DNA bending by retinoid X receptor-containing retinoid and thyroid hormone receptor complexes

Differential Effects of Hepatocyte Nuclear Factor 4α Isoforms on Tumor Growth and T-Cell Factor 4/AP-1 Interactions in Human Colorectal Cancer Cells

The nuclear receptor hepatocyte nuclear factor 4α (HNF4α) is tumor suppressive in the liver but amplified in colon cancer, suggesting that it also might be oncogenic. To investigate whether this discrepancy is due to different HNF4α isoforms derived from its two promoters (P1 and P2), we generated Tet-On-inducible human colon cancer (HCT116) cell lines that express either the P1-driven (HNF4α2) or P2-driven (HNF4α8) isoform and analyzed them for tumor growth and global changes in gene expression (transcriptome sequencing [RNA-seq] and chromatin immunoprecipitation sequencing [ChIP-seq]). The results show that while HNF4α2 acts as a tumor suppressor in the HCT116 tumor xenograft model, HNF4α8 does not. Each isoform regulates the expression of distinct sets of genes and recruits, colocalizes, and competes in a distinct fashion with the Wnt/β-catenin mediator T-cell factor 4 (TCF4) at CTTTG motifs as well as at AP-1 motifs (TGAXTCA). Protein binding microarrays (PBMs) show that HNF4α and TCF4 share some but not all binding motifs and that single nucleotide polymorphisms (SNPs) in sites bound by both HNF4α and TCF4 can alter binding affinity in vitro, suggesting that they could play a role in cancer susceptibility in vivo. Thus, the HNF4α isoforms play distinct roles in colon cancer, which could be due to differential interactions with the Wnt/β-catenin/TCF4 and AP-1 pathways.

General molecular biology and architecture of nuclear receptors

Nuclear receptors (NRs) regulate and coordinate multiple processes by integrating internal and external signals, thereby maintaining homeostasis in front of nutritional, behavioral and environmental challenges. NRs exhibit strong similarities in their structure and mode of action: by selective transcriptional activation or repression of cognate target genes, which can either be controlled through a direct, DNA binding-dependent mechanism or through crosstalk with other transcriptional regulators, NRs modulate the expression of gene clusters thus achieving coordinated tissue responses. Additionally, non genomic effects of NR ligands appear mediated by ill-defined mechanisms at the plasma membrane. These effects mediate potential therapeutic effects as small lipophilic molecule targets, and many efforts have been put in elucidating their precise mechanism of action and pathophysiological roles. Currently, numerous nuclear receptor ligand analogs are used in therapy or are tested in clinical trials against various diseases such as hypertriglyceridemia, atherosclerosis, diabetes, allergies and cancer and others.

Proposed mechanism for the stabilization of nuclear receptor DNA binding via protein dimerization

Hepatocyte nuclear factor 4 (HNF-4) defines a new subgroup of nuclear receptors that exist in solution and bind DNA exclusively as homodimers. We recently showed that the putative ligand binding domain (LBD) of HNF-4 is responsible for dimerization in solution and prevents heterodimerization with other receptors. In this report, the role of the LBD in DNA binding by HNF-4 is further investigated by using electrophoretic mobility shift analysis. A comparison of constructs containing either the DNA binding domain (DBD) alone or the DBD plus the LBD of HNF-4 showed that dimerization via the DBD was sufficient to provide nearly the full DNA binding affinity of the full-length HNF-4. In contrast, dimerization via the DBD was not sufficient to produce a stable protein-DNA complex, whereas dimerization via the LBD increased the half-life of the complex by at least 100-fold. Circular permutation analysis showed that full-length HNF-4 bent DNA by approximately 80 degrees while the DBD bent DNA by only 24 degrees. Nonetheless, analysis of other constructs indicated that the increase in stability afforded by the LBD could be explained only partially by an increased ability to bend DNA. Coimmunoprecipitation studies, on the other hand, showed that dimerization via the LBD produced a protein-protein complex that was much more stable than the corresponding protein-DNA complex. These results led us to propose a model in which dimerization via the LBD stabilizes the receptor on DNA by converting an energetically favorable two-step dissociation event into an energetically unfavorable single-step event. Implications of this one-step model for other nuclear receptors are discussed.

