Ets-dependent regulation of target gene expression during megakaryopoiesis.

Megakaryopoiesis is the process by which hematopoietic stem cells in the bone marrow differentiate into mature megakaryocytes. The expression of megakaryocytic genes during megakaryopoiesis is controlled by specific transcription factors. Fli-1 and GATA-1 transcription factors are required for development of megakaryocytes and promoter analysis has defined in vitro functional binding sites for these factors in several megakaryocytic genes, including GPIIb, GPIX, and C-MPL. Herein, we utilize chromatin immunoprecipitation to examine the presence of Ets-1, Fli-1, and GATA-1 on these promoters in vivo. Fli-1 and Ets-1 occupy the promoters of GPIIb, GPIX, and C-MPL genes in both Meg-01 and CMK11-5 cells. Whereas GPIIb is expressed in both Meg-01 and CMK11-5 cells, GPIX and C-MPL are only expressed in the more differentiated CMK11-5 cells. Thus, in vivo occupancy by an Ets factor is not sufficient to promote transcription of some megakaryocytic genes. GATA-1 and Fli-1 are both expressed in CMK11-5 cells and co-occupy the GPIX and C-MPL promoters. Transcription of all three megakaryocytic genes is correlated with the presence of acetylated histone H3 and phosphorylated RNA polymerase II on their promoters. We also show that exogenous expression of GATA-1 in Meg-01 cells leads to the expression of endogenous c-mpl and gpIX mRNA. Whereas GPIIb, GPIX, and C-MPL are direct target genes for Fli-1, both Fli-1 and GATA-1 are required for formation of an active transcriptional complex on the C-MPL and GPIX promoters in vivo. In contrast, GPIIb expression appears to be independent of GATA-1 in Meg-01 cells.

Ets family members each contain a conserved winged helixloop-helix DNA binding (ETS) domain that allows recognition of purine-rich DNA sequences with a core GGA(A/T) consensus, designated EBS 1 (Ets binding sequence) (1)(2)(3). These transcription factors have critical roles in the transcriptional control of genes important in development, morphogenesis, proliferation, and angiogenesis. Ets factors can function as either positive or negative transcriptional regulators (4,5) and additional binding of other transcription factors to cis-elements located near EBSs contributes to regulated transcription of specific target genes. Thus, in addition to binding to DNA, Ets transcription factors participate in protein interactions that affect their functions (6,7).
Fli-1, a member of the Ets gene family of transcription factors, performs functions critical for normal development and oncogenesis (for a review, see Ref. 8). Fli-1 is preferentially expressed in cells of hematopoietic lineages and vascular endothelial cells, and has been shown to transcriptionally activate genes, including the stem cell leukemia gene (9), tenascin (10), the stress response gene GADD153 (11), the anti-apoptotic gene Bcl-2 (12) and several megakaryocytic specific genes (see below). Fli-1 also forms ternary complexes through interaction with SRF to bind to SRE elements of fos and Egr-1 promoters (13,14). Fli-1 protein interaction with other regulatory proteins modulates their activities (15,16) and the Ets family member TEL has been shown to inhibit Fli-1 transcriptional and biological activity (17)(18)(19). In addition, we (20,21) and others (22) have shown that Fli-1 binding can result in transcriptional repression, dependent on promoter and cell context.
Overexpression of Fli-1 in transgenic mice results in death from progressive immunological renal disease associated with an increased number of autoreactive T-and B-lymphocytes (23). These mice also have an increased number of mature B-cells, which have a reduced activation-induced apoptotic response compared with B-cells from wild type animals (23). The possible role of Fli-1 in autoimmunity is further supported by our observation of elevated expression of Fli-1 mRNA in lymphocytes from patients with systemic lupus erythematosus (24). In addition, it has been demonstrated that Fli-1 expression promotes a megakaryocytic phenotype in K562 cells and increases the expression of megakaryocytic genes (25)(26)(27). Viral integration and insertional activation of Fli-1 is associated with hematological cancers, including erythroleukemia by Friend-MuLV (28), granulocytic leukemia induced by the Graffi virus (29), primitive stem cell tumors by the 10A1 isolate of MuLV (30), and non-T, non-B lymphomas by the Cas-Br virus (31). Taken together, these reports suggest that Fli-1 plays a crucial role in normal hematopoietic differentiation and lineage selection. To clarify the physiological role of Fli-1 in hematopoiesis, we (32) and others (33) generated mice with the targeted disruption of Fli-1. The Fli-1 homozygous mutant (Fli-1Ϫ/Ϫ) embryos showed hemorrhage from the dorsal aorta into the lumen of the neural tube and the ventricles of the brain beginning on embryonic day 11.0 (E11.0) and were dead on or before day E12.0. In addition, severe dysmegakaryopoiesis (33,34) and vascular defects (33) were found. Analysis of cultured cells from day 10.0 embryos demonstrated absence of megakaryocytes and aberrant red blood cell development (34).
