BIOLOGICAL OVERVIEW

A Drosophila corepressor mediates transcriptional silencing of the Ecdysone receptor:Ultraspiracle heterodimer. SMRT-related ecdysone receptor-interacting factor (Smrter), also known as SANT domain protein, is a large nuclear protein that, surprisingly, shows only limited homology to the vertebrate corepressors SMRT and N-CoR. Nevertheless, the fact that EcR:USP associates with Smrter and Smrter associates with murine Sin3A and Drosophila Sin3A, co-repressors known to form a complex with the histone deacetylase Rpd3/HDAC (see Drosophila Rpd3), indicates a conserved mechanism underlying transcriptional repression by vertebrate and invertebrate nuclear receptors. Given the genetic and biochemical evidence that Sin3A associates with Rpd3/HDAC in both yeast and mammalian cells, and the likelyhood for a similar association in Drosophila, it is expected that Smrter also recruits a histone deacetylase complex to EcR. The linkage of EcR to Rpd3 is a potential explaination for the role of histone deacetylase in triggering the regression of chromosome puffs. Yet the presence of Smrter in puffed loci of polytene chromosomes indicates that a complete dissociation of Smrter complex may not be a prerequisite step for the formation of chromosomal puffs. Rather, other factors, such as coactivators with histone acetyltransferase activity, may play a significant role in triggering the formation of chromosome puffs (Tsai, 1999 and references).

Genetics experments have provided the first evidence for the existence of a corepressor for EcR:Usp heterodimers. Previous genetic studies have shown that mutations in the ecdysone receptor give rise to lethal and developmental defects, indicating that this signaling pathway is essential for development. However, the means by which these mutations cause the observed defects remain unclear. The fact that the EcR:USP heterodimer can carry out its regulatory functions as both an activator and a potent repressor in vertebrate cultured cells (CV-1 cells) enabled the use mammalian cells to carry out initial studies on the molecular consequences of these EcR mutations. One of these EcR mutations, EcR with a mutation in alanine 483 (A483T), is of particular interest as it displays increased reporter activity both in the absence and the presence of hormone. In this experiment, EcR or its mutant derivative, EcR^A483T, was cotransfected along with vp16-USP and a luciferase reporter that contains multimerized Ecdysone receptor response elements (hsp27EcREs). A potential explanation for its elevated activity is that EcR A483T fails to recruit an endogenous (vertebrate) repressor complex in CV-1 cells. To investigate whether repression by EcR in CV-1 cells is mediated by its association with a vertebrate corepressor and whether such an interaction, if it does occur, is impaired by the A483T mutation, a mammalian two-hybrid assay with Gal4-c-SMRT, the vertebrate corepressor, was conducted. No significant induction of the Gal4 reporter has been observed for Gal4-c-SMRT alone or in cells cotransfected with EcR or with vp16-USP. However, in the presence of both EcR and vp16-USP, the resulting heterodimerization shows strong association with Gal4-c-SMRT. The addition of hormone appears to completely dissociate SMRT from the heterodimer complex, therefore eliminating reporter activity. In contrast, EcR A483T alone or in the presence of vp16-USP is unable to associate with Gal4-c-SMRT, resulting in minimal activity in the presence or absence of ligand. Not only does this result suggest that A483 of EcR is critical for corepressor binding, but it also reveals that EcR:USP is a preferred binding complex for SMRT (Tsai, 1999).

Sequence alignment of EcR with the vertebrate TR, RAR, and v-erbA, an oncogenic TR variant, reveals that this alanine 483 is located within a highly conserved 23-amino acid (aa) loop region connecting helices 3 and 4, termed the LBD signature motif. Based on structural studies of vertebrate nuclear receptors, this alanine residue appears to be on the exposed surface, consistent with it being a potential corepressor binding site for nuclear receptors. In vivo studies indicate that EcR^A483T is a semilethal allele. When EcR^A483T is placed in trans with EcR^E261st (an allele that removes both the DBD and LBD domains of EcR), lethality (greater than 95%) ensues. However, the few surviving EcR^A483T/EcR^E261st flies display significant delays in development, blistered wings, and defective tergites, indicating that EcR is involved in the development of these tissues. The ability of EcR to bind a vertebrate corepressor and the loss of this property in EcR A483T suggests that the defects observed in EcR^A483T flies may result from the disruption of its interaction with a Drosophila corepressor (Tsai, 1999).

Although EcR readily interacts with vertebrate SMRT in both mammalian and yeast cells, repeated low-stringency hybridization screens failed to identify a Drosophila SMRT homolog. Given that no SMRT/N-CoR homolog is found in C. elegans, it was speculated that either a SMRT/N-CoR-like corepressor is not conserved in invertebrates or, alternatively, invertebrate corepressors may have diverged significantly from their vertebrate counterparts. To pursue the isolation of an EcR corepressor, a yeast interaction screen of a Drosophila embryonic cDNA library using EcR as bait was conducted. This screen resulted in the isolation of a clone, E52, whose protein product interacts with EcR as well as with the vertebrate RAR and TR, but notably not with USP. Intriguingly, unlike the interaction between E52 and RAR, which can be dissociated by all-trans retinoic acid, the interaction between E52 and EcR or the interaction between SMRT and EcR is not dissociated by Muristerone A (MurA). This result suggests that other factors essential for the dissociation of E52 from EcR, such as USP, are missing in yeast (Tsai, 1999).

