquaporin-2) as a novel group of p53-target genes (32). The current study provides new evidence that p53 is bound to the B2R enhancer in vivo during the process of terminal nephron differentiation, forming a higher-order multiprotein complex with two neighboring transcription factors, CREB and KLF-4, and the transcriptional coactivator CBP. Importantly, KLF-4, p53, and CREB are enriched in differentiating renal epithelia. Immunolocalization studies performed in serial sections confirmed that p53, CREB, and KLF-4 are coexpressed in the same tubular segments. Closer examination revealed that p53 is expressed at higher levels in distal nephron segments, whereas CREB is more enriched in proximal nephron segments. We previously reported that B2R expression in the newborn kidney is predominantly found in collecting ducts and to a lesser extent in proximal tubules (15). This study confirms these findings and provides further evidence for the expression of p53, KLF-4, and CREB in B2R-positive cells. Functional studies performed in cultured cells further demonstrated that the three transcription factors cooperate in the activation of the B2R promoter. Accordingly, the highly disorganized pattern of B2R expression in the renal cortex of p53-null pups (32, 34) may reflect disrupted coassembly and cooperation among the three transcription factors.
KLF-4 is a zinc-finger transcription factor that is highly abundant in terminally differentiated cells (5, 24, 28). p53 Stimulates KLF-4 expression, resulting in cell-cycle arrest in the G1 phase secondary to induction of the cyclin-dependent kinase inhibitor p21 (43, 45). In addition, p53 and KLF-4 exhibit direct protein-protein interactions (45). KLF-4-deficient mice die perinatally as a result of dehydration due to impaired epithelial development (36). Involvement of KLF-4 in differentiation of the colon (24, 46) and epidermis (36) and in the regulation of the p21 gene (43, 45), combined with its restricted expression pattern in differentiating renal epithelial cells (this study), makes KLF-4 an attractive candidate for a similar contribution in the renal epithelia and warrants additional studies to examine the renal phenotype in KLF-4-deficient mice.
Our data also show developmentally regulated expression of p53, P-Ser15/20p53, acetylatedLys372/382p53, CREB, and Ser133phospho-CREB during terminal nephron differentiation. The EMSA results indicate cooperative binding between p53 and CREB at the B2R promoter. Interestingly, the mutation of p53 site in p53-CRE had a more dramatic effect on binding than the mutation of CRE site in p53-CRE, suggesting that the DNA binding affinity of the p53-CREB complex and/or its overall stability are more dependent on p53 contact with DNA than CREB. ChIP analysis further revealed that that these transcription factor/DNA interactions occur in vivo on the endogenous B2R promoter and that promoter occupancy by p53 and CREB is higher in developing than adult kidneys. In keeping with EMSA results, ChIP assays demonstrated that CREB binding adjacent to p53-binding site on chromatin at the B2R promoter is compromised in p53-null kidney tissue. Furthermore, that the assembly of these factors is productive is suggested by the presence of CBP at this region, again higher in newborn than adult kidneys. CBP acts as a transcriptional coactivator via several mechanisms: 1) bridging the KLF-4/p53/CREB complex with the basal transcription machinery; 2) acetylation of nucleosomal histones, thus allowing better access of DNA-binding proteins to DNA; and 3) direct acetylation and activation of p53. The ChIP assays revealed that H4 acetylation at the B2R promoter is only modestly higher in the newborn that adult kidney, despite dramatically higher binding of transcription factors and CBP to the same region. The more limited changes in acetylation may be due to the fact that the entire kidney is used as a source of chromatin rather than the differentiation zone. Recent studies showed that higher p53 binding to sites in target promoters does not always correlate with increased histone H3 and H4 acetylation (22) and that other histone modifications (e.g., methylation) may be involved (35).
In transient transfection assays, KLF-4, p53, and CREB cooperatively activated B2R transcription. Although transcriptional output was compromised when either the p53 or CREB binding sites were mutated, the CRE mutation had a more profound effect on B2R transcription. Moreover, interference with CREB binding to CRE by expression of a non-DNA-binding mutant of CREB abrogates B2R promoter responsiveness to p53. Interestingly, p53 and CREB have been shown to interact with each other directly via the bZIP domain of CREB and the NH2 terminus of p53 (18). This interaction strongly enhances association of p53 with CBP, thereby recruiting CBP to p53-responsive promoters. In this regard, we suggest that the levels of CBP at the B2R promoter are not entirely dependent on recruitment by CREB because the promoter is activated by the CREB mutant lacking the ability to interact with CBP. Because both KLF-4 and p53 are capable of interacting with CBP, it is possible that CBP recruitment in this case is p53/KLF-4 dependent. Indeed, interference with p53-CBP interaction by E1A inhibits B2R promoter activation.
In summary, we propose a combinatorial model for the transcriptional regulation of B2R expression during terminal nephron differentiation (Fig. 9). According to this model, p53 induces KLF-4 expression in differentiating epithelial cells (45). Subsequently, p53 and KLF-4 cooperate to induce downstream differentiation genes (e.g., p21) (45). The concomitant presence of CREB, p53, and KLF-4 cooperatively activates B2R transcription. Under these circumstances, the coactivator, CBP, bridges the CREB-p53-KLF4 complex and potentially induces acetylation of p53 (20) and CREB (9). Bending of DNA, which helps in the formation and stabilization of the multiprotein complex, can conceivably be mediated by p53, and the degree of DNA bending directly correlates to the binding affinity of p53 to a site (30). It is interesting to note the placement of the p53-binding site in the B2R promoter between the KLF-4 and CRE sites. Presumably, introduction of a 5-bp spacer DNA hinders p53-CREB interaction by altering the helical phasing and the precise three-dimensional specificity of the transcription factor binding sites, resulting in decreased transcription.
Support for this study was provided by National Institutes of Health Grants DK-56264 and DK-62250 (to S. E. Dahr), and 1F32 DK-61137 (Z. Saifudeen), and the American Heart Association, Southeast Affiliate (H. Fan). Digital images were obtained at the Tulane Hypertension and Renal Center of Excellence Imaging Core Facility HEF2001–06-07, Louisiana Board of Regents, through the Millennium Trust Health Excellence Fund.
We thank Drs. B. Vogelstein, R. Goodman, R. Kwok, M. Montminy, M. Greenberg, G. Morris, P. Brindle, and V. Yang for the generous contributions of plasmids and cell lines.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society