The exosome complex establishes a barricade to erythroid maturation

Abstract

Complex genetic networks control hematopoietic stem cell differentiation into progenitors that give rise to billions of erythrocytes daily. Previously, we described a role for the master regulator of erythropoiesis, GATA-1, in inducing genes encoding components of the autophagy machinery. In this context, the Forkhead transcription factor, Foxo3, amplified GATA-1-mediated transcriptional activation. To determine the scope of the GATA-1/Foxo3 cooperativity, and to develop functional insights, we analyzed the GATA-1/Foxo3-dependent transcriptome in erythroid cells. GATA-1/Foxo3 repressed expression of Exosc8, a pivotal component of the exosome complex, which mediates RNA surveillance and epigenetic regulation. Strikingly, downregulating Exosc8, or additional exosome complex components, in primary erythroid precursor cells induced erythroid cell maturation. Our results demonstrate a new mode of controlling erythropoiesis in which multiple components of the exosome complex are endogenous suppressors of the erythroid developmental program.

Figure 1

GATA-1/Foxo3-dependent genetic network. (A) Venn diagrams depicting genes regulated uniquely or coregulated by GATA-1 and Foxo3. (B) Venn diagrams demonstrating relationships between GATA-1– and Foxo3-activated and -repressed genes. (C) Gene ontology analysis of GATA-1/Foxo3-coactivated genes. (D) Gene ontology analysis of GATA-1– and Foxo3-corepressed genes. The top 10 gene ontology categories are displayed and ordered by P value. Redundant gene ontology categories were curated and removed.

Figure 2

GATA-1/Foxo3-mediated repression of Exosc8 expression. (A) Real-time RT-PCR analysis of Foxo3, Exosc8 mRNA and primary transcripts upon GATA-1 activation in G1E-ER-GATA-1 cells (mean ± SE; 3 independent experiments). Values were normalized to 18S rRNA expression. (B) Semiquantitative western blot analysis of Foxo3 in G1E-ER-GATA-1 cells after 0, 24, or 48 hours estradiol treatment (left). The representative image is shown from 3 independent experiments. Quantitation of band intensity (mean ± SE; 3 independent experiments) (right). (C) Real-time RT-PCR analysis of Foxo3 and Exosc8 mRNA upon Foxo3 knockdown in G1E-ER-GATA-1 cells after 42 hours activation of GATA-1 (mean ± standard error [SE]; 4 independent experiments). Values were normalized to 18S rRNA expression and expression is shown relative to the control. (D) ChIP-seq profile of GATA-1 occupancy at EXOSC8 in primary human erythroblasts. (E) ChIP analysis of Foxo3 occupancy at the Exosc8 promoter in untreated and estradiol-treated (24 hours) G1E-ER-GATA-1 cells (mean ± SE; 4 independent experiments). (F) Real-time RT-PCR analysis of mRNA levels in control vs Exosc8 knockdown G1E-ER-GATA-1 cells (mean ± SE; 3 independent experiments). Values were normalized to 18S rRNA expression. mRNA levels are shown relative to the control siRNA with estradiol treatment of activated genes (left) and without estradiol treatment of repressed genes (middle). Fold changes upon Exosc8 knockdown relative to control (right). (G) Real-time RT-PCR analysis of primary transcripts in control vs Exosc8 knockdown G1E-ER-GATA-1 cells (mean ± SE; 3 independent experiments). Values were normalized to 18S rRNA expression, and the expression is shown relative to estradiol-treated control siRNA. Fold changes are also depicted upon Exosc8 knockdown relative to control (right). *P < .05, **P < .01, ***P < .001.

Figure 3

Repression of exosome-complex component expression during erythroid maturation. (A) Schematic diagram and crystal structure of the human exosome complex. Solid line, direct interactions; dashed line, indirect interactions; red, associated components with catalytic activity. (B) Expression profile of mRNA levels for exosome complex components in primary mouse bone marrow erythroblasts during distinct maturation stages mined from the Erythron DB. P, proerythroblast; B, basophilic erythroblast; O, orthochromatic erythroblast; R, reticulocyte. (C) Expression profile of mRNAs encoding exosome-complex components during primary human erythroid differentiation mined from the Human Erythroblast Maturation (HEM) Database. C, colony-forming unit-erythroid (CFU-E); P, proerythroblast; I, intermediate-stage erythroblast; L, late-stage erythroblast. (D) Expression profile of mRNAs encoding exosome-complex components during ex vivo differentiation of primary mouse fetal liver–derived erythroid precursor cells. 1, R1 population: progenitor cells; 2, R2 population: proerythroblasts and early basophilic erythroblasts; 3, R3 population: early and late basophilic erythroblasts. 4, R4/5 population: polychromatophilic, orthochromatic erythroblasts, and reticulocytes.

