The AML1-MTG8 leukemic fusion protein forms a complex with a novel member of the MTG8(ETO/CDR) family, MTGR1

Abstract

The AML1-CBFbeta transcription factor complex is essential for the definitive hematopoiesis of all lineages and is the most frequent target of chromosomal rearrangements in human leukemia. In the t(8;21) translocation associated with acute myeloid leukemia (AML), the AML1(CBFA2/PEBP2alphaB) gene is juxtaposed to the MTG8(ETO/CDR) gene. We show here that the resultant AML1-MTG8 gene product specifically and strongly interacts with an 85-kDa phosphoprotein. Molecular cloning of cDNA indicated that the AML1-MTG8-binding protein (MTGR1) is highly related to MTG8 and similar to Drosophila Nervy. Comparison of amino acid sequences among MTGR1, MTG8, and Nervy revealed four evolutionarily conserved regions (NHR1 to NHR4). Ectopic expression of AML1-MTG8 in L-G murine myeloid progenitor cells inhibits differentiation to mature neutrophils and induces cell proliferation in response to granulocyte colony-stimulating factor (G-CSF). Analysis with C-terminal deletion mutants of AML1-MTG8 indicated that the region of 51 residues (488 to 538), which contains NHR2, is essential for the induction of G-CSF-dependent cell proliferation. Immunoprecipitation analysis indicates that this region is required for AML1-MTG8 to form a stable complex with MTGR1. Overexpression of MTGR1 stimulates AML1-MTG8 to induce G-CSF-dependent proliferation of L-G cells and to interfere with AML1-dependent transcription. These results suggest that AML1-MTG8 could function as a complex with MTGR1 and that the complex might be important in promoting leukemogenesis.

FIG. 1

Expression of AML1-MTG8 induces G-CSF-dependent proliferation of L-G cells and inhibits their differentiation to mature neutrophils. (A) Structures of AML1-MTG8, AML1b, and AML1a. AML1-MTG8 retains the runt homology domain (Runt) of AML1 and almost all of the region of MTG8. AML1-MTG8 and AML1a lack the C-terminal transactivation domain (PST) which is present in AML1b. (B and C) Immunoprecipitation of AML1-MTG8, AML1b, and AML1a. L-G cells which expressed HA-AML1-MTG8, HA-AML1b, and HA-AML1a were labeled with [³⁵S]methionine (B) or [³²P]orthophosphate (C), and immunoprecipitations were performed with the anti-HA monoclonal antibody 12CA5. The immunoprecipitates were subjected to electrophoresis on SDS–10% polyacrylamide gels. The proteins were visualized with a BAS2000 phosphorimager (Fuji). The positions of bands of AML1a, AML1b, AML1-MTG8, CBFβ, p85, and p35 are indicated on the right. (D and E) Growth curve of the infected L-G cells in response to IL-3 (D) or G-CSF (E). The growing cells which express AML1a, AML1b or AML1-MTG8 were washed twice and cultured in the presence of 2 ng of G-CSF per ml (D) or 0.1 ng of IL-3 per ml (E). The relative numbers of viable cells are indicated. (F and G) Morphology of the cells. The vector-control (F) or AML1-MTG8-expressing cells (G) were exposed to G-CSF for 7 days and stained with May-Gruenwald’s and Giemsa’s solutions.

FIG. 2

Identification of the region of AML1-MTG8 required for induction of G-CSF-dependent cell proliferation. (A) Schematic representation of the structure of AML1-MTG8 deletion mutants. The runt homology domain (Runt), the proline-rich regions (P), and the Nervy homology regions (numbered 1 through 4) are indicated. The numbers above the top bar indicate the positions of amino acid residues. (B) Immunoblot analysis of lysate of infected L-G cells which express either wild-type (WT) or mutant AML1-MTG8 with anti-MTG8 polyclonal antibody. The asterisk indicates the position of intrinsic MTG8 or MTG8-like proteins. (C) Growth curve of the L-G cells in response to G-CSF. The growing cells which express wild-type (WT) or mutant AML1-MTG8 were washed twice and cultured in the presence of 2 ng of G-CSF per ml. (D) Repression of AML1-dependent transcriptional activation. P19 cells were cotransfected with 1.0 μg of TCRβ-TK-CAT, 1.0 μg of either pLNSX vector (−) or pLNSX-AML1b (+), 1.0 μg of either wild-type (WT) or mutant pLNSX-AML1-MTG8, and 0.5 μg of TK-luciferase in a 6-cm-diameter plate. The results represent the mean of relative CAT activity from three experiments which were normalized with luciferase expressed from thymidine kinase-luciferase as an internal control.

