Portal de Periódicos da CAPES

Role of Acidic and Phosphorylated Residues in Gene Activation by the Glucocorticoid Receptor

1995; Elsevier BV; Volume: 270; Issue: 29 Linguagem: Inglês

10.1074/jbc.270.29.17535

1083-351X

Tova Almlöf, Anthony P. H. Wright, Jan-Ακε Gustafsson,

Cytokine Signaling Pathways and Interactions

To investigate the role of acidic and phosphorylated amino acids in the function of the major transactivation domain (τ1) of the glucocorticoid receptor, we have performed a mutagenesis study. Aspartic and glutamic acid residues were neutralized in clusters of 2 to 4 amino acids throughout the τ1 domain. The activity of the mutant proteins was determined using transactivation assays in yeast and mammalian cells. Some acidic residues in the core region of τ1 appear to play a minor role in τ1 activity, but, generally, individual acidic residues are not critical for activity. Mutagenesis of five serine residues that are phosphorylated in the mouse glucocorticoid receptor and which are conserved in the human receptor did not affect the transactivation activity of the τ1 domain in yeast. As in mouse cells, these serine residues are the predominant sites of phosphorylation for ectopically expressed receptor in yeast, since the mutant protein lacking all five sites had a severely reduced phosphorylation level. Mutant proteins in which larger numbers of acidic residues are neutralized show a progressive decrease in activity indicating that acidity in general is important for τ1 function. However, our results are not consistent with the “acid blob” theory of transactivator function that has been suggested for some other activator proteins. Other putative roles for the acidity of τ1 are discussed. To investigate the role of acidic and phosphorylated amino acids in the function of the major transactivation domain (τ1) of the glucocorticoid receptor, we have performed a mutagenesis study. Aspartic and glutamic acid residues were neutralized in clusters of 2 to 4 amino acids throughout the τ1 domain. The activity of the mutant proteins was determined using transactivation assays in yeast and mammalian cells. Some acidic residues in the core region of τ1 appear to play a minor role in τ1 activity, but, generally, individual acidic residues are not critical for activity. Mutagenesis of five serine residues that are phosphorylated in the mouse glucocorticoid receptor and which are conserved in the human receptor did not affect the transactivation activity of the τ1 domain in yeast. As in mouse cells, these serine residues are the predominant sites of phosphorylation for ectopically expressed receptor in yeast, since the mutant protein lacking all five sites had a severely reduced phosphorylation level. Mutant proteins in which larger numbers of acidic residues are neutralized show a progressive decrease in activity indicating that acidity in general is important for τ1 function. However, our results are not consistent with the “acid blob” theory of transactivator function that has been suggested for some other activator proteins. Other putative roles for the acidity of τ1 are discussed. The glucocorticoid receptor (GR)1 1The abbreviations used are: GRglucocorticoid receptorhGRhuman glucocorticoid receptorCMVcytomegalovirusDBDDNA binding domainPBSphosphate-buffered salineIgimmunoglobulinPAGEpolyacrylamide gel electrophoresismAbmonoclonal antibodyCATchloramphenicol acetyltransferase. belongs to a large family of nuclear receptors with a related structure (1Evans R.M. Science. 1988; 240: 889-895Crossref PubMed Scopus (6341) Google Scholar, 2Jensen E.V. Parker M.G. Nuclear Hormone Receptors. Academic Press, London1991: 1-10Google Scholar, 3Beato M. Cell. 1989; 56: 335-344Abstract Full Text PDF PubMed Scopus (2852) Google Scholar, 4O'Malley B. Mol. Endocrinol. 1990; 4: 363-369Crossref PubMed Scopus (387) Google Scholar, 5Wahli W. Martinez E. FASEB J. 1991; 5: 2243-2249Crossref PubMed Scopus (255) Google Scholar). Binding of glucocorticoid steroid hormones transforms the receptor to an active form that is able to bind to specific glucocorticoid response elements within target genes and subsequently to activate or repress gene activity(1Evans R.M. Science. 1988; 240: 889-895Crossref PubMed Scopus (6341) Google Scholar, 2Jensen E.V. Parker M.G. Nuclear Hormone Receptors. Academic Press, London1991: 1-10Google Scholar, 3Beato M. Cell. 1989; 56: 335-344Abstract Full Text PDF PubMed Scopus (2852) Google Scholar, 4O'Malley B. Mol. Endocrinol. 