(A) wt and Stat5^−/− E13.5 embryos (left) and newborn animals (right). (B) Ter-119⁺ erythroid cells (left) and fetal liver cellularity (right; data are presented as mean ± SD; n = 6) of wt versus Stat5^−/− fetal livers. (C) ELISA for Epo from serum of wt and knockout E16.5 embryos and newborns. Data are presented as means ± SD; n = 5. *P < .05; **P < .01.

(A) Western blot of lysates from freshly isolated fetal livers or Ter-119⁺ and Ter-119⁻ subfractions (separated using magnetic beads; “Methods”) for Bcl-x_L (left; ERK, loading control). Western blot for Mcl-1 of 2 individual freshly isolated fetal livers of each genotype (right; eIF4E, loading control). (B) Quantitative PCR analysis for Mcl-1 mRNA isolated from the Ter-119⁺ subfraction of freshly isolated fetal liver cells (data are presented as means ± SD; n = 3). (C) wt and Stat5^−/− primary erythroblasts expanded for 5 days under self renewal conditions (“Methods”) were deprived of factors for 3 hours, followed by a 30-minute restimulation with 10 U/mL Epo. qPCR analysis for Mcl-1 (representative experiment; error bars are SD of experimental triplicates). (D) wt or Stat5^−/− fetal liver cells were infected with retroviruses encoding GFP alone (murine stem cell virus [MSCV]), or Bcl-x_L plus GFP, or Mcl-1 plus GFP from bicistronic constructs. After retroviral infection (72 hours), primary erythroblasts were cultivated for another 48 hours under self-renewal conditions (“Methods”) in the presence or absence of Epo. Rates of apoptosis were determined by flow cytometry for annexin V. One representative set of histograms from 3 independently performed experiments of GFP-gated annexin V⁺ cells at 48 hours of treatment is depicted. Arrowheads indicate increased levels of apoptosis. *P < .05; **P < .01.

(A) Western blot analysis of primary wt and Stat5^−/− erythroblast lysates for IRP-1 (left panel) and IRP-2 (right panel). (B) Quantification of Western blot analysis from panel A (data are presented as means ± SD). Samples were normalized on eIF4E as levels and quantified using the Odyssey infrared imaging system. (C) Two representative lysates each of wt and Stat5^−/− erythroblasts were subjected to RNA-EMSAs for IRP-1 and IRP-2 using an IRE-RNA probe corresponding to the IRE of mouse ferritin heavy chain64 (for experimental details, see Document S1). (D) Quantification of IRP-2–binding activity (data are presented as means ± SD; n = 4). *P < .05; **P < .01.

Left side shows that in iron-replete cells, IRP-1 is converted into cytosolic aconitase (catalyzes isomerization of citrate to isocitrate in the citric acid cycle and exhibits no mRNA-binding affinity; green asterisk) and IRP-2 is degraded. Therefore, both cannot bind to IREs in the 3′UTR of TfR-1 mRNA. Free unprotected IREs in turn enhance degradation rates of TfR-1 mRNA, resulting in reduced iron uptake. In iron-depleted cells, IRP-1 and IRP-2 bind to the respective IREs, thereby stabilizing TfR-1 mRNA, resulting in increased iron uptake. The right side shows Stat5^−/−. Due to lack of Stat5, basal TfR-1 transcript abundance is reduced in comparison to wt cells. In addition, Stat5 deficiency further results in decreased levels of IRP-2 and, in consequence, a reduction of binding to IREs in the 3′UTR of TfR-1 mRNA and decreased transcript stabilization. Together, this constitutes a double-negative effect on erythroid iron uptake even in a situation of high iron demand, as in iron-deficiency anemia. *P < .05; **P < .01.

(A) Freshly isolated fetal livers from E13.5 to E15.5 were stained for Ter-119 and annexin V to determine rates of apoptosis (data are presented as means ± SD; n = 3) for each genotype and time point. (B) CFU-E colonies derived from wt or Stat5^−/− fetal liver cells using the indicated Epo concentrations (data are presented as means ± SD; n = 4). (C) E13.5 fetal liver cells of wt and Stat5^−/− embryos were infected with a retrovirus encoding GFP. TfR-1^high/c-Kit⁺/GFP⁺ cells were isolated by FACS after 7 days under self-renewal conditions (cytospin; right panel). (D) A total of 1.5 × 10⁷ TfR-1^high/c-Kit⁺/GFP⁺ cells were injected into the tail vein of lethally irradiated mice (950 rad) and Ter-119⁺/GFP⁺ bone marrow cells were scored 3 days later (mean ± SD; n = 4). *P < .05; **P < .01.

