Introduction

GATA-1, the prototypical member of the GATA family of transcription factors, recognizes a consensus target DNA sequence of WGATAR found in promoters, enhancers and locus control regions 1. The expression of GATA-1 is restricted to erythroid cells, megakaryocytes, eosinophils and mast cells, as well as to Sertoli cells in the testes 2, 3, 4. GATA-1 is essential for hematopoietic cell differentiation. Loss of GATA-1 leads to erythroid maturation arrest 5 and embryonic lethality that is due to anemia in mice 6. In vitro differentiation assays have shown that murine GATA-1− embryonic stem cells can differentiate into erythroid precursors, but undergo cell-cycle arrest and death at the proerythroblast stage 7. Selective loss of GATA-1 expression in megakaryocytes of mutant mice results in a reduction in the number of platelets and produces hyperproliferation of megakaryocytes, indicating a role for GATA-1 in regulating the differentiation and maturation of these cell types 8.

Given the importance of GATA-1 in specifying the erythroid lineage, it is of particular interest to define how the GATA-1 gene is transcriptionally regulated. GATA-1 is transcribed from two promoters, IT and IE. IT is active predominantly in Sertoli cells, whereas IE is primarily utilized in hematopoietic cells 9. The IE promoter has maximal activity in erythroid cells in transient transfection assays 10, 11, 12. In a previous study, a lacZ reporter gene under the control of a GATA-1 promoter was expressed in a mouse lacking the GATA-1 gene 13. In addition, two regions within the IE promoter, double GATA-1 and CACCC sites, are conserved in mice and are important for GATA-1 expression 12, 14. Surprisingly, a mutation of the double GATA-1 site did not abrogate GATA-1 expression in erythroid cells, but instead eliminated the eosinophil lineage 15. Although the CACCC box is critical for GATA-1 IE promoter activity 16 and multiple proteins such as Sp1/Sp3 and KLF family factors have been reported as binding to the CACCC box, no work has been carried out to identify the proteins involved. Therefore, a detailed investigation will indicate which factors play critical roles in GATA-1 gene expression through binding of the CACCC motif 17.

To clarify this issue, we performed a functional analysis of the human GATA-1 gene IE promoter through a series of deletions. We identified a 120-bp DNA element (from −221 to −102 bp), which is essential to drive reporter gene expression in megakaryotic K562 cells. By site-specific mutation, we confirmed that both the CACCC and CGCCC box are important for the activity of the 120-bp element, while the double GATA-1 site is dispensable for GATA-1 gene expression. Electrophoretic mobility shift (EMSA) and chromatin immuno-precipitation assays (ChIP) revealed that proteins binding to the IE promoter are the ubiquitous transcription factors Sp1/Sp3, instead of EKLF, in K562 cells, while EKLF, Sp1 and Sp3 are all found enriched at the mouse GATA-1 promoter in adult mouse erythrocytes. The nucleosomes in the IE promoter are also shown to be highly acetylated. Co-transfection of Sp1 expression plasmids enhanced reporter gene activity in a dose-dependent manner.

Results

Sequence analysis of the GATA-1 IE promoter

Sequences of the human and mouse GATA-1 IE promoters were analyzed, and transcription factor binding sites were identified. Sequences are denoted as reported 14. A double GATA motif is located at around −680 bp in both promoters. In addition, four GATA sites (at −847, −551, −357 and −340 bp) found in the human GATA-1 promoter were not found in the mouse promoter. The functions of these GATA sites have not been characterized. Two CACCC boxes in the mouse GATA-1 promoter are critical for promoter activity, whereas only one CACCC box was identified in the human GATA-1 promoter 16. Corresponding to the second mouse, CACCC site is a CGCCC site in the human GATA-1 promoter, which might be less important since it is not as conserved. Moreover, one YY1 site at −786 bp and two Ap1 sites at −561 and −369 bp were also found in the human GATA-1 promoter.

Reporter gene analysis of the proximal promoter

The activity of reporter genes containing the human GATA-1 promoter and a series of deletions from the 5′ end was tested and compared with the pGL3-Basic (Figure 1). The GATA-1 IE promoter (from −860 to −1 bp) exhibited strong transcription initiation activity in K562 cells. By contrast, the sequence from −102 to −1 bp displayed very low transcription activity. The sequence from −221 to −1 bp retained most of the activity of the −860 bp fragment, indicating that this sequence is the core of the GATA-1 erythroid promoter.