1990; 4: 363-369Crossref PubMed Scopus (387) Google Scholar, 5Wahli W. Martinez E. FASEB J. 1991; 5: 2243-2249Crossref PubMed Scopus (255) Google Scholar). The GR has a modular structure such that different functions are often performed by independent domains within the protein. Thus, protein domains that are competent for steroid binding, DNA binding, and transactivation can be separated on independent protein segments. Two regions of the human glucocorticoid receptor that are important for transcriptional activation by DNA-bound receptor have been identified. These transactivation domains were named τ1 (residues 77-262) and τ2 (residues 526-556)(6Hollenberg S.M. Evans R.M. Cell. 1988; 55: 899-906Abstract Full Text PDF PubMed Scopus (549) Google Scholar, 7Giguère V. Hollenberg S.M. Rosenfeld M.G. Evans R.M. Cell. 1986; 46: 645-652Abstract Full Text PDF PubMed Scopus (678) Google Scholar). Transient transfection experiments using truncated receptor derivatives suggest that τ1 represents the major transactivation activity of the GR(6Hollenberg S.M. Evans R.M. Cell. 1988; 55: 899-906Abstract Full Text PDF PubMed Scopus (549) Google Scholar, 8Danielsen M. Northrop J.P. Jonklaas J. Ringold G.M. Mol Endocrinol. 1987; 1: 816-822Crossref PubMed Scopus (131) Google Scholar, 9Godowski P.J. Picard D. Yamamoto K.R. Science. 1988; 241: 812-816Crossref PubMed Scopus (167) Google Scholar). glucocorticoid receptor human glucocorticoid receptor cytomegalovirus DNA binding domain phosphate-buffered saline immunoglobulin polyacrylamide gel electrophoresis monoclonal antibody chloramphenicol acetyltransferase. Transactivation domains have been classified into different classes depending on their amino acid composition. Thus, transactivation domains rich in glutamine, proline, and acidic amino acids have been identified(10Mitchell P.J. Tjian R. Science. 1989; 245: 371-378Crossref PubMed Scopus (2210) Google Scholar). The τ1 domain of the GR contains a high proportion of aspartic and glutamic acid residues suggesting that it belongs to the acidic class of activators. Indeed, some degree of correlation between acidity and activity of different GR derivatives has been reported(8Danielsen M. Northrop J.P. Jonklaas J. Ringold G.M. Mol Endocrinol. 1987; 1: 816-822Crossref PubMed Scopus (131) Google Scholar). Furthermore, squelching studies suggest a functional similarity between the τ1 domain and the classical acidic activator VP16, from herpes simplex virus(11Tasset D. Tora L. Fromental C. Scheer E. Chambon P. Cell. 1990; 62: 1177-1187Abstract Full Text PDF PubMed Scopus (298) Google Scholar). Identification of phosphorylated amino acids in the mouse GR (12Bodwell J.E. Orti E. Coull J.M. Pappin D.J.C. Smith L.I. Swift F. J. Biol. Chem. 1991; 266: 7549-7555Abstract Full Text PDF PubMed Google Scholar) suggests that the acidity of the τ1 domain is increased further by phosphorylation. A role for acidity in transactivation by the τ1 domain is also suggested by the high density of acidic and phosphorylated residues found in the recently identified functional core of the τ1 domain(13Dahlman-Wright K. Almlöf T. McEwan I.J. Gustafsson J.-Å. Wright A.P.H. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 1619-1623Crossref PubMed Scopus (85) Google Scholar). The lack of precise transactivation domain end points in deletion studies of the GAL4 (14Ma J. Ptashne M. Cell. 1987; 48: 847-853Abstract Full Text PDF PubMed Scopus (604) Google Scholar) and GCN4 (15Hope I.A. Mahadevan S. Struhl K. Nature. 1988; 333: 635-640Crossref PubMed Scopus (220) Google Scholar) yeast activators led to the “acid blob” concept(16Sigler P.B. Nature. 1988; 333: 210-212Crossref PubMed Scopus (300) Google Scholar), according to which acidic transactivation domains do not adopt a defined structure but rather function by general ionic interactions with target proteins. More recent studies question this model. First, mutagenesis studies on the VP16 activator showed that while acidic residues were important for function, key hydrophobic amino acids were also crucial for activity(17Cress W.D. Triezenberg S.J. Science. 1991; 251: 87-90Crossref PubMed Scopus (324) Google Scholar, 18Regier J.L. Shen F. Triezenberg S.J. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 883-887Crossref PubMed Scopus (226) Google Scholar), suggesting that general ionic interactions are not sufficient for transactivation activity. Second, mutagenesis studies of the adenovirus E1A activator also showed that acidity is not the most important determinant of activity(19Martin J.K. Lillie J.W. Green M.R. Nature. 