We have used a transient transfection system with a cytomegalovirus-based vector expressing high levels of biologically active human estrogen receptor (ER) in COS-1 cells to study the phosphorylation of human ER and to identify major hormone-regulated phosphorylation sites. The features of phosphorylation of the wild-type ER were very similar to those previously observed for the endogenous ER in uterine cells: The ER exhibited a basal level of phosphorylation which was increased approximately 3-4-fold by estrogen (estradiol) and by antiestrogens (hydroxytamoxifen and ICI164,384), and phosphorylation was increased to an almost similar extent by activation of either protein kinase A or C signal transduction pathways with cholera toxin plus isobutyl methylxanthine (CT+IBMX) or phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA), respectively. Phosphoamino acid analysis revealed that the phosphorylation occurred exclusively on serine residues in all cases. Tryptic phosphopeptide analysis of ER, using a two-dimensional peptide mapping procedure, revealed similar patterns for ER in cells treated with estradiol, antiestrogens or TPA; with CT+IBMX treatment, the same phosphopeptides were seen, but the relative phosphorylation of the different ER phosphotryptic peptides differed. In ER deleted of the NH2-terminal A and B (A/B) domains, estrogen and antiestrogen-stimulated phosphorylations were abolished, while the phosphorylation induced by CT+IBMX was maintained. This suggests that sites of phosphorylation enhanced by estradiol and antiestrogen, but not those induced by CT+IBMX, are located in the A/B domain. These results were further confirmed by comparing the tryptic phosphopeptide patterns of wild-type and A/B-deleted receptor upon estradiol and CT+IBMX treatments, and then by site-directed mutagenesis, by substituting alanines for the serine residues in the A/B domain (Ser104, Ser106, Ser118, Ser154, and Ser167) involved in known protein kinase consensus sequences. Comparison of the tryptic phosphopeptide patterns of wild-type ER and these mutant ERs allowed us to identify serine 104 and/or serine 106 and serine 118, all three being part of a serine-proline motif, the preferred substrate of proline-directed protein kinase, as major ER phosphorylation sites. When tested with two estrogen-responsive reporter gene constructs in several cell types, the mutant S104A, S106A, S118A showed a approximately 40% reduction in transactivation activity in response to E2, while the mutants S118A and S104A, S106A alone showed a approximately 15% decrease in transactivation. Our studies identify several serines in the NH2-terminal portion of the human ER as being major hormone-regulated phosphorylation sites.(ABSTRACT TRUNCATED AT 250 WORDS)