| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Biology, Indiana University, Bloomington, Indiana 47405-3700
Address all correspondence and requests for reprints to: Peter Cherbas, Department of Biology, Jordan Hall, 1001 East Third Street, Bloomington, Indiana 47405. E-mail: cherbas{at}indiana.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
): an N-terminal A/B domain of variable length and nonconserved sequence, a highly conserved C or DNA-binding domain (DBD) 66 residues long comprising two zinc fingers, a short nonconserved hinge region rich in basic amino acids, a moderately conserved E or ligand-binding domain (LBD) of about 250 residues, and a C-terminal F domain that is variable in length and nonconserved in sequence.
|
Although molecular studies of EcR and USP are much less extensive than those of vertebrate NRs, it is already clear that the ecdysone receptor has some unusual properties. For example, EcR by itself cannot bind hormone; it must heterodimerize with USP to do so (2). This requirement for heterodimerization complicates the functional analysis of ligand binding, but mutational analysis (e.g. Ref.3) suggests that EcR is the hormone-binding moiety and that dimerization with USP stabilizes a hormone-binding conformation of EcR. EcR is very poor at forming homodimers; thus, EcR/USP is certainly the predominant and probably the only species bound to EcREs in nuclear extracts of Drosophila cells. Despite the ecdysone receptor’s other similarities to TR/RXR and retinoic acid receptor/RXR, it does not share their preference for binding sites composed of directly repeated half-sites; virtually all the known natural EcREs and the strongest synthetic EcREs are palindromes. Finally, Drosophila EcR contains an unusually long (223 residues) F domain, for which no function is known.
Like its vertebrate cognates (4, 5, 6), unliganded EcR/USP is a repressor of transcription (7, 8). Thus, in a Kc cell assay system, unliganded receptor depresses reporter gene expression 3- to 4-fold below the reference level (expression from a control plasmid lacking an EcRE), whereas liganded receptor stimulates expression to at least 20 times the same reference level (7). The unliganded EcR/USP complex interacts with corepressor proteins such as SMRT-related ecdysone receptor-interacting factor (SMRTER; Ref.9), and it appears that one important role of hormone binding is to provoke the release of these corepressors (10). In vertebrate systems dissociation of corepressors reveals the activities of the transcriptional activation functions (AFs) of the heterodimeric receptor. Typically each receptor polypeptide has two such functions, AF1 in the A/B domain (11) and AF2 associated with the LBD (12, 13, 14). Conserved AF2 sequences occupy the carboxy-terminal helix of the LBD. Upon ligand binding, this helix flips inward, sealing the hormone-binding pocket and providing a new surface for protein-protein contacts (15, 16, 17, 18). Conserved AF2 residues have been shown to be required for hormone-dependent transcription activation (11, 12), and for relief of repression (19, 20, 21). These residues interact directly, in a ligand-dependent manner, with a variety of coactivators such as glucocorticoid receptor-interacting protein 1 (GRIP1; Ref.22), steroid receptor coactivator 1 (SRC1; Ref.23), CBP-interacting protein (p/CIP; Ref.24), CREB-binding protein (CBP/p300; Ref.25), or vitamin D receptor (VDR)-interacting proteins (26). By contrast, the AF1 elements are not conserved in sequence and appear to function in diverse ways. AF1 may activate transcription by binding to the same coactivators that interact with AF2 (27, 28, 29, 30, 31, 32), to basal transcription factors (33), or to a variety of other proteins (34, 35, 36, 37, 38).
In Drosophila, a single EcR gene produces three isoforms that differ only in their amino-terminal domains (39). The isoform-specific A/B domains have no sequence similarity and vary widely in length (197 residues for isoform A, 226 for isoform B1, and 17 residues for isoform B2). Immunostaining and genetic studies show that the three isoforms have different spatial and temporal expression patterns and distinct functions in development (39, 40, 41, 42, 43, 44, 45). Each of the three isoform-specific A/B domains has previously been shown in at least one test system to confer transcriptional activation (46, 47).
Here, we report a systematic analysis of the transcriptional activation functions of the EcR/USP heterodimer in a single cell line using a single test promoter. Our goals have been to begin to decipher the roles of EcR and USP in activation, to identify AFs active in at least this one setting, and to lay the groundwork for identifying coactivators.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
) contains a luciferase reporter gene driven by a promoter derived from the upstream region of the ecdysone-responsive Eip71CD gene. This promoter contains no EcREs. When a functional EcRE is inserted and the vector introduced into normal (EcR-containing) Drosophila Kc167 cells, reporter expression is repressed in the absence of hormone and induced in the presence of hormone (7). The overall effect of hormone depends on the EcRE sequence; Fig. 2 shows a typical 70-fold induction using a single copy of a 23-bp EcRE derived from the hsp27 gene (panel A) and a 180-fold induction using a consensus EcRE (panel B).
|
).
When full-length EcR protein was supplied by transfecting L57-3-11 cells with EcR-coding sequences under the control of a strong promoter, the ecdysone response was restored (Refs.49 and50 and Fig. 2A). As the amount of transfected receptor plasmid was increased, reporter induction rose to a plateau and became saturated (Fig. 2A, inset); cotransfecting with an USP-expressing construct did not affect this plateau (data not shown). Thus, the measurements of reporter induction in Fig. 2A were made under conditions that were saturating for both EcR and USP. Each of the three EcR isoforms produced a hormone dose-response curve similar to that produced by the mixture of receptor isoforms found in normal Kc167 cells (EC50 
10-7 M (Fig. 2B).
