Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Homo- and heterodimers of the retinoid X receptor (RXR) activate transcription in yeast

Homo- and heterodimers of the retinoid X receptor (RXR) activate transcription in yeast Downloaded from https://academic.oup.com/nar/article-abstract/22/5/726/1047624 by DeepDyve user on 05 August 2020 726-731 Nucleic Acids Research, 1994, Vol. 22, No. 5 © 1994 Oxford University Press Homo- and heterodimers of the retinoid X receptor (RXR) activate transcription in yeast David M.Heery, Benoit Pierrat, Hinrich Gronemeyer, Pierre Chambon* and R6gine Losson Laboratoire de Gene"tique Moleculaire des Eucaryotes du Centre National de la Recherche Scientifique, Unite 184 de Biologie Moleculaire et de G6nie Genetique de I'lnstitut National de la Sante et de la Recherche M6dicale, Institut de Chimie Biologique, 11 rue Humann, 67085 Strasbourg Cedex, France Received December 22, 1993; Revised and Accepted January 31, 1994 ABSTRACT The polymorphic nature of sequences which act as RARs and RXRs are members of the steroid/thyroid hormone retlnolc acid response elements (RAREs and RXREs) receptor superfamily, and exhibit the modular protein structure in transactivatlon assays In mammalian cells, suggests typical to this group, including domains which function in DNA that elements consisting of a direct repetition of a half binding, dimerisation, ligand binding and transactivation. Ligand site motif, separated by 1 to 5 base pairs (DR1 to DR5), competition experiments and binding studies revealed that, while the three RAR types show strong affinities for both the all-trans are targets for retinolc acid (RA) signalling. In a and 9-cis isoforms of RA, the RXRs showed a marked specificity previous report we showed that in yeast cells, hetero- for the latter molecule (6-8). Ligand binding appears to be dimers of the retinoic acid receptors RARa and RXRa required to induce transactivation functions (AFs) which overlap were required for efficient transcription of a reporter the N-terminal and ligand binding domains of these proteins gene containing a DR5 element [Heery et al., (1993); (9-11) . It has recently been shown that the affinities of RARs Proc. Natl. Ac ad. Scl. USA, 90: 4281-4285] . Here we for their target sequences is strongly increased when they are report that DR1 to DR5 elements containing a direct complexed as heterodimers with RXRs (12—18), and a series repeat of the 5'-AGGTCA-3' motif, and an Inverted of heptad repeats present in the E domains of both receptors, repeat of the same sequence with no spacer (IRO), are largely responsible for heterodimer formation in solution (Ref. behave as RAREs in yeast cells coexpressing RARa and 19 for review). RXRs exhibit promiscuous heterodimerisation RXRa, albeit with different efficacies. Heterodimer properties in vitro, forming complexes with other factors activity was strongest on a DR5 reporter gene, and the including the thyroid receptor (TR), Vitamin D receptor (VDR), strength of activation of the reporter series (DR5 > DR1 peroxisome proliferator activated receptor (PPAR), the COUP- > DR3 > DR2 = IRO = DR4) correlated with the ability TF receptor, which stimulates their cooperative and selective of the heterodimer to bind to the corresponding binding to cognate hormone response elements in vitro (Refs. sequences in vitro. Significantly, a reporter containing 5, 20 and references therein). a DR1 element was selectively and efficiently activated In yeast cells expressing only RXRa. This activity was Naturally occurring RA response elements (RAREs) have been dependent on the Induction by 9-^ls retinoic acid of an identified in the promoters of a number of genes, and generally activation function (AF-2) located in the RXRa llgand consist of direct repetitions (DR) of one or more copies of a binding domain. In addition, a strong synerglstlc moderately conserved hexamer sequence, 5'-PuG(G/T)TCA-3'. activity of RXRa was observed on a reporter containing The length of the spacer sequence between the hexameric motifs the putative RXR element (RXRE) from the rat CRBPII is variable, being 5 bp in RAREs found in the genes encoding RARa2 (21), RARjS2 (22, 23), RAR 2 (24) and alcohol gene promoter. Thus, RXRa can function Independ- dehydrogenase 3 (ADH3; 25), 4 bp in the Laminin Bl gene (26), ently as a transcription factor, in the absence of RARs 2 bp as in the cellular retinol binding protein (CRBPI; 27) and or other heteromerlc partners. Similarly, homodimers cellular retinoic acid binding protein (CRABPII; 28) genes, or of RARa selectively stimulated the transcription of a 1 bp as found in the genes encoding CRABPII (28), phospho- DR5 reporter in a ligand-dependent manner, but less enolpyruvate carboxykinase (PEPCK; 29), apolipoprotein Al efficiently than RARa/RXRa heterodimers. (ApoAl; 30) and medium chain acyl-CoA dehydrogenase (MC- AD; 31). A systematic study by Mader et al., (32) showed that INTRODUCTION the sequences of the half sites, the length and sequence of the spacer between them and the flanking sequences, all influence Retinoic acid (RA) signalling involves at least two classes of the capacity of these elements to function as RAREs in transient proteins, the retinoic acid receptors (RARa, RAR/3, RAR7) and transfection experiments. In addition, it has been proposed that retinoid X receptors (RXRa, RXRjS, RXR7) (reviewed in 1 -5) . *To whom correspondence should be addressed Downloaded from https://academic.oup.com/nar/article-abstract/22/5/726/1047624 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 5 727 a distinct signalling pathway governed by RXR homodimers using an adaptor containing the RXRE sequence from the rat mediated through RXR-specific response elements (RXREs) CRBPII gene promoter (33). exists in vivo (33). Evidence to support this hypothesis came from transient transfection studies in animal cells where transfection RESULTS of plasmids expressing RXRs, but not RARs, resulted in RA- Direct repeats of the motif 5-AGGTCA-3' with different dependent activation of reporter genes driven by putative RXREs spacings and an inverted repeat of the same motif function from the rat CRBPII (33), human ApoAl (30) and mouse CRBPII as RAREs in yeast (34) promoters, and a synthetic RXRE (32). These elements share a common arrangement of one or more half sites repeated at In a previous study, we demonstrated that efficient RA induction single base pair intervals (i.e. DRl-type elements). Additionally, of a reporter gene driven by a RARE sequence (DR5-URA3) homodimers of RXRs show a preference for binding to DRl-type in yeast required heterodimers of RARs and RXRs (10), as elements in vitro (32, 34, 35). However, transient transfection observed in animal cells (28). As considerable polymorphism exists in the interval lengths found in sequences which function studies (even in insect cell-lines) are prone to interference from as RAREs in animal cells (20-31), a logical extension of this endogenous factors which may contribute to the apparent RXR- work was to test the effect of varying the distance between the specific activity. For this reason, we have chosen the yeast system half site motifs in the response element on responsiveness to to address the questions on the apparent polymorphic nature of RARs and RXRs, in yeast. The RARE used previously (10) RAREs and the possible independent function of RXR in vivo. consisted of a direct repetition of the motif 5'-AGGTCA-3' separated by a 5 bp sequence identical to that found in the RARE MATERIALS AND METHODS from the RARa2 promoter (21, 39). Reporters consisting of the yeast URA3 gene preceded by direct repeats of the same motif Yeast strains and expression plasmids separated by spacers of 1 to 4 base pairs (DR1-URA3 through S.cerevisiae strain YPH250 (MATa um3-52 Iys2-8O1 ade2-101 DR4-URA3), and also a reporter containing an inverted repeat trpl-Al his3-A200 Ieu2-Al) (36) was used for all yeast of the same motif with no spacing (IR0-URA3), were constructed experiments and was a gift from P.Hieter. The human RARal (See Materials and Methods). The sequences of these response (aa 1 -462) , mouse RXRa (aa 1 -467) , and mouse dnRXRa (aa elements and the DR5 element are indicated in Fig. 1A. The 1-448) cDNAs (10) and the cDNA encoding the mouse reporters were cloned into a centromeric plasmid, as