Figure 1
figure 1

Reporter gene analysis of the proximal promoter. (A) Structure of the promoter constructs. (B) Corresponding transcriptional activity of the promoter constructs in K562 cells. Results are expressed as the ratio of luciferase activity, normalized to an internal control (Renilla luciferase), to the promoter-less plasmid pGL3-Basic set as 1. All histograms represent the mean ± standard error of at least three independent experiments.

Full size image

The activity of pGL3-331-luc is only 60.5% of that of pGL3-221-luc, suggesting that the sequence between −331 and −222 bp may be a negative control element. So far, no binding sites for known transcription factors have been identified in this region. When the sequence was extended to −441 bp, reporter gene activity recovered to the level of the −221-luc fragment, implying that the sequence between −441 and −332 bp may be a positive regulatory element in which there are two GATA-1-binding sites and one AP1-binding site. The data demonstrate that these sites might be necessary for the activity of this regulatory element.

Neither deletion of the YY1 site at −786 bp nor deletion of the AP1 site at −561 bp resulted in any significant effect on the expression of the reporter gene. Likewise, no significant effect on reporter gene activity was observed after deleting the double GATA-1 site (around −680 bp). Although GATA-1 sites are supposed to be important for self-regulation, those sites seem not to be important for GATA-1 gene expression in K562 cells.

Functional analysis of the CACCC box

There are two CACCC boxes in the mouse GATA-1 IE promoter, which are separated by 17 nucleotides, whereas there is one CACCC box (from −173 to −169 bp) and one CGCCC box (from −195 to −191 bp) in the human GATA-1 promoter. The activity level after deletion of the CGCCC box (−180 bp fragment) was 36% less than that of the −221 bp construct. Deletion of both CGCCC and CACCC boxes (−160 bp fragment) reduced the reporter gene activity to 13.2% of that of the −221 bp construct and 22% of the activity of the −180 bp construct with only the CGCCC box deleted (Figure 2A and 2B). These results reveal that the conserved CACCC box is more crucial for the activity of the human GATA-1 promoter than the CGCCC box, which is not conserved between the human and mouse genomes.

Figure 2
figure 2

Functional analysis of the CACCC box. (A) Structures of the pGL3-221-luc promoter deletion constructs. (B)Transcriptional activity of the deletion constructs in K562 cells. Results are expressed as the ratio of luciferase activity, normalized to an internal control (Renilla luciferase), to the promoter-less plasmid pGL3-Basic set as 1. All histograms represent the mean ± standard error of at least three independent experiments. (C) Structures of promoter constructs with a mutated CACCC box. (D) Transcriptional activity of the constructs with a mutated CACCC site in K562 cells. Results are expressed as the ratio of luciferase activity, normalized to an internal control (Renilla luciferase), to the promoter-less plasmid pGL3-Basic set as 1. All histograms represent the mean ± standard error of at least three independent experiments.

Full size image

The CACCC box was mutated to further test its importance for promoter activity. The data revealed that mutation of CACCC box reduced reporter activity to ∼30% of the wild type −221 construct, confirming the importance of the CACCC box for the activity of the human GATA-1 promoter (Figure 2C and 2D).

Identification of the transcription factor that binds the CACCC box

EMSA and ChIP assays were carried out to determine which transcription factors bind the CACCC box in the human GATA-1 promoter. K562 nuclear extract was incubated with a CACCC box probe synthesized from double −stranded (ds) WT oligonucleotides (−180 to −161 bp, see Materials and Methods) and labeled with [γ]-32P-ATP. Cold ds MP, Sp1W and Sp1M oligonucleotides (see Materials and Methods) were also synthesized and used as competitors. The data revealed that one band was shifted (S) (Figure 3A, upper), which was eliminated by the addition of cold ds WT and Sp1W DNA fragments at a 100× molar excess, but not by ds MP and Sp1M DNA fragments (Figure 3A). A monoclonal antibody against Sp1 (sc-420X) was added in the supershift assay. The shift band S was weakened and another retarded band (SS) was formed (Figure 3B), indicating that the transcription factor binding the CACCC box might be Sp1. CGCCC box was also shown to be shifted to a similar level (Figure 3A, lower). To confirm the proteins that bind to the IE promoter in vivo, we carried out ChIP assays with antibodies against Sp1, Sp3 and EKLF. Sp1 and Sp3 but not EKLF were shown to be enriched at the GATA-1 promoter only in K562 cells (Figure 3C). However, also by ChIP analysis, Sp1/Sp3 and EKLF were all shown to be enriched at the mouse GATA-1 promoter in mouse anemic spleen adult erythrocytes (Figure 3D).