1990; 346: 147-152Crossref PubMed Scopus (142) Google Scholar). Third, similar conclusions have been drawn from genetic studies of the yeast GAL4 transcriptional activator protein(20Leuther K.K. Salmeron J.M. Johnston S.A. Cell. 1993; 72: 575-585Abstract Full Text PDF PubMed Scopus (105) Google Scholar). Finally, structural studies have recently shown that the GCN4(21Van Hoy M. Leuther K.K. Kodadek T. Johnston S.A. Cell. 1993; 72: 587-594Abstract Full Text PDF PubMed Scopus (121) Google Scholar), GAL4(21Van Hoy M. Leuther K.K. Kodadek T. Johnston S.A. Cell. 1993; 72: 587-594Abstract Full Text PDF PubMed Scopus (121) Google Scholar), VP16(22Donaldson L. Capone J.P. J. Biol. Chem. 1992; 267: 1411-1414Abstract Full Text PDF PubMed Google Scholar), NF-κB (p65)(23Schmitz M.L. dos Santos Silvia M.A. Altmann H. Czisch M. Holak T.A. Baeuerle P.A. J. Biol. Chem. 1994; 269: 25613-25620Abstract Full Text PDF PubMed Google Scholar), and GR (τ1 core) (24Dahlman-Wright K. Baumann H. McEwan I.J. Almlöf T. Wright A.P.H. Gustafsson J.-Å. Härd T. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 1699-1703Crossref PubMed Scopus (140) Google Scholar) transactivation domains have the propensity to form defined secondary structures. The purpose of this study was to investigate the role of acidic and phosphorylated amino acids in the transactivation activity of the τ1 domain of the GR. A SacI fragment encoding residues 77-262 of the human GR was removed from plasmid pKV-GRDBD (25Zilliacus J. Dahlman-Wright K. Wright A. Gustafsson J.-Å. Carlstedt-Duke J. J. Biol. Chem. 1991; 266: 3101-3106Abstract Full Text PDF PubMed Google Scholar) and cloned into the SacI site of M13mp18. Single-stranded DNA was prepared as a substrate for in vitro mutagenesis. Point mutations were created by oligonucleotide-directed in vitro mutagenesis using the methods of Eckstein (26; Amersham, plc) or Kunkel(27Kunkel T.A. Roberts J.D. Zakour R.A. Methods Enzymol. 1987; 154: 367-382Crossref PubMed Scopus (4560) Google Scholar, 28Geisselsoder M.C. Clary J.A. Witney F. BioTechniques. 1989; 7: 282-289PubMed Google Scholar). Mutants were identified and checked by sequencing the entire τ1 insert. The τ1 insert was cloned into SacI-digested pKVXE(29Wright A.P.H. McEwan I.J. Dahlman-Wright K. Gustafsson J.-Å. Mol. Endocrinol. 1991; 5: 1366-1372Crossref PubMed Scopus (35) Google Scholar). The resulting plasmids express mutant τ1 domains (residues 77-262) coupled to the GR DBD (residues 370-500). Expression plasmids were transformed into Saccharomyces cerevisiae, strain W303-IA(30Wright A.P.H. Carlstedt-Duke J. Gustafsson J.-Å. J. Biol. Chem. 1990; 265: 14763-14769Abstract Full Text PDF PubMed Google Scholar), according to Beggs(31Beggs J. Nature. 1978; 275: 104-108Crossref PubMed Scopus (793) Google Scholar). Transformants were mated with yeast strain K396-11B (30Wright A.P.H. Carlstedt-Duke J. Gustafsson J.-Å. J. Biol. Chem. 1990; 265: 14763-14769Abstract Full Text PDF PubMed Google Scholar) containing the reporter plasmid pLGZ-TAT(30Wright A.P.H. Carlstedt-Duke J. Gustafsson J.-Å. J. Biol. Chem. 1990; 265: 14763-14769Abstract Full Text PDF PubMed Google Scholar). Diploid strains were grown to stationary phase at 30°C in minimal medium containing 3% glycerol and 1% ethanol as carbon sources, lacking uracil and leucine. Cells were diluted with the same medium containing 2% galactose to a density of about A600 = 0.3 to induce expression of τ1-DBD proteins. Cells were harvested after 5 h. Protein extracts were prepared and assayed for β-galactosidase activity and protein concentration(30Wright A.P.H. Carlstedt-Duke J. Gustafsson J.-Å. J. Biol. Chem. 1990; 265: 14763-14769Abstract Full Text PDF PubMed Google Scholar). β-Galactosidase activity was measured as nanomoles of o-nitrophenyl-β-O-galactoside substrate converted per min per mg of protein and expressed relative to the wild type level. Equivalent amounts of each protein extract were resolved by SDS-PAGE and transferred to nitrocellulose filters (32Burnette W.N. Anal. Biochem. 