Over 200 Ets target genes have been identified by the presence of functional EBS (5), based upon transient transfection studies. However, few of these have been characterized on chromatin templates in vivo. To better understand how Ets and GATA-1 proteins regulate the transcription of megakaryocytic target genes, we examined the in vivo binding of Fli-1, Ets-1, and GATA-1 proteins to the endogenous GPIIb, GPIX, and C-MPL promoters. The differential expression of these genes in megakaryocytic cell lines allowed us to correlate Ets factor occupancy, histone acetylation status, and/or presence of phosphorylated RNA polymerase II with transcription from these promoters. Co-expression of GATA-1 and Fli-1 is critical for expression of endogenous C-MPL and GPIX genes. We also demonstrate for the first time that simultaneous in vivo promoter occupancy by these factors is directly correlated with the transcriptional status of specific megakaryocytic genes.

EXPERIMENTAL PROCEDURES
Megakaryocytic Cell Lines-Meg-01 cells (48) were obtained from Dr. Steven Lentz (University of Iowa, Iowa). CMK11-5 cells (49,50) were provided by Dr. Nicholas Greco (The American Red Cross, Rockville, MD) with permission from Dr. T. Sato (Chiba University, Chiba, Japan), and Dr. Masao Kobayashi (Mochida Pharmaceutical Co., Tokyo, Japan). These two cell lines were maintained in RPMI 1640 media plus 10% fetal bovine serum, according to the ATCC protocol for Meg-01 cells. Briefly, cells were subcultured by dilution into fresh media at a split ratio of 1:2 to 1:3, 2 to 3 times weekly and maintained between 1 ϫ 10 5 and 5 ϫ 10 5 cells/ml. For Meg-01 cells, adherent cells (less than 10% of the population) were scraped into the media prior to subculture. All cell lines were propagated at 37°C in an atmosphere containing 5% CO 2.
Plasmid Constructs-The pXM GATA-1 expression vector was kindly provided by Dr. S. Orkin (Harvard Medical School, Boston, MA) (54,55). Mouse GATA-1 cDNA was excised by XhoI digestion of pXM GATA-1 and cloned into the eukaryotic expression vector, pSGNeoKS (modification of pSG5 (Stratagene, La Jolla, CA) to contain a neomycin/G418resistance cassette and the multiple cloning site from pBluescript II KS vector) at the XhoI restriction site to generate pSGNeo GATA-1. Orientation and sequence were verified by analysis on an ABI 373 automated sequencer (Applied Biosystems, Foster City, CA).
RNA Preparation and Northern Blot Analysis-Total RNA was isolated using the TRIzol reagent (Invitrogen) according to the manufacturer's instructions. Gel electrophoresis of RNA and transfer to nylon membrane (Duralon-UV membrane, Stratagene) was performed according to Sambrook et al. (56). Prehybridization and hybridization were performed at 65°C with QuikHyb Hybridization Solution (Stratagene) according to the manufacturer's specifications. Membranes were washed once at room temperature in 2ϫ SSPE, 0.2% SDS for 15 min, once at room temperature in 1ϫ SSPE, 0.2% SDS for 15 min, and then twice at 55°C in 0.2ϫ SSPE, 0.2% SDS for 30 min.
Probes for Northern and Southern Blots-Each probe was labeled with [␣-32 P]dCTP using Rediprime II (Amersham Biosciences). Ets-1 cDNA was excised by ApaI-SmaI digestion of pSG5-Ets-1 (10). GATA-1 cDNA was excised by XhoI digestion of pXM GATA-1 (54,55). Fli-1 was prepared by PCR using the previously described human Fli-1 cDNA vector (57) as a template with the primers and conditions provided in Table I. For the c-mpl, gpIIb, and gpIX probes, a first strand cDNA generated from CMK11-5 mRNA was used in PCR utilizing the conditions provided in Table I. The resultant PCR products were cloned into the pCR2.1-TOPO vector and sequence-verified clones served as templates in PCR to generate the DNA for labeling.