The various EcR isoforms clearly differ in their activities in this test system (Fig. 2A). When EcR-B1 was expressed at saturating levels, it mediated reporter gene induction at very high levels—levels about half those seen in Kc167 cells. Interestingly, the artificial isoform EcR-C was fully 50% as effective as EcR-B1. EcR-C contains only those sequences common to all three natural isoforms and has no A/B domain. We conclude that a large fraction of the activity of EcR-B1 is due to a common region activation function (i.e. AF2). EcR-A was no more active than EcR-C; its long, unique A/B domain apparently does not contribute to activation in this setting. Surprisingly, EcR-B2 was nearly as active as EcR-B1, suggesting that the very short A/B sequence of EcR-B2 contains a powerful activating function. Saturation curves for all of the EcR isoforms were essentially parallel, with each isoform reaching its characteristic maximum at approximately 10 µg transfected plasmid (Fig. 2A, inset). This implies that when EcR is present at saturating concentration, the three isoforms mediate qualitatively different responses.
In all of the transfection experiments that follow, we used 40 µg of the receptor expression plasmid per transfection. This plasmid concentration is sufficient to ensure that the exogenous full-length receptor is present at saturating concentration (Fig. 2A). Fusion proteins used in assays reported later in this paper were tested in an EMSA to confirm that the protein was present in transfected cells at approximately the same concentration for all constructs used (data not shown). Consequently, we can assume that the receptor proteins are present at saturating concentration in all of the experiments reported here. When L57-3-11 cells contain a saturating level of any wild-type EcR isoform, the regulation of an EcRE-linked reporter is similar to that of wild-type Kc cells in the shape of the 20E dose-response curve, in the level of reporter induction (Fig. 2), and in the magnitude of basal inhibition (data not shown). We infer that the presence of a saturating level of exogenous EcR does not substantially distort the properties of the hormone response.
Intrinsic Activation Functions of the Isoform-Specific Amino Termini of EcR
To test their intrinsic abilities to activate transcription, A/B domains from each EcR isoform and from USP were expressed as fusions with the DBD of GAL4, using the plasmids CMA-GBD-EcR-B1-N, CMA-GBD-EcR-A-N, CMA-GBD-EcR-B2-N, and CMA-GBD-USP-N. Each plasmid was cotransfected into L57-3-11 cells with a reporter plasmid containing copies copies of a GAL4 binding sequence (UAS; Fig. 1A). Transcriptional activation was estimated by comparing reporter activity in cells transfected with a fusion protein expression plasmid and cells transfected with CMA-GBD, which expresses the GAL4-DBD alone. EcR-B1-N activated reporter expression about 100-fold, and EcR-B2-N about 60-fold (Fig. 3). By contrast, the amino-terminal domains of EcR-A and of USP yielded no significant transcriptional activation. These results suggest that the activation by full-length EcR isoforms in L57-3-11 cells (Fig. 2A) can be considered as the sum of an AF2 activation conferred by the common region and AF1 activations associated with the amino-terminal regions of isoforms B1 and B2. They also suggest that the A/B domain of USP does not contribute to transcriptional activation.
|
). Amino-terminal deletions of the EcR-B1 A/B domain showed that the region downstream of residue 181 was dispensable, but transcriptional activation was progressively decreased by further deletion. The data imply contributions from at least three regions: residues 56–118, 118–144, and 158–181. The region 115–181 appears to be necessary but not sufficient for activation because deletion from 181 to 118 reduced transcriptional activation by about 80%, but 115–181 alone had very little ability to activate transcription. Residues 1–56 of EcR-B1 had little detectable function, although this region is strongly conserved among dipteran (Refs.51, 52, 53, 54 ; and Berlinger, M., I. A. Hansen, and S. Meyer, submission to GenBank, accession no. AF325360), lepidopteran (55, 56, 57, 58, 59), and coleopteran (60) EcRs.
The fusion protein encoded by CMA-GBD-B2-N includes the 17 residues specific to isoform B2 plus 6 residues from the common region. Deletion of the six common region residues did not decrease transcriptional activation (GBD-EcR-B2-NS). The primary sequence of the 17-residue B2-specific region is compatible with a short amphipathic helix (Fig. 3B). We therefore tested three point mutations, designed to decrease the hydrophobicity of a residue on the putative hydrophobic face (V7A, I14A) or reverse the charge of a residue on the putative hydrophilic face (E9K). The mutation V7A had little or no effect, but I14A and E9K dramatically decrease transcriptional activation. These observations are consistent with the notion that the B2-specific AF1 functions as an amphipathic helix.
The Extended F Domain of EcR
Sequence alignments indicate that the LBD of EcR comprises residues 400–655 of the B1 sequence. After the LBD (E domain), most NRs contain an F domain of 10–50 residues. The F domain is variable in length and unconserved in sequence, and no clear function has so far been ascribed to it. Drosophila EcR has an extended F domain of 223 residues, composed largely of simple, highly repetitive sequences. The contribution of the EcR F domain to transcriptional activation was assessed as follows:
C-terminal deletions of EcR-B1 were tested for their ability to mediate reporter activation in L57-3-11 cells. Figure 4 shows a comparison of the activities of EcR-B1 and three deletion mutants: EcR
C689 (i.e. 1–689 of EcR-B1 retained), EcRC655, and EcRC649. Deleting the entire 223 F domain of EcR-B1 (EcRC655) reduced induction only marginally. By contrast, EcRC649, in which the deletion removes several residues of the conserved AF2 sequence region (see below), had no detectable EcR activity. We shall return later to the properties of EcRC649.
|
below). We conclude that the F domain of EcR is not very important for activation, although it may contribute modestly to AF2 function.
|
illustrates several important features of this assay:
|
, top panel and inset).