Figure 3
figure 3

Identification of the transcription factors that bind to the CACCC box. Electrophoresis mobility shift and chromatin immuno-precipitation assays of the human promoter. (A) Upper panel: the ds WT oligonucleotide (from −180 to −161 bp) labeled with [γ]-32P-ATP was incubated with K562 nuclear extracts. Cold ds WT, MP, Sp1W and Sp1M fragments were used as competitors and run on a 4% 0.5× TBE non-denaturing polyacrylamide gel. Lane 1, free labeled probe (negative control); lanes 2-10, labeled probe with 1 μg nuclear extract of K562 (band S indicates the specific binding band of the probe); lanes 3 and 4, unlabeled WT was used as a competitor (lane 3, 10-fold molar excess; lane 4, 100-fold molar excess); lanes 5 and 6, unlabeled MP was used as a competitor (lane 5, 10-fold molar excess; lane 6, 100-fold molar excess); lanes 7 and 8, unlabeled Sp1W was used as a competitor (lane 7, 10-fold molar excess; lane 8, 100-fold molar excess); lanes 9 and 10, unlabeled Sp1M was used as a competitor (lane 9, 10-fold molar excess; lane 10, 100-fold molar excess). Lower panel: CGCCC and CACCC oligos were used as probes in an EMSA assay. Lane 1, free labeled CACCC probe (negative control); lanes 2-5, labeled probes (lanes 2 and 3, CACCC probe; lanes 4 and 5, CGCCC probe) with 1 μg nuclear extract of K562 (band S indicates the specific binding band of the probe); lane 3, unlabeled CACCC oligo was used as a competitor (100-fold molar excess); lane 5, unlabeled CGCCC oligo was used as a competitor (100-fold molar excess). (B) Supershift assay reveals the effect of commercial Sp1 antibody on the shift of K562 nuclear extracts that bind the labeled probe. Band SS is the retarded band above the S band. (C) Identification of Sp1/Sp3 in vivo binding to the promoter region of human GATA-1. A ChIP assay was performed using anti-Sp1, anti-Sp3 and anti-EKLF antibodies as described in Materials and Methods. Lanes 1-5 on non-hematopoietic 293T cells: lane 1, input DNA; lane 2, non-specific IgG control; lanes 3-5, anti-EKLF, anti-Sp1 and anti-Sp3 antibodies. Lanes 6-10 on K562 cells: lane 6, input DNA; lane 7, non-specific IgG control; lanes 8-10, anti-EKLF, anti-Sp1 and anti-Sp3 antibodies. PCR amplified a 171-bp fragment from the human GATA-1 promoter region. The lowest marker size is 100 bp. (D) ChIP on mouse adult erythrocytes from anemic spleen. Lane 1, input DNA; lane 2, non-specific IgG control; lanes 3-5, anti-EKLF, anti-Sp1 and anti-Sp3 antibodies. PCR amplified a 124-bp fragment from the mouse GATA-1 promoter region. The lowest marker size is 200 bp.

Full size image

Acetylation of histones H3 and H4 in the human GATA-1 IE promoter

Actively transcribed genes are enriched with acetylated nucleosomes that facilitate the remodeling of chromatin and increase the accessibility of DNA to transcription factors. An enrichment of acetylated histones H3 and H4 at the human GATA-1 promoter region was seen in K562 but not in 293T cells using ChIP analysis (Figure 4). The presence of acetylated histones in the Sp1/Sp3 binding region indicates that Sp1/Sp3 might have a role in the recruitment of histone acetyltransferase (HAT) to the human GATA-1 IE promoter.