FIG. 3

p85, which interacts with AML1-MTG8, is recognized by anti-MTG8 antibody. (A) Immunoprecipitation of AML1-MTG8. Infected L-G cells which express either wild-type (WT) or mutant HA-AML1-MTG8 were labeled with [³²P]orthophosphate, and immunoprecipitations were performed with the anti-HA monoclonal antibody 12CA5. The immunoprecipitates were subjected to electrophoresis on SDS–10% polyacrylamide gels. The proteins were visualized with a BAS2000 phosphorimager. The position of the p85 band is indicated on the right. (B) Immunoblot of AML1-MTG8 complexes with anti-MTG8 antibody. The immunoprecipitates with the HA antibody, using lysates of cells which express wild-type (WT) or mutant HA-AML1-MTG8, were separated on SDS-polyacrylamide gels and analyzed by immunoblotting with anti-MTG8 polyclonal antibody. The position of the p85 band is indicated on the right.

FIG. 4

Analysis of MTG8 and MTGR1 transcripts in L-G and Kasumi-1 cells. Poly(A)⁺ RNAs were prepared from L-G cells, L-G cells infected with LNSX-HA-AML1-MTG8, and Kasumi-1 cells which have the t(8;21) translocation and express AML1-MTG8 chimeric transcripts. A 2-μg sample of poly(A)⁺ RNAs was analyzed by Northern blotting with human MTG8 (A) or human MTGR1 (B) cDNAs as probes. Note that the human MTGR1 transcript (7.5 kb) in Kasumi-1 cells is longer than the mouse MTGR1 transcript (6.7 kb) in L-G cells.

FIG. 5

Nucleotide sequence of the MTGR1 cDNA and the deduced amino acid sequence of the MTGR1 protein. The 3′-noncoding region (nucleotide 1921 to the 3′ end) is not shown.

FIG. 6

Structure of MTGR1. (A) Comparison of the amino acid sequences of MTGR1, MTG8, and Drosophila Nervy proteins. Residues identical among two or more proteins are shaded. The conserved regions (NHR1 to NHR4) are indicated above the sequence. Cysteine residues of zinc finger motifs are marked underneath the sequence. Epitopes for anti-MTG8 and anti-MTGR1 antibodies are underlined. (B) Schematic representation of the structures of human MTGR1, human MTG8b, and Drosophila Nervy. The conserved regions (NHR1 to NHR4) among these proteins are indicated by shaded boxes. The percentages represent identity to MTGR1. Numbers above bars indicate the positions of amino acid residues. (C) Alignments of TAF homology regions of MTGR1, MTG8, Drosophila Nervy, and human and Drosophila TAFs. Residues identical among four or more proteins are shaded. (D) A helical wheel of the NHR2 of MTG8. The numbers indicate the positions of amino acid residues, where the first residue of the NHR2 is position 1. The wheel of positions 3 to 22 is shown. Hydrophobic and hydrophilic amino acids are indicated by gray and black, respectively. The hydrophobic and hydrophilic sides of the helix are indicated.

FIG. 6

Structure of MTGR1. (A) Comparison of the amino acid sequences of MTGR1, MTG8, and Drosophila Nervy proteins. Residues identical among two or more proteins are shaded. The conserved regions (NHR1 to NHR4) are indicated above the sequence. Cysteine residues of zinc finger motifs are marked underneath the sequence. Epitopes for anti-MTG8 and anti-MTGR1 antibodies are underlined. (B) Schematic representation of the structures of human MTGR1, human MTG8b, and Drosophila Nervy. The conserved regions (NHR1 to NHR4) among these proteins are indicated by shaded boxes. The percentages represent identity to MTGR1. Numbers above bars indicate the positions of amino acid residues. (C) Alignments of TAF homology regions of MTGR1, MTG8, Drosophila Nervy, and human and Drosophila TAFs. Residues identical among four or more proteins are shaded. (D) A helical wheel of the NHR2 of MTG8. The numbers indicate the positions of amino acid residues, where the first residue of the NHR2 is position 1. The wheel of positions 3 to 22 is shown. Hydrophobic and hydrophilic amino acids are indicated by gray and black, respectively. The hydrophobic and hydrophilic sides of the helix are indicated.