previously RXRaAAB (aa 140-467) (11) were cloned into the unique described (10), and introduced into a yeast strain expressing EcoR 1 site of the yeast PGK expression cassette present in the RARa, RXRa or both receptors from multicopy (2/x) plasmid yeast expression vectors YEp90, YCp90, YEplO or YCplO as vectors (See Materials and Methods). Reporter (orotidine-5'- required. The construction of the 2/i containing expression vector monophosphate decarboxylase; OMPdecase) activity in cell-free YEp90, which contains the yeast HIS3 gene as a selectable extracts of cultures grown to exponential phase in the presence marker, has been described previously (37). The plasmid YEplO or absence of 500nM all-trans RA or 9-cis RA, was determined (a gift from D.Metzger) is identical to YEp90 except that the as previously described (40). As shown in Fig.IB, little or no HIS3 marker is replaced by a BglH fragment containing the yeast reporter activities were observed in the absence of RA receptors TRP1 gene. YCp90 and YCplO are ARS CEN containing vectors (control), indicating that endogenous yeast factors were incapable which carry HIS3 or TRP1 markers respectively, and which were of significant activation of these reporters. In the absence of derived from the plasmids pRS313 and pRS314 (36) as follows. ligand, coexpression of RARa and RXRa led to a 3-fold First the unique EcoRl sites in the polylinker sequences of the constitutive increase in the activities of the DR2, DR3 and DR4 plasmids pRS313 and pRS314 were destroyed by £coRI reporters, and 10-fold and 16-fold increases in the activities of digestion, treatment with Klenow fragment and religation. The the DR1 and DR5 reporters, respectively. Addition of all-trans yeast PGK expression cassette from the plasmid pTG848 (38) RA (500 nM) to the growth medium further induced the activities was then cloned as a Clal fragment into the unique Clal site in of all reporters (5- to 6-fold above the control for DR2, DR3 the polylinkers of both plasmids. and DR4, 18-fold for DR1 and 27-fold for DR5). Addition of 9-cis RA (500 nM) to the growth medium induced the DR2, DR3 Reporter plasmids and DR4 reporters approximately 10-fold, the DR1 reporter The URA3 reporter system and the construction of the 36-fold and the DR5 reporter 60-fold above their respective pRS315-DR5-URA3 reporter plasmid have been described control activities. Thus, direct repetitions of the motif 5'-AGGT- previously (10, 37). To construct a URA3 reporter series driven CA-3' with spacings of 1 to 5 base pairs render a reporter gene by different response elements (DR1—DR4 and IRO), we inducible by RA in yeast, albeit with different efficacies, which synthesised a set of adaptors containing single response element is consistent with similar observations in animal cells (5, 21 -32) . sequences as shown in Fig. 1A. All elements were flanked by Coexpression of RARa and RXRa was also found to increase identical sequences on their 5' (5'-GATCC-3') and 3 ' (5'-G-3') the activity of a reporter containing an inverted repeat of 5'-A- sides which were designed to generate ends compatible with Nhe\ GGTCA-3'. The basal activity in the absence of receptors and BglU. One copy of each adaptor was inserted into the unique (control) of the IR0-URA3 reporter was increased 15-fold by the Nhe\ and BglU restriction sites present in an engineered URA3 heterodimer in the absence of ligand, and induced 20-fold and promoter, carried in pBluescript as a HinDUJ—Pstl fragment 40-fold above the control activity in the presence of all-trans RA (10). The DRn-URA3 and IR0-URA3 promoters were then and 9-cis RA, respectively (Fig. IB). However, note that despite excised as HinDIU—Pstl fragments and used to replace the the apparently high activity in terms of fold stimulation, the actual corresponding fragment in the pRS315-DR5-URA3 reporter levels of activity for this reporter are quite low when compared plasmid, to create the new series of URA3 reporter genes. The to the DR5 reporter. Other groups have reported RAR/RXR CRBPII-URA3 reporter was constructed in an identical fashion Downloaded from https://academic.oup.com/nar/article-abstract/22/5/726/1047624 by DeepDyve user on 05 August 2020 728 Nucleic Acids Research, 1994, Vol. 22, No. 5 DR1-URA3 DR1-URA3 DR1 AGGTCAgAGGTCA RARa + RXRa RXRa 1.2- a a EC 2.0- DR2 AGGTCAgcAGGTCA DR3 AGGTCAgcgAGGTCA DR4 AGGTCAgcgaAGGTCA 1 0.8- DR5 AGGTCAgcgagAGGTCA 0) CO 1.0- IRO AGGTCATGACCT a £ 0.4- RARa + RXRa c—o-o-o-o-o 0 0- 0.0- 2.5 -9 -8 -7 -6 -5 -9 -8 -7 -6 -5 DR2-URA3 DR3-URA3 DR1-URA3 Log Ligand (M) Log Ligand (M) 2.0- 0 All-trans RA • 9-cis RA 1.5- URA3 DR1-URA3 CRBPII RXRE- 1.0- 1.2- 0.5- L-^rrlf 0.0 | 0.8- 4.2P 10- <o DR4-URA3 DR5-URA3 tf I IR0-URA3 2.0 <D • o | 0.4- 1.5- 1.0- 0.0- o B o m o 8 8 ~ 5 5 0.5 § oc oc cc 8 K cc | tr 0.0 • No ligand All-trans RA • 9-cis RA < X X Figure 2. Transactivation of the DRI-URA3 reporter gene by RXRa and RARa/RXRa heterodimers, synergistic activation of the CRBPII RXRE-URA3 D noLigand 0 All-trans RA 9-cis RA by RXRa, and effect of deletions in RXRa on reporter activation. A. Ligand dose responses of DR1 reporter activity in the presence of RXRa and RARa/RXRa heterodimers. Activation of the DR1-URA3 reporter measured as OMPdecase activities in cell-free extraas of yeast expressing RXRa. or RARa and RXRa together, grown in the presence of all-trans RA (empty circles) and 9-cis RA (filled circles) at the indicated concentrations, as described previously Figure 1. Direct repeats of a half site motif with different spacuigs and an inverted (10). B. Activation of DR1-URA3 and CRBPII RXRE-URA3 reporters in yeast repeat of the same sequence with no spacing, function as RAREs in yeast A. expressing RARa. RXRa. RXRaAAB and dnRXRa or no receptor (control) Sequences of the response elements used in this study consisting of a direct repeat in the presence or absence of ligand. of a hexamer sequence spaced by one to five base pairs (DR1 -DR5) , or an inverted repeat of the same hexamer with no spacing (IRO). The hexamer repeats are indicated by arrows and the spacer sequences are presented in lower case. activity on a similar response element (TREpal) in animal cells B. Transactivation of URA3 reporter genes containing the different elements by (33, 41) and in yeast (42). homodimers and heterodimers of RARa and RXRa. The sequences represented in Fig. IA were cloned into the promoter of a URA3 reporter gene carried on To determine how the observed transactivation data correlated a centromertc plasmid (See Materials and Methods) to generate the DRn-URA3 with the ability of the receptors to bind to the corresponding reporter series and IR0-URA3. The reporters were introduced into a yeast strain sequences, we performed in vitro DNA binding assays. Extracts containing multicopy vectors expressing RARa, RXRa or no receptor. Reporter from yeast cells expressing both receptors efficiently retarded activities in the presence or absence of ligand (500nM) were determined by the mobility of the DR1 and DR5 probes, and bound more weakly measuring the OMPdecase activity (38). 'Control' indicates the basal reporter activity in the absence of receptors (using 'empty' expression vectors), and the to the DR2, DR3, DR4 probes under identical conditions (Fig. experiments using only one receptor were performed in the presence of the 1C, lanes 1-5). A weak band shift was also observed for the corresponding 'empty' vector. The white, hatched and black columns indicate IRO element (lane 6) after prolonged autoradiography. The the reporter activity in the presence of no ligand, all-trans RA and 9-cis RA, presence of RARa and RXRa in these complexes was verified respectively. The reporter activities are given as units of OMPdecase activity per minute per mg protein, and the values represent the average of at least 2 by supershift assays using specific monoclonal antibodies against experiments using at least 2 different clones per experiment. Deviation of values the receptors (not shown). Thus, the efficiency of reporter was less than 10%. C . GeJ retardation assays using radiolabelled probes (containing activation correlates closely with the capacity of the heterodimer the sequences described in 1A) and cell-free extracts from yeast coexpressing to bind to the corresponding element in vitro. Note that the RARa and RXRa