The transcriptional repressor GFI-1 antagonizes PU.1 activity through protein-protein interaction

Abstract

Mice lacking the zinc finger transcriptional repressor protein GFI-1 are neutropenic. These mice generate abnormal immature myeloid cells exhibiting characteristics of both macrophages and granulocytes. Furthermore, Gfi-1(-/-) mice are highly susceptible to bacterial infection. Interestingly, Gfi-1(-/-) myeloid cells overexpress target genes of the PU.1 transcription factor such as the macrophage colony-stimulating factor receptor and PU.1 itself. We therefore determined whether GFI-1 modulates the transcriptional activity of PU.1. Our data demonstrate that GFI-1 physically interacts with PU.1, repressing PU.1-dependent transcription. This repression is functionally significant, as GFI-1 blocked PU.1-induced macrophage differentiation of a multipotential hematopoietic progenitor cell line. Retroviral expression of GFI-1 in primary murine hematopoietic progenitors increased granulocyte differentiation at the expense of macrophage differentiation. We interbred Gfi-1(+/-) and PU.1(+/-) mice and observed that heterozygosity at the PU.1 locus partially rescued the Gfi-1(-/-) mixed myeloid lineage phenotype, but failed to restore granulocyte differentiation. Our data demonstrate that GFI-1 represses PU.1 activity and that lack of this repression in Gfi-1(-/-) myeloid cells contributes to the observed mixed lineage phenotype.

FIGURE 1. Gfi-1 inhibits M-CSFR promoter activity induced by PU.1

(A) Transient transfection of U937 cells with a 540bp M-CSFR promoter luciferase construct in the presence or absence of rv-Gfi-1 plasmid. The decrease in transcriptional activity induced by Gfi-1 is statistically significant (p<0.0001). (B) M-CSFR promoter activity induced in 293T cells with co-transfection of MigR1-PU.1 in the presence or absence of increasing amounts of rv-Gfi-1 plasmid. The decrease in transcriptional activity induced by Gfi-1 is statistically significant (p<0.05). The differences in reporter activity between M-CSFR reporter transfected in the presence or absence of Gfi-1 (no PU.1 expression plasmid) was not statistically significant. (C) M-CSFR promoter activity induced in 293T cells by co-transfected MigR1-C/EBPα in the presence or absence of rv-Gfi-1 plasmid. The observed increase in transcriptional activity induced by Gfi-1 was statistically significant (p<0.02). (D) 293T cells co-transfected with MigR1-PU.1, MigR1-C/EBPα, and rv-Gfi-1. Gfi-1 represses transcription mediated by both PU.1 and C/EBPα. This repression was statistically significant (p<0.002). For (A) luciferase activity is reported as relative light units. For (B, C and D) luciferase activity is reported as fold-induction above the activity seen in 293T cells transfected with only the reporter plasmid and rv-GFP. Total DNA content for all transfections was kept constant with rv-GFP plasmid. All transfections also contained the renilla luciferase plasmid pRL-tk. Luciferase values were determined 48h post-transfection, and were normalized to renilla luciferase values. Luciferase values are the mean +/- standard error of the mean of three independent transfections.

FIGURE 2. Gfi-1 inhibits the activity of a fusion protein between full-length PU.1 and the GAL4 DNA binding domain

Transient transfections of 293T cells with a reporter construct containing five GAL4 sites upstream of a minimal tk promoter driving luciferase expression. (A) Cells were co-transfected with either GAL4 DNA binding domain (G4DBD), GAL4-full length PU.1 (G4PU.1) or GAL4-Hif3 activation domain (G4HIF3AD) expression plasmid in the presence or absence of rv-Gfi-1 plasmid. Gfi-1's repression of transcriptional activity induced by GAL4-PU.1 was statistically significant (p<0.0001). (B) Cells were co-transfected with a GAL4 DNA binding domain or GAL4-PU.1 transactivation domain (G4PUAD) expression plasmid in the presence or absence of rv-Gfi-1. The slight enhancement of transcription by Gfi-1 was statistically significant (p<0.05) (C) GAL4-full-length PU.1 was co-transfected either with either rv-Gfi-1 or rv-Gfi-1 P to A plasmid. Rv-Gfi-1 P to A encodes a form of Gfi-1 that contains an inactivating mutation in the repressive SNAG domain. The difference in activity observed between Gfi-1 and Gfi-1 P to A transfected cells was statistically significant (p<0.001). Total DNA content for all transfections was kept constant with rv-GFP plasmid. All transfections contained the renilla luciferase plasmid pRL-tk. Luciferase values were determined 48h post-transfection, and were normalized to renilla luciferase values. Luciferase activity is the mean +/- standard error of the mean of three independent transfections, and luciferase activity is reported as fold-induction above the activity seen in 293T cells transfected with the GAL4 DNA binding domain expression plasmid.

