Physical interaction between retinoic acid receptor and the oncoprotein myb inhibits retinoic acid-dependent transactivation

Abstract

The c-myb protooncogene is predominantly expressed in hematopoietic cells and plays a vital role in hematopoiesis. Retinoic acid (RA) is able to induce differentiation of several hematopoietic cells. This differentiation is linked to decreased c-myb expression, suggesting that retinoid receptors (RAR/RXR) may down-regulate c-myb gene expression. Furthermore, recent data indicate that RAR inhibits the function of the Myb protein itself. In addition, the Myb-Ets oncogenic fusion protein has been shown to inhibit transcriptional activation by RAR and thyroid hormone receptor. Myb-Ets also antagonizes the biological response of erythrocytic progenitor cells to RA and thyroid hormone. This prompted us to investigate a possible cross talk between RAR and Myb. Here, we demonstrate that RA inhibits the expression of the endogenous Myb target gene tom-1. Conversely, Myb functions as a potent inhibitor of RA-induced biological responses. Functional analysis of Myb mutants in transfection studies revealed that the Myb DNA-binding domain (DBD) is necessary for repression whereas the transactivation domain is dispensable. Furthermore, we show that v-Myb and RAR interact in vitro and in vivo. This interaction requires the DBD of RAR. In contrast, glutathione S-transferase-pulldown assays with v-Myb mutants indicate that the DBD and the C terminus of Myb directly interact with RAR. Our results suggest that the physical interaction between Myb and RAR may play a role in the regulation of hematopoietic gene expression.

Figure 1

v-Myb-ER fusion proteins block RA-induced morphological changes in 10.4 macrophage-like cells. (A) 10.4 cells were grown with solvent (EtOH) in the absence of RA and E₂, (B) treated with 1 μM RA for 18 h, and (C) treated with 1 μM RA plus 2 μM E₂ for 18 h. Light field photomicrographs of representative fields are shown. (D) RA antagonizes the induction of tom-1 expression by E₂ in cells expressing a v-Myb-ER fusion protein. Cells (10.4) were grown without ligands (lane 1) or treated for 18 h with 1 μM RA (lane 2) 2 μM E₂ (lane 3), or incubated simultaneously with both ligands (lane 4). Expression of tom-1 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was analyzed by reverse transcription–PCR.

Figure 2

Myb inhibits RA-mediated transcriptional activation in a dose-dependent manner in different cell lines. (A) (TREp)₂-TKLUC reporter plasmids or the control plasmids TKLUC were cotransfected in CV1 cells together with constant amounts of plasmids expressing hRARα and hRXRα and various amounts of expression vectors encoding full-length c-Myb (pCM100) or with control plasmids containing the c-Myb sequence in antisense orientation (pCM101). The cells were either untreated (filled bars) or treated with 10⁻⁶ M RA (striped bars). Numbers indicate the amount of cotransfected plasmid DNA in μg. (B) Transcriptional activity of reporter plasmids (TREp)₂-TKLUC and βRE2-TKLUC in NIH 3T3 cells, cotransfected with 2.5 μg of expression vectors pCM100 or pCM101. Cells were treated with 10⁻⁶ M RA. (C) CV1 cells were transfected with 5 μg of expression vectors encoding hRARα (lane 1) or hRARα and full-length c-Myb. Whole-cell extracts were prepared 48 h after transfection and used in immunoprecipitation experiments with αRAR antibodies. (D) The TA of Myb is dispensable for repression of RA-mediated gene activation. Reporter plasmids βRE2-TKLUC were cotransfected in NIH 3T3 cells with 2.5 μg of various expression vectors coding for c-Myb (pCM100), v-Myb (pVM116), or the indicated mutants of v-Myb (pVM130 and DIN1). As controls, the expression vectors pCM101or pVM111 containing a frame-shift mutation in the v-Myb coding region close to its 5′ end were transfected. Reporter activity is shown as fold activation. Cells were treated with 10⁻⁶ M RA. The structure of the c-Myb, v-Myb, and v-Myb mutant proteins are illustrated schematically. L, leucine zipper. (E) Whole-cell extracts from cells transiently transfected with expression vectors coding for v-Myb (pVM116) (lanes 1 and 4) or the v-Myb mutants pVM130 (lanes 2 and 5) and DIN1 (lanes 3 and 6) were analyzed by Western blotting with the αMyb antibodies 5E11 that recognize the DBD of Myb (lanes 1–3) and 2-2-79, recognizing the C terminus of Myb (lanes 4–6).

Figure 3

v-Myb inhibits RAR/RXR-DNA complex formation. EMSAs were performed with nuclear extracts from COS-7 cells either untreated (lanes 1–3) or transiently transfected with v-Myb expression vectors (pVM116) (lanes 4–6). A ³²P-labeled oligonucleotide containing the βRARE was used as a probe. The composition of DNA/protein complexes were determined by adding the following antibodies to the reaction mixtures: αRAR (lanes 2 and 5) and αMyb (lanes 3 and 6). In addition, two nonspecific complexes (NS) were observed.

Figure 4

v-Myb and RAR physically interact in vitro. (A) v-Myb physically interacts with RAR but not with RXR. The ability of RAR and RXR proteins to interact with v-Myb was evaluated by GST-pulldown assays. RAR and RXR were expressed as GST-fusion proteins, immobilized to glutathione-linked Sepharose beads and tested for interaction with in vitro-translated, [³⁵S]methionine-labeled v-Myb. In the reciprocal experiment GST-v-Myb was tested with in vitro-translated, [³⁵S]methionine-labeled RAR and RXR. The same amounts of either GST or GST fusion proteins were linked to glutathione Sepharose beads. The input control in lanes 1, 5, and 8 reflects 10% of the total amount of [³⁵S]methionine-labeled proteins used for the pulldown experiments. (B) The DBD of RAR is necessary and sufficient for interaction with v-Myb. To localize the domain within RAR necessary for interaction with v-Myb either full-length RAR or various mutants thereof were expressed as GST-fusion proteins and tested for interaction with in vitro-translated, [³⁵S]methionine-labeled v-Myb. In the reciprocal set of experiments GST-v-Myb was tested with in vitro-translated, [³⁵S]methionine-labeled RAR or mutants thereof. The input control reflects 10% of the total amount of [³⁵S]methionine-labeled proteins used for the pulldown assay. The full-length hRAR protein is composed of the N-terminal (A/B) domain, the DBD, and the C-terminal ligand binding domain (Ligand). (C) Several domains of v-Myb are involved in interaction with RAR. v-Myb and various mutants of v-Myb were either expressed as GST-fusion proteins or in vitro translated in the presence of [³⁵S]methionine and tested in GST-pulldown experiments. v-Myb and v-Myb mutant proteins are illustrated schematically. L, leucine zipper.

Figure 5

v-Myb interacts with RAR in vivo. Galp₃TKLUC reporter plasmids were cotransfected with 250 ng of plasmids expressing Gal4-v-myb or Gal4 alone or in combination with 50 ng of expression vectors encoding RAR-VP16 in NIH 3T3 cells. The cells were either untreated (filled bars) or treated with 10⁻⁶ M RA (striped bars).