were performed as described previously (10). Similar amounts relatively low affinity of RAR/RXR heterodimers for the DR2 of each probe (50,000 cpm) were used in the assays. ivity RXRa/ Downloaded from https://academic.oup.com/nar/article-abstract/22/5/726/1047624 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 5 729 and DR4 probes in vitro, and the weak activation of the dominant negative RXRa (dnRXRaAAB; refs. 10, 28) confirmed corresponding reporter genes in yeast, is a consequence of the that RARa is responsible for this stimulation by all-trans RA spacer and flanking sequences of the elements used in this study. (D.M.H. and B.P., unpublished results). Thus, the DR1 element For example, a DR2 probe with different spacer and flanking appears to bind both RAR/RXR heterodimers and RXR alone sequences was bound 6-fold more efficiently by RARs and RXRs (presumably as homodimers). This result is in contrast to a recent produced in yeast, or E.coli (Z.P.Chen, personal communi- report by Hall et al., (42) who were unable to demonstrate significant activity of RXR on a DR1 element. Note that these cation). This is consistent with the observations that changes in authors used another receptor isoform (RXR7), a different the spacer, half site motif or the flanking sequences can reporter (based on the yeast CYC1 promoter), and a different significantly affect binding and activation by RARs and RXRs spacer nucleotide in the DR1 sequence as compared to this study. produced in mammalian cells (21, 32) and in yeast (our To confirm that the RXRa activity we observed on DR1 was unpublished results). not simply due to high level expression of the receptor in yeast, The ability of the isolated receptors to activate the reporter we expressed RXRa from a centromeric expression vector. series was also tested. As previously reported (10) expression Significant 9-cis RA induced activation was maintained on the of RARa alone in yeast resulted in a 5- to 7-fold activation of DR1 and IRO reporters (10- to 15-fold), despite a strong decrease DR5 reporter in the presence of either all-trans RA or 9-cis RA (30- to 50-fold) in the level of receptor detectable in western blot (Fig. IB). However, RARa alone showed little if any ability to analyses (data not shown). stimulate the activity of the DR1, DR2, DR3, DR4 and IRO reporters, in the presence or absence of ligand. Similarly RXRa The rat CRBPII gene promoter contains a sequence consisting had little effect on the DR2, DR3, DR4 and DR5 reporter of four almost perfect repeats of the hexamer motif 5'-AGGT- activities (2-3 fold 9-cis RA-dependent stimulation), but CA -3 ' with lbp spacers, which has been reported to act as a RXRE in animal cells (33, 34). We determined if this element stimulated the activity of the DR1 reporter at least 30-fold in would also function as a target for transactivation by RXRa in response to 9-cis RA, but not all-trans RA (Fig. IB). Western yeast. As shown in Fig. 2B, the rat CRBPII RXRE reporter was blot experiments using specific monoclonal antibodies (gifts from very efficiently activated (116-fold over the control) by RXRa C.Egly and M.P.Gaub) detected similar levels of RARa or in the presence of 9-cis RA in yeast, i.e. almost 20-fold higher RXRa in extracts from cells expressing either one or both than that of the DR1 reporter in the same system. This result receptors (data not shown), thus eliminating the possibility that indicates a synergistic activation by RXRa molecules bound to the observed increases in DNA binding and transactivation by the multiple binding sites in this promoter. As observed with the heterodimers were due to altered levels of receptor proteins. DR1 reporter, the activity of the rat CRBPII RXRE reporter was RXRa also activated the IRO reporter 20-fold above background also significantly induced (12-fold) in the presence of 9-cis RA, in a ligand dependent manner (Fig. IB), although as stated above, when RXRa was expressed from a centromeric vector (data not the absolute value of this activity was low (6-fold lower than the shown). These results indicate a synergistic activation by RXRa DR1 reporter). Cell-free extracts from transformants expressing molecules bound to the multiple binding sites in this promoter. no receptor, or either of the receptors alone, failed to retard the Coexpression of RARa and RXRa in yeast increased the activity DR5 probe (10) or any of the other DR probes in gel shift of the rat CRBPII RXRE reporter 14-fold above the control in experiments under conditions identical to those used in the the absence of ligand, and 21-fold and 70-fold above the control experiment shown in Fig. 1C (data not shown). However, after in the presence of all-trans RA and 9-cis RA, respectively (data prolonged autoradiography, a weak band shift was detected for not shown). Accordingly, gel shift experiments showed that RXRa homodimers which was specific for the DR1 probe. These results are consistent with the lack of significant constitutive RARa strongly increased the binding of RXRa to the rat CRBPII activation of the reporter gene series by homodimers. In our RXRE element in vitro (not shown), and thus coexpression of hands, the inclusion of all-trans RA or 9-cis RA (500 nM) in both receptors in yeast did not have the same dramatic negative the growth medium, extraction buffer and assay buffer did not effect on activation of the rat CRBPII RXRE reporter as further stimulate the binding of either homodimers or previously observed in some (33), but not all (34), animal cells. heterodimers to any of the probes (data not shown), although Our previous study (10) established that a 9-cis RA-inducible we did not determine if the conditions used in the gel shift activation function (AF-2) present in the ligand binding domain is largely responsible for the activity of RXRa in yeast, as we experiments were optimal for the binding of ligand to the failed to detect a strong activation function in the N-terminal (AB) receptors. region of RXRa by linking it to a heterologous DBD. 9-cis retinoic acid induced activity of RXRa on putative Accordingly, deletion of the A/B region of RXRa (RXRaAAB) RXRE sequences in yeast did not adversely affect the levels of activation achieved on either Ligand dose response curves (Fig. 2A) revealed that the the DR1 or rat CRBPII RXRE reporters, in the presence or concentration of 9-cis RA required for half maximal activity of absence of ligand (Fig. 2B). In contrast, a truncated receptor in RXRa on the DR1 reporter in yeast was approximately lOOnM. which the 19 most carboxy-terminal amino acids were deleted In contrast, no induction was observed in the presence of all- (dnRXRa), which behaves as a dominant negative receptor in trans RA, even at a concentration of ifiM. This is in agreement mammalian cells (28), failed to mediate a response of either the with our previous data for the chimeric receptors RXRa-ER(C) DR1 or the rat CRBPII RXRE reporters to 9-cis RA in yeast and RXRa(DEF)-ER.Cas, whose activities were induced (Fig. 2B), despite being expressed to a similar level as the full exclusively by 9-cis stereoisomers of RA in yeast (10). The dose length receptor (not shown). This result indicates that the carboxyl response curves in the presence of both RARa and RXRa terminal region is required for ligand-induced transcription by confirmed the modest stimulation of the DR1 reporter activity RXRa. It is also of note that unliganded RXRa stimulated the by the heterodimer in the presence of all-trans RA (Fig. 2A). rat CRBPII RXRE reporter activity 8- to 10-fold over the activity In addition, experiments using a transcriptionally inactive of the control (Fig. 2B). This constitutive activity was observed Downloaded from https://academic.oup.com/nar/article-abstract/22/5/726/1047624 by DeepDyve user on 05 August 2020 730 Nucleic Acids Research, 1994, Vol. 22, No. 5 plasmids. Monoclonal antibodies raised against RARa and RXRa even when the A/B domain was deleted, and thus may indicate were generously provided by C.Egly, M.P.Gaub and Y.Lutz. a weak constitutive activity of AF-2. We also thank the secretariat, illustration and photography services for help in preparing the manuscript. This work was DISCUSSION supported by funds from the Centre National de la Recherche Scientifique, the Institut National de la Same" et de la Recherche We have demonstrated that RXRa activates transcription M&licale, the Association pour la Recherche sur le Cancer, the efficiently in a biological system devoid of other members of the Fondation pour la Recherche Mddicale and the Centre Hospitalier nuclear receptor superfamily. This activity of RXRa was found Universitaire Regional. D.M.H. was awarded fellowships from to be specific for elements having a DRl-type structure, so-called the European Molecular Biology Organisation and the Association 'RXREs', and to a lesser extent, an inverted repeat similar to pour la Recherche sur le Cancer, and B.P. from the Ministere TRE-pal. DRl-type elements have been found in the promoters de la Recherche et de la Technologie, France. of genes including the rat CRBPII (33), human ApoAl (30) and mouse CRABPII (28) genes, which are involved in vitamin A metabolism, and it has been suggested that this may signify the REFERENCES existence of a 9-cis RA autoregulatory network which operates through RXR (43). However, DR1 elements are also recognised 1. Green. S and Chambon. P. (1988) Trends Genet. 