Erythroid Kruppel-like factor (EKLF) is recruited to the gamma-globin gene promoter as a co-activator and is required for gamma-globin gene induction by short-chain fatty acid derivatives

Abstract

Objectives: The erythroid Kruppel-like factor (EKLF) is an essential transcription factor for beta-type globin gene switching, and specifically activates transcription of the adult beta-globin gene promoter. We sought to determine if EKLF is also required for activation of the gamma-globin gene by short-chain fatty acid (SCFA) derivatives, which are now entering clinical trials.

Methods: The functional and physical interaction of EKLF and co-regulatory molecules with the endogenous human globin gene promoters was studied in primary human erythroid progenitors and cell lines, using chromatin immunoprecipitation (ChIP) assays and genetic manipulation of the levels of EKLF and co-regulators.

Results and conclusions: Knockdown of EKLF prevents SCFA-induced expression of the gamma-globin promoter in a stably expressed microLCRbeta(pr)R(luc) (A)gamma(pr)F(luc) cassette, and prevents induction of the endogenous gamma-globin gene in primary human erythroid progenitors. EKLF is actively recruited to endogenous gamma-globin gene promoters after exposure of primary human erythroid progenitors, and murine hematopoietic cell lines, to SCFA derivatives. The core ATPase BRG1 subunit of the human SWI/WNF complex, a ubiquitous multimeric complex that regulates gene expression by remodeling nucleosomal structure, is also required for gamma-globin gene induction by SCFA derivatives. BRG1 is actively recruited to the endogenous gamma-globin promoter of primary human erythroid progenitors by exposure to SCFA derivatives, and this recruitment is dependent upon the presence of EKLF. These findings demonstrate that EKLF, and the co-activator BRG1, previously demonstrated to be required for definitive or adult erythropoietic patterns of globin gene expression, are co-opted by SCFA derivatives to activate the fetal globin genes.

Figures

Figure 1

Effect of erythroid Kruppel-like factor (EKLF) knockdown in murine cells on transcriptional induction of a human γ-globin promoter. Panel (A) immunoblot showing knockdown of endogenous EKLF protein levels in GM979 cells. Cells were transfected with either a scrambled siRNA (siC) or with siRNA directed towards the open reading frame of EKLF (siEKLF), and treated with RB7 or vehicle. β-actin protein levels (lower panel) serve as a loading control. Panel (B) dual luciferase reporter assay. γ-globin gene/β-globin gene induction measured as a ratio of firefly luciferase to renilla luciferase activities in GM979 cells treated with control- or EKLF-specific siRNA, with or without exposure to RB7. Values are normalized to results obtained from cells that were transfected with a control siRNA (siC), and represent an average of four experiments. Error bars represent the standard error of the mean (SEM). *P < 0.01, compared with siRNA control.

Figure 2

Effects of knockdown of endogenous erythroid Kruppel-like factor (EKLF) in cultured human erythroid precursors on globin gene induction and promoter occupancy by SCFA derivatives. Panel (A) immunoblot showing EKLF protein levels in cells cultured under control conditions (C), and in the presence of RB7 at 20 μm (RB7) with 400 nm of siRNA control (RB7 siControl), or with 400 nm siEKLF (RB7 siEKLF). The blot was stripped and re-probed with an antibody to β-actin to serve as a loading control. Panel (B) EKLF transcript (top panel) and γ-globin transcript (bottom panel) levels from cells treated under the same conditions described for panel A. γ-globin gene induction measured by γ-globin mRNA transcript levels in primary human erythroid progenitor cells by quantitative RT-PCR, relative to β-actin mRNA levels. Values are normalized to results obtained from cells that were transfected with a control siRNA (siC), and represent an average of four experiments. Panel (C) ChIP assay for EKLF at the endogenous globin gene promoters under the conditions described in panel A. Photograph of an ethidium-stained agarose gel showing amplification of DNA fragments purified following ChIP using primers for γ-globin, β-globin, and β-actin promoters. ‘Input’ shows starting material prior to chromatin IP. The column on the right side of this panel shows amplified products from input DNA, without immunoprecipitation, using the primers for the γ-globin, β-globin or β-actin promoter regions. Panel (D) quantitative PCR following ChIP assay under the experimental conditions used for panel C. Relative association of EKLF with endogenous γ- and β-globin promoters. Values are normalized to cells cultured under control (vehicle) conditions, and are averaged from independent experiments. In panels B and D, error bars represent the standard error of the mean (SEM). *P < 0.01, compared with siRNA control. As specificity controls, parallel ChIP assays using pooled IgG, or anti-β-actin, anti-Ikaros or anti-TR2/TR4 antibodies produced no amplified products by quantitative PCR at 40 amplification cycles (Fig. 3).