ChIP Assays-ChIP assays were adapted from Boyd and Farnham (58) with modifications. Cells (10 7 ) were resuspended in phosphatebuffered saline and cross-linked with 1% formaldehyde at room temperature for 15 min and the reaction was stopped by the addition of glycine to a final concentration of 200 mM. Cells were then washed with ice-cold phosphate-buffered saline and lysed in 1 ml of ChIP cell lysis buffer (5 mM PIPES, pH 8.0, 85 mM KCl, 0.5% Nonidet P-40, 25 l protease inhibitor mixture, Sigma). Nuclei were obtained by centrifugation at 3,500 ϫ g and lysed in ChIP nuclei lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl, pH 8.1, 50 l of protease inhibitor mixture, Sigma). DNA was sheared by sonication to yield an average fragment size of 500 bp and centrifuged for 15 min at 14,000 ϫ g. The supernatants were stored at Ϫ70°C. For immunoprecipitation, supernatants were diluted 5-fold in IP buffer (16.7 mM Tris-HCl, pH 8.1, 1.1% Triton X-100, 1.2 mM EDTA, 167 mM NaCl, 0.01% SDS, protease inhibitors) and pre-cleared with 80 l of a 50% protein A-Sepharose slurry (equilibrated in 50 mM Tris-HCl, pH 7.0, 20 g/ml salmon sperm DNA, 0.5 mg/ml bovine serum albumin, 100 g/ml preimmune IgG) for 2 h at 4°C. After centrifugation at 14,000 ϫ g for 2 min, specific antibodies (10 g) were added to the supernatants. Immunocomplexes were formed overnight at 4°C and collected with 60 l of 50% protein A-Sepharose (equilibrated as above, except no IgG was added) for 1 h at 4°C. Beads were washed for 5 min on ice in buffer A (20 mM Tris-HCl, pH 8.1, 1% Triton X-100, 2 mM EDTA, 150 mM NaCl, 0.1% SDS), buffer B (buffer A with 500 mM NaCl), buffer C (10 mM Tris-HCl, pH 8.1, 0.25 M LiCl, 1% Nonidet P-40, 1% sodium deoxycholate, 1 mM EDTA), buffer D (buffer C with 500 mM LiCl) and twice in TE buffer (10 mM Tris-HCl, pH 8, 1 mM EDTA). Immunocomplexes were eluted off the beads with 150 l of 1% SDS, 0.1 M NaHCO 3 and cross-links were reversed by incubation for 4 h at 65°C after addition of NaCl to the final concentration of 0.3 M. Proteins were digested with proteinase K (40 g/ml) for 1 h at 50°C. DNA was purified using the QIAquick PCR Purification Kit (Qiagen, Valencia, CA). Of the total yield of 30 l of PCR template, 4 l was used in each PCR using AmpliTaq Gold DNA polymerase (Applied Biosystems). Gene-specific primer sequences and predicted amplicon sizes are summarized in Table I. In all experiments, the specificity of the ChIP reaction was monitored by PCR utilizing primers derived from the coding region of the C-MPL gene.
Sequential ChIP Assays-After primary immunoprecipitation, the cross-linked immunocomplexes were eluted from the beads by incubation in elution buffer at room temperature, and the eluate then diluted 1:10 in IP buffer, followed by re-immunoprecipitation with the second antibody. All the subsequent steps were carried out as described above.
DNA Transfection and Selection of Stable Clones-Meg-01 cells were transfected with pSGNeo GATA-1 or pSGNeo control vector using the FuGENE 6 Transfection Reagent (Roche Molecular Biochemicals) according to the manufacturer's instructions. After transfection, cells were allowed to recover for 16 h, and G418 (Calbiochem, San Diego, CA) was added to a final concentration of 0.8 mM to allow selection of clones with stable expression. After 3 weeks, pools of selected cells were used for RNA isolation.
Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR)-Total RNA was isolated from pools of transfected Meg-01 cells using TRIzol reagent (Invitrogen). RNA (1 g) was reverse transcribed in first strand synthesis buffer according to the manufacturer's recommendations (Invitrogen). 14 ng of the cDNA products were used as template in subsequent PCR using the primers and conditions listed in Table I. The products were electrophoretically separated on a 1.5% agarose gel. After ethidium bromide staining of the molecular weight marker, the gel was used for Southern blotting and hybridization.