4. 309-314 by other members of the steroid receptor superfamily, including 2 Evans. R M. (1988) Science 240. 889-895. PPAR, COUP-TF, Arp-1, HNF4, and ear2 (20, and references 3. DeLuca. L. M (1991) FASEB J. 5. 2924-2933. 4. Linney. E. (1992) In Current Topics in Dewlopmental Biolog\\ Vol. 27. therein). Thus it was important to test the ability of RXR to Academic Press. activate transcription from a DR1 element in a system which has 5. Leid, M.. Kastner. P. and Chambon. P. (1992) Trends Biodiem. Sri. 17. no known homologues of nuclear receptors. Transactivation was 427-433 . very potent on the rat CRBPII RXRE reporter, which contains 6. Heyman. R.A.. Mangelsdorf. D J.. Dyck. J A.. Stein. R B.. Eichele. multiple DR1 repeats, suggesting synergistic activation by RXRa G.. Evans. R. M and Thaller. C. (1992) Cell 68. 397-406. 7. Levin. A. A.. Sturzenbecker. L. J., Kazmer, S.. Bosakowski. T . Huselton. molecules bound to the multiple target sites. Our findings that C Allenby, G.. Speck. J., Kratzeisen. Cl . Rosenberger. M.. Lovey. A. RXRa strongly activates reporter genes driven by DR 1 elements and Grippo. J. F. (1992) Nature (London) 355. 359-361. in yeast are consistent with the hypothesis that, in addition to 8. Allenby, G.. Bocquel. M -T.. Saunders. M.. Kazmer. S.. Speck. J.. Grippo. their role as heterodimeric partners of RAR, VDR, PPAR and J. F . Chambon. P. and Levin. A A (1992) Proc. Natl. Acad. Set. USA TR, RXRs may also function independently as ligand-dependent 89. 30-34. 9. Nagpal, S.. Saunders. M., Kastner. P.. Durand. B.. Nakshatri. H. and transcription factors, in the more classical fashion of steroid Chambon. P (1992) Cell 70. 1007-1019. receptors, i.e. as homodimers. Similarly, homodimers of RARa 10. Heery. D. M.. Zacharewski. T.. Pierral. B.. Groncmcyer. H.. Chambon. were found to activate transcription in yeast (albeit less efficiently P. and Losson. R (1993) Proc. Natl. Acad. Sci. USA 90. 4281-4285. than heterodimers) in a ligand-dependent manner, but only on 11 Nagpal. S . Friant. S . Nakshaln. H. and Chambon (1993) EMBO J. 12. 2349-2360. a DR5 reporter. As with steroid receptors such as the ER, 12. Leid. M.. Kastner. P . Lyons. R.. Nakshatn. H.. Saunders. M.. Zacharewski. transactivation by homodimers of RARs and RXRs appears to T.. Chen. J.-Y.. Staub. A.. Gamier, J.-M . Madcr. S. and Chambon. P. be largely ligand-dependent, and in the case of RXRa is due to (1992) Cell 68, 377-395. the presence of an activation function (AF-2) in the ligand-binding 13. Yu. V.C.. Delsert. C . Andersen. B.. Holloway. J. M., Devary. O. V.. Naar. Kim, S. Y.. Boutin, J -M., Glass. C. K and Roscnfcld. M. G. (1991) domain which is specifically induced by 9-cis stereoisomers of CW/67. 1251-1266. RA. In addition AF-2 may also be responsible for the weak 14. Kliewer, S A.. Umesono, K.. Mangelsdorf. D.J. and Evans. R M. (1992) constitutive activity of RXRa observed on the rat CRBPII RXRE Nature (London) 355. 446-449. and DR1 reporters. In contrast, the A/B region of RXRa appears 15 Zhang, X K.. Hoffmann. B . Tran. P B V , Graupner. G. and Pfahl. M. to be dispensable for the transcriptional activity of this receptor (1992) Nature (London) 355. 441 -446 . 16. Marks. M S . Hallenbeck, P. L . Nagata. T . Segars. J H . Appella. E . in our experiments. Nikodem. V. M. and Ozata. K. (1992) EMBO J. 11. 1419-1435. RA-dependent transactivation in animal cells appears to be 17. Bugge, T. H . Pohl. J., Lonnoy. O and Stunncnberg. H. G (1992) EMBO mediated through DRs consisting of half-site motifs spaced by 7.11. 1409-1418 18 Berrodin, T. J . Marks. M S., Ozato. K , Linney. E and Lazar. M. A 1 to 5 base pairs. Our results in yeast corroborate this observation, (1992) Mol. Endocrmol. 6. 1468-1478 showing a strong correlation between the affinity of the 19 Forman, B.. M. and Samuels. H H. (1990) Tl\e New Biologist 2. 587-594. heterodimer complex for a sequence and the strength of activation 20. Green, S. (1993) Nature (London). 361. 590-591 of the corresponding reporter gene in yeast. Although in this 21. Leroy, P., Nakshatn. H. and Chambon. P. (1991) Proc. Nail. Acad. Sci. USA 88. 10138-10142. report we have only considered the effect of spacer length on 22 de The\ H . Vivanco-Ruiz. M . P. Tiollais. H. Stunnenberg and Dcjean. DNA binding and activation, additional experiments have A. (1990) Nature (London) 343. 177-180. confirmed that the sequences of the half sites, spacers and flanking 23. Sucov. H M., Murakami, K K and Evans. R. (1990) Proc. Nail Acad. DNA all clearly influence the strength of binding and activation Sci. USA 87. 5392-5396. 24. Lehmann. J M.. Zhang. X. K. and Pfahl. M (1992) Mol Cell. Biol. 12, by heterodimers in yeast (our unpublished results), in agreement 2976-2985. with the findings of Mader et al., (32). 25. Duester, G., Shean, M. L., McBride, M. S and Stewart. M. J. (1991) Mol. Cell. Biol. 11. 1638-1648 26. Vasios. G., Madcr. S., Gold. J. D., Leud. M.. Lulz. Y.. Gaub. M. P . ACKNOWLEDGMENTS Chambon, P. and Gudas, L (1991) EMBO J. 10. 1149- 1158. 27 Smith. W. C . Nakshatri. H.. Leroy, P.. Rees. J. and Chambon. P. (1991) D.M.H. and B.P. made equal contributions to this work. We EMBO J. 10, 2223-2230. are grateful to all our colleagues at LGME for the provision of 28. Durand. B., Saunders. M.. Leroy. P., Leid. M. and Chambon. P. (1992) research materials and for helpful discussions. In particular we Cell 71. 73-85 . acknowledge the excellent assistance of J.-M.Garnier and 29. Hall. R. K.. Scott, D. K., Noisin. E. L.. Lucas. P. C , and Granncr, D. coworkers in the construction of several receptor expression K (1992) Mol. Cell Biol. 12. 5527-5535 Downloaded from https://academic.oup.com/nar/article-abstract/22/5/726/1047624 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 5 731 30. Rottman, J. N., Widom, R. J., Nadal-Gidard, B., Mahdavi, V., and Karathanasis, S. K. (1991) Mol. Cell. Biol. 11, 3814-3820. 31. Raisher, B. D., Gulkk, T., Zhang, Z., Strauss, A. W., Moore, D. D. and Kelly, D. P. (1992) J. Biol. Chem. 267, 20264-20269. 32. Mader, S., Leroy, P., Chen, J-Y. and Chambon, P. (1993) J. Biol. Chem. 268, 591-600. 33. Mangelsdorf, D. J., Umesono, K., Kliewer, S. A., Borgmeyer, U., Ong, E. S. and Evans, R. M. (1991) Cell 66, 555-561. 34. Nakshatri, H. and Chambon, P. (1993) J. Biol. Chem. (in press). 35. Kliewer, S. A., Umesono, K., Heymann, R. A., Mangelsdorf, D. J., Dyck, J. A. and Evans, R. M. (1992). Proc. Natl Acad. Sd. USA 89, 1448-1452. 36. Sikorski, R. S. and Hieter, P. (1989) Genetics 122, 19-27. 37. Pierrat, B., Heery, D. M., Lemoine, Y. and Losson, R. (1992) Gene 119, 237-245 . 38. Loison, G., Vidal, A., Findeli, A., Roitsch, C , Bellou, J. M. and Lemoine, Y. (1989) Yeast 5, 497-507. 39. Brand, N. J., Petkovich, M. and Chambon, P. (1990) Nuc. Acids Res. 18, 6799-6806. 40. Wolcott, J.H. and Ross, C. (1966) Biochem. Biophys. Acta 122, 532-534. 41. Umesono, K., Giguere, V., Glass, C. K., Rosenfeld, M. G. and Evans, R. M. (1988) Nature 345, 262-265. 42. Hall, B. L, Smit-McBride, Z. and Privalsky, M. L. (1993) Proc. Nail. Acad. Sci. USA 90, 6929-6933. 43. Lucas, P. C. and Granner, D. K. (1992) Annu. Rev. Biochem. 61, 1131-1173. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nucleic Acids Research Oxford University Press