2) Because L57-3-11 cells are deficient in EcR but have a normal usp gene, we expected endogenous USP to substitute, and perhaps compete, for GAD-UL as a partner for GBT-EL. In fact, transfection with GBT-EL in the absence of GAD-UL gave about 50-fold ecdysone induction, a level of induction comparable to that seen with an AF1-deleted EcR in these cells (Fig. 2). We attribute this induction to GBT-EL/USP heterodimers because EcR in the absence of USP is unable to bind ecdysone. [A low level of binding by reticulocyte lysate-translated EcR reported previously (2) is apparently due to heterodimerization with RXR because EcR LBD produced in bacteria does not bind hormone (61).] Under the conditions of our assay, endogenous USP is in excess because inclusion of a USP expression plasmid did not increase the level of reporter induction (data not shown). Cotransfection of GBD-EL and GAD-UL caused a slightly higher transcriptional activation than that produced by GBD-EL alone; the difference is probably attributable to the GAL4 activation domain present in GAD-UL. Therefore, endogenous USP can substitute for GAD-UL. It does not, however, compete effectively with the high levels of GAD-USP produced in our transfections, since the combinations GBT-UL/GAD-EL and GBT-EL/GAD-UL gave similar activations (Fig. 5B). We infer that endogenous USP does not significantly affect the outcome of these two-hybrid assays.
3) As expected, GBT-UL alone gave virtually no hormone induction.
Transcriptional inhibition by the unliganded receptor (basal inhibition) was also detected in the two-hybrid assay (Fig. 6). Basal inhibition is assessed by comparing reporter expression from a UAS-containing plasmid and from a similar plasmid lacking the UAS sequences. The same combinations of EcR and USP LBD fusions that gave rise to ecdysone induction (Fig. 5B) also decreased reporter expression about 3-fold in the absence of ecdysone (Fig. 6). These results imply that 1) basal inhibition is conferred by the receptor LBDs; 2) like hormone induction, basal inhibition requires EcR/USP heterodimerization; and 3) dimerization of the EcR and USP LBDs takes place in the absence of hormone under the conditions of this assay.
|
). Because the GAL4 activation domain is present in the latter complex, but not the former, we conclude that the unliganded EcR/USP conformation does not permit transcriptional activation by the GAD domain. This is most easily interpreted by postulating that corepressor binding masks both the AF2 function of the nuclear receptor and the activation function of the GAD fragment.
Transcriptional Activation by Helix 12 Mutants in EcR and USP
AF2 has been well characterized in the LBDs of several NRs (62). It resides largely in helix 12, the C-terminal helix of the LBD. Sequence alignments identify helix 12 of EcR as residues 638–655 and that of USP as residues 483–500. To test the role of each AF2 in transcriptional activation by the LBD heterodimer, we introduced into GBT-EL and GBT-UL a series of C-terminal deletions extending into helix 12 and a series of substitutions for helix 12 residues that are well conserved among species. Each mutant protein was assayed in the Kc cell two-hybrid system described above (Fig. 7A). A C-terminal deletion removing the last six residues of helix 12 from EcR (GBT-EL649) abolished transcriptional induction, as did a similar deletion from full-length EcR (Fig. 4). Mutation of a conserved phenylalanine on the outer surface of EcR helix 12 (F645A) or of a conserved tryptophan on the inner surface of the same helix (W650A) also destroyed the ability of the EcR-LBD to mediate ecdysone-dependent transcriptional activation. GBT-EL (F645A) and GBT-EL (W650A) bind normally to GAD-UL as does GBT-EL649 (Fig. 9 and data not shown). By contrast, a C-terminal deletion removing the last eight residues of helix 12 from USP caused only a modest effect (GBT-UL492), and none of the point mutations we tested (E493K, E493Q, L490A, L487A, L495A) had any detectable effect on the induction of transcription. Of all the USP mutations we tested, only deletion to residue 486, removing almost all of helix 12, abolished hormone-dependent transcriptional activation (Fig. 7B), and this deletion severely decreased the ability of the USP-LBD to heterodimerize with EcR-LBD (EMSA data not shown).
|
|
, mutations affecting the EcR LBD helix 12 (GBT-EL F645A, W650A, and truncation to position 649) blocked transcriptional activation. By contrast, they did not affect the ability of the LBD to mediate basal repression (Fig. 7C). Basal inhibition was also unaffected by our point mutations in USP, but it was abolished by terminal deletions of USP helix 12. In fact, the diminished induction ratio previously noted for UL492 (Fig. 7B) appears to be consist entirely of a failure of basal inhibition, combined with normal hormone-activated transcription.
Further Characterization of Helix 12 Mutations in EcR
The ligand-binding capacities of EcR and USP helix 12 mutants were tested, both in the context of the full-length proteins and as LBD-GAL4 fusions. Mutant proteins were combined with wild-type heterodimer partners, and their ability to bind the radiolabeled ecdysone analog I125-iodoponasterone (63) was compared with binding by the wild-type heterodimer. Both EcR/USP-E493A and EcR-F645A/USP bound ligand and, taking into account the probable variations in translation efficiency, both were essentially indistinguishable from wild type (Fig. 8A). EcR-F645A/USP affinity for ecdysone did not differ appreciably from that of the wild-type heterodimer in a competitive binding assay (Fig. 8B). By contrast, EcR-W650A/USP and EcR-C649/USP did not bind hormone detectably. We estimate that our assay would detect binding with affinity at least 1–2 orders of magnitude lower than that of the wild-type receptor.