1981; 112: 195-203Crossref PubMed Scopus (5927) Google Scholar). Filters were blocked with 5% milk powder and incubated with monoclonal antibodies against the GR (1 εg/ml) in PBS + 0.5% Tween 80 and 1% milk powder for 60 min at room temperature. After washing (PBS + 0.5% Tween 80), the filters were incubated with alkaline phosphatase-conjugated anti-mouse rabbit IgG (0.22 εg/ml); DaKopatts A/S, Glostrup, Denmark) in PBS + 0.5% Tween 80 for 30 min. After washing, the blot was developed using a colorimetric substrate (Promega Corp.) according to the manufacturer's instructions. Mutant τ1 domains were amplified by polymerase chain reaction using Vent DNA polymerase (New England Biolabs) from pKV-τ1GRDBD as BglII fragments and cloned into a derivative of CMVhGRα from which a BglII fragment encoding τ1 had been deleted. The pCMVhGRα (gift from Katrin Hecht, Dept. of Medical Nutrition, Karolinska Institute) is a derivative of CMV4 (33Andersson S. Davis D.L. Dahlbäck H. Jörnvall H. Russell D.W. J. Biol. Chem. 1988; 264: 8222-8229Abstract Full Text PDF Google Scholar) in which the intact hGR protein is expressed from the CMV promoter. COS7 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and antibiotics. Nonsaturating amounts of expression plasmids (0.3 εg) were cotransfected together with 5 εg of the reporter plasmid, p19LUC-TK, into COS7 cells at 80% confluency using the transfection reagent DOTAP (Boehringer Mannheim) according to the manufacturer's instructions. Transfections were performed in triplicate. p19LUC-TK (gift from Paul T. van der Saag, Hubrecht Laboratory, Netherlands Institute for Developmental Biology) is a modified version of pG29LtkCAT (34Schule R. Muller M. Kaltschmidt C. Renkawitz R. Science. 1988; 242: 1418-1420Crossref PubMed Scopus (388) Google Scholar) containing 2 glucocorticoid response element binding sites upstream of a truncated thymidine kinase promoter linked to the luciferase gene. 1 εM dexamethasone (Sigma) was added 24 h after transfection, and the cells were then incubated for an additional 24 h prior to measurement of luciferase activity(35Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual. 2nd Ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY1989Google Scholar). Extracts containing wild type and mutant τ1-DBD proteins, extracted in buffer (40 mM Tris-HCl, pH 8.5, 10% glycerol, 10 mM dithiothreitol, 50 mM phenylmethylsulfonyl fluoride, 10 mM leupeptin, 10 mM trypsin inhibitor), were incubated in the presence or absence of 2 to 4 units of alkaline phosphatase (Boehringer Mannheim) and with or without phosphatase inhibitors (5 mM EDTA and 5 mM Na2HPO4) at 37°C for 60 min. Extracts were resolved by SDS-PAGE and immunoblotted as described above. Cells from an overnight culture, grown in minimal medium with 2% glucose as carbon source and lacking leucine and uracil, were diluted in the same medium, except that 3% glycerol and 1% ethanol were used as the carbon sources. The cultures were incubated for 24 h at 30°C to give a final cell density of about A600 = 0.5. The cells were washed in 20 ml of phosphate-free YP medium (1% yeast extract, 1% Bacto-peptone) and resuspended in 10 ml of the same medium, containing 3% glycerol, 1% ethanol, and 2% galactose, to a cell density of A600 = 0.3. Phosphate-free medium was produced by the addition of 130 mM NH4OH and 10 mM MgSO4. After stirring for 30 min, the resulting precipitate was removed by filtration and the pH of the medium was adjusted to 6.25 with HCl prior to autoclaving. The cultures were shaken at 30°C for 1.5 h. 100 εCi of [32P]orthophosphate (Amersham) was then added to the medium and incubation continued for an additional 3.5 h. The cells were harvested, washed three times with water, and resuspended in 0.15 ml of buffer (120 mM sodium phosphate, pH 7.0, 10 mM KCl, 1 mM MgSO4, 2 mM phenylmethylsulfonyl fluoride, 10 εg/ml leupeptin, 100 εg/ml DNase I, 100 εg/ml RNase A). Cell-free protein extracts were prepared by shaking vigorously with an equal volume of glass beads (0.5 mm diameter) for 20 min, mixing with 100 εl of fresh buffer, and centrifugation in a Microfuge for 10 min. 50-100 εg of extract was mixed with an equal volume of a slurry, containing Sepharose beads coupled to receptor-specific or control monoclonal antibodies and incubated on ice for 1 h with occasional mixing. Beads were pelleted by centrifugation at 2000 rpm in a Microfuge for 5 s and washed three times with 50 mM sodium phosphate buffer, pH 7.4, containing increasing NaCl concentrations (50 mM, 300 mM, and 1000 mM, respectively). Proteins were eluted from the beads by addition of 50-100 εl of 1 × SDS-PAGE loading buffer followed by incubation at 100°C for 10 