Homo- and heterodimers of the retinoid X receptor (RXR) activate transcription in yeast

Download PDF
6 pages

Loading...

Page 2

Loading...

Page 3

Loading...

Page 4

Loading...

Page 5

Loading...

Page 6

 
/lp/oxford-university-press/homo-and-heterodimers-of-the-retinoid-x-receptor-rxr-activate-VWWo9ONIs0

References (1)

Publisher
Oxford University Press
Copyright
© Oxford University Press
ISSN
0305-1048
eISSN
1362-4962
DOI
10.1093/nar/22.5.726
Publisher site
See Article on Publisher Site

Abstract

Downloaded from https://academic.oup.com/nar/article-abstract/22/5/726/1047624 by DeepDyve user on 05 August 2020 726-731 Nucleic Acids Research, 1994, Vol. 22, No. 5 © 1994 Oxford University Press Homo- and heterodimers of the retinoid X receptor (RXR) activate transcription in yeast David M.Heery, Benoit Pierrat, Hinrich Gronemeyer, Pierre Chambon* and R6gine Losson Laboratoire de Gene"tique Moleculaire des Eucaryotes du Centre National de la Recherche Scientifique, Unite 184 de Biologie Moleculaire et de G6nie Genetique de I'lnstitut National de la Sante et de la Recherche M6dicale, Institut de Chimie Biologique, 11 rue Humann, 67085 Strasbourg Cedex, France Received December 22, 1993; Revised and Accepted January 31, 1994 ABSTRACT The polymorphic nature of sequences which act as RARs and RXRs are members of the steroid/thyroid hormone retlnolc acid response elements (RAREs and RXREs) receptor superfamily, and exhibit the modular protein structure in transactivatlon assays In mammalian cells, suggests typical to this group, including domains which function in DNA that elements consisting of a direct repetition of a half binding, dimerisation, ligand binding and transactivation. Ligand site motif, separated by 1 to 5 base pairs (DR1 to DR5), competition experiments and binding studies revealed that, while the three RAR types show strong affinities for both the all-trans are targets for retinolc acid (RA) signalling. In a and 9-cis isoforms of RA, the RXRs showed a marked specificity previous report we showed that in yeast cells, hetero- for the latter molecule (6-8). Ligand binding appears to be dimers of the retinoic acid receptors RARa and RXRa required to induce transactivation functions (AFs) which overlap were required for efficient transcription of a reporter the N-terminal and ligand binding domains of these proteins gene containing a DR5 element [Heery et al., (1993); (9-11) . It has recently been shown that the affinities of RARs Proc. Natl. Ac ad. Scl. USA, 90: 4281-4285] . Here we for their target sequences is strongly increased when they are report that DR1 to DR5 elements containing a direct complexed as heterodimers with RXRs (12—18), and a series repeat of the 5'-AGGTCA-3' motif, and an Inverted of heptad repeats present in the E domains of both receptors, repeat of the same sequence with no spacer (IRO), are largely responsible for heterodimer formation in solution (Ref. behave as RAREs in yeast cells coexpressing RARa and 19 for review). RXRs exhibit promiscuous heterodimerisation RXRa, albeit with different efficacies. Heterodimer properties in vitro, forming complexes with other factors activity was strongest on a DR5 reporter gene, and the including the thyroid receptor (TR), Vitamin D receptor (VDR), strength of activation of the reporter series (DR5 > DR1 peroxisome proliferator activated receptor (PPAR), the COUP- > DR3 > DR2 = IRO = DR4) correlated with the ability TF receptor, which stimulates their cooperative and selective of the heterodimer to bind to the corresponding binding to cognate hormone response elements in vitro (Refs. sequences in vitro. Significantly, a reporter containing 5, 20 and references therein). a DR1 element was selectively and efficiently activated In yeast cells expressing only RXRa. This activity was Naturally occurring RA response elements (RAREs) have been dependent on the Induction by 9-^ls retinoic acid of an identified in the promoters of a number of genes, and generally activation function (AF-2) located in the RXRa llgand consist of direct repetitions (DR) of one or more copies of a binding domain. In addition, a strong synerglstlc moderately conserved hexamer sequence, 5'-PuG(G/T)TCA-3'. activity of RXRa was observed on a reporter containing The length of the spacer sequence between the hexameric motifs the putative RXR element (RXRE) from the rat CRBPII is variable, being 5 bp in RAREs found in the genes encoding RARa2 (21), RARjS2 (22, 23), RAR 2 (24) and alcohol gene promoter. Thus, RXRa can function Independ- dehydrogenase 3 (ADH3; 25), 4 bp in the Laminin Bl gene (26), ently as a transcription factor, in the absence of RARs 2 bp as in the cellular retinol binding protein (CRBPI; 27) and or other heteromerlc partners. Similarly, homodimers cellular retinoic acid binding protein (CRABPII; 28) genes, or of RARa selectively stimulated the transcription of a 1 bp as found in the genes encoding CRABPII (28), phospho- DR5 reporter in a ligand-dependent manner, but less enolpyruvate carboxykinase (PEPCK; 29), apolipoprotein Al efficiently than RARa/RXRa heterodimers. (ApoAl; 30) and medium chain acyl-CoA dehydrogenase (MC- AD; 31). A systematic study by Mader et al., (32) showed that INTRODUCTION the sequences of the half sites, the length and sequence of the spacer between them and the flanking sequences, all influence Retinoic acid (RA) signalling involves at least two classes of the capacity of these elements to function as RAREs in transient proteins, the retinoic acid receptors (RARa, RAR/3, RAR7) and transfection experiments. In addition, it has been proposed that retinoid X receptors (RXRa, RXRjS, RXR7) (reviewed in 1 -5) . *To whom correspondence should be addressed Downloaded from https://academic.oup.com/nar/article-abstract/22/5/726/1047624 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 5 727 a distinct signalling pathway governed by RXR homodimers using an adaptor containing the RXRE sequence from the rat mediated through RXR-specific response elements (RXREs) CRBPII gene promoter (33). exists in vivo (33). Evidence to support this hypothesis came from transient transfection studies in animal cells where transfection RESULTS of plasmids expressing RXRs, but not RARs, resulted in RA- Direct repeats of the motif 5-AGGTCA-3' with different dependent activation of reporter genes driven by putative RXREs spacings and an inverted repeat of the same motif function from the rat CRBPII (33), human ApoAl (30) and mouse CRBPII as RAREs in yeast (34) promoters, and a synthetic RXRE (32). These elements share a common arrangement of one or more half sites repeated at In a previous study, we demonstrated that efficient RA induction single base pair intervals (i.e. DRl-type elements). Additionally, of a reporter gene driven by a RARE sequence (DR5-URA3) homodimers of RXRs show a preference for binding to DRl-type in yeast required heterodimers of RARs and RXRs (10), as elements in vitro (32, 34, 35). However, transient transfection observed in animal cells (28). As considerable polymorphism exists in the interval lengths found in sequences which function studies (even in insect cell-lines) are prone to interference from as RAREs in animal cells (20-31), a logical extension of this endogenous factors which may contribute to the apparent RXR- work was to test the effect of varying the distance between the specific activity. For this reason, we have chosen the yeast system half site motifs in the response element on responsiveness to to address the questions on the apparent polymorphic nature of RARs and RXRs, in yeast. The RARE used previously (10) RAREs and the possible independent function of RXR in vivo. consisted of a direct repetition of the motif 5'-AGGTCA-3' separated by a 5 bp sequence identical to that found in the RARE MATERIALS AND METHODS from the RARa2 promoter (21, 39). Reporters consisting of the yeast URA3 gene preceded by direct repeats of the same motif Yeast strains and expression plasmids separated by spacers of 1 to 4 base pairs (DR1-URA3 through S.cerevisiae strain YPH250 (MATa um3-52 Iys2-8O1 ade2-101 DR4-URA3), and also a reporter containing an inverted repeat trpl-Al his3-A200 Ieu2-Al) (36) was used for all yeast of the same motif with no spacing (IR0-URA3), were constructed experiments and was a gift from P.Hieter. The human RARal (See Materials and Methods). The sequences of these response (aa 1 -462) , mouse