Figure 3

Panel (A) chromatin immunoprecipitation (ChIP) assays at the endogenous β-type globin gene promoters: Relative association of erythroid Kruppel-like factor (EKLF) at the β-globin promoter (left side) and γ-globin promoter (right side) after human erythroid progenitors were cultured in the presence of short-chain fatty acid derivatives RB7 (at 20 μm) or RB14 (20 μm). Results are normalized to the vehicle-treated control culture sample (C) for either the β-globin or γ-globin primers sets, arbitrarily set at 100% for each control. and represent an average of independent experiments. Panel (B) as a specificity control, parallel ChIP assays using pooled IgG or anti-β-actin antibodies produced no amplified products by quantitative PCR at 40 cycles. Similarly, parallel ChIP assays using anti-Ikaros or anti-TR2/TR4 antibodies produced no amplified products by quantitative PCR at 40 cycles (not shown). Panel (C) relative EKLF transcript levels in erythroid progenitors cultured under control conditions (C), or in the presence of short-chain fatty acid derivatives RB7 (20 μm) or RB14 (20 μm), measured by quantitative RT-PCR relative to β-actin transcript levels. Values are normalized to control conditions. Results are averaged from independent experiments. In all panels, error bars represent the standard error of the mean (SEM). *P < 0.01, compared with control.

Figure 4

Panel (A) inducible co-immunoprecipitation of EKLF/BRG1. Human erythroid progenitors were cultured either under control (vehicle) conditions (C) or in the presence of RB7 or RB14, then lysed and proteins were immunoprecipitated with an anti-BRG1 antibody, separated by PAGE, and immunoblotted with an anti-BGR1 or an anti-EKLF antibody. BRG1 protein levels in the immunoblot demonstrate equal loading. Relative amounts of EKLF protein co-immunoprecipitating with BRG1 under each condition are plotted, after quantitation, in arbitrary units. Panel (B) effects of knockdown of endogenous BRG1 protein in human erythroid progenitors on γ-globin gene induction. Immunoblot showing BRG1 protein levels in RB7- and RB14-treated cultures. Cells were treated with either a scrambled siRNA (siControl) or BRG1-specific siRNA (siBRG1) (both at 400 nm). Immunoblot of β-actin levels serves as a loading control. γ-globin mRNA transcripts were quantified by real-time quantitative RT-PCR after primary erythroid cells were cultured either in control conditions (C), or in presence of short-chain fatty acid derivatives RB7 (left) or RB14 (right). Values were normalized to levels under control conditions. In both panels, error bars represent the standard error of the mean (SEM). *P < 0.01, compared with siRNA control.

Figure 5

Chromatin immunoprecipitation (ChIP) assays of BRG1 at the endogenous β-type globin gene promoters in human primary erythroid progenitor cells. Quantitative PCR of immunoprecipitated complexes showing relative association of BRG1 with the γ-globin promoter (top left plot) and β-globin promoter (bottom left plot); RB7: Chromatin from RB7-treated cultures (20 μm); RB7 (siControl): Chromatin from RB7-treated cultures, transfected with scrambled siRNA (400 nm); RB7 (siEKLF): Chromatin from RB7-treated cultures, transfected with an siRNA specific for EKLF (400 nm). As specificity controls, no products were amplified at 40 PCR cycles after IP with the BRG1-specific antibody, using primers specific for the c-fos promoter (right plot). Results represent an average of two independent experiments. Error bars indicate the standard error of the mean (SEM). *P < 0.01.