Figure 4
figure 4

Acetylation of histones H3 and H4 in the GATA-1 promoter region in K562 cells. A ChIP assay was carried out using antibodies against acetylated histones H3 and H4 as described in Materials and Methods. Lanes 1-4 on non-hematopoietic 293T cells: lane 1, input DNA; lane 2, non-specific IgG control; lanes 3 and 4, anti-acetylated histones H3 and H4. Lanes 5-8 on K562 cells: lane 5, input DNA; lane 6, non-specific IgG control; lanes 7 and 8, anti-acetylated histones H3 and H4 and PCR amplification of anti-acetylated histones H3 and H4. PCR amplified a 171-bp fragment from human GATA-1 promoter region. The lowest marker size is 100 bp.

Full size image

Sp1 upregulates reporter gene expression

To further confirm the effect of Sp1 on GATA-1 promoter activity, Sp1 expression plasmids were co-transfected with the wild-type (−180 to −1 bp) construct into K562 cells. As shown in Figure 5, reporter gene activity was upregulated in an Sp1-dependent manner, indicating that Sp1 in K562 cells is involved in the transcriptional initiation of the human GATA-1 promoter through binding to the CACCC box.

Figure 5
figure 5

Effect of Sp1 on reporter gene expression. The Sp1 expression plasmids were co-transfected with the reporter gene construct pGl3-180-luc. Results are expressed as the ratio of luciferase activity, normalized to the internal control (Renilla luciferase), to the promoter-less plasmid pGL3-Basic set as 1. All histograms represent the mean ± standard error of at least three independent experiments.

Full size image

Discussion

The GATA-1 transcription factor has a critical role in the differentiation and maturation of erythroid cell lines by regulating the expression of downstream hematopoietic genes 18, 19, 20. However, the transcriptional regulation of GATA-1 itself is still open to further investigation. Characterization of the cis-acting elements that control the expression of the GATA-1 gene should lead to the identification of factors that act upstream of GATA-1 and shed light on the transcriptional regulatory network formed by GATA-1. To date, most knowledge of the regulation of GATA-1 gene expression has been obtained from studies of mouse, chicken and zebrafish promoters. In this report, we carried out serial deletion and mutation assays in the human GATA-1 erythroid promoter.

Studies in mice and chickens have shown that a functional double GATA motif in the proximal region of the GATA-1 promoter, which is important for full activity, is bound by GATA-1 14, 11, 12. Based on such observations, a positive autoregulatory mechanism was proposed for increasing and maintaining GATA-1 expression during differentiation and cellular maturation. McDevitt et al. 13 showed that GATA-1 is not required for activation and maintenance of GATA-1 expression because a GATA-1/lacZ transgene could be expressed on a GATA-1− background Our results reveal that the deletion of the double GATA motif in the human GATA-1 proximal promoter does not have a significant effect on the full activity of the promoter in transient transfection assays. In addition, the deletion of GATA motifs at −847 and −551 bp do not produce an effect on the promoter activity. Deletion of a sequence (from −441 to −331 bp) containing two GATA motifs (at −357 and −340 bp) and one AP1 binding site (at −369 bp) reduces promoter activity by approximately 40% compared with the −860 bp construct, indicating that all three motifs might be important in regulating the expression of human GATA-1. Therefore, the importance of the double GATA-1 sites should be considered carefully, and other GATA sites in the promoter should be examined to determine their role in the proposed GATA-1 autoregulatory mechanism. Putative GATA sites could be found frequently in the sequence of a promoter. However, whether those sites are occupied by GATA-1 family proteins during hematopoietic lineage differentiation remain to be discovered. And if bound by GATA-1 family proteins, how they collaborate with other transcription factors is also interesting to investigate. Although the changes in promoter activity using deletion or mutation promoters in vitro may reflect the importance of specific elements in vivo, the importance of the regulatory elements identified in vitro needs to be tested in vivo.

The double CACCC boxes are important for the full activity of mouse and zebrafish GATA-1 promoters 16, 21. The functional difference between the two CACCC boxes has not been analyzed and proteins binding to those boxes have not been identified. Whether both CACCC boxes are equally important in the expression of the GATA-1 gene remains to be seen. We found that the human GATA-1 promoter has only one CACCC box and a CGCCC box that corresponds to the second mouse CACCC box. Deletion and mutation assays showed that the non-conserved CGCCC box, although necessary for full promoter activity, is less important than the conserved CACCC box for activity of the human GATA-1 promoter.