(A) Representative flow cytometry histograms of E13.5 wt and Stat5^−/− fetal liver cells stained for the erythroid markers TfR-1 and Ter-119 (left). The sequence from gate R1 (TfR-1^low Ter-119^low) to gate R5 (TfR-1⁻ Ter-119^high) represents development from the most immature erythroid progenitors (late BFU-E; CFU-E) to mature erythroid cells (orthochromatic erythroblasts; reticulocytes).58 Quantification of gates R1 to R5 (data are presented as means ± SD; n = 4) (right). (B) Cell-surface expression of TfR-1 of Ter-119^high gated wt (dark gray line) or Stat5^−/− (light gray line) fetal liver cells (left). Quantification of TfR-1 cell-surface expression of wt and Stat5^−/− fetal livers (right; data are presented as means ± SD; n = 4). (C) Expression of TfR-1 mRNA from lysates of freshly isolated wt or Stat5^−/− fetal liver cells (data are presented as means ± SD; n = 3). Expression was normalized on hypoxanthine-guanine phosphoribosyltransferase (HPRT) levels. (D) Western blot analysis of freshly isolated wt and Stat5^−/− fetal liver cell lysates for TfR-1. ERK was used as loading control. (E) Densitometric quantification of TfR-1 Western blot in 3D (data are presented as means ± SD; n = 4). (F) Primary wt fetal liver erythroblasts were stimulated with Epo for 30 minutes and ChIP for Stat5 was performed. DNA from Epo-stimulated primary wt erythroblasts was recovered using 2 different antisera directed against N- or C-terminal epitopes (αS5 C, αS5 N). Specific PCR products from Stat5-binding sites GAS 1, GAS 2, and GAS 3 in TfR-1 intron 159 were only obtained with Stat5-specific antibodies but not with control IgGs. PCR for the genuine Stat5 site in the CIS promoter was used as positive control. (G) Primary wt fetal liver–derived erythroblasts were factor depleted for 2.5 hours, followed by 1.5 hours of Epo stimulation (10 U/mL). TfR-1 mRNA expression was scored by qPCR normalized on HPRT (data are presented as means ± SD; n = 4). (H) Iron concentration in freshly isolated fetal liver lysates determined via atomic absorption spectrometry (data are presented as means ± SD; n = 4; for experimental details, see Document S1). *P < .05; **P < .01.

(A) Sequence of the IRP-2 promoter −1030 bp to −1100 bp upstream of the transcription start, showing one perfect GAS site (boxed in black) and 2 GAS sites with one mismatch (boxed in gray). (B) Multiple perfect Stat5 sites taken from Yao et al26 and Ehret et al65 together with the IRP-2-I. (C) Luciferase reporter assay using a DNA fragment ranging from 2 kb immediately upstream of the predicted IRP-2 transcription start site. Vertical lines indicate the approximate positions of the putative Stat5 response elements. 293T cells were cotransfected with constructs encoding IRP-2-firefly-luciferase, renilla-luciferase, Stat5a, and EpoR. Cells were treated with 10 U/mL Epo or left untreated, and luminescence was scored 3 hours later. Transfection efficiencies were normalized to renilla-luciferase activity (representative experiment; error bars are SD of experimental triplicates). (D) Epo-dependent induction of endogenous IRP-2 and oncostatin M analyzed via quantitative PCR in murine erythroid leukemia cells serum deprived for 3 hours, followed by stimulation with 10 U/mL Epo (1 hour). Quantitative PCR was normalized on HPRT (data are presented as means ± SD; n = 4). (E) 293T cells were cotransfected with constructs for EpoR and wt Stat5a, followed by 30 minutes of stimulation with 10 U Epo/mL. Whole-cell extracts of these cells were subjected to EMSAs using either the IRP-2-I oligonucleotide (left) or β-casein oligonucleotide as positive control (right). Respective arrows indicate Stat5 DNA complexes and Stat5 DNA complex supershifts. (F) Primary wt fetal liver erythroblasts were stimulated with Epo for 30 minutes, and ChIP for Stat5 was performed. Recovered DNA was analyzed for the presence of promoters of IRP-2 and CIS (positive control) by PCR. *P < .05; **P < .01.