min. The eluates were analyzed by SDS-PAGE followed by autoradiography using XAR film (Kodak). Quantification was performed using a BAS 2000 phosphoimager (Fuji). To test the importance of acidic residues for the activity of the τ1 transactivation domain of the GR, we have used in vitro mutagenesis to change glutamate and aspartate residues to glutamine and asparagine, respectively. The acidic residues were mutated in clusters that could be spanned by individual mutagenic primers giving rise to a series of τ1 derivatives (TA1 to TA8) in which between two and four adjacent acidic residues have been neutralized (Fig. 1). The mutated τ1 domains were expressed as fusion proteins with the GR DBD and tested for their ability to activate a GR responsive lacZ reporter gene in yeast(29Wright A.P.H. McEwan I.J. Dahlman-Wright K. Gustafsson J.-Å. Mol. Endocrinol. 1991; 5: 1366-1372Crossref PubMed Scopus (35) Google Scholar). Fig. 2A shows that the TA1 to TA8 mutations cause a relatively small, if any, loss of transactivation activity. Furthermore, Fig. 2B shows that none of the mutants are substantially overexpressed compared to the wild type protein which might otherwise have led to an overestimation of their transactivation activity. The TA4 mutation causes the most pronounced defect, showing only 40% of the wild type level. Interestingly, the residues mutated in TA4 (Asp-187, Asp-192, Asp-196, and Glu-198) lie within the functional core of the τ1 domain(13Dahlman-Wright K. Almlöf T. McEwan I.J. Gustafsson J.-Å. Wright A.P.H. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 1619-1623Crossref PubMed Scopus (85) Google Scholar). The activity of the remaining mutants ranges between 55% (TA7) and 125% (TA3) of the wild type level. From this analysis we conclude that individual acidic amino acids are not critical for the transactivation activity of the τ1 domain although some acidic residues appear to play a minor role. It remained a possibility that acidity in general is an important component of τ1 activity, but, due to the redundancy of acidic amino acids, this requirement was not exposed using relatively limited mutations such as those in TA1 to TA8. To address this question, we made more extreme mutant derivatives of τ1 containing different combinations of the mutant residues that were mutated in TA1 to TA8. These mutants have a much more severe effect on activity (Fig. 3A). Transactivation activity is lost progressively as the number of acidic residues that are neutralized increases. This downward trend tails off at about 10% of wild type activity in mutants in which 60% or more of the acidic residues in τ1 have been neutralized. Thus, we conclude that overall acidity is important for the transactivation activity of τ1. Possible interpretations of this result are given under “Discussion.” To confirm that results obtained for the isolated τ1 domain in yeast cells are relevant for the intact receptor in mammalian cells, we wanted to test the effect of acidic residue mutations in the τ1 domain in the context of the full length GR. Wild type GR protein and a set of GR proteins with mutated τ1 domains were analyzed in COS7 cells for their ability to transactivate a cotransfected luciferase reporter gene controlled by two glucocorticoid response element binding sites. Fig. 3B shows that the mutant GR proteins exhibit essentially the same rank order of transactivation activity in mammalian cells as was observed for the isolated τ1 domain in yeast. However, the mutations reduce transactivation activity to a lesser extent in the intact receptor, presumably due to the activity of other transactivation domains in the intact GR that are not affected by mutations in τ1. The close correspondence of the results from yeast and mammalian cells supports the notion that similar transactivation mechanisms are employed by the GR in both systems. Five of the seven phosphorylation sites that have been identified in the mouse GR are conserved in mammals and are located within the τ1 domain(12Bodwell J.E. Orti E. Coull J.M. Pappin D.J.C. Smith L.I. Swift F. J. Biol. Chem. 1991; 266: 7549-7555Abstract Full Text PDF PubMed Google Scholar). Since phosphorylation of these serines would contribute considerably to the acidity of τ1, we wished to determine the extent to which phosphorylation is important for transactivation activity. We first determined whether τ1 was phosphorylated in yeast. This was done by pulse labeling strains