FIG. 7

Characterization of MTGR1. (A) MTGR1 is an 85-kDa phosphoprotein. L-G cells were infected with LNSX vector, LNSX-HA-AML1-MTG8, or LNSX-Flag-MTGR1, selected with G418. The infected cells were labeled with [³²P]orthophosphate, and immunoprecipitations were performed with the anti-HA monoclonal antibody or the anti-Flag monoclonal antibody. The immunoprecipitates were subjected to electrophoresis on SDS–10% polyacrylamide gels. The proteins were visualized with a BAS2000 phosphorimager. The positions of bands of AML1-MTG8 and MTGR1 are indicated on the right. (B) Interaction between AML1-MTG8 and MTGR1. Cell lysates were prepared from infected L-G cells which express the HA-tagged AML1-MTG8 and/or the Flag-tagged MTGR1. Immunoprecipitations were performed with anti-HA monoclonal antibody (lanes 1 to 8) or anti-Flag monoclonal antibody (lanes 9 to 12). The immunoprecipitates were separated on SDS–10% polyacrylamide gels and analyzed by immunoblotting with anti-MTGR1 polyclonal antibody (lanes 1 to 4) or anti-MTG8 polyclonal antibody (lanes 5 to 12).

FIG. 8

Determination of the region of AML1-MTG8 which is required for interaction with MTGR1. L-G cells were infected with LXSH-Flag-MTGR1 and LNSX-retrovirus, which encodes either wild-type or mutant HA-tagged AML1-MTG8. Control cells (vector) expressed only the Flag-tagged MTGR1. Immunoprecipitations were performed with anti-HA monoclonal antibody (C and D) or anti-Flag monoclonal antibody (B). The immunoprecipitates and total-cell lysates (A) were separated on SDS–10% polyacrylamide gels and analyzed by immunoblotting with anti-MTGR1 polyclonal antibody (D) or anti-MTG8 polyclonal antibody (A to C).

FIG. 9

Interaction of AML1-MTG8 and MTGR1 in Kasumi-1 cells. Cell lysates were prepared from L-G cells which were infected with LNSX-HA-AML1-MTG8 (lane 1) and Kasumi-1 cells (lanes 2 to 9). Immunoprecipitations were performed with anti-HA monoclonal antibody (lane 1), anti-AML1 antibody (lanes 3 and 8), anti-MTGR1 middle region antibody (α-MTGR1M; lane 4), anti-MTGR1 C-terminal region antibody (α-MTGR1C; lane 5), anti-MTG8 antibody (lane 6), and rabbit immunoglobulin G (IgG) (lanes 7 and 9). The immunoprecipitates were separated on SDS–7.5% polyacrylamide gels and analyzed by immunoblotting with anti-MTG8 antibody (lanes 1 to 7) or anti-MTGR1M antibody (lanes 8 and 9).

FIG. 10

No detectable homocomplex of AML1-MTG8 in L-G cells. Cell lysates were prepared from L-G cells which were infected with LNSX-HA-AML1-MTG8 and/or LXSH-D582 without any tag. Immunoprecipitations were performed with anti-HA monoclonal antibody. The immunoprecipitates were separated on SDS–10% polyacrylamide gels and analyzed by immunoblotting with anti-MTG8 polyclonal antibody.

FIG. 11

MTGR1 enhances the activities of AML1-MTG8. (A and B) Growth curve of the infected L-G cells in response to G-CSF (A) or IL-3 (B). The growing cells, which were infected with LNSX-AML1-MTG8 and/or LXSH-MTGR1, were washed twice and cultured in the presence of 2 ng of G-CSF per ml (A) or 0.1 ng of IL-3 per ml (B). Relative numbers of viable cells are indicated. (C) Repression of AML1-dependent transcriptional activation. P19 cells were cotransfected with 1.0 μg of TCRβ-TK-CAT, 1.0 μg of either pLNSX vector or pLNSX-AML1b, the indicated amounts (in micrograms) of pLNSX-AML1-MTG8 and pLNSX-MTGR1, and 0.5 μg of thymidine kinase-luciferase in a 6-cm-diameter plate. Results represent the mean relative CAT activity from three experiments which were normalized with luciferase expressed from thymidine kinase-luciferase as an internal control.

FIG. 12

Model for the action of AML1-MTG8 on repression of transcription. AML1-MTG8 forms a heterotrimer complex with MTGR1 and CBFβ/PEBP2β through the regions containing the NHR2 and the rhd, respectively. The complex binds to the DNA sequence for AML1 and represses transcription from the adjacent target genes.