Isolation of overlapping cDNA and genomic clones led to the identification of a full-length sequence encoding a large protein of 3446 amino acids. This protein contains several unusually long stretches of Gln, Ala, Gly, and Ser repeats. Comparative analysis reveals it to be a novel protein with limited regions of clear homology with the vertebrate nuclear receptor corepressors SMRT and N-CoR. This protein was named Smrter, SMRT-related ecdysone receptor–interacting factor. Northern blot analysis indicates that Smrter encodes large transcripts (greater than 12 kb) expressed broadly throughout the embryonic stage and three larvae stages, as well as in adult flies (Tsai, 1999)

To fulfill criteria necessary to establish Smrter as a corepressor of EcR, the Gal4-Smrter constructs were each examined for their ability to repress basal transcription. This resulted in the identification of three autonomous repressor domains termed SMRD1, SMRD2, and SMRD3, which respectively decrease basal activity of Gal4-DBD by approximately 17-fold, approximately 5-fold, and approximately 14-fold. Interestingly, SMRD1 contains the SNOR and SANT domains while SMRD2 overlaps with Ecdysone receptor interaction domain 1 (ERID1). Since repression by SMRT and N-CoR is mediated, at least in part, by the mSin3/HDAC complex and Smrter functions as a potent repressor in mammalian cells, it was asked whether the SMRD1, 2, or 3 could associate with vertebrate mSin3A. In the yeast two-hybrid assays, SMRD3 showed a strong association with mSin3A, while the interactions between SMRD1 or 2 and mSin3A are essentially undetectable. Using serial deletion constructs, a minimal Smrter-interacting domain has been localized to the first amphipathic helix of mSin3A. This region, PAH1, is also the primary site in mSin3 for the interaction with SMRT and N-CoR. These results indicate that at least one of the repression domains of Smrter can interact with vertebrate mSin3A and suggest that, despite their differences, insect and vertebrate corepressors employ a similar mechanism to recruit the repressor complex. However, the lack of significant interactions among SMRD1, SMRD2 and mSin3A raises the possibility that they may repress transcription using different strategies. The Drosophila mSin3A homolog (dSin3A) (Neufeld, 1998 and Pennetta, 1998 ) shows substantial sequence similarity to mSin3A in the four PAH domains as well as in the region between PAH3 and PAH4, which is required for interaction with histone deacetylase. The yeast two-hybrid screen was used in pairwise assays to identify Smrter interaction domains in dSin3A. The results reveal again that Smrter also interacts with dSin3A and that the region, including the PAH1 of dSin3A, is critical for associating with SMRD3 (Tsai, 1999).

Sequence comparison of SMRD3 with the defined Sin3-interacting domains of N-CoR and SMRT reveals a potential shared motif (the DALA motif) with N-CoR (aa 1835-1859), but not with SMRT. Deletion of a predicted helical secondary structure in N-CoR (aa 1833-1845), which includes the DALA motif, severs the interaction with mSin3A. The significance of this DALA motif is further strengthened by the finding that replacement of several conserved residues within the DALA motif of SMRD3 (mutations M2 and M3) destroys the interaction between SMRD3 and mSin3A/dSin3A in yeast two-hybrid experiments. In functional assays, these two mutated SMRD3s consistently fail to repress basal activity of Gal4-DBD, supporting the idea that association with Sin3A is a necessary component of Smrter-mediated repression (Tsai, 1999).

In keeping with the evidence that dSin3A is a component in the EcR regulatory pathway, whether dSin3A interacts genetically with EcR was examined using several previously characterized EcR and dSin3A mutants. In the experiment, in which female dSin3A^K07401 mutants were crossed with male EcR^E261st mutants, only (approximately) 14% of the scored EcR^E261st/dSin3A^K07401 progenies survived; this is significantly lower than the expected 33.3%. This suggests that a large portion of the EcR^E261st/dSin3A^K07401 animals either die prior to eclosion or fail to eclose. Additionally, surviving EcR^E261st/dSin3A^K07401 escapers show delayed development and wing defects, which are held horizontally at 45°-90° angle from the body axis. These results suggest that dSin3A shares an overlapping regulatory pathway with EcR. Strikingly, in a reverse genetic cross, in which female EcR^E261st flies were crossed with male dSin3A^K07401 flies, none of the EcR^E261st/dSin3A^K07401 flies survived to adulthood. Apparently, EcR^E261st/dSin3A^K07401 results in a genetically sensitized background. By halving the maternally deposited EcR in embryos descended from female EcR^E261st/SM6b, the lethality for EcR^E261st/dSin3A^K07401 is further increased. These results reveal that, in addition to its previously known zygotic function, EcR also contributes maternally to Drosophila development (Tsai, 1999).