expressing either τ1 (Fig. 4A, lanes 1 and 2) or τ1-DBD (lanes 3 and 4) with [32P]orthophosphate. The expressed proteins were immunoprecipitated with the GR specific monoclonal antibody mAb 250 (lanes 1 and 3) or a nonspecific antibody (lanes 2 and 4) and analyzed by SDS-PAGE. The autoradiograph shows that both τ1-DBD and τ1 are phosphorylated in yeast cells, and, thus, phosphorylation could contribute to the transactivation activity as measured in yeast. To determine whether phosphorylation of the five conserved serine residues within τ1 are important for the activity of τ1, we constructed a mutant in which each of the five serines is mutated to alanine (TP1-5, Fig. 1). We also constructed three mutants, TP3, TP4, and TP5, in which individual serine residues were replaced by alanine. The only mutant that shows a significantly reduced transactivation activity is TP3 (Fig. 4B). However, this is probably due to the low level of expressed τ1-DBD protein seen in this strain (Fig. 4C, lane 2). Consistent with this, the TP1-5 mutant, which contains the TP3 mutation, is expressed well and is not defective in activity. We conclude that phosphorylation of the five conserved serines plays little, if any, role in the transactivation activity measured. A role for phosphorylation remained a possibility since the identity of the τ1 residues phosphorylated in yeast had not been determined and in general it cannot be excluded that phosphorylation of other residues occurs if the preferred phosphorylation sites are mutated. However, Fig. 4C shows that the migration of the TP1-5 mutant protein is faster than the wild type protein during SDS-PAGE. This would be consistent with underphosphorylation of the mutant protein. To investigate this further, yeast extracts containing wild type and TP1-5 mutant τ1-DBD proteins were incubated with increasing concentrations of alkaline phosphatase in the presence or absence of phosphatase inhibitors (Fig. 5A). Alkaline phosphatase treatment of τ1-DBD converts it to a faster migrating species in a dose-dependent fashion (Fig. 5A, lanes 1-3). This is not observed in the presence of phosphatase inhibitors (Fig. 5A, lanes 4-6), strongly suggesting that the migration changes result from dephosphorylation of the τ1-DBD protein. The migration of the TP1-5 mutant protein did not change upon alkaline phosphatase treatment (lanes 7-12), suggesting that it was largely devoid of phosphate residues that could serve as substrates for phosphatase action. To measure the phosphorylation level of the TP1-5 mutant protein directly, cells expressing either the wild type protein or the mutant TP1-5 protein were labeled with [32P]orthophosphate, and the expressed proteins were immunoprecipitated using the GR specific monoclonal mAb 250. The autoradiograph in Fig. 5B shows a much lower level of 32P incorporation for the TP1-5 mutant protein compared to the wild type protein. The relatively high background in Fig. 5B makes it difficult to quantify the level of the mutant protein but even according to our most conservative estimate, it contains only about 30% of the wild type phosphate level. Fig. 5C shows that the amount of protein precipitated was very similar for the wild type and TP1-5 mutant proteins. In summary, our data show that mutations of the five serines that are phosphorylated in the region of the mouse GR corresponding to the τ1 domain severely reduces phosphorylation of the τ1-DBD protein in yeast. However, the reduced phosphorylation does not result in reduced transactivation activity, and, thus, phosphorylation is not crucial for maximal transactivation by the τ1 domain in yeast. The first important conclusion from this study is that neutralization of 2 to 4 adjacent acidic amino acid residues throughout the τ1 domain have minimal if any effects on its transactivation activity. The greatest effect was caused by the TA4 mutant that maps to the recently identified core region of the τ1 domain(13Dahlman-Wright K. Almlöf T. McEwan I.J. Gustafsson J.-Å. Wright A.P.H. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 1619-1623Crossref PubMed Scopus (85) Google Scholar). A similar effect of the TA4 mutant is also seen in COS7 cells indicating that the core region is also important in mammalian cells. Thus, one or more of the acidic residues at position Asp-187, Asp-192, Asp-196, and Glu-198 might play a role in the activity of the τ1 core domain, which has been shown to include interaction with the general transcriptional machinery(36McEwan I.J. Wright A.P.H. Dahlman-Wright K. Carlstedt-Duke J. Gustafsson J.