The protein that binds to the CACCC/CGCCC box has been speculated as Sp1, and possibly EKLF, but there is little concrete evidence to show exactly which protein it is. These motifs are high-affinity binding sites for the ubiquitous Sp1/Sp3 protein family and Krüppel-like factors (KLF). Through EMSA and ChIP assays we show for the first time that the proteins that bind to the CACCC box in the human GATA-1 erythroid promoter at least include Sp1/Sp3 in K562 cells. Moreover, co-transfection of Sp1 expression plasmids enhanced the transcription of a reporter gene with the minimal promoter (−180 to −1 bp) in an apparently dose-dependent manner. We cannot rule out the possibility that erythroid KLF (EKLF) binds the CACCC/CGCCC box at other hematopoietic developmental stages or in other specific cell types. We tested mouse adult erythrocytes from anemic spleen and showed that Sp1/Sp3 and EKLF are enriched at the mouse GATA-1 promoter. However, it is unlikely that EKLF is required for GATA-1 expression because GATA-1 is still expressed in murine EKLF−/− embryos 22. This observation could also be explained by the substitution of EKLF by the ubiquitous factors Sp1/Sp3 that bind to the CACCC/CGCCC box. In addition, recent studies indicate that EKLF expression is not detected in K562 cells 23, 24. But the relationship between Sp1/Sp3 and EKLF in adult mouse erythrocytes remains another interesting question. If erythroid EKLF is not necessary for GATA-1 promoter activity, the hematopoietic-specific expression of GATA-1 may be due to other transcription factors. Since GATA-2 is expressed earlier than GATA-1 and multiple GATA sites are identified in the GATA-1 promoter, we speculate that GATA-2 could be important at the beginning to specify the transcription of GATA-1 in a hematopoietic-specific manner.

Sp1 physically interacts with GATA-1 to synergistically activate reporter transcription in transient transfection assays 25, 26. Moreover, the proximal promoter region of the human β-globin gene also contains two GATA sites and one CACCC box, an organization similar to that in the human GATA-1 proximal promoter. Although there is no direct evidence that the Sp1 bound to the CACCC box in the human GATA-1 promoter interacts with the GATA-1 protein, we propose that they may form a large regulatory complex on the promoter and recruit other factors that are involved in regulating the expression of this gene. For example, the HAT, like CBP/p300, was reported to be recruited by both GATA-1 and Sp1 to target sites, since HATs generally do not interact with DNA directly 27, 28. Our preliminary data also suggest that the erythroid specificity of human GATA-1 gene expression may also depend on some other regulatory elements beyond the proximal promoter. Therefore, how the enhancer and promoter communicate, and the role of Sp1/Sp3 binding, are subjects for further investigation. Mouse and human GATA-1 genes are highly conserved. It is predicted that similar regulatory mechanisms are involved in both promoters. Studying both human and mouse promoters helps in our understanding of how the GATA-1 gene is transcriptionally regulated.

Materials and Methods

Cloning of the human GATA-1 promoter

Human genomic DNA was extracted from cultured K562 cells with the Genomic DNA Purification Kit (BioColor Co., Shanghai, China). The 860-bp (from −860 to −1 bp) human GATA-1 IE promoter was amplified with the upstream primer HGFPa and the downstream primer HGRP (Table 1) and cloned into the pGL3-Basic plasmid (Promega) after digestion with HindIII and KpnI (Invitrogen) to generate the reporter construct pGL3-860-luc. The insert was sequenced bi-directionally and then analyzed for putative regulatory elements with TRANSFAC 4.0 29. Position −1 denotes the last nucleotide of exon-1 transcribed from the IT promoter.

Table 1 Primers used in the analysis of the human GATA-1 IE promoter
Full size table

Generation of reporter gene constructs

The pGL3-751-luc, pGL3-629-luc, pGL3-528-luc, pGL3-441-luc, pGL3-331-luc, pGL3-221-luc, pGL3-102-luc, pGL3-a-luc, pGL3-b-luc and pGL3-c-luc reporter constructs were obtained with the upstream forward primers HGFPb, HGFPc, HGFPd, HGFPe, HGFPf, HGFPg, HGFPh, HGFGa, HGFGb and HGFGc, in corresponding order, and the downstream reverse primer HGRP by PCR and cloned (see Results). The pGL3-m-luc mutant construct was generated with the upstream primer mHGFP (Table 1) and the downstream primer HGRP. All plasmids were prepared using a Plasmid Midi Prep Kit (Novagen) and verified by sequencing.