RXRa (aa 1 -467) , and mouse dnRXRa (aa elements and the DR5 element are indicated in Fig. 1A. The 1-448) cDNAs (10) and the cDNA encoding the mouse reporters were cloned into a centromeric plasmid, as previously RXRaAAB (aa 140-467) (11) were cloned into the unique described (10), and introduced into a yeast strain expressing EcoR 1 site of the yeast PGK expression cassette present in the RARa, RXRa or both receptors from multicopy (2/x) plasmid yeast expression vectors YEp90, YCp90, YEplO or YCplO as vectors (See Materials and Methods). Reporter (orotidine-5'- required. The construction of the 2/i containing expression vector monophosphate decarboxylase; OMPdecase) activity in cell-free YEp90, which contains the yeast HIS3 gene as a selectable extracts of cultures grown to exponential phase in the presence marker, has been described previously (37). The plasmid YEplO or absence of 500nM all-trans RA or 9-cis RA, was determined (a gift from D.Metzger) is identical to YEp90 except that the as previously described (40). As shown in Fig.IB, little or no HIS3 marker is replaced by a BglH fragment containing the yeast reporter activities were observed in the absence of RA receptors TRP1 gene. YCp90 and YCplO are ARS CEN containing vectors (control), indicating that endogenous yeast factors were incapable which carry HIS3 or TRP1 markers respectively, and which were of significant activation of these reporters. In the absence of derived from the plasmids pRS313 and pRS314 (36) as follows. ligand, coexpression of RARa and RXRa led to a 3-fold First the unique EcoRl sites in the polylinker sequences of the constitutive increase in the activities of the DR2, DR3 and DR4 plasmids pRS313 and pRS314 were destroyed by £coRI reporters, and 10-fold and 16-fold increases in the activities of digestion, treatment with Klenow fragment and religation. The the DR1 and DR5 reporters, respectively. Addition of all-trans yeast PGK expression cassette from the plasmid pTG848 (38) RA (500 nM) to the growth medium further induced the activities was then cloned as a Clal fragment into the unique Clal site in of all reporters (5- to 6-fold above the control for DR2, DR3 the polylinkers of both plasmids. and DR4, 18-fold for DR1 and 27-fold for DR5). Addition of 9-cis RA (500 nM) to the growth medium induced the DR2, DR3 Reporter plasmids and DR4 reporters approximately 10-fold, the DR1 reporter The URA3 reporter system and the construction of the 36-fold and the DR5 reporter 60-fold above their respective pRS315-DR5-URA3 reporter plasmid have been described control activities. Thus, direct repetitions of the motif 5'-AGGT- previously (10, 37). To construct a URA3 reporter series driven CA-3' with spacings of 1 to 5 base pairs render a reporter gene by different response elements (DR1—DR4 and IRO), we inducible by RA in yeast, albeit with different efficacies, which synthesised a set of adaptors containing single response element is consistent with similar observations in animal cells (5, 21 -32) . sequences as shown in Fig. 1A. All elements were flanked by Coexpression of RARa and RXRa was also found to increase identical sequences on their 5' (5'-GATCC-3') and 3 ' (5'-G-3') the activity of a reporter containing an inverted repeat of 5'-A- sides which were designed to generate ends compatible with Nhe\ GGTCA-3'. The basal activity in the absence of receptors and BglU. One copy of each adaptor was inserted into the unique (control) of the IR0-URA3 reporter was increased 15-fold by the Nhe\ and BglU restriction sites present in an engineered URA3 heterodimer in the absence of ligand, and induced 20-fold and promoter, carried in pBluescript as a HinDUJ—Pstl fragment 40-fold above the control activity in the presence of all-trans RA (10). The DRn-URA3 and IR0-URA3 promoters were then and 9-cis RA, respectively (Fig. IB). However, note that despite excised as HinDIU—Pstl fragments and used to replace the the apparently high activity in terms of fold stimulation, the actual corresponding fragment in the pRS315-DR5-URA3 reporter levels of activity for this reporter are quite low when compared plasmid, to create the new series of URA3 reporter genes. The to the DR5 reporter. Other groups have reported RAR/RXR CRBPII-URA3 reporter was constructed in an identical fashion Downloaded from https://academic.oup.com/nar/article-abstract/22/5/726/1047624 by DeepDyve user on 05 August 2020 728 Nucleic Acids Research, 1994, Vol. 22, No. 5 DR1-URA3 DR1-URA3 DR1 AGGTCAgAGGTCA RARa + RXRa RXRa 1.2- a a EC 2.0- DR2 AGGTCAgcAGGTCA DR3 AGGTCAgcgAGGTCA DR4 AGGTCAgcgaAGGTCA 1 0.8- DR5 AGGTCAgcgagAGGTCA 0) CO 1.0- IRO AGGTCATGACCT a £ 0.4- RARa + RXRa c—o-o-o-o-o 0 0- 0.0- 2.5 -9 -8 -7 -6 -5 -9 -8 -7 -6 -5 DR2-URA3 DR3-URA3 DR1-URA3 Log Ligand (M) Log Ligand (M) 2.0- 0 All-trans RA • 9-cis RA 1.5- URA3 DR1-URA3 CRBPII RXRE- 1.0- 1.2- 0.5- L-^rrlf 0.0 | 0.8- 4.2P 10- <o DR4-URA3 DR5-URA3 tf I IR0-URA3 2.0 <D • o | 0.4- 1.5- 1.0- 0.0- o B o m o 8 8 ~ 5 5 0.5 § oc oc cc 8 K cc | tr 0.0 • No ligand All-trans RA • 9-cis RA < X X Figure 2. Transactivation of the DRI-URA3 reporter gene by RXRa and RARa/RXRa heterodimers, synergistic activation of the CRBPII RXRE-URA3 D noLigand 0 All-trans RA 9-cis RA by RXRa, and effect of deletions in RXRa on reporter activation. A. Ligand dose responses of DR1 reporter activity in the presence of RXRa and RARa/RXRa heterodimers. Activation of the DR1-URA3 reporter measured as OMPdecase activities in cell-free extraas of yeast expressing RXRa. or RARa and RXRa together, grown in the presence of all-trans RA (empty circles) and 9-cis RA (filled circles) at the indicated concentrations, as described previously Figure 1. Direct repeats of a half site motif with different spacuigs and an inverted (10). B. Activation of DR1-URA3 and CRBPII RXRE-URA3 reporters in yeast repeat of the same sequence with no spacing, function as RAREs in yeast A. expressing RARa. RXRa. RXRaAAB and dnRXRa or no receptor (control) Sequences of the response elements used in this study consisting of a direct repeat in the presence or absence of ligand. of a hexamer sequence spaced by one to five base pairs (DR1 -DR5) , or an inverted repeat of the same hexamer with no spacing (IRO). The hexamer repeats are indicated by arrows and the spacer sequences are presented in lower case. activity on a similar response element (TREpal) in animal cells B. Transactivation of URA3 reporter genes containing the different elements by (33, 41) and in yeast (42). homodimers and heterodimers of RARa and RXRa. The sequences represented in Fig. IA were cloned into the promoter of a URA3 reporter gene carried on To determine how the observed transactivation data correlated a centromertc plasmid (See Materials and Methods) to generate the DRn-URA3 with the ability of the receptors to bind to the corresponding reporter series and IR0-URA3. The reporters were introduced into a yeast strain sequences, we performed in vitro DNA binding assays. Extracts containing multicopy vectors expressing RARa, RXRa or no receptor. Reporter from yeast cells expressing both receptors efficiently retarded activities in the presence or absence of ligand (500nM) were determined by the mobility of the DR1 and DR5 probes, and bound more weakly measuring the OMPdecase activity (38). 