|
). Briefly, a GBT fusion (GBT-EL or GBT-UL) was mixed with the complementary GAD fusion (GAD-UL or GAD-EL, respectively) and with a labeled UAS probe. The dimer band is slower migrating and easily resolvable from the monomer band. Under the conditions used (i.e. in the presence of 0.1% Nonidet P-40), the ability of ligand to promote heterodimerization is especially obvious. We also observe a small shift in the mobility of the heterodimer when hormone is present, an effect that we attribute to conformational differences. We find that the choice of fusion pairs (i.e. GBT-EL and GAD-UL v. GAD-EL and GBT-UL) does not affect the outcome (data not shown). Figure 9, lanes 9–11, illustrates these observations for the wild-type LBDs. As shown in lanes 1–8, neither of the helix 12 mutations F645A and W650A had a significant effect on the ability of the EcR-LBD to heterodimerize or on the ability of hormone to increase the amount and change the conformation of the heterodimer. Ecdysone is known to promote the heterodimerization of EcR and USP (2); the fact that it does so, even in the presence of the W650A mutation, indicates that this mutant EcR binds hormone, although at an affinity too low to be detected by the assays used in Fig. 8.
Intrinsic Activation Functions of Helix 12 from EcR and USP
Helix 12 regions of NRs harbor sequences that are sufficient to activate transcription independent of context (11). We tested several fusions containing EcR helix 12 residues without finding evidence for such an autonomous activation function. The most active of these fusions, GBT-EAF2, yielded an extremely modest 2-fold effect (Fig. 10).
By contrast, USP helix 12 activated transcription 10-fold, and this activation was sensitive to mutations known to be dispensable in the context of the receptor heterodimer (cf. Figs. 7A and 10). Evidently, the full-length wild-type structures in the liganded EcR/USP dimer create an AF2 that requires EcR helix 12 while masking the intrinsic activation function of USP helix 12.
Dominant-Negative Phenotypes of EcR AF2 Mutants
Two point mutations (F645A and W650A) and the deletion mutation 649 abolish the ability of EcR-LBD to mediate an ecdysone response, without loss of their ability to heterodimerize with USP-LBD (Fig. 9). Because LBD mutations are not expected to interfere with DNA binding, we predicted that these mutant EcRs would compete with wild-type EcR for binding to USP and to EcREs; at sufficiently high concentration, the mutant EcR would be expected to displace wild-type EcR at ecdysone-regulated promoters, causing a dramatic decrease in ecdysone-induced transcription. To test this prediction, L57-3-11 cells were transfected with a limiting concentration of the wild-type EcR expression plasmid CMA-EcR-B1, in combination with various concentrations of each helix 12 mutant EcR expression plasmid (Fig. 11). In each case, ecdysone induction of a reporter was modestly reduced when the ratio of mutant EcR to wild-type EcR was 1:1, and reduced about 4-fold when the ratio was 10:1. Thus, each of these EcR-AF2 mutants exhibits a dominant-negative phenotype.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
In the experiments reported here, we have examined the activation functions of the ecdysone receptor functioning at a single promoter in a single cell type under conditions of transient expression. Our intent is to analyze the contributions of the potential transcriptional activation domains in the EcR/USP heterodimer to the ecdysone response in a defined system. We emphasize that while some of the conclusions reported here may reflect general properties of the ecdysone receptor, some properties of EcR/USP undoubtedly would be different when tested on a different promoter or in a different environment. An example is the function of the isoform-specific amino-terminal region of EcR-A, which we discuss below.
The A/B Domain
Our data show a strong activation function (AF1) in the A/B domains of EcR-B1 and EcR-B2, but no detectable AF1 in EcR-A or in USP. We note that the A/B domain of EcR-A was reported to cause transcriptional activation in assays in yeast (46). Furthermore, the A/B domain of EcR-A is required to support a normal ecdysone response in certain cell types in the fly, and that of B2 in other cell types (45). It seems clear, therefore, that each isoform-specific A/B domain of EcR has a functionally distinct AF1; two of them are able to activate transcription in our assay system. The A/B domain of USP does not activate transcription in our assay system; whether it can do so on any promoter or in any cell remains unknown.
Deletion analysis of EcR-B1 AF1 reveals several interdependent elements that occupy much of the 227-residue A/B domain. It seems likely that this domain binds to multiple coactivators. It is surprising that the N-terminal residues of EcR-B1, which are highly conserved in insect EcRs (see above), are not included among those we find to be essential. Still more surprising is our observation that the 17-residue A/B domain of EcR-B2 functions as a potent activator. It seems likely that it forms complexes as an amphipathic helix, its helicity being stabilized by its interaction partners because its ability to activate transcription can be destroyed by alteration of a charged residue on the ionic face or of a hydrophobic residue on the nonionic face.
The E Domain
Transcriptional activation by the AF2 region of EcR is revealed by the effects of mutations in helix 12, although helix 12 in isolation has little activity. By contrast, helix 12 of USP has an easily demonstrable activation function, but mutations in this sequence have no consequence in the context of the intact LBD; thus, the transcription activation ability intrinsic to helix 12 of USP is masked in the intact EcR/USP heterodimer. This conclusion is consistent with structural studies of USP which suggest that helix 12 is locked in an inactive conformation (70).