Cell culture, transfection and luciferase assay

K562 cells were cultured in RPMI 1640 medium (Invitrogen) supplemented with 10% fetal bovine serum (FBS). In each transfection, 1 μg of reporter plasmid and 0.01 μg of pRL-CMV (plasmid encoding Renilla luciferase as an internal control, Promega) were co-transfected. Plasmid DNA (1 μg) and DMRIE-C reagent (2.4 μL) were each diluted in 100 μL of Opti-MEMI reduced serum medium (Invitrogen). The plasmid DNA and DMRIE-C reagent were then mixed and incubated at room temperature for 30 min on a 24-well-plate. Approximately 4 × 105 cells in 40 μL of serum-free medium were added to each well and mixed gently. The cells were incubated at 37 °C with 5% CO2 for 4 h and 400 μL of growth medium containing 15% FBS were added to each well. After 48 h, luciferase activity was measured with the Dual-Luciferase Reporter Assay System (Promega) on a Lumat LB 9507 luminometer (EG&G), normalized to the internal control. Relative luciferase activity was calculated by comparing the luciferase activity of the reporter construct with that of the promoter-less pGL3-Basic. In the co-transfection assay, 0.2 μg of reporter plasmid was used with 0.4 or 0.8 μg of Sp1 expression plasmid (pcDNA3.1-Sp1) and with the addition of pcDNA3.1 to the total amount of 1 μg. All transfections were performed in duplicate plates for three independent experiments.

Electrophoretic mobility shift and supershift assays

Following is a list of the synthesized ds oligonucleotides corresponding to bases −221 to −161 of the human GATA-1 IE promoter. Oligos for the wild-type CACCC box (WT) were 5′-ACTGCCCCACCCACTGAGGC-3′ (WT#1) and 5′-GCCTCAGTGGGTGGGGCAGT-3′ (WT#2). The CACCC mutated (MP) oligonucleotides were 5′-ACTGCCCCAAACACTGAGGC-3′ (MP#1) and 5′-GCCTCAGTGTTTGGGGCAGT-3′ (MP#2). The wild-type CACCC box is underlined and the mutated base pairs are shown in bold. Oligos for the wild-type CGCCC box were 5′-AAGCATGGGACCCGCCCCCTCCCCTGGACT-3′ and 5′-AGTCCAGGGGAGGGGGCGGGT-CCCATGCTT-3′. The CGCCC mutated CACCC oligonucleotides were 5′-AAGCATGGGACCCACCCCCTCCCCTGGACT-3′ and 5′-AGTCCAGGGGAGGGGGTGGGTCCCATGCTT-3′. The wild-type CGCCC box is underlined and the mutated base pairs are shown in bold. The following Sp1 consensus sequence (Sp1W) and mutated Sp1 (Sp1M) ds oligonucleotides were also synthesized: 5′-ATTCGATCGGGGCGGGGCGAGC-3′ (Sp1W#1), 5′-GCTCGCCCCGCCCCGATCGAAT-3′ (Sp1W#2), 5′-ATTCGATCGGTTCGGGGCGAGC-3′ (Sp1M#1) and 5′-GCTCGCCCCGAACCGATCGAAT-3′ (Sp1M#2). The wild-type Sp1 site is underlined and the mutated base pairs are shown in bold.

The oligonucleotides were mixed, boiled for 5 min and allowed to slowly cool and anneal specifically. The WT ds oligonucleotide was 5′ end labeled with [γ]-32P-ATP (Amersham, Piscataway, NJ, USA) using T4 kinase (Invitrogen) and purified over a Sephadex G-50 column. All reactions contained the following components: 2 μL of labeled probe (0.025 pmol), 5 μL of 5× binding buffer, 0 or 1 μL of K562 nuclear extract (prepared as previously described 30, 31); cold ds WT, MP, Sp1W and Sp1M oligonucleotides were added as competitors. Water was added to a final volume of 25 μL (if needed). All reactions were incubated for 30 min on ice, and the samples were resolved on a 4% non-denaturing 0.5× TBE polyacrylamide gel. The gel was dried and exposed to an X-ray film overnight 31. In competition and supershift assays, prior to the addition of a probe, the unlabeled competitors (10- to 100-fold molar excess) or 1 μL of Sp1 monoclonal antibody (sc420X, Santa Cruz) were pre-incubated with K562 nuclear extract on ice for 30 min.