'Control' indicates the basal reporter activity in the absence of receptors (using 'empty' expression vectors), and the to the DR2, DR3, DR4 probes under identical conditions (Fig. experiments using only one receptor were performed in the presence of the 1C, lanes 1-5). A weak band shift was also observed for the corresponding 'empty' vector. The white, hatched and black columns indicate IRO element (lane 6) after prolonged autoradiography. The the reporter activity in the presence of no ligand, all-trans RA and 9-cis RA, presence of RARa and RXRa in these complexes was verified respectively. The reporter activities are given as units of OMPdecase activity per minute per mg protein, and the values represent the average of at least 2 by supershift assays using specific monoclonal antibodies against experiments using at least 2 different clones per experiment. Deviation of values the receptors (not shown). Thus, the efficiency of reporter was less than 10%. C . GeJ retardation assays using radiolabelled probes (containing activation correlates closely with the capacity of the heterodimer the sequences described in 1A) and cell-free extracts from yeast coexpressing to bind to the corresponding element in vitro. Note that the RARa and RXRa were performed as described previously (10). Similar amounts relatively low affinity of RAR/RXR heterodimers for the DR2 of each probe (50,000 cpm) were used in the assays. ivity RXRa/ Downloaded from https://academic.oup.com/nar/article-abstract/22/5/726/1047624 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 5 729 and DR4 probes in vitro, and the weak activation of the dominant negative RXRa (dnRXRaAAB; refs. 10, 28) confirmed corresponding reporter genes in yeast, is a consequence of the that RARa is responsible for this stimulation by all-trans RA spacer and flanking sequences of the elements used in this study. (D.M.H. and B.P., unpublished results). Thus, the DR1 element For example, a DR2 probe with different spacer and flanking appears to bind both RAR/RXR heterodimers and RXR alone sequences was bound 6-fold more efficiently by RARs and RXRs (presumably as homodimers). This result is in contrast to a recent produced in yeast, or E.coli (Z.P.Chen, personal communi- report by Hall et al., (42) who were unable to demonstrate significant activity of RXR on a DR1 element. Note that these cation). This is consistent with the observations that changes in authors used another receptor isoform (RXR7), a different the spacer, half site motif or the flanking sequences can reporter (based on the yeast CYC1 promoter), and a different significantly affect binding and activation by RARs and RXRs spacer nucleotide in the DR1 sequence as compared to this study. produced in mammalian cells (21, 32) and in yeast (our To confirm that the RXRa activity we observed on DR1 was unpublished results). not simply due to high level expression of the receptor in yeast, The ability of the isolated receptors to activate the reporter we expressed RXRa from a centromeric expression vector. series was also tested. As previously reported (10) expression Significant 9-cis RA induced activation was maintained on the of RARa alone in yeast resulted in a 5- to 7-fold activation of DR1 and IRO reporters (10- to 15-fold), despite a strong decrease DR5 reporter in the presence of either all-trans RA or 9-cis RA (30- to 50-fold) in the level of receptor detectable in western blot (Fig. IB). However, RARa alone showed little if any ability to analyses (data not shown). stimulate the activity of the DR1, DR2, DR3, DR4 and IRO reporters, in the presence or absence of ligand. Similarly RXRa The rat CRBPII gene promoter contains a sequence consisting had little effect on the DR2, DR3, DR4 and DR5 reporter of four almost perfect repeats of the hexamer motif 5'-AGGT- activities (2-3 fold 9-cis RA-dependent stimulation), but CA -3 ' with lbp spacers, which has been reported to act as a RXRE in animal cells (33, 34). We determined if this element stimulated the activity of the DR1 reporter at least 30-fold in would also function as a target for transactivation by RXRa in response to 9-cis RA, but not all-trans RA (Fig. IB). Western yeast. As shown in Fig. 2B, the rat CRBPII RXRE reporter was blot experiments using specific monoclonal antibodies (gifts from very efficiently activated (116-fold over the control) by RXRa C.Egly and M.P.Gaub) detected similar levels of RARa or in the presence of 9-cis RA in yeast, i.e. almost 20-fold higher RXRa in extracts from cells expressing either one or both than that of the DR1 reporter in the same system. This result receptors (data not shown), thus eliminating the possibility that indicates a synergistic activation by RXRa molecules bound to the observed increases in DNA binding and transactivation by the multiple binding sites in this promoter. As observed with the heterodimers were due to altered levels of receptor proteins. DR1 reporter, the activity of the rat CRBPII RXRE reporter was RXRa also activated the IRO reporter 20-fold above background also significantly induced (12-fold) in the presence of 9-cis RA, in a ligand dependent manner (Fig. IB), although as stated above, when RXRa was expressed from a centromeric vector (data not the absolute value of this activity was low (6-fold lower than the shown). These results indicate a synergistic activation by RXRa DR1 reporter). Cell-free extracts from transformants expressing molecules bound to the multiple binding sites in this promoter. no receptor, or either of the receptors alone, failed to retard the Coexpression of RARa and RXRa in yeast increased the activity DR5 probe (10) or any of the other DR probes in gel shift of the rat CRBPII RXRE reporter 14-fold above the control in experiments under conditions identical to those used in the the absence of ligand, and 21-fold and 70-fold above the control experiment shown in Fig. 1C (data not shown). However, after in the presence of all-trans RA and 9-cis RA, respectively (data prolonged autoradiography, a weak band shift was detected for not shown). Accordingly, gel shift experiments showed that RXRa homodimers which was specific for the DR1 probe. These results are consistent with the lack of significant constitutive RARa strongly increased the binding of RXRa to the rat CRBPII activation of the reporter gene series by homodimers. In our RXRE element in vitro (not shown), and thus coexpression of hands, the inclusion of all-trans RA or 9-cis RA (500 nM) in both receptors in yeast did not have the same dramatic negative the growth medium, extraction buffer and assay buffer did not effect on activation of the rat CRBPII RXRE reporter as further stimulate the binding of either homodimers or previously observed in some (33), but not all (34), animal cells. heterodimers to any of the probes (data not shown), although Our previous study (10) established that a 9-cis RA-inducible we did not determine if the conditions used in the gel shift activation function (AF-2) present in the ligand binding domain is largely responsible for the activity of RXRa in yeast, as we experiments were optimal for the binding of ligand to the failed to detect a strong activation function in the N-terminal (AB) receptors. region of RXRa by linking it to a heterologous DBD. 9-cis retinoic acid induced activity of RXRa on putative Accordingly, deletion of the A/B region of RXRa (RXRaAAB) RXRE sequences in yeast did not adversely affect the levels of activation achieved on either Ligand dose response curves (Fig. 2A) revealed that the the DR1 or rat CRBPII RXRE reporters, in the presence or concentration of 9-cis RA required for half maximal activity of absence of ligand (Fig. 2B). In contrast, a truncated receptor in RXRa on the DR1 reporter in yeast was approximately lOOnM. which the 19 most carboxy-terminal amino acids were deleted In contrast, no induction was observed in the presence of all- (dnRXRa), which behaves as a dominant negative receptor in trans RA, even at a concentration of ifiM. This is in agreement mammalian cells (28), failed to mediate a response of either the with our previous data for the chimeric receptors RXRa-ER(C) DR1 or the rat CRBPII RXRE reporters to 9-cis RA in yeast and RXRa(DEF)-ER.Cas, whose activities were induced (Fig. 2B), despite being expressed to a similar level as the full exclusively by 9-cis stereoisomers of RA in yeast (10). The dose length receptor (not shown). This result indicates that the carboxyl response curves in the presence of both RARa and RXRa terminal region is required for ligand-induced transcription by confirmed the modest stimulation of the DR1 reporter activity RXRa. It is also of note that unliganded RXRa stimulated the by the heterodimer in the presence of all-trans RA (Fig. 2A). rat CRBPII RXRE reporter activity 8- to 10-fold over the activity In addition, experiments using a transcriptionally inactive of the control (Fig. 2B). This constitutive activity was observed Downloaded from https://academic.oup.com/nar/article-abstract/22/5/726/1047624 by DeepDyve user on 05 August 2020 730 Nucleic Acids Research, 1994, Vol. 22, No. 5 plasmids. Monoclonal antibodies raised against RARa and RXRa even when the A/B domain was deleted, and thus may indicate were generously provided by C.Egly, M.P.Gaub and Y.Lutz. a weak constitutive activity of AF-2. We also thank the secretariat, illustration and photography services for help in preparing the manuscript. This work was DISCUSSION supported by funds from the Centre National de la Recherche Scientifique, the Institut National de la Same" et de la Recherche We have demonstrated that RXRa activates transcription M&licale, the Association pour la Recherche sur le Cancer, the efficiently in a biological system devoid of other members of the Fondation pour la Recherche Mddicale and the Centre Hospitalier nuclear receptor superfamily. This activity of RXRa was found Universitaire Regional. D.M.H. was awarded fellowships from to be specific for elements having a DRl-type structure, so-called the European Molecular Biology Organisation and the Association 'RXREs', and to a lesser extent, an inverted repeat similar to pour la Recherche sur le Cancer, and B.P. from the Ministere TRE-pal. DRl-type elements have been found in the promoters de la Recherche et de la Technologie, France. of genes including the rat CRBPII (33), human ApoAl (30) and mouse CRABPII (28) genes, which are involved in vitamin A metabolism, and it has been suggested that this may signify the REFERENCES existence of a 9-cis RA autoregulatory network which operates through RXR (43). However, DR1 elements are also recognised 1. Green. S and Chambon. P. (1988) Trends Genet. 