The F Domain
In isolation, the C-terminal tail can produce a modest activation of transcription, but it seems unlikely that it performs this role in the context of the intact receptor. Deletion of the C-terminal domain of EcR had little detectable effect in the assays described here. Furthermore, EcR-B1 lacking the C-terminal domain is able to support the same developmental functions as full-length EcR-B1 when expressed in individual fly tissues (Ref.45 ; and Cherbas, L., unpublished observations). The C-terminal sequence in Drosophila EcR as well as in EcR from other insect species consists largely of repeated residues, and there is no apparent conservation of sequence among EcRs from different species. By contrast with the 223-residue C-terminal domain in Drosophila EcR, the C-terminal region in EcRs from other flies are typically around 50–100 residues (Refs.51, 52, 53, 54 and71 ; and Berlinger, M., I. A. Hansen, and S. Meyer, submission to GenBank, accession no. AF325360), in moths around 20–30 residues (Refs.55, 56, 57 ; and Berlinger, M., I. A. Hansen, and S. Meyer, submission to GenBank, accession no. AF325360), and less than 5 residues in EcR sequences reported for ticks (72), crabs (73), locusts (74), and beetles (60). Even so, moth EcRs are active in L57-3-11 cells (Ref.50 ; and Hu, X., and P. Cherbas, unpublished). For all of these reasons, we think it unlikely that the C-terminal tail plays an essential role in Drosophila EcR.
Properties of EcR-AF2 Mutations
The EcR sequence can be aligned with the sequences of vertebrate NRs for which crystal structures are known; these alignments predict that EcR-F645 lies on the outside surface of helix 12, potentially able to interact with coactivators, while W650 lies on the inner surface of the same helix, in a position to interact with the ligand. It is therefore not surprising that the mutation W650A dramatically reduces affinity for ecdysone, while the mutation F645A abolishes transcriptional activation without affecting ligand binding. Both mutant proteins can mediate basal inhibition. If binding of hormone to F645A caused the release of a corepressor, we should see a 3- to 4-fold induction of reporter expression as a consequence of relieving basal inhibition; in fact, there is no detectable increase of reporter expression upon addition of the hormone. This observation suggests that binding of hormone does not lead to release of a corepressor in this mutant. Our results are completely consistent with previous observations on vertebrate NRs: residues homologous to EcR W650 have been implicated in ligand binding in peroxisome proliferator-activated receptor-
(T473; Ref.75) and vitamin D receptor (F422; Ref.76). In TR, mutations in a residue homologous to EcR F645 (TRß L454S, L454V, L454A) reduce the receptor’s affinity for coactivators and increase its affinity for corepressors while preserving ligand binding (77, 78). Residues at this position are known to constitute part of the binding interface for p160 coactivators [TRß L454 (79), peroxisome proliferator-activated receptor- L468 (75), estrogen receptor L539 (80)]. Because no coactivators for the ecdysone receptor have yet been identified, and a crystal structure for EcR is not yet available, it is not currently possible to test directly the effects of mutation on coactivator binding; nonetheless, it seems very likely that the point mutation F645A has effects very similar to mutations in the corresponding residues of vertebrate receptors.
The ability of EcR-F645A and EcR-W650A to function as dominant negative EcRs has now been confirmed in a variety of tissues in intact Drosophila (45). In each tested tissue, there is little or no effect when the wild-type and mutant EcRs are expressed from identical promoters, but when the mutant EcR is expressed from a strong promoter and wild-type EcR from its own (weak) promoter, striking defects in the ecdysone response are detected; this is completely consistent with the cell culture data reported here. The requirement for a large excess of mutant EcR expression to achieve the dominant negative phenotype may have any of several causes: 1) It is possible that the mutant proteins are less stable than the wild type. 2) Because ecdysone binding stabilizes the EcR/USP heterodimer, it is likely that the ligand binding-defective mutants W650A and 649 compete relatively poorly for binding to USP. F645A binds hormone, but its apparent failure to release corepressor upon binding to ecdysone suggests that it fails to undergo the normal conformational changes associated with hormone binding; it is therefore quite possible that F645A also competes relatively poorly with wild-type EcR for binding to USP in the presence of hormone. 3) Binding of an NR to its response element is probably stabilized by interaction with other factors bound at the same promoter (81). To the extent that heterodimers containing mutant EcR fail to participate in such interactions, they should compete relatively poorly for binding to EcREs.
The Role of USP
In our assay system, hormone-dependent activation does not depend directly on any USP sequences. The USP A/B domain contains no intrinsic activation function that we can detect, and the weak function in helix 12 is masked. Studies of other promoters or other cell types may reveal a different situation. In addition, we note that if USP has a ligand, it is possible that its activation functions could be revealed only in response to that ligand. Still, even in the system under study here, USP clearly plays a critical role that goes beyond its contribution to DNA binding. Our studies of LBD fusions bypass the DNA binding functions of the receptor components, yet they are consistent with the idea that USP is required for hormone-dependent transcriptional activation. In view of in vitro studies demonstrating that EcR does not bind ligand except as the heterodimer EcR/USP (2), we suspect that the role of USP is largely allosteric, i.e. it is required to stabilize EcR in a conformation capable of binding hormone. Because the USP helix 12 mutation 492 does not support basal repression, it is plausible that USP plays a more direct role in the formation of corepressor complexes; a role for RXR helix 12 in mediating corepressor interactions has been reported (82).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Kc167 and L57-3-11 cells were grown as described previously (49, 83). L57-3-11 is an EcR-deficient clone derived from Kc167 by parahomologous targeting (49).
Two days before a transfection, the cells were transferred to fresh medium at approximately 106 cells/ml. Electroporation was carried out as described previously (83, 84). After electroporation, the cells were diluted to their original density in M3 supplemented with bactopeptone, yeast extract, and 5% fetal calf serum and dispensed in 1-ml aliquots into 12-well plates. Hormones in ethanol carrier were added 3 h after the transfection (7) and cells were harvested for reporter assay 45–49 h after the transfection.