Image Analysis-The Kodak Digital Science 1D version 2.0.2 software running on a PowerMac G3 machine was used to calculate the mean intensity of each ethidium bromide or SYBR Green-stained PCR band or the sum intensity of bands on autoradiograms.

Expression of Transcription Factors and Megakaryocytic
Marker Genes-The Meg-01 and CMK11-5 cell lines were selected as model systems representing early and late stages of megakaryocytic differentiation, respectively. Meg-01 is a human megakaryoblastic leukemia cell line derived from bone marrow cells of a patient in megakaryoblastic crisis of chronic myelogeneous leukemia. Meg-01 is characterized by a mixed population of adherent and floating cells, does not display B or T lymphocyte markers, whereas platelet glycoprotein gpIIb/ IIIa is present on their surface (48). CMK is a cell line of the megakaryocyte/platelet lineage (49,60) and CMK11-5 is its more differentiated subclone. CMK11-5 cells are larger than CMK cells, and contain multinucleated giant cells. CMK11-5 cells express the platelet antigens gpIb, gpIIb/IIIa, and gpIV and shed platelet-like particles (49,50).
To characterize molecular events during megakaryopoiesis, we first determined the expression level of two Ets transcription factors, Ets-1 and Fli-1, in these cells. Fli-1 mRNA and protein levels were not significantly lower in Meg-01 cells, compared with CMK11-5 (91 and 83%, respectively). Whereas Ets-1 mRNA in Meg-01 cells is 15% of that detected in CMK11-5 cells, Ets-1 protein levels are more similar (Fig. 1). Next we analyzed the GATA-1 expression in the two cell lines. CMK11-5 cells contain six times more GATA-1 mRNA than detected in Meg-01 cells; more importantly, no GATA-1 protein is detected in Meg-01 cells, even after prolonged film exposure (data not shown).
We next assessed the expression of three megakaryocytic genes, GPIIb, GPIX, and C-MPL, in Meg-01 and CMK11-5 cells. As shown in Fig. 1A, Northern blot analysis of RNA prepared from Meg-01 and CMK11-5 cells demonstrates that gpIIb is expressed in both cell lines at essentially equivalent levels. In contrast, whereas gpIX and c-mpl mRNA are each expressed in both cells, the mRNA level found in Meg-01 cells is at least 1 order of magnitude less than that detected in CK11-5 cells. This expression pattern is consistent with the model that CMK11-5 cells are further differentiated along the megakaryocytic lineage than Meg-01 Cells.
Fli-1 and Ets-1 Bind to Endogenous Megakaryocytic Promoters in Meg-01 and CMK11-5 Cell Lines-We next examined whether Fli-1 and/or Ets-1 occupies the GPIIb, GPIX, and C-MPL promoters in vivo, using Meg-01 and CMK11-5 cells for ChIP assays. After immunoprecipitation with antibodies against Fli-1 or Ets-1, enrichment of the endogenous promoter fragments in each sample was monitored by PCR amplification using primers specific for each promoter (Fig. 2). As a positive control, 5% of the input chromatin was used in a PCR (Fig. 2,  input). As a negative control, a reaction lacking precipitating antibody was also performed (Fig. 2, no Ab). Both Ets-1 and Fli-1 occupy the Ϫ237 to ϩ35 and Ϫ1133 to Ϫ951 regions of the GPIIb promoter in both cell lines. However, only Fli-1 was bound to the Ϫ625 to Ϫ404 region and this was observed only in CMK11-5 cells. With the C-MPL promoter, both Ets-1 and Fli-1 occupy the Ϫ108 to ϩ122 and the Ϫ715 to Ϫ505 region and no difference was observed between the two cell lines. At the GPIX promoter, Ets-1 and Fli-1 were found only at the Ϫ147 to ϩ98 region in both cell lines. In summary, we found that both Fli-1 and Ets-1 bind, simultaneously or individually, to EBSs present in these megakaryocytic promoters in both Meg-01 and CMK11-5 cells (Fig. 6).