Figure 4

Exosc8-dependent barrier to erythroid maturation. (A) Real-time RT-PCR analysis of Exosc8 and selected GATA-1 target gene mRNA levels in control vs Exosc8 knockdown primary murine erythroid precursor cells cultured under expansion (–) or differentiation conditions (+) (mean ± SE; 3 independent experiments). Values were normalized to 18S rRNA expression and the expression is shown relative to control shRNA under expansion conditions. (B) Flow cytometric quantitation of erythroid developmental stage by CD71 and Ter119 staining upon Exosc8 knockdown in primary erythroid precursor cells. Representative flow cytometry data, with the R1-R5 gates denoted from 3 independent experiments. The percentage of live cells from each condition and the cell populations in R1-R5 stages (mean ± SE; 3 independent experiments). E, expansion; D, differentiation. (C) Representative images of Wright-Giemsa staining in control vs Exosc8 shRNA–infected primary erythroid precursor cells cultured in expansion or differentiation media (scale bar = 10 µm) and quantitation of cell size by measuring forward scatter using flow cytometry. (D) Flow cytometric cell-cycle analysis of primary erythroid precursor cells infected with retrovirus expressing control or Exosc8 2 shRNA. Representative cell-cycle profile is shown from 3 independent experiments. The percentage of the cell population in each cell-cycle stage is from 3 independent experiments (mean ± SE). Blue, S phase; red, G₀/G₁ or G₂/M phase. (E) Quantitative ChIP analysis of serine 5-phosphorylated RNA Polymerase II occupancy at Exosc8-regulated GATA-1 target and control genes in control and Exosc8-knockdown primary murine erythroid precursor cells (mean ± SE; 3 independent experiments). *P < .05, **P < .01, ***P < .001.

Figure 5

Evidence that the exosome complex creates a barrier to erythroid maturation. (A) Crystal structure of the human exosome complex demonstrating the interaction between Exosc8 and Exosc9. (B) Real-time RT-PCR analysis of Exosc8 and Exosc9 mRNA levels in control vs Exosc8 or Exosc9 knockdown in expanding primary murine erythroid precursor cells sorted into distinct, R1, R2, R3, and R4/5 cell populations (mean ± SE; 5 independent experiments). Values were normalized to 18S rRNA level and the expression relative to the control R1 shRNA condition. (C) Semiquantitative western blot analysis of Exosc9 protein in expanding primary murine erythroid precursor cells 24 hours postinfection with Exosc8- or Exosc9-specific shRNA retroviruses. Representative image from 3 independent experiments. (D) Real-time RT-PCR analysis of selected GATA-1 target genes and erythroid transcription factors mRNA levels in control vs Exosc8 or Exosc9 knockdown in expanding primary murine erythroid precursor cells sorted into distinct, R1, R2, R3, and R4/5 cell populations (mean ± SE; 5 independent experiments). Values were normalized to 18S rRNA level and the expression relative to the control R1 shRNA condition. (E) Flow cytometric quantitation of erythroid maturation stage by CD71 and Ter119 staining upon Exosc8 or Exosc9 knockdown in primary erythroid precursor cells. Representative flow cytometry data, with the R1-R5 gates denoted from 4 independent experiments. The percentage of the cell populations in R1-R5 stages (mean ± SE; 4 independent experiments). (F) Representative images of Wright-Giemsa–stained cells infected with control vs Exosc8 or Exosc9 shRNA-expressing virus. Cells were cultured under expansion conditions (scale bar = 10 µm). (G) Cell number fold change during the 72-hour expansion of primary murine erythroid precursor cells postinfection with control, Exosc8, or Exosc9 shRNA-expressing retroviruses (mean ± SE; 8 independent experiments). (H) Flow cytometric quantitation of erythroid maturation stage by CD71 and Ter119 staining upon Exosc8 or Dis3 knockdown in primary erythroid precursor cells. Representative flow cytometry data, with the R1-R5 gates denoted. The percentage of the cell populations in R1-R5 stages (mean ± SE; 6 independent experiments. *P < .05, **P < .01, ***P < .001.

Figure 6

GATA-1/Foxo3/Exosc8 coregulate genes important for erythroid maturation. (A) Classification of GATA-1– and Exosc8-regulated genes based on activation or repression. (B) Gene ontology analysis of GATA-1–activated, Exosc8-repressed genes. Redundant Gene Ontology categories were curated and removed. (C) Classification of Foxo3- and Exosc8-regulated genes based on activation or repression. (D) Gene ontology analysis of Foxo3-activated, Exosc8-repressed genes. Redundant Gene Ontology categories were curated and removed.

Figure 7

The exosome complex suppresses cell-cycle arrest genes during erythroid maturation. (A) Name and function^-,,,, of genes in “cell cycle arrest” category derived from GO term analysis from GATA-1–activated and Exosc8-repressed genes. (B) ChIP-seq profiles of GATA-1 occupancy at cell-cycle regulatory genes in primary human erythroblasts. All genes were orientated left to right. (C) Real-time RT-PCR analysis of genes in “cell cycle arrest” category upon Exosc8 or Exosc9 knockdown in primary murine erythroid precursor cells under expansion conditions, sorted into distinct, R1, R2, R3, and R4/5 cell populations (mean ± SE; 5 independent experiments). Values were normalized to 18S rRNA and the expression relative to the control R1 population. (D) Flow cytometric cell cycle analysis of primary erythroid precursor cells, within the R3 population, infected with retrovirus expressing control, Exosc8, or Exosc9 shRNA. Representative cell-cycle profile from 2 independent experiments. The percentage of the cell population in each cell cycle stage (mean ± SE; 2 independent experiments). Shaded, S phase; red, G₀/G₁ or G₂/M phase. *P < .05, **P < .01, ***P < .001. (E) Flow cytometric cell cycle analysis of primary erythroid precursor cells treated with 25 μM HU for either 24 or 48 hours. Representative cell-cycle profile. The percentage of the cell population in each cell cycle stage (mean ± SE; 3 independent experiments). Blue, S phase; red, G₀/G₁ or G₂/M phase. (F) Flow cytometric quantitation of erythroid maturation stage by CD71 and Ter119 staining upon 25 μM HU treatment of 24 or 48 hours in primary erythroid precursor cells. Representative flow cytometry data, with the R1-R5 gates denoted (3 independent experiments). (G) Model of GATA-1/Foxo3 function to overcome the exosome complex–dependent erythroid maturation barricade, which involves multiple alterations in erythroid cell function.