-Å. Mol. Cell. Biol. 1993; 13: 399-407Crossref PubMed Google Scholar). These data are analogous to previous findings for the viral and yeast activator proteins, VP16 (17Cress W.D. Triezenberg S.J. Science. 1991; 251: 87-90Crossref PubMed Scopus (324) Google Scholar) and GAL4(20Leuther K.K. Salmeron J.M. Johnston S.A. Cell. 1993; 72: 575-585Abstract Full Text PDF PubMed Scopus (105) Google Scholar), respectively. In each case, individual acidic residues could also be mutated without causing major effects on activity. In VP16, mutations of specific hydrophobic residues caused much greater effects(17Cress W.D. Triezenberg S.J. Science. 1991; 251: 87-90Crossref PubMed Scopus (324) Google Scholar, 18Regier J.L. Shen F. Triezenberg S.J. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 883-887Crossref PubMed Scopus (226) Google Scholar). Phosphorylation also contributes to acidity. The τ1 domain contains the 5 conserved serine residues that are phosphorylated in mouse GR. Our mutants, in which serines have been mutated to alanine, suggest that phosphorylation of these serine residues contributes little, if anything, to the transactivation activity of the isolated τ1 domain in yeast. This is consistent with a recent report in which equivalent mutations did not affect the activity of the intact mouse GR in mammalian cells(37Mason S.A. Housley P.R. J. Biol. Chem. 1993; 268: 21501-21504Abstract Full Text PDF PubMed Google Scholar). However, as shown in recent studies of the 1,25-dihydroxyvitamin D3 receptor (38Hilliard IV, G.M. Cook R.G. Weigel N.L. Pike J.W. Biochemistry. 1994; 33: 4300-4311Crossref PubMed Scopus (65) Google Scholar), one cannot exclude that phosphorylation occurs on alternative residues when the primary sites have been mutated. Furthermore, although we showed that τ1 is phosphorylated in yeast, we had no initial expectation that phosphorylation of the same serine residues would occur in yeast as in mouse. Two lines of evidence suggest that the phosphorylation pattern in yeast is similar to that reported for mouse. First, the migration of τ1-DBD protein during electrophoresis is increased in the TP1-5 mutant in which all 5 serine residues are mutated to alanine. A similar migration change for the wild type protein can be specifically induced by phosphatase treatment. Second, 32P incorporation experiments show directly that the phosphate content of the TP1-5 mutant protein is severely reduced in comparison to wild type. Thus, the majority of phosphorylation occurs on the same residues in both yeast and mammalian cells. A similar observation has recently been reported for the chicken progesterone receptor where the phosphorylation pattern for ectopically expressed receptor in yeast closely mimicked that seen in chick oviduct(39Poletti A. Conneely O.M. Mc Donnell D.P. Schrader W.T. O'Malley B.W. Weigel N.L. Biochemistry. 1993; 32: 9563-9569Crossref PubMed Scopus (20) Google Scholar). Our data showing that phosphorylation levels of τ1 do not correlate with its transactivation activity suggest a role for phosphorylation in some other aspect of τ1 structure or function. A role of phosphorylation in nuclear cytoplasmic shuttling of the GR has been suggested(40Orti E. Mendel D.B. Smith L.I. Bodwell J.E. Munck A. J. Steroid Biochem. 1989; 34: 85-96Crossref PubMed Scopus (42) Google Scholar). Our data would suggest that evolutionarily conserved protein kinases might be involved, and, interestingly, some of the phosphorylation sites are potential substrates for p34cdc2 protein kinase which is critical in cell cycle regulation and is functionally conserved between yeasts and humans (41Lee M.G. Nurse P. Nature. 1987; 327: 31-35Crossref PubMed Scopus (765) Google Scholar). The conclusion that the acidity of individual amino acid residues, resulting from either the acidity of their side chains or from phosphorylation, is not critical for the molecular interactions underlying transactivation does not exclude the acid blob mechanism of acidic activator action. According to this model, acidic activators do not have a defined structure and interact with target factors via general ionic interactions(16Sigler P.B. Nature. 