Hoxa9 transduction induces hematopoietic stem and progenitor cell activity through direct down-regulation of geminin protein

Abstract

Hoxb4, a 3'-located Hox gene, enhances hematopoietic stem cell (HSC) activity, while a subset of 5'-located Hox genes is involved in hematopoiesis and leukemogenesis, and some of them are common translocation partners for Nucleoporin 98 (Nup98) in patients with leukemia. Although these Hox gene derivatives are believed to act as transcription regulators, the molecular involvement of the Hox gene derivatives in hematopoiesis and leukemogenesis remains largely elusive. Since we previously showed that Hoxb4 forms a complex with a Roc1-Ddb1-Cul4a ubiquitin ligase core component and functions as an E3 ubiquitin ligase activator for Geminin, we here examined the E3 ubiquitin ligase activities of the 5'-located Hox genes, Hoxa9 and Hoxc13, and Nup98-Hoxa9. Hoxa9 formed a similar complex with the Roc1-Ddb1-Cul4a component to induce ubiquitination of Geminin, but the others did not. Retroviral transduction-mediated overexpression or siRNA-mediated knock-down of Hoxa9 respectively down-regulated or up-regulated Geminin in hematopoietic cells. And Hoxa9 transduction-induced repopulating and clonogenic activities were suppressed by Geminin supertransduction. These findings suggest that Hoxa9 and Hoxb4 differ from Hoxc13 and Nup98-Hoxa9 in their molecular role in hematopoiesis, and that Hoxa9 induces the activity of HSCs and hematopoietic progenitors at least in part through direct down-regulation of Geminin.

Figure 1. Purification of the recombinant RDCOXA9 complex from Sf9, and E3 ubiquitin ligase activity for Geminin.

(A) Left panel: Crude extracts. Each member of the complex was detected in the crude extract by means of immunoblot analysis. Hox derivatives were detected with an anti-Flag antibody, GST-Roc1 with an anti-Roc1 antibody, Ddb1 with an anti-Ddb1 antibody and HA-Cul4a with an anti-HA antibody. Right panel: the affinity-purified. The complex with GST-Roc1 was pulled down with a glutathione affinity column chromatography. Each member was detected by means of immunoblot analysis in the pulled-down complex. Schematic representation of the complex is shown in the lower panel. The samples include GST-Roc1, Ddb1, HA-Cul4a and either of Flag-Hoxb4, Flag-Hoxa9, Flag-Hoxc13 or Flag-Nup98-Hoxa9. (B) E3 ubiquitin ligase activity for Geminin. The affinity-purified recombinant complex was subjected to in vitro ubiquitination reaction (myc-Geminin+E1+E2+ubiquitin), and the reaction product was analyzed by immunoblot analysis with an anti-myc antibody. The amount of GST-Roc1 in the RDCOXA9 complex was adjusted to that of the RDCOXB4 complex (1 µg). Ub1-Geminin, mono-ubiquitinated Geminin; Ubn-Geminin, poly-ubiquitinated Geminin. The mono-ubiquitinated Geminin bands were detected in this in vitro assay system even in the absence of E3 ubiquitin ligase (second lane). (C) In vitro ubiquitination reaction with biotin-tagged ubiquitin. myc-Geminin was immunoprecipitated with an anti-myc antibody after the reaction, and ubiquitinated Geminin was detected through biotin-avidin interaction. *, unspecified bands. **biotin-ubiquitin was detected with avidin vectastain.

Figure 2. Effects of Hoxa9, Hoxc13 and Nup98-Hoxa9 transfection in HEK-293 cells.

Either Flag-Hoxb4, Flag-Hoxa9, Flag-Hoxc13 or Flag-Nup98-Hoxa9 was transfected. Effect of Flag-Hoxb4 transfection was used as positive control. (A) Effect on the cell cycle. The percentages of cell subpopulations in each phase of the cell cycle are shown. (B) Effect on expression of mRNA for Geminin. mRNA for Geminin was detected by real-time PCR analysis. The amount of transfected Hox derivatives are shown in the x-axis. (C) Protein expression of transfected Hox derivatives and the effect on Geminin protein expression examined by immunoblot analysis. Flag-Hoxa9- or Flag-Hoxb4-induced down-regulation of Geminin was suppressed by treatment with MG132. (D) Quantitative analysis of Geminin protein expression in HEK-293 cells transfected with either Flag-Hoxb4, Flag-Hoxa9, Flag-Hoxc13 or Flag-Nup98-Hoxa9. Bands detected by immunoblot analysis of the three independent experiments were scanned with the Image J program (NIH). Relative expression of Geminin protein to that of ß-actin was subjected to statistical analysis. Representative findings of the immunoblot analysis were shown in Fig. 2C. (E) Ubiquitination of Geminin and Flag-Hoxa9. Geminin, and either Flag-Hoxa9 or HA-Ub, or both, were co-transfected. The cells were then treated with or without MG132, and were subjected to immunoblot analysis with anti-Geminin (upper panel) and anti-Flag antibodies (lower panel). (F) Effects of Flag-Hoxb4, Flag-Hoxa9, Flag-Hoxc13 and Flag-Nup98-Hoxa9 transfection on the stability of Geminin. [³⁵S]-labeled Geminin was traced by means of autoradiography after immunoprecipitation (lower panel). The detected bands were scanned with the Image J program (NIH), and the half-life of Geminin protein was calculated (insets in the upper panel). -, an empty control vector.

