The is combined with 8-bromoadenosine 3',5'-cyclic monophosphate (8-Br-cAMP), MMTV-CAT is induced to levels approaching that stimulated by R5020 alone. Also, RU486 in the presence of 8-Br-cAMP is only partially effective in antagonizing R5020 action. The agonist activity exhibited under these conditions appears to be due to RU486 acting through hPR as evidenced by the fact that 8-Br-cAMP alone has no effect on MMTV-CAT, whereas induction by the combination of 8-BrcAMP and RU486 is dose responsive to RU486 in a saturable manner and can be inhibited by the type I antiprogestin (prevents hPR-DNA binding) ZK98299, which does not exhibit positive functional cooperation with cAMP. Acquisition of agonist activity in the presence of 8-Br-cAMP also extends to the type II antiprogestin (permits hPR-DNA binding) ZK112993. Since RU486 is also a type II antagonist, these results suggest that detection of functional synergism between cAMP and antiprogestins may require binding of the hPRantagonist complex to DNA. We propose that cross-talk between second messenger and steroid receptor signal transduction pathways may be one mechanism for resistance to steroid antagonists that frequently develops in breast cancer.

The human progesterone receptor (PR) is a member of the steroid/thyroid hormone superfamily of nuclear receptors. The receptor is expressed as two forms, PR-B and the shorter PR-A, which lacks the NH2-terminal 164 amino acids of PR-B; whereas PR-B seems to be predominantly a transcriptional activator, PR-A also functions as a repressor. Our previous studies of PR expressed in T47D breast cancer cells have shown that PR is a phosphoprotein whose phosphorylation is enhanced in response to hormone. There is an initial rapid (minutes) increase in phosphorylation followed by a slower, less substantial increase, which results in decreased mobility of the receptor on sodium dodecyl sulfate gels. We now report the identification of three phosphorylation sites, which are predominantly phosphorylated during the later phase of the response to hormone. These sites, Ser102, Ser294, and Ser345, are all found in Ser-Pro consensus sequences. Whereas Ser294 and Ser345 are common to PR-A and PR-B, Ser102 is unique to PR-B. Finally, we demonstrate that phosphorylation of Ser345 is associated with the altered mobility on sodium dodecyl sulfate gels.

Human progesterone receptors (PR) were overexpressed in Spodoptera frugiperda (Sf9) insect cells using a recombinant baculovirus system. Recombinant viruses were constructed that produced either full-length A (94K) or B (120K) forms of human PR, and each was expressed as a functional protein. Steroid and DNA binding activities were found to be indistinguishable from that of endogenous human PR in T47D breast cancer cells. Moreover, as analyzed by gel-mobility shift, recombinant PR-A and PR-B each bound to specific progesterone response elements in a strictly hormone-dependent manner. Native receptors expressed in Sf9 cells also exhibited structural properties similar to that of endogenous PR. Cytosolic PR (PR-A or PR-B), prepared in low salt buffer, sedimented on density gradients as an 8S oligomeric complex that was converted largely to 4S by treatment with 0.4 M NaCl. Immune isolation of the 8S cytosol PR complex and analysis of protein composition revealed the presence of two specific copurifying proteins of approximately 90K and 70K. The 90-K component was identified immunologically as heat shock protein 90. The 70-K component was not identified but is likely to be the insect equivalent of heat shock protein 70. Immune isolation of PR from Sf9 cells metabolically labeled with [32Pi], revealed that expressed PR was capable of being phosphorylated in insect cells. Hormone addition to Sf9 cells, however, did not stimulate the same increase in PR phosphorylation or upshift in mobility on sodium dodecyl sulfate gels that occurs with endogenous receptors in T47D cells. Thus some, but not all, phosphorylations occur with human PR expressed in Sf9 cells. These phosphorylation data, together with the fact that expressed PR required hormone for DNA binding, indicate that the hormone-dependent phosphorylation step responsible for PR upshifts on sodium dodecyl sulfate-polyacrylamide gel electrophoresis is not required for receptor binding to DNA. The baculovirus expression system, therefore, may prove valuable in dissecting the functional role(s) for both hormone-dependent and hormone-independent PR phosphorylation.

In this study we have demonstrated that dimerization of mammalian progesterone receptors (PR) occurs in the absence of DNA. A specific immune coisolation assay was performed on extracts ofT-47D human breast cancer cells with a monoclonal antibody specific for the full-length B form of progesterone receptor (PR-B). This resulted in coisolation of significant amounts of truncated form-A receptors (PR-A), indicating the presence of stable PR-APR-B dimers in solution. A positive correlation was observed between the

Human progesterone receptors (PR) in T47D breast cancer cells are synthesized as two different sized proteins, PR-A [94 kilodaltons (kDa)] and PR-B (120 kDa). Progestin addition to cells (in vivo) causes a 2-fold increase in total phosphorylation of PR and an increase in the apparent mol wt of both PR-A and PR-B on sodium dodecyl sulfate (SDS)-gels. Time-course experiments showed that increased PR phosphorylation that results from hormone addition is a multistep process and involves a rapid increase into total 32P labeling that takes place before the more slowly occurring phosphorylation(s) responsible for the change in electrophoretic mobility of PR on SDS-gels. As an approach to test whether phosphorylation is involved in regulating PR activity, we have examined the effects of cellular modulators of protein phosphorylation on PR-mediated target gene transcription in vivo using a T47D cloned cell line containing a stably transfected mouse mammary tumor virus-chloramphenicol acetyltransferase construct. Treatment with 8-bromo-cAMP (activator of cAMP-dependent protein kinases) or okadaic acid (protein phosphatase-1 and -2A inhibitor) did not stimulate target gene expression in the absence of progestin. When added together with progestin, either compound augmented PR-mediated target gene transcription by 3- to 4-fold. The cyclic nucleotide-dependent protein kinase inhibitor H8 completely blocked target gene responsiveness to hormone. Neither 8-bromo-cAMP, okadaic acid, nor H8 altered the hormone- or DNA-binding activities of PR, as measured in vitro or affected cellular concentrations of PR. These agents, therefore, appeared to selectively modulate PR transcriptional activity. Moreover, none of these compounds altered expression from a control reporter gene, pSV2CAT, indicating that these agents affect PR-mediated processes directly and are not acting through a general effect on transcription. Effects on PR phosphorylation were assessed by measuring 32P labeling of PR in vivo. None of these treatments had a substantial effect on the extent of total 32P labeling of immune isolated PR or on the phosphorylation(s) responsible for PR up-shifts on SDS-gels. This suggests that these agents modulate PR transcriptional activity either through phosphorylation of another protein intimately involved in PR-mediated transcription or through modification of a key site(s) not measurable as a change in total PR phosphorylation or electrophoretic mobility on SDS gels.