1988; 333: 210-212Crossref PubMed Scopus (300) Google Scholar). Alternatively, acidic activation domains may form defined structures and the acidity might then play a role in their structural/functional integrity. The data in Fig. 3A showing a progressive loss of τ1 activity as more acidic residues are neutralized are consistent with both models. In an attempt to discriminate between these two models, we have compared the activity of the mutants from Fig. 3A with that of a series of fragments derived from τ12 2K. Dahlman-Wright, unpublished observations. (13) in relation to their acidity (Fig. 6). The neutralization mutants from Fig. 3A show an approximately linear decrease in activity as the number of acidic residues decreases (Fig. 6A). According to the acid blob theory, a similar pattern would be expected for the τ1 fragments, which also vary in acidity. However, the activity of these fragments correlates much less well with their acidity (Fig. 6B), and, thus, the acid blob model is at best an incomplete explanation for the mechanism of τ1 action. A number of other observations support this conclusion. (i) Although unstructured in aqueous buffer, the functional core of τ1 does form α-helices under hydrophobic conditions in vitro(24Dahlman-Wright K. Baumann H. McEwan I.J. Almlöf T. Wright A.P.H. Gustafsson J.-Å. Härd T. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 1699-1703Crossref PubMed Scopus (140) Google Scholar). Analogous observations have been reported for the acidic activation domains of VP16(22Donaldson L. Capone J.P. J. Biol. Chem. 1992; 267: 1411-1414Abstract Full Text PDF PubMed Google Scholar), GAL4(21Van Hoy M. Leuther K.K. Kodadek T. Johnston S.A. Cell. 1993; 72: 587-594Abstract Full Text PDF PubMed Scopus (121) Google Scholar), GCN4(21Van Hoy M. Leuther K.K. Kodadek T. Johnston S.A. Cell. 1993; 72: 587-594Abstract Full Text PDF PubMed Scopus (121) Google Scholar), and NF-κB (p65)(23Schmitz M.L. dos Santos Silvia M.A. Altmann H. Czisch M. Holak T.A. Baeuerle P.A. J. Biol. Chem. 1994; 269: 25613-25620Abstract Full Text PDF PubMed Google Scholar). (ii) That τ1 might be structured in vivo is suggested by the pattern of τ1 degradation in cell-free extracts which produces defined degradation products that are inconsistent with indiscriminate proteolysis expected for an unfolded peptide.2 (iii) Deletion analysis defined relatively clear-cut borders for the functional core of τ1(13Dahlman-Wright K. Almlöf T. McEwan I.J. Gustafsson J.-Å. Wright A.P.H. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 1619-1623Crossref PubMed Scopus (85) Google Scholar) rather than the gradual loss of activity with progressive deletion that has been associated with the acid blob hypothesis(14Ma J. Ptashne M. Cell. 1987; 48: 847-853Abstract Full Text PDF PubMed Scopus (604) Google Scholar, 15Hope I.A. Mahadevan S. Struhl K. Nature. 1988; 333: 635-640Crossref PubMed Scopus (220) Google Scholar, 16Sigler P.B. Nature. 1988; 333: 210-212Crossref PubMed Scopus (300) Google Scholar). In conclusion, we favor a model in which the high density of charged residues in τ1, particularly in the functional core region, is important in the context of a structured transactivation domain. One possibility is that the charged residues are important to ensure location of the transactivation region on the solvent-exposed surface of the receptor protein. Consistent with this, the mutants studied here have a pronounced effect on the intact GR as well as the isolated τ1 domain. Also, τ1 appears to be an important feature on the GR surface since most monoclonal antibodies raised against intact GR recognize epitopes within τ1(42McEwan I.J. Almlöf T. Wikström A.-C. Dahlman-Wright K. Wright A.P.H. Gustafsson J.-Å. J. Biol. Chem. 1994; 269: 25629-25636Abstract Full Text PDF PubMed Google Scholar). We thank members of the Steroid Receptor unit at the Center for Biotechnology for helpful discussions and comments during this work, Beatrix Vecsey for technical assistance, Jacqueline Ford for help with mammalian cell techniques, Ann-Charlotte Wikström and Marika Rönnholm (Dept. of Medical Nutrition, Karolinska Institute) for generously providing the monoclonal antibodies against GR, George G. J. M. Kuiper (Erasmus University of Rotterdam) for advice on alkaline phosphatase treatment, Katrin Hecht (Dept. of Medical Nutrition, Karolinska Institute), Paul T. van der Saag (Hubrecht Laboratory, Netherlands Institute for Developmental Biology), and Delta Biotechnology Ltd. for generously providing plasmids. We also thank Karin Dahlman-Wright for critical reading of the manuscript.