The reporter plasmids were designed for use with the Dual-Luciferase reporter system (Promega Corp., Madison, WI), with firefly luciferase as the primary reporter and Renilla luciferase as an internal transfection control. Each transfection contained 1 µg/cuvette of the Renilla luciferase plasmid (S-188-cc-RLuc) and the same amount of an experimental reporter plasmid (based on firefly luciferase). Plasmids expressing EcR, USP, or their derivatives were used at 40 µg/cuvette unless otherwise indicated. Cells were extracted and reporter activities measured with the Dual-Luciferase Assay kit (Promega Corp.), according to a slightly modified version of the manufacturer’s protocol. Unless otherwise stated, data represent the mean of at least three independent experiments, with error bars indicating SE.
The measurement of hormone effects is complicated by the fact that when reporter constructs lacking EcREs are tested, the ratio of firefly luciferase/Renilla luciferase decreases slightly (
2-fold) but reproducibly after hormone treatment. This effect is promoter independent but EcR dependent; it is not due to an endogenous luciferase. Evidently it reflects some minor effects of the changed cellular environment on the relative stabilities and/or activities of the two enzymes or their RNAs. Given this interpretation, we included, as a control in every experiment, cells transfected with only the EcRE-null vector S-188-cc-Luc. Adjusted reporter activity values were calculated as the ratio: raw reporter activity (experimental)/raw reporter activity (S-188-cc-Luc). These are the values reported in the figures. This correction has minor quantitative effects on the data; no conclusion reported here would be affected qualitatively by use of the unadjusted data.
Binding Proteins and EMSAs
Proteins were synthesized in vitro using the TNT-coupled reticulocyte lysate system (Promega Corp.) according to the manufacturer’s instructions. Whole-cell extracts were made by three cycles of freezing/thawing in 20 mM HEPES (pH 7.9), 0.4 M KCl, 1 mM dithiothreitol (DTT), 0.1 mM EDTA, and 0.5 mM phenylmethylsulfonyl fluoride.
An 80-bp DNA fragment containing a single GAL4 binding site was excised from UASX1-188-cc-Luc using BspDI and HindIII and purified by gel electrophoresis. The probe was end-labeled using 
-32P-dATP (Amersham Pharmacia Biotech, Arlington Heights, IL; >3000 Ci/mmol) and Klenow fragment and purified by gel filtration or spin column. Typical probe specific activity was approximately 105 cpm/ng DNA.
DNA binding reactions were carried out in 20 µl total volume. Proteins were incubated with probe and poly(dI-dC) (0.5 µg/20 µl) at 22 C in DNA binding buffer (20 mM HEPES, pH 7.5; 100 mM KCl; 7.5% glycerol; 2 mM DTT; 40 µM ZnSO4) for 1 h. Nonidet P-40 (0.1%) was included in reactions involving EcR/USP heterodimers. When the binding reactions involved crude cell extracts, we substituted UAS binding buffer (20 mM HEPES, pH 7.5; 50 mM KCl; 5 mM MgCl2; 10% glycerol; 1 mM DTT; 250 µg/ml BSA; 50 µg/ml salmon sperm DNA; and 0.5 mM spermidine) for DNA binding buffer. PAGE was run according to standard procedures.
Each of the fusion proteins that was assayed in cells for transcriptional activation activity was also tested for stability by extracting 1 ml of transfected cells and subjecting the extract to an EMSA, with and without added antibody to GAL4-DBD or GAL4-AD (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) to establish the presence of the protein. In each case, the protein was detected at the approximate expected titer (data not shown).
Ligand Binding Assays
Hormone binding assays were performed essentially as described previously (63). To make the ligand binding conditions identical to those used for DNA binding assays, the salt concentrations of the binding and washing buffer were increased from 40 mM to 100 mM KCl, and the pH of the binding buffer was changed from 7.0 to 7.5. These changes proved to have no detectable effect on levels of ligand binding.
Plasmids
The structures of plasmids used in these experiments are described below. For full details of their construction, see the supplemental data to this paper.
Reporter Plasmids
S-188-cc-Luc contains (in order) a polylinker, the Eip71CD promoter region (-188 to +11, relative to the transcriptional start), and coding sequence for firefly luciferase. This plasmid contains no detectable EcREs (see Fig. 1A). S-188-cc-RLuc is identical to S-188-cc-Luc save that it contains the enzymologically distinct Renilla luciferase. In Pal1SX-188-cc-Luc, a consensus EcRE has been inserted into the polylinker of S-188-cc-Luc (Fig. 1A). hsp27-188-cc-Luc is identical, except that the EcRE sequence derives from the hsp27 gene (85). In UASX4-188-cc-Luc four tandem copies of a GAL4 binding site have been inserted into the polylinker of S-188-cc-Luc (Fig. 1A); UASX1-188-cc-Luc is identical except that there is only a single copy of the UAS GAL4 binding site.
Generic Expression Vectors
pCMA was made from CMX (86) by inserting a Drosophila actin5C promoter fragment. The resulting plasmid contains promoters from cytomegalovirus, T7, and Drosophila actin 5C. The actin promoter gives strong expression in Drosophila cultured cells, where the cytomegalovirus promoter is weak (Cherbas, L., unpublished observations). In pCMA-GBD and pCMA-GBT, a fragment encoding residues 1–147 of GAL4 is inserted in the polylinker of pCMA. The two plasmids differ in the origin of the GAL4-DBD fragment. pCMA-GBD contains a fragment from pAS1 (87), which encodes an epitope tag in addition to the GAL4 fragment, whereas pCMA-GBT contains a fragment from pGBT9 (CLONTECH Laboratories, Inc., San Diego, CA), which has no epitope tag; polylinker-encoded residues also differ between the two plasmids.