Heterodimeric interaction between retinoid X receptor alpha and orphan nuclear receptor OR1 reveals dimerization-induced activation as a novel mechanism of nuclear receptor activation

OR1 is a member of the steroid/thyroid hormone nuclear receptor superfamily which has been described to mediate transcriptional responses to retinoids and oxysterols. On a DR4 response element, an OR1 heterodimer with the nuclear receptor retinoid X receptor alpha (RXR alpha) has been described to convey transcriptional activation in both the absence and presence of the RXR ligand 9-cis retinoic acid, the mechanisms of which have remained unclear. Here, we dissect the effects of RXR alpha and OR1 ligand-binding domain interaction on transcriptional regulation and the role of the respective carboxy-terminal activation domains (AF-2s) in the absence and presence of the RXR ligand, employing chimeras of the nuclear receptors containing the heterologous GAL4 DNA-binding domain as well as natural receptors. The results show that the interaction of the RXR and OR1 ligand-binding domains unleashes a transcription activation potential that is mainly dependent on the AF-2 of OR1, indicating that interaction with RXR activates OR1. This defines dimerization-induced activation as a novel function of heterodimeric interaction and mechanism of receptor activation not previously described for nuclear receptors. Moreover, we present evidence that activation of OR1 occurs by a conformational change induced upon heterodimerization with RXR.

Prebending the estrogen response element destabilizes binding of the estrogen receptor DNA binding domain

Binding of many eukaryotic transcription regulatory proteins to their DNA recognition sequences results in conformational changes in DNA. To test the effect of altering DNA topology by prebending a transcription factor binding site, we examined the interaction of the estrogen receptor (ER) DNA binding domain (DBD) with prebent estrogen response elements (EREs). When the ERE in minicircle DNA was prebent toward the major groove, which is in the same direction as the ER-induced DNA bend, there was no significant effect on ER DBD binding relative to the linear counterparts. However, when the ERE was bent toward the minor groove, in a direction that opposes the ER-induced DNA bend, there was a four- to eightfold reduction in ER DBD binding. Since reduced binding was also observed with the ERE in nicked circles, the reduction in binding was not due to torsional force induced by binding of ER DBD to the prebent ERE in covalently closed minicircles. To determine the mechanism responsible for reduced binding to the prebent ERE, we examined the effect of prebending the ERE on the association and dissociation of the ER DBD. Binding of the ER DBD to ERE-containing minicircles was rapid when the EREs were prebent toward either the major or minor groove of the DNA (k(on) of 9.9 x 10(6) to 1.7 x 10(7) M(-1) s(-1)). Prebending the ERE toward the minor groove resulted in an increase in k(off) of four- to fivefold. Increased dissociation of the ER DBD from the ERE is, therefore, the major factor responsible for reduced binding of the ER DBD to an ERE prebent toward the minor groove. These data provide the first direct demonstration that the interaction of a eukaryotic transcription factor with its recognition sequence can be strongly influenced by altering DNA topology through prebending the DNA.

Direct modulation of simian virus 40 late gene expression by thyroid hormone and its receptor

Transcription of the late genes of simian virus 40 (SV40) is repressed during the early phase of the lytic cycle of infection of primate cells by the binding of cellular factors, called IBP-s, to the SV40 late promoter; repression is relieved after the onset of viral DNA replication by titration of these repressors (S. R. Wiley, R. J. Kraus, F. R. Zuo, E. E. Murray, K. Loritz, and J. E. Mertz, Genes Dev. 7:2206-2219, 1993). Recently, we showed that IBP-s consists of several members of the steroid/thyroid hormone receptor superfamily (F. Zuo and J. E. Mertz, Proc. Natl. Acad. Sci. USA 92:8586-8590, 1995). Here, we show that the thyroid hormone receptor TRalpha1, in combination with retinoid X receptor alpha (RXRalpha), is specifically bound at the transcriptional initiation site of the major late promoter of SV40. This binding repressed transcription from the SV40 late promoter by preventing the formation of pre-initiation complexes. Addition of the thyroid hormone 3,5,3'-L-triiodothyronine (T3) resulted in reversal of this repression in cotransfected CV-1 cells. Interestingly, repression did not occur when this thyroid response element (TRE) was translocated to 50 bp upstream of the major late initiation site. Binding of TRalpha1/RXRalpha heterodimers to this TRE induced bending of the promoter DNA. We conclude that hormones and their receptors can directly affect the expression of SV40, probably by affecting protein-protein and protein-DNA interactions involved in the formation of functional preinitiation complexes.

In simple synthetic promoters YY1-induced DNA bending is important in transcription activation and repression