Fli-1 and GATA-1 Proteins Are Present Simultaneously on the C-MPL and GPIX Promoters in CMK11-5 Cells-Based on
previous studies demonstrating that Ets and GATA-1 proteins bind and transcriptionally activate the GPIIb (36, 37), GPIX (27, 40 -42), and C-MPL (43,44) promoters in vitro and the observation that Fli-1 binds to the GPIIb, GPIX, and C-MPL promoters in vivo (Fig. 2), we next determined whether GATA-1 is bound to these promoters in vivo. ChIP analysis was performed using Meg-01 and CMK11-5 cell lines. These experiments revealed that GATA-1 is bound to the C-MPL and GPIX promoters (Ϫ108 to ϩ122 and Ϫ147 to ϩ98 regions, respectively) in CMK11-5 cells (Fig. 3). Consistent with low GATA-1 protein level, these regions were not enriched in ChIP fragments from Meg-01 cells. Furthermore, we did not detect GATA-1 at the GPIIb promoter in either cell line. Other regions of the analyzed promoters were not detected in the GATA-1 ChIP fragments from either cell line.
To investigate whether both Fli-1 and GATA-1 co-occupy these promoters in vivo, we performed sequential ChIP analysis. We first used Meg-01 cells and anti-Fli-1 antibody in both the first and second round of immunoprecipitation. The C-MPL promoter region was successfully detected in the ChIP-enriched DNA fragments using PCR (data not shown). Next, we used chromatin prepared from both Meg-01 and CMK11-5 cells

FIG. 2. Both Fli-1 and Ets-1 bind to the promoters of the megakaryocytic target genes in vivo.
PCR was performed on chromatin fragments enriched by immunoprecipitation with or without the indicated antibodies. The antibodies used in these experiments are shown above each lane. Input represents 5% total cross-linked, reversed chromatin before immunoprecipitation. To show the efficiency of washing, no antibody was added to one sample, marked as no Ab. Primers specific for the promoter regions of the genes indicated on the left were used in PCR.

FIG. 3. Fli-1 and GATA-1 co-occupy the C-MPL and GPIX promoters in vivo in CMK11-5 cells.
ChIP experiments were carried out to investigate whether Fli-1 and GATA-1 bind to the C-MPL, GPIIb, or GPIX promoter regions in the indicated cell lines. Input represents 5% total cross-linked, reversed chromatin before immunoprecipitation. In lanes marked Fli-1/GATA-1, anti-Fli-1 antibody was used in the first round and anti-GATA-1 antibody was used in the second round of immunoprecipitation. As a negative control, isotype matched pre-immune IgG was added to one sample, marked as IgG.
in the sequential ChIP assays, employing anti-Fli-1 antibody in the first round and anti-GATA-1 antibody in the second round of immunoprecipitation. The Ϫ108 to ϩ122 C-MPL promoter region was detected in the ChIP-enriched fragments from CMK11-5 cells, but not from Meg-01 cells (Fig. 3). This is consistent with our results that GATA-1 is not bound to the C-MPL promoter in Meg-01 cells (Fig. 3). Moreover, these results show that Fli-1 and GATA-1 proteins are simultaneously present on this C-MPL promoter region in CMK11-5 cells in vivo. Similarly, sequential ChIP analysis demonstrated simultaneous occupancy by Fli-1 and GATA-1 proteins on the GPIX promoter (Ϫ147 to ϩ98 region) in vivo in CMK11-5 cells. In contrast, we were not able to detect GPIIb promoter fragments in ChIP-enriched chromatin from either cell line.
Histone Acetylation and RNA Polymerase II Phosphorylation on Fli-1 Target Gene Promoters-We have shown above that both Ets-1 and Fli-1 occupy the GPIIb, GPIX, and C-MPL promoters in Meg-01 cells. However, in these cells the C-MPL and GPIX mRNA levels are less than 10% of that detected in CMK11-5 cells. Additional criteria for transcriptionally active Fli-1 target genes, independent of GATA-1 status, were investigated. One hallmark of many transcriptionally active genes is the presence of acetylated histones at the respective promoters (61). Therefore, we analyzed the acetylation status of histone H3 at Fli-1 occupied megakaryocytic promoters (C-MPL, GPIIb, and GPIX) by performing sequential ChIP assays using anti-Fli-1 and anti-acetylated-H3 antibodies. ChIP enriched C-MPL-, GPIIb-, and GPIX-derived fragments were detected in chromatin prepared from CMK11-5 cells, indicating that Fli-1 and acetylated histone H3 are both present on these transcribed promoters (Fig. 4A). Similarly, ChIP demonstrated that Fli-1 and acetylated H3 are both present on the GPIIb promoter in Meg-01 cells. The ratio of band intensity of the GPIIb ChIP fragments for Meg-01 and CMK11-5 was 0.78, which correlates with the relative mRNA expression between the two cell lines (81%, Fig. 1). Consistent with reduced GPIX and C-MPL expression in Meg-01 cells, ChIP analyses demonstrated that acetylated H3 on these Fli-1 occupied promoters in Meg-01 cells is significantly reduced (17% of control compared with 110% of control) or absent (0%), respectively (Fig. 4A).