Figure 3. Immunoprecipitation analysis of Hox derivatives and effect of Cul4a knock-down on Hoxa9-mediated down-regulation of Geminin protein in HEK-293 cells.

(A) Either of Flag-Hoxa9, Flag-Hoxc13 or Flag-Nup98-Hoxa9 was transfected in HEK-293 cells, and the complex formation with endogenous Cul4a, Ddb1 and Roc1 was examined by means of immunoprecipitation analysis using an anti-Flag antibody. (B) Cul4a siRNA was transfected, and the effect on Hoxa9-mediated down-regulation of Geminin protein was examined. Down-regulation of Cul4a by siRNA was confirmed by immunoblot analysis, and the level was restored by transfection of myc-tagged Cul4a. Endogenous Cul4a was also detected in myc-tagged Cul4a-transfected cells even if cells were pre-treated with siRNA for Cul4a probably because exogenously overexpressed mRNA for Cul4a prevented siRNA from affecting endogenous Cul4a.

Figure 4. Effects of Hoxb4, Hoxa9, Hoxc13 and Nup98-Hoxa9 transduction on BM.

Retrovirally transduced cells were subjected to flow cytometry analysis as well as to real-time PCR analysis. (A) Cell populations in each phase of the cell cycle were analyzed by cell sorting analysis. (B) Relative expression of mRNA for Geminin. (C) Geminin protein expression in each phase of the cell cycle. Geomean of the fluorescence intensity was shown. MEP, an empty control vector.

Figure 5. Expression of Hoxa9 in hematopoietic cells, and effect of siRNA-mediated Hoxa9 knock-down on Geminin expression in FL.

(A) Expression of Hoxa9 mRNA in BM and FL. (B) Efficiency of siRNA transfection into FL. Hoxa9siRNA was co-transfected with either non-labeled non-targeting (dotted line) or green fluorescent dye-labeled non-targeting siRNA (green line) as an indicator to monitor the transfection efficiency. High transfection efficiency was confirmed by detecting expression of the indicator by flow cytometry. Rate of a green-fluorescence⁺ or green-fluorescence⁻ subpopulation (%) in the transfectants is shown in the upper part of the panel. (C) Effect of Hoxa9 siRNA transfection on expression of mRNA for Hoxa9, Hoxa10, Hoxb4, Hoxd13, Geminin, Cdt1 and Cyclin A2. Specific knock-down of Hoxa9 mRNA expression was confirmed by real-time PCR analysis. Hoxa9siRNA, siRNA for Hoxa9; C, non-targeting control siRNA. (D) Effect of siRNA transfection on Geminin protein expression in each phase of the cell cycle. Geomean of fluorescence intensity for Geminin protein was shown. (E) Effect of siRNA transfection on cell cycle. (F) Effect of siRNA transfection on apoptosis. The data from three independent experiments were subjected to the statistical analysis. The representative cell sorting data are shown.

Figure 6. Effect on clonogenic activity.

(A) Effect of transduction of Hox derivatives on clonogenic activity. Numbers and types of colonies are shown in the upper panel, and close-up photographs of representative colonies in the lower panel. (B) Effect of Geminin supertransduction on Hoxa9 transduction-mediated induction of clonogenic activity. Numbers and types of colonies are shown in the upper panel. Close-up photographs of representative colonies and images obtained with an inverted microscopy are shown in the lower panel. A cell cluster with more than 20 cells was counted as a colony under microscopy. (C) Effect of Geminin supertransduction on Hoxa9 transduction-mediated induction of replating activity. MEP and MPI, empty control vectors.

Figure 7. Effect of Geminin supertransduction on Hoxa9 transduction-mediated hematopoietic induction in reconstituted BM.

(A) Repopulating activities of retrovirally transduced BM. Repopulating activities in the reconstituted mice were examined 1 month (1M) and 3 months (3M) after the injection. Since the repopulating activity was low in MEP+MPI-transduced BM, mice injected with 5-fold higher numbers of retrovirally transduced cells were used in subsequent analyses as controls. The number of recipient mice is indicated above each bar. (B) Cell populations in each phase of the cell cycle were analyzed by cell sorting analysis. (C) Geminin protein expression in each phase of the cell cycle. (D) Cell sorting procedure for analyzing the primitive hematopoietic cells. BM were subjected to cell sorting analysis. Primitive hematopoietic cells in EYFP⁺ cells were examined by the immunophenotype analysis. *, cells with a non-specific Sca1 signal due to the spectrum overlap (E) Cell numbers of Lin⁻, progenitor, CD34⁺KSL and CD34⁻KSL subpopulations in BM from the reconstituted mice. The data from three independent mice were subjected to the statistical analysis. MEP and MPI, empty control vectors.