Depending on promoter context, YY1 can activate or repress transcription, or provide a site for transcription initiation. To investigate whether the ability of YY1 to induce DNA bending influenced its ability to activate and repress transcription, simple synthetic promoters were constructed in which the YY1 binding site was inserted between the TATA box and either the NF1 or AP1 recognition sequences. In transient transfections of COS cells, the NF1YY1TATA and NF1RYY1TATA promoters exhibited a dramatic 15-20-fold increase in correctly initiated transcription. These promoters exhibited even larger 60-80-fold increases in transcription in HeLa cells. Neither multiple copies of the YY1 binding site alone, nor placement of a YY1 site upstream of the NF1 site activated transcription. Deletion of 4 bp between the NF1 and YY1 sites, which changes the phase of the DNA bends, abolished the 16-fold activation of transcription by NF1YY1TATA. Insertion of the YY1 site between the AP1 site and the TATA box decreased transcription approximately 3-fold. Replacing the YY1 binding site with an intrinsic DNA bending sequence mimicked this transcription repression. Sequences of similar length which do not bend DNA fail to repress AP1-mediated transcription. Gel mobility shift assays were used to show that binding of YY1 to its recognition sequence did not repress binding of AP1 to its recognition sequences. Our data indicate that YY1-induced DNA bending may activate and repress transcription by changing the spatial relationships between transcription activators and components of the basal transcription apparatus.

Precise alignment of sites required for mu enhancer activation in B cells

The lymphocyte-specific immunoglobulin mu heavy-chain gene intronic enhancer is regulated by multiple nuclear factors. The previously defined minimal enhancer containing the muA, muE3, and muB sites is transactivated by a combination of the ETS-domain proteins PU.1 and Ets-1 in nonlymphoid cells. The core GGAAs of the muA and muB sites are separated by 30 nucleotides, suggesting that ETS proteins bind to these sites from these same side of the DNA helix. We tested the necessity for appropriate spatial alignment of these elements by using mutated enhancers with altered spacings. A 4- or 10-bp insertion between muE3 and muB inactivated the mu enhancer in S194 plasma cells but did not affect in vitro binding of Ets-1, PU.1, or the muE3-binding protein TFE3, alone or in pairwise combinations. Circular permutation and phasing analyses demonstrated that PU.1 binding but not TFE3 or Ets-1 bends mu enhancer DNA toward the major groove. We propose that the requirement for precise spacing of the muA and muB elements is due in part to a directed DNA bend induced by PU.1.

The receptor-DNA complex determines the retinoid response: a mechanism for the diversification of the ligand signal

To obtain insights into the principles governing the complex biological responses to retinoids, we have analyzed the ligand sensitivities of various retinoid receptor-DNA complexes. We find that different retinoid receptor heterodimers show distinct activation patterns with various response elements while a given heterodimer can be activated at different retinoic acid concentrations on different response elements. In vitro binding experiments suggest that the same retinoic acid receptor-retinoid X receptor (RAR-RXR) heterodimer can have different ligand affinities, depending on the response element it is bound to. The differential responses of a particular receptor heterodimer with various retinoic acid responsive elements can be enhanced through the use of conformationally restricted retinoids. RAR- and RXR-selective retinoids can also synergistically activate the receptor heterodimers, indicating that both partners in the heterodimer can contribute to ligand-induced transcriptional activation. However, the relative influence of the RAR or RXR partner is specific for each response element. Together, our data demonstrate that it is the receptor-DNA complex and not the receptor alone that determines the ligand response. This flexibility allows for a highly pleiotropic retinoid response. Furthermore, conformationally restricted retinoids can accentuate the differential responses and exhibit a certain degree of gene selectivity by differentially activating the RAR or RXR component in the context of a given response element.

DNA bending by the silencer protein NeP1 is modulated by TR and RXR

NeP1 binds to the F1 silencer element of the chicken lysozyme gene and, in the presence of TR, v-ERBA or RAR, synergistically represses transcriptional activity. This repression involves a silencing mechanism acting independently of the relative promoter position. Here we show that NeP1 alone can induce a significant directed bend on DNA. The chicken homologue of human NeP1, CTCF, shows identical binding and bending properties. In contrast, the isolated DNA binding domain of CTCF efficiently binds DNA, but fails to confer bending. Similarly, the TR-RXR hetero- or homodimer, binding adjacent to NeP1 at the F2 sequence, do not show significant DNA bending. The binding of the T3 ligand to TR changes neither the magnitude nor the direction of the NeP1 induced bend. However, when all factors are bound simultaneously as a quaternary complex, the TR-RXR heterodimer changes the location of the bend center, the flexure angle and the bending direction.

Retinoid receptors cause distortion of the retinoic acid response element in the phosphoenolpyruvate carboxykinase gene promoter