Chromatin immuno-precipitation assay

5 × 106 cells (K562 cells, 293T cells or mouse adult erythrocytes from the anemic spleen) were crosslinked by adding formaldehyde to a cell culture medium to a final concentration of 1% and incubated for 10 min at room temperature. Approximately 1 × 106 cells were used for each immuno-precipitation. The crosslinking reaction was stopped by the addition of glycine to a final concentration of 0.125 M. Cells were washed twice with ice-cold phosphate-buffered saline, then re-suspended in cell lysis buffer (10 mM Tris, 10 mM NaCl, 0.2% NP-40 (pH 8.0)) containing protease inhibitors (100 ng/ml aprotinin and 100 ng/ml leupeptin) and 0.5 mM AEBSF and kept on ice for 15 min. Nuclei were pelleted by centrifugation at 5 000 rpm for 5 min at 4 °C. The pellets were re-suspended in nuclei lysis buffer (50 mM Tris, 10 mM EDTA, 1% SDS) containing protease inhibitors and AEBSF and kept on ice for 10 min. The nuclear lysates were sonicated on ice to an average chromatin length of 0.3-1.0 kb and centrifuged at 12 000 rpm for 10 min at 4 °C. The supernatants were incubated (with rotation) at 4 °C overnight with Protein A-agarose beads (KPL) in IP buffer (50 mM Tris-HCl (pH 8.1), 10 mM EDTA, 0.1% SDS, 0.5% deoxycholic acid and 500 mM LiCl) containing protease inhibitors and AEBSF. After removal of the Protein A-agarose beads, the pre-cleared lysates were used as soluble chromatin for ChIP.

Chromatin was incubated at 4 °C for 3 h with 2 μL of anti-Sp1 (sc-420X, Santa Cruz), anti-Sp3 (sc-28305, Santa Cruz), anti-EKLF (ab2483, Abcam), anti-acetylated histone H3 (06-599, Upstate) or anti-acetylated histone H4 (06-866, Upstate) antibodies. Control samples with non-specific IgG (sc-2027, Santa Cruz) and no antibody were included. Immuno-precipitates were recovered by incubation with Protein A-agarose beads previously pre-cleared in IP buffer for 2 h at 4 °C. The beads were washed twice with 500 μL of IP wash buffer 1 (20 mM Tris, 50 mM NaCl, 2 mM EDTA, 0.1% SDS, 1% Triton X-100 (pH 8.1), once with IP wash buffer 2 (10 mM Tris, 0.25 M LiCl, 1 mM EDTA, 1% Nonidet P-40, 1% desoxycholate (pH 8.1)) and twice with TE (pH 8.0). Immuno-precipitates were eluted from the beads with two passages of 200 μL of freshly prepared IP elution buffer (0.1 M NaHCO3, 1% SDS). NaCl was added to a final concentration of 0.3 M and incubated at 65 °C overnight to reverse the crosslinks. Proteinase K (to 0.25 mg/ml) was added and incubated at 50 °C for another hour. DNA was extracted with phenol, phenol:chloroform and chloroform, and then precipitated with ethanol. DNA was dissolved in 50 μL of TE for PCR. Primers were designed in the human or mouse GATA-1 IE promoter DNA adjacent to the CGCCC and CACCC boxes to amplify ChIP DNA. The following primers were used for ChIP PCR analysis of the sequence between −196 and −26 bp of the human GATA-1 IE promoter: 5′-AGC CAG GAA GAC GCA CAT AC-3′ for hGCFP and 5′-CTA GGC TCC CGT GGA CTG TA-3′ for hGCRP. The mouse GATA-1 promoter primers that were used were 5′-AGG TCC AAC CCA CAC ATA GC-3′ for mGCFP and 5′-TGG AGT CCA GTG TCC AGT CA-3′ for mGCRP.