4. 309-314 by other members of the steroid receptor superfamily, including 2 Evans. R M. (1988) Science 240. 889-895. PPAR, COUP-TF, Arp-1, HNF4, and ear2 (20, and references 3. DeLuca. L. M (1991) FASEB J. 5. 2924-2933. 4. Linney. E. (1992) In Current Topics in Dewlopmental Biolog\\ Vol. 27. therein). Thus it was important to test the ability of RXR to Academic Press. activate transcription from a DR1 element in a system which has 5. Leid, M.. Kastner. P. and Chambon. P. (1992) Trends Biodiem. Sri. 17. no known homologues of nuclear receptors. Transactivation was 427-433 . very potent on the rat CRBPII RXRE reporter, which contains 6. Heyman. R.A.. Mangelsdorf. D J.. Dyck. J A.. Stein. R B.. Eichele. multiple DR1 repeats, suggesting synergistic activation by RXRa G.. Evans. R. M and Thaller. C. (1992) Cell 68. 397-406. 7. Levin. A. A.. Sturzenbecker. L. J., Kazmer, S.. Bosakowski. T . Huselton. molecules bound to the multiple target sites. Our findings that C Allenby, G.. Speck. J., Kratzeisen. Cl . Rosenberger. M.. Lovey. A. RXRa strongly activates reporter genes driven by DR 1 elements and Grippo. J. F. (1992) Nature (London) 355. 359-361. in yeast are consistent with the hypothesis that, in addition to 8. Allenby, G.. Bocquel. M -T.. Saunders. M.. Kazmer. S.. Speck. J.. Grippo. their role as heterodimeric partners of RAR, VDR, PPAR and J. F . Chambon. P. and Levin. A A (1992) Proc. Natl. Acad. Set. USA TR, RXRs may also function independently as ligand-dependent 89. 30-34. 9. Nagpal, S.. Saunders. M., Kastner. P.. Durand. B.. Nakshatri. H. and transcription factors, in the more classical fashion of steroid Chambon. P (1992) Cell 70. 1007-1019. receptors, i.e. as homodimers. Similarly, homodimers of RARa 10. Heery. D. M.. Zacharewski. T.. Pierral. B.. Groncmcyer. H.. Chambon. were found to activate transcription in yeast (albeit less efficiently P. and Losson. R (1993) Proc. Natl. Acad. Sci. USA 90. 4281-4285. than heterodimers) in a ligand-dependent manner, but only on 11 Nagpal. S . Friant. S . Nakshaln. H. and Chambon (1993) EMBO J. 12. 2349-2360. a DR5 reporter. As with steroid receptors such as the ER, 12. Leid. M.. Kastner. P . Lyons. R.. Nakshatn. H.. Saunders. M.. Zacharewski. transactivation by homodimers of RARs and RXRs appears to T.. Chen. J.-Y.. Staub. A.. Gamier, J.-M . Madcr. S. and Chambon. P. be largely ligand-dependent, and in the case of RXRa is due to (1992) Cell 68, 377-395. the presence of an activation function (AF-2) in the ligand-binding 13. Yu. V.C.. Delsert. C . Andersen. B.. Holloway. J. M., Devary. O. V.. Naar. Kim, S. Y.. Boutin, J -M., Glass. C. K and Roscnfcld. M. G. (1991) domain which is specifically induced by 9-cis stereoisomers of CW/67. 1251-1266. RA. In addition AF-2 may also be responsible for the weak 14. Kliewer, S A.. Umesono, K.. Mangelsdorf. D.J. and Evans. R M. (1992) constitutive activity of RXRa observed on the rat CRBPII RXRE Nature (London) 355. 446-449. and DR1 reporters. In contrast, the A/B region of RXRa appears 15 Zhang, X K.. Hoffmann. B . Tran. P B V , Graupner. G. and Pfahl. M. to be dispensable for the transcriptional activity of this receptor (1992) Nature (London) 355. 441 -446 . 16. Marks. M S . Hallenbeck, P. L . Nagata. T . Segars. J H . Appella. E . in our experiments. Nikodem. V. M. and Ozata. K. (1992) EMBO J. 11. 1419-1435. RA-dependent transactivation in animal cells appears to be 17. Bugge, T. H . Pohl. J., Lonnoy. O and Stunncnberg. H. G (1992) EMBO mediated through DRs consisting of half-site motifs spaced by 7.11. 1409-1418 18 Berrodin, T. J . Marks. M S., Ozato. K , Linney. E and Lazar. M. A 1 to 5 base pairs. Our results in yeast corroborate this observation, (1992) Mol. Endocrmol. 6. 1468-1478 showing a strong correlation between the affinity of the 19 Forman, B.. M. and Samuels. H H. (1990) Tl\e New Biologist 2. 587-594. heterodimer complex for a sequence and the strength of activation 20. Green, S. (1993) Nature (London). 361. 590-591 of the corresponding reporter gene in yeast. Although in this 21. Leroy, P., Nakshatn. H. and Chambon. P. (1991) Proc. Nail. Acad. Sci. USA 88. 10138-10142. report we have only considered the effect of spacer length on 22 de The\ H . Vivanco-Ruiz. M . P. Tiollais. H. Stunnenberg and Dcjean. DNA binding and activation, additional experiments have A. (1990) Nature (London) 343. 177-180. confirmed that the sequences of the half sites, spacers and flanking 23. Sucov. H M., Murakami, K K and Evans. R. (1990) Proc. Nail Acad. DNA all clearly influence the strength of binding and activation Sci. USA 87. 5392-5396. 24. Lehmann. J M.. Zhang. X. K. and Pfahl. M (1992) Mol Cell. Biol. 12, by heterodimers in yeast (our unpublished results), in agreement 2976-2985. with the findings of Mader et al., (32). 25. Duester, G., Shean, M. L., McBride, M. S and Stewart. M. J. (1991) Mol. Cell. Biol. 11. 1638-1648 26. Vasios. G., Madcr. S., Gold. J. D., Leud. M.. Lulz. Y.. Gaub. M. P . ACKNOWLEDGMENTS Chambon, P. and Gudas, L (1991) EMBO J. 10. 1149- 1158. 27 Smith. W. C . Nakshatri. H.. Leroy, P.. Rees. J. and Chambon. P. (1991) D.M.H. and B.P. made equal contributions to this work. We EMBO J. 10, 2223-2230. are grateful to all our colleagues at LGME for the provision of 28. Durand. B., Saunders. M.. Leroy. P., Leid. M. and Chambon. P. (1992) research materials and for helpful discussions. In particular we Cell 71. 73-85 . acknowledge the excellent assistance of J.-M.Garnier and 29. Hall. R. K.. Scott, D. K., Noisin. E. L.. Lucas. P. C , and Granncr, D. coworkers in the construction of several receptor expression K (1992) Mol. Cell Biol. 12. 5527-5535 Downloaded from https://academic.oup.com/nar/article-abstract/22/5/726/1047624 by DeepDyve user on 05 August 2020 Nucleic Acids Research, 1994, Vol. 22, No. 5 731 30. Rottman, J. N., Widom, R. J., Nadal-Gidard, B., Mahdavi, V., and Karathanasis, S. K. (1991) Mol. Cell. Biol. 11, 3814-3820. 31. Raisher, B. D., Gulkk, T., Zhang, Z., Strauss, A. W., Moore, D. D. and Kelly, D. P. (1992) J. Biol. Chem. 267, 20264-20269. 32. Mader, S., Leroy, P., Chen, J-Y. and Chambon, P. (1993) J. Biol. Chem. 268, 591-600. 33. Mangelsdorf, D. J., Umesono, K., Kliewer, S. A., Borgmeyer, U., Ong, E. S. and Evans, R. M. (1991) Cell 66, 555-561. 34. Nakshatri, H. and Chambon, P. (1993) J. Biol. Chem. (in press). 35. Kliewer, S. A., Umesono, K., Heymann, R. A., Mangelsdorf, D. J., Dyck, J. A. and Evans, R. M. (1992). Proc. Natl Acad. Sd. USA 89, 1448-1452. 36. Sikorski, R. S. and Hieter, P. (1989) Genetics 122, 19-27. 37. Pierrat, B., Heery, D. M., Lemoine, Y. and Losson, R. (1992) Gene 119, 237-245 . 38. Loison, G., Vidal, A., Findeli, A., Roitsch, C , Bellou, J. M. and Lemoine, Y. (1989) Yeast 5, 497-507. 39. Brand, N. J., Petkovich, M. and Chambon, P. (1990) Nuc. Acids Res. 18, 6799-6806. 40. Wolcott, J.H. and Ross, C. (1966) Biochem. Biophys. Acta 122, 532-534. 41. Umesono, K., Giguere, V., Glass, C. K., Rosenfeld, M. G. and Evans, R. M. (1988) Nature 345, 262-265. 42. Hall, B. L, Smit-McBride, Z. and Privalsky, M. L. (1993) Proc. Nail. Acad. Sci. USA 90, 6929-6933. 43. Lucas, P. C. and Granner, D. K. (1992) Annu. Rev. Biochem. 61, 1131-1173.

Journal

Nucleic Acids Research – Oxford University Press

Published: Mar 11, 1994

There are no references for this article.