Functional retinoic acid response elements (RAREs) have been described wherein the direct repeats are separated by 1, 2 or 5 bp (termed DR1, DR2 and DR5 respectively). We have previously shown that retinoic acid receptor/retinoid X receptor (RAR/RXR) binds a DR1 RARE within the phosphoenolpyruvate carboxykinase (PEPCK) gene promoter and is the trans-acting complex that mediates the retinoic acid (RA) response. However, the mechanism of trans-activation is unknown. The consequences of RAR/RXR binding to the PEPCK RARE were examined using a circular permutation analysis as a first step to explore the possible role of DNA conformational changes in the RA response. The RAR/RXR heterodimer produced a distortion angle of 78 degrees. The DNA distortion was shown to be at the centre of the PEPCK RARE; RA did not affect the severity of the distortion angle or the location of the distortion centre. Monomers and homodimers of RAR also distorted the DNA, but to a lesser extent than did RAR/RXR. The results of a phasing analysis demonstrated that RAR/RXR heterodimers did not induce a static DNA bend, in either the presence or the absence of RA. A cyclization kinetics assay was employed to show that RAR/RXR binding affected DNA ring closure in a phase-sensitive, RA-insensitive, manner. Taken together, these observations support the idea that RAR/RXR heterodimers distort the structure of the PEPCK RARE, at least in part, by altering DNA flexibility. The conformational change in the PEPCK RARE upon RAR/RXR binding has implications for how RAR/RXR heterodimers recognize various RARE structures.

DNA bending by thyroid hormone receptor: influence of half-site spacing and RXR

Transcriptional activation by thyroid hormone (T3) requires interactions between the T3 receptor (TR) and T3 response elements (TREs) composed of two copies of sequences related to AGGTCA. Direct repeats of this sequence are a functional TRE when spaced by 4 but not by 5 bp (DR4 versus DR5). TR bound as monomers, homodimers and heterodimers with retinoid X receptor (RXR) to both DR4 and DR5, with an approximately 10-fold greater affinity for DR4 due to reduced dissociation of the protein-DNA complex. We explored DNA bending as an additional variable which could influence the transcriptional outcome of the TR-TRE interaction. Circular permutation indicated a large distortion of the DNA following TR binding, but phasing analysis strongly suggested that this was due only in small part to DNA bending. Phasing analysis indicated that both TR/RXR and TR homodimer induced bends of approximately 10 degrees in DR4, but caused little bending of DR5. Moreover, the TR homo- and heterodimers bent DR4 in opposite directions. These results indicate that in addition to regulating the affinity and spacing requirement for DNA binding by TR, the TR dimer partner may also modulate transcription by influencing the direction of the bending induced by TR binding to DNA, although this effect may be subtle, due to the modest degree of bending.

Homodimer of p50 (NF kappa B1) does not introduce a substantial directed bend into DNA according to three different experimental assays

Transcription factors can distort the conformation of the DNA double helix upon binding to their target sites. Previously, studies utilizing circular permutation--electrophoretic mobility shift assay suggested that the homodimer of p50 (NF kappa B1), canonical NF-kappa B (p65-p50), as well as several non-canonical NF-kappa B/Rel complexes, may induce substantial DNA bending at the binding site. Here we have applied three additional experimental approaches, helical phasing analysis, minicircle binding and cyclization kinetics, and conclude that the homodimer of p50 introduces virtually no directed bend into the consensus kappa B sequences GGGACTTTCC or GGGAATTCCC.

The nonconserved hinge region and distinct amino-terminal domains of the ROR alpha orphan nuclear receptor isoforms are required for proper DNA bending and ROR alpha-DNA interactions

ROR alpha 1 and ROR alpha 2 are two isoforms of a novel member of the steroid-thyroid-retinoid receptor superfamily and are considered orphan receptors since their cognate ligand has yet to be identified. These putative receptors have previously been shown to bind as monomers to a DNA recognition sequence composed of two distinct moieties, a 3' nuclear receptor core half-site AGGTCA preceded by a 5' AT-rich sequence. Recognition of this bipartite hormone response element (RORE) requires both the zinc-binding motifs and a group of amino acid residues located at the carboxy-terminal end of the DNA-binding domain (DBD) which is referred to here as the carboxy-terminal extension. In this report, we show that binding of ROR alpha 1 and ROR alpha 2 to the RORE induces a large DNA bend of approximately 130 degrees which may be important for receptor function. The overall direction of the DNA bend is towards the major groove at the center of the 3' AGGTCA half-site. The presence of the nonconserved hinge region which is located between the DBD and the putative ligand-binding domain (LBD) or ROR alpha is required for maximal DNA bending. Deletion of a large portion of the amino-terminal domain (NTD) of the ROR alpha protein does not alter the DNA bend angle but shifts the DNA bend center 5' relative to the bend induced by intact ROR alpha. Methylation interference studies using the NTD-deleted ROR alpha 1 mutant indicate that some DNA contacts in the 5' AT-rich half of the RORE are also shifted 5', while those in the 3' AGGTCA half-site are unaffected. These results are consistent with a model in which the ROR alpha NTD and the nonconserved hinge region orient the zinc-binding motifs and the carboxy-terminal extension of the ROR alpha DBD relative to each other to achieve proper interactions with the two halves of its recognition site. Transactivation studies suggest that both protein-induced DNA bending and protein-protein interactions are important for receptor function.