FIGURE 3. Gfi-1 interacts with PU.1

(A) 293T cells were transfected with PU.1 and Gfi-1 expression plasmids as indicated. Whole cell lysates were prepared and subjected to immunoprecipitation with either PU.1 or Gfi-1 antibody. Immunoprecipitated complexes were separated by SDS-PAGE. Immunoblots were prepared and probed with anti-PU.1 antibody or anti-Gfi-1 as indicated. (B) U937 cell extracts were prepared and immunoprecipitated with either anti-PU.1 antibody or non-specific antibody (anti-tubulin). 10% input and immunoprecipitates were immunoblotted and probed with anti-Gfi-1 antibody. (C) Full-length ³⁵S-methionine labeled PU.1 was incubated with GST, GST full-length (FL) Gfi-1, GST N-terminal (NT) Gfi-1, GST Gfi-1 lacking the last 4 zinc fingers (deltaZnfs3-6) GST Gfi1 zinc finger 3, 4, and 5 (Znfs3-5) or GST C-terminal (CT) Gfi-1 Binding of PU.1 to different GST-Gfi1 proteins was detected by autoradiography. (D) ³⁵S-methionine labeled in vitro translated PU.1 deletion mutants were incubated with GST-Gfi-1 protein. Protein complexes were washed and separated by SDS-PAGE. Binding was detected as described above.

FIGURE 4. Gfi-1 blocks macrophage differentiation of PUER myeloid progenitor cells

(A) Immunoblot of Gfi-1, PU.1 and β actin expression in PUER cells that were superinfected with rv-GFP or rv-Gfi-1 virus. Cell lines were either undifferentiated (no OHT) or differentiated (100nM OHT) for eight days in the presence of IL-3. (B) PUER (GFP and Gfi-1) cell lines were induced to differentiate with 100nM OHT for eight days and subsequently cytocentrifuged. Cells were Wright stained to evaluate morphology. Magnification 40× (C) Immunoblot analysis of M-CSFR expression in whole cell lysates obtained from OHT treated PUER cells infected with either rv-GFP or rv-Gfi-1 retrovirus. Whole cell lysates were prepared from PUER cells at 0,1,2,4, and 8 days after OHT addition. (D) Chromatin immunoprecipitation (ChIP) of the M-CSFR promoter from PUER cells using anti-PU.1 antibody. PUER cell lines were either untreated or treated for 7 days with OHT as indicated. (E) ChIPs as described above except anti-Gfi-1 was used to immunoprecipitate. (F) RT-PCR analysis of RNA obtained from control PUER cells (GFP) and cells expressing Gfi-1. Cells were either untreated or treated with 100nM OHT for 8 days.

FIGURE 5. Gfi-1 does not block neutrophil differentiation induced by PU.1

(A) rv-GFP and rv-Gfi-1 infected PUER cells were grown under neutrophil differentiating conditions. Cells were subsequently cytocentrifuged and stained with May-Grunwald-Giemsa. (B) RT-PCR and (C) northern blot analysis were performed on RNA prepared from rv-GFP and rv-Gfi-1 infected PUER cells. IL-3 indicates cells grown in the cytokine IL-3. A + mark indicates that cell were treated with 100nm OHT. G indicates cells Grown in the cytokine G-CSF.

FIGURE 6. Gfi-1 increases granulocyte differentiation of primary progenitors

Nucleated bone marrow cells were lineage depleted to enrich for hematopoietic progenitors and infected with (A) rv-Gfi-1 or (B) rv-Gfi-1 P to A retrovirus for 2 days in the presence of SCF, IL-6, and IL-3. Cells were cultured 4 additional days in G-CSF and analyzed by FACS for expression of GFP, CD11b, Gr-1 and F4/80. Cells were gated on two populations: GFP^low and GFP^high. Upper panels show expression of CD11b and Gr-1, and the bottom panels show expression of CD11b and F4/80. Sorting of populations has shown previously that CD11b⁺Gr-1⁺ cell are granulocytes and CD11b⁺F4/80⁺ cells are macrophages (Dahl et al., 2003).

FIGURE 7. PU.1 heterozygosity reduces the mixed lineage phenotype of Gfi-1^-/- hematopoietic cells

(A) Hematopoietic colony assays were performed with nucleated bone marrow harvested from 8-week old mice. 25,000 cells from each animal were plated into methylcellulose media containing SCF, IL-3, IL-6, and Epo. Duplicate cultures were prepared and after 7 days of culture hematopoietic colonies were scored for number and types of colonies. Average number of colonies and the standard error are shown for each genotype assayed. 6 wildtype, 3 Gfi-1^-/-, and 6 Gfi-1^-/-PU.1^+/- animals were examined in this assay. (B) Realtime RT-PCR analysis of M-CSFR, CD11b, and ELA2 expression in nucleated bone marrow obtained from WT, PU.1^+/-, Gfi-1^-/-, and Gfi-1^-/-PU.1^+/- mice. Expression was normalized to β-actin and show as fold difference from levels detected in wildtype mice. 3 mice for each genotype were examined. * indicates a statistical difference between the levels of M-CSFR detected in Gfi-1^-/- and Gf-1^-/-PU.1^+/- samples (P<0.05) (C) Flow cytometry analysis of nucleated bone marrow isolated from 8 week old mice. Representative FACs plots obtained from Mac3 and Gr-1 analysis of wildtype, Gfi-1^-/- and Gfi-1^-/-PU.1^+/- cells is shown.