The selectivity of the point mutation A483T in disrupting Smrter association is noteworthy because it still allows heterodimer formation, DNA binding, ligand binding, and activation. Based on the crystal structure of vertebrate nuclear receptors, residue 483 of EcR is located at the surface exposed loop between helices 3 and 4 of the LBD. Given that this region is well conserved among nuclear receptors and that the EcR^A483T mutation disrupts the binding of both Smrter and SMRT, this L3-4 represents a likely component of the binding site. If this is indeed the case, it is expected that the ability of TR and RAR to interact with SMRT or N-CoR may also be impaired by mutations at the corresponding residue of these two vertebrate nuclear receptors. In this study, the phenotype of the lethal EcR^A483T mutant could be linked to a selective deficiency in corepressor binding, thus providing strong evidence as to the importance of nuclear receptor:corepressor complex in animal development. This result reveals the outcome of a constitutive dissociation of corepressor from nuclear receptor; this complements previous studies that have linked several oncogenic diseases to the constitutive association of corepressor to nuclear receptors. For example, v-erbA, an oncogenic TR variant, acts as a constitutive repressor due to its constant unregulated association with SMRT (Chen, 1995), and two translocation mutant forms of the RAR in humans, RAR-PLZF and RAR-PML, which are involved in acute promyelocytic leukemia, also repress transcription because of unregulated association with corepressor (Hong, 1997; He, 1998 and Lin, 1997). Together, these results indicate that a coordinated interaction between corepressors and nuclear receptors is an essential feature of metazoan development and hormone action (Tsai, 1999).

GENE STRUCTURE

PROTEIN STRUCTURE

Amino Acids - 3446

Structural Domains

While the underlying nuclear receptors' core structure and the Sin3:RPD3 deacetylase complexes in Drosophila and vertebrates are well conserved, the adapter protein Smrter that links these two components in Drosophila is highly divergent from its mammalian counterpart. Overall, Smrter is only approximately 13.5% identical to SMRT and N-CoR. Adapter proteins such as Smrter, SMRT, and N-CoR are evidently the most variable component of the nuclear receptor signaling complex. Oddly, even though the nuclear receptor-interacting domains of Smrter are divergent from those of N-CoR and SMRT both in primary sequence and location in the proteins, Smrter binds vertebrate nuclear receptors, and, reciprocally, vertebrate SMRT also binds EcR. Evidently, critical docking sites in the receptors can productively interact with widely varying motifs. This is surprising and suggests that comparison of cocrystal structures of the receptor LBD, together with the various interacting domains will be revealing in terms of how corepressors interact with nuclear receptors. The following shared features between Smrter and SMRT/N-CoR have been found: (1) an extended highly charged 294-amino acid region (aa 573-847) termed SNOR (Smrter, SMRT, and N-CoR) is 35%/36% identical and 54%/55% similar to SMRT and N-CoR, respectively. (2) C-terminal to SNOR is a recently identified 50-amino acid segment referred to as the SANT domain (SWI3, ADA2, N-CoR, and TFIIIB-B', Aasland, 1996), which shares structural similarity with the Myb domain found in a wide variety of transcription factors. A growing repertoire of SANT domains are found in heterogeneous proteins, including human (KIAA0071), C. elegans (C14B9.6, F53H10.2, cMTA-1), and yeast (YCR592). In this short domain, Smrter is most related to SMRT and N-CoR (64%/82% and 68%/80% in identity/similarity, respectively). This domain is repeated in SMRT and N-CoR, while only one copy is found in Smrter. (3) The next 1500 amino acids of Smrter are highly diverged from its vertebrate counterparts. Included in this region are two octapeptides (GSI motifs), which currently possess no known function. (4) Adjacent to the two GSI motifs is a stretch of 16 amino acids, termed ITS. In SMRT, the ITS motif is located within the region that interacts with CBF1/RBP-Jk, the mammalian homolog of Drosophila Suppresser of Hairless [Su(H)], a component involved in the Notch signaling pathway. In N-CoR, the ITS motif maps to the region that associates with c-Ski, a protooncogene, involved in both transformation and differentiation of avian embryo fibroblasts. (5) Approximately 700 amino acids from the C terminus of Smrter, SMRT, and N-CoR, a segment composed largely of alternating acidic Asp or Glu and basic Arg residues, are found (D/ER repeat). (6) A conserved short motif, LSD, of 17-18 amino acids is found at the extreme C terminus of each of these corepressors. Surprisely, two defined receptor interaction domains and several identified repression domains in SMRT and N-CoR are not conserved in Smrter. The remarkable divergence suggests that Smrter may achieve these functions in a distinct fashion as compared to its vertebrate counterparts. (Tsai, 1999 and references).

date revised: 18 February 2024

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.