The presence of phosphorylated RNA polymerase II is another property of most actively transcribed genes (62). We therefore analyzed the same megakaryocytic promoters using sequential ChIP experiments utilizing anti-Fli-1 and antiphosphorylated RNA polymerase II antibodies. Our results show that the C-MPL and GPIX promoters were detected in the ChIP-enriched fragments from CMK11-5, but not from Meg-01 cells (Fig. 4B). In contrast, the GPIIb promoter fragments were found to be enriched by ChIP using chromatin prepared from both Meg-01 or CMK11-5 cells. Moreover, the relative band intensity for GPIIb ChIP fragments for Meg-01 and CMK11-5 (0.69) correlated with the observed mRNA ratio (Fig. 1). Collectively, these ChIP results are consistent with the differential expression of these three genes and demonstrate the utility of sequential ChIP for identification of transcriptionally active Fli-1 target genes.
Exogenous Expression of GATA-1 in Meg-01 Cells Leads to Increased Expression of the C-MPL and GPIX Genes-Because GATA-1 has been shown to be a key regulator of megakaryocytic gene expression, we next investigated whether expression of GATA-1 in Meg-01 cells would be sufficient to promote C-MPL and GPIX expression. We transfected Meg-01 cells with the GATA-1 expression plasmid, enriched for stable GATA-1 transfected Meg-01 cells and analyzed the c-mpl, gpIX, gpIIb, and GATA-1 mRNA levels by RT-PCR. Total RNA was prepared from GATA-1 or pSG5/neo-transfected cells and used for FIG. 4. ChIP analysis of histone modification status and occupancy by phosphorylated RNA polymerase II on Fli-1 occupied megakaryocytic promoters. A, histone H3 is acetylated at the Fli-1 occupied transcriptionally active promoters. Anti-Fli-1 and anti-acetylated-H3 antibodies were used in sequential ChIP assays to precipitate fragments representing promoters of Fli-1-regulated active genes. Primers specific for C-MPL, GPIIb, and GPIX promoter regions were used in PCR. Input and No Ab controls were performed as described above (Fig. 2). Numbers below bands indicate percent intensity relative to input. B, phosphorylated RNA polymerase II is present on the Fli-1occupied active promoters. Anti-Fli-1 and anti-phosphorylated RNA polymerase II antibodies were used in sequential ChIP experiments. input and no Ab controls were as described above. the preparation of cDNA. PCR was performed using specific primers for c-mpl, gpIX, gpIIb, GATA-1, and S26. Relative to S26, significantly higher expression of c-mpl and gpIX mRNA was detected in the cells with exogenous expression of GATA-1 (Fig. 5). Thus, exogenous expression of GATA-1 mRNA (over 3-fold) led to a significant increase of c-mpl and gpIX mRNA transcripts in the transfected cells, consistent with a direct role for regulation of these promoters by this transcription factor. In contrast, GATA-1 expression did not significantly alter gpIIb expression.