In pCMA-GAD a fragment from pGAD424 (CLONTECH Laboratories, Inc.) is inserted into the pCMA polylinker. This fragment encodes the GAL4 activation domain (residues 768–881) preceded by an SV40 nuclear localization signal.
EcR Expression Vectors
pCMA-EcR-B1, pCMA-EcR-B2, and pCMA-EcR-A each contain the complete coding sequence for the corresponding wild-type EcR isoform. pCMA-EcR-C encodes only the residues common to all three isoforms; its product lacks an A/B domain.
pCMA-EcRC689, pCMA-EcRC655, and pCMA-EcRC649 encode carboxy-truncated versions of EcR-B1; in each case the number indicates the carboxy-terminal residue.
pCMA-EcR-F645A and pCMA-EcR-W650A encode EcR-B1 truncated after residue 655, and carrying the indicated single-residue substitution.
USP Expression Vectors
pCMA-USP contains the full coding sequence of wild-type USP. In pCMA-USPC492, the coding sequence is truncated after residue 492.
GAL4 DBD Fusion Expression Vectors
Fragments of EcR and USP were inserted into CMA-GBD or CMA-GBT to generate coding sequences for fusion proteins in which the DBD of GAL4 was placed at the amino terminus of the EcR fragment. In CMA-GBD-EcR-B1-N, CMA-GBD-EcR-B2-N, and CMA-GBD-EcR-A-N, the EcR fragment encodes the isoform-specific A/B region of isoform B1, B2, or A, plus the first six residues of the common region. CMA-GBD-EcR-B2-NS is identical to CMA-GBD-EcR-B2-N, save that only the first residue of the common region is included. In CMA-GBD-EcR-B1-N-C181, CMA-GBD-EcR-B1-N-C158, CMA-GBD-EcR-B1-N-C144, CMA-GBD-EcR-B1-N-C118, and CMA-GBD-EcR-B1-N-C56, the EcR fragment encodes the amino-terminal portion of the isoform B1-specific A/B domain, truncated after the indicated residue. The EcR fragment of CMA-GBD-EcR-B1-N-(115–181) encodes only residues 115–181 of the isoform B1-specific A/B region. CMA-GBD-EcR-B2-NS-V7A, GBD-EcR-B2-NS-E9K, and CMA-GBD-EcR-B2-NS-I14A are all identical to CMA-GBD-EcR-B2-NS except for the indicated point mutations. CMA-GBD-USP-N contains a fragment encoding USP residues 1–104.
CMA-GBT-EL contains an EcR fragment encoding residues 362–689 of EcR-B1; this region corresponds to the hinge region, the LBD, and part of the F region. CMA-GBT-EL655 and CMA-GBT-EL649 encode carboxy-truncated versions of the same fragment, ending at the indicated residues. CMA-GBT-EL-F645A and CMA-GBT-EL-W650A are identical to CMA-GBT-EL655 save for the indicated point mutations.
CMA-GBT-UL contains a fragment encoding residues 206–508 of USP, encompassing a portion of the hinge region, and the entire LBD. CMA-GBT-UL492 and CMA-GBT-UL486 contain carboxy-terminal deletions of this fragment, ending at the indicated residues. CMA-GBT-UL-L487A, CMA-GBT-UL-L490A, CMA-GBT-UL-E493K, GBT-UL-E493Q, GBD-UAF2-E493K, GBD-UAF2-E493Q, and CMA-GBT-UL-L495A are identical to CMA-GBT-UL save for the indicated point mutations.
CMA-GBD-ECT contains an fragment encoding the F region of EcR (residues 656–878 of EcR-B1).
A fragment encoding residues 638–655 of EcR-B1, corresponding to helix 12 of the LBD, is contained in CMA-GBD-EAF2. In CMA-GBD-EAF2-C689 and CMA-GBD-EAF2-C878, the EcR fragment is extended to residue 689 and 878, respectively. CMA-GBD-UAF2 contains the coding sequence for USP residues 483–508, corresponding to helix 12. CMA-GBD-UAF2-L487A, CMA-GBD-UAF2-L490A, CMA-GBD-UAF2-E493K, CMA-GBD-UAF2-E493Q, and CMA-GBD-UAF2-L495A are identical to CMA-GBD-UAF2 except for the indicated point mutations.
GAL4 Activation Domain Fusion Expression Vectors
Fragments of EcR and of USP were inserted into pCMA-GAD to encode proteins in which the transcriptional activation domain of GAL4 was fused to the amino terminus of a portion of EcR or USP. CMA-GAD-EL contains coding sequence for residues 362–689 of EcR-B1, comprising a portion of the hinge region, the complete LBD, and a portion of the F domain of EcR. CMA-GAD-UL contains the coding sequences for residues 206–508 of USP, extending from within the hinge region to the carboxy terminus of USP. CMA-GAD-UL-L490A is identical to CMA-GAD-UL save for the indicated point mutation.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Present address: Pharmacia Corp., Chesterfield, Missouri 63017. 
Abbreviations: AF, Activation function; DBD, DNA-binding domain; DTT, dithiothreitol; 20E, 20-hydroxyecdysone; EcR, ecdysone receptor; EcRE, ecdysone response element; GAD, GAL4 activation domain; GBT, GAL 4 DBD; LBD, ligand-binding domain; NR, nuclear receptor; TR, thyroid receptor; RXR, retinoid X receptor; UAS, GAL4 binding sequence; USP, ultraspiracle.
Received for publication August 19, 2002. Accepted for publication December 31, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
ligand-binding domain bound to all-trans retinoic acid. Nature 378:681–689[CrossRef][Medline]
activation functions AF1 and AF2 mediated by transcription intermediary factor TIF2. EMBO Rep 1:151–157[CrossRef][Medline]
transactivation function 1 (AF-1) and AF-2 mediated by steroid receptor coactivator protein-1: requirement for the AF-1 -helical core and for a direct interaction between the N- and C-terminal domains. Mol Endocrinol 15:1953–1970 is essential for transcriptional activation and interaction with the general transcription factor TFIIB. Mol Cell Biol 15:4507–4517[Abstract]
. Biochem J 15:253–258
. J Biol Chem 277:28624–28630 and RAR is indispensable for differentiation of F9 cells upon retinoic acid and cAMP treatment. EMBO J 16:6452–6465[CrossRef][Medline]
(RAR) and their phosphorylation are differentially involved in parietal endodermal differentiation of F9 cells and retinoic-induced expression of target genes. Mol Endocrinol 14:1398–1410. Nature 395:137–143[CrossRef][Medline]
NURSA Molecule Pages Link:
This article has been cited by other articles:
 |
|
 |
|
  H. L. D. Brown and J. W. Truman Fine-tuning of secondary arbor development: the effects of the ecdysone receptor on the adult neuronal lineages of the Drosophila thoracic CNS Development, October 1, 2009; 136(19): 3247 - 3256. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. K. Mirth, J. W. Truman, and L. M. Riddiford The Ecdysone receptor controls the post-critical weight switch to nutrition-independent differentiation in Drosophila wing imaginal discs Development, July 15, 2009; 136(14): 2345 - 2353. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Li, Z. Zhang, G. E. Robinson, and S. R. Palli Identification and Characterization of a Juvenile Hormone Response Element and Its Binding Proteins J. Biol. Chem., December 28, 2007; 282(52): 37605 - 37617. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Browning, E. Martin, C. Loch, J.-M. Wurtz, D. Moras, R. H. Stote, A. P. Dejaegere, and I. M. L. Billas Critical Role of Desolvation in the Binding of 20-Hydroxyecdysone to the Ecdysone Receptor J. Biol. Chem., November 9, 2007; 282(45): 32924 - 32934. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Kato, K. Kobayashi, S. Oda, N. Tatarazako, H. Watanabe, and T. Iguchi Cloning and characterization of the ecdysone receptor and ultraspiracle protein from the water flea Daphnia magna J. Endocrinol., April 1, 2007; 193(1): 183 - 194. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Zhu, L. Chen, G. Sun, and A. S. Raikhel The Competence Factor {beta}Ftz-F1 Potentiates Ecdysone Receptor Activity via Recruiting a p160/SRC Coactivator Mol. Cell. Biol., December 15, 2006; 26(24): 9402 - 9412. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-H. Park, G. M. Attardo, I. A. Hansen, and A. S. Raikhel GATA Factor Translation Is the Final Downstream Step in the Amino Acid/Target-of-Rapamycin-mediated Vitellogenin Gene Expression in the Anautogenous Mosquito Aedes aegypti J. Biol. Chem., April 21, 2006; 281(16): 11167 - 11176. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. S. Seto and H. J. Bellen Internalization is required for proper Wingless signaling in Drosophila melanogaster. J. Cell Biol., April 10, 2006; 173(1): 95 - 106. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. L. D. Brown, L. Cherbas, P. Cherbas, and J. W. Truman Use of time-lapse imaging and dominant negative receptors to dissect the steroid receptor control of neuronal remodeling in Drosophila Development, January 15, 2006; 133(2): 275 - 285. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Badenhorst, H. Xiao, L. Cherbas, S. Y. Kwon, M. Voas, I. Rebay, P. Cherbas, and C. Wu The Drosophila nucleosome remodeling factor NURF is required for Ecdysteroid signaling and metamorphosis Genes & Dev., November 1, 2005; 19(21): 2540 - 2545. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Sun, J. Zhu, L. Chen, and A. S. Raikhel Synergistic action of E74B and ecdysteroid receptor in activating a 20-hydroxyecdysone effector gene PNAS, October 25, 2005; 102(43): 15506 - 15511. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H.-W. Kim, E. S. Chang, and D. L. Mykles Three calpains and ecdysone receptor in the land crab Gecarcinus lateralis: sequences, expression and effects of elevated ecdysteroid induced by eyestalk ablation J. Exp. Biol., August 15, 2005; 208(16): 3177 - 3197. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L Chen, J Zhu, G Sun, and A S Raikhel The early gene Broad is involved in the ecdysteroid hierarchy governing vitellogenesis of the mosquito Aedes aegypti J. Mol. Endocrinol., December 1, 2004; 33(3): 743 - 761. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. B. Kumar, D. W. Potter, R. E. Hormann, A. Edwards, C. M. Tice, H. C. Smith, M. A. Dipietro, M. Polley, M. Lawless, P. R. N. Wolohan, et al. Highly Flexible Ligand Binding Pocket of Ecdysone Receptor: A SINGLE AMINO ACID CHANGE LEADS TO DISCRIMINATION BETWEEN TWO GROUPS OF NONSTEROIDAL ECDYSONE AGONISTS J. Biol. Chem., June 25, 2004; 279(26): 27211 - 27218. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. A. Sullivan and C. S. Thummel Temporal Profiles of Nuclear Receptor Gene Expression Reveal Coordinate Transcriptional Responses during Drosophila Development Mol. Endocrinol., November 1, 2003; 17(11): 2125 - 2137. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