DISCUSSION
Specific combinations of cell-restricted and ubiquitous transcription factors control the maturation of pluripotent hematopoietic stem cells into multiple differentiated cell lineages. During megakaryopoiesis, the expression of megakaryocytic genes, such as GPIIb, GPIX, and C-MPL, is strictly regulated. An early marker of the megakaryocytic lineage, gpIIb (CD41) (63,64), combines with the more broadly expressed ␤ 3 (gpIIIa, CD61) subunit to form the gpIIb/IIIa heterodimer receptor that regulates cell adhesion and functions in aggregation of activated platelets by binding fibrinogen, fibronectin, vitronectin, collagen, and von Willebrand factor (63,64). gpIX is required for assembly of the gpIb⅐V⅐IX complex (65) that provides the binding site for the von Willebrand receptor and is critical for initial adhesion of platelets to damaged blood vessels. The thrombopoietin receptor, c-mpl, is required for megakaryocytic differentiation and mice lacking either the c-mpl receptor or its ligand, thrombopoietin, have reduced megakaryocytes and are severely thrombocytopenic (66,67). Consistent with the importance of these genes in megakaryopoiesis and the dysmegakaryopoiesis observed in Fli-1 mutant mice, c-mpl (34) and gpIX (33) mRNA levels are significantly reduced in Fli-1 mutant embryos. Ets factors and GATA-1 functionally regulate the transcription of these (27, 36, 37, 40 -44) and other (26,27,35,38,39,45)

megakaryocytic genes in vitro.
To better understand the molecular basis for Ets-1, Fli-1, and GATA-1 target gene regulation, we utilized Meg-01 and CMK11-5 cells. Whereas Ets-1 and Fli-1 proteins were detected in both cells, GATA-1 protein was restricted to the more differentiated CMK11-5 cell line (Fig. 1). Although Northern blot analysis revealed no significant difference in gpIIb expression level between these two cell lines, gpIX and c-mpl were expressed at a significantly lower level (7%) in Meg-01 cells.
We next examined the in vivo binding of Ets factors and GATA-1 to these megakaryocytic genes (Fig. 6). Our ChIP studies demonstrate that Fli-1 and Ets-1 occupy the GPIIb, GPIX, and C-MPL proximal promoter regions in vivo in both Meg-01 and CMK11-5 cells, consistent with previous in vitro studies (36,40,44). The only Ets-specific and cell-specific in vivo interaction was found on the GPIIb promoter (Ϫ625 to Ϫ404). All other cell-specific interactions were directly correlated with GATA-1 expression. Furthermore, we also observed Fli-1 and Ets-1 occupancy of the GPIIb promoter at the Ϫ1133 to Ϫ951 region and of the C-MPL promoter at the Ϫ714 to Ϫ505 region. Although factor-binding sites within these regions were not identified by previous in vitro studies, potential EBSs were predicted by computer analysis. In addition, we determined that potential Ets and GATA-1 sites of C-MPL previously defined by in vitro studies are not bound in vivo by Ets-1, Fli-1, or GATA-1. The chromatin structure around these sites may account for these disparate observations. Alternatively, other Ets or GATA family members may bind to these regions. Overall, we conclude that Fli-1 and Ets-1 may be part of the complex that mediates transcription of these genes in vivo. In addition, based upon the absence of GATA-1 in Meg-01 cells, our results indicate that Ets-1 and Fli-1 are able to bind to these megakaryocytic promoters independently of GATA-1.
Northern analysis demonstrated that gpIIb mRNA is present at similar levels (81% relative ratio) in Meg-01 and CMK11-5 cells, independent of GATA-1 expression. Thus, although GATA-1 and Fli-1 have been shown to synergistically activate the GPIIb promoter in vitro (37), GPIIb transcription appears to be independent of GATA-1 in Meg-01 cells. In contrast, both gpIX and c-mpl mRNA are highly abundant in CMK11-5 cells, with only 7% relative expression level in Meg-01 cells, suggesting that transcription of these genes requires GATA-1 and/or other factors that are differentially expressed between these two cell lines. Either GPIIb transcriptional regulation is celltype specific, or it is possible that the GPIIb promoter has a higher affinity for the limited amount of GATA-1 protein ex- pressed in Meg-01 cells, compared with that for the GPIX and C-MPL promoters. However, the failure to enrich for GPIIb promoter fragments by our ChIP experiments argues against this model. Furthermore, the level of gpIIb mRNA was not significantly affected by exogenous expression of GATA-1 in Meg-01 cells (Fig. 5). It is likely, however, that gpIIb transcription requires additional factors. Cell-type specific transcription of gpIIb is also controlled by silencer elements at Ϫ120 to Ϫ116 present near the proximal EBS in both the human (68,69) and rat (70) promoters. Whether Fli-1 presence on the promoter modulates the binding of a silencer complex to this element remains to be directly evaluated.
We did not detect significant GATA-1 bound to the GPIX or C-MPL promoter in Meg-01 cells in vivo, whereas this was readily observed in CMK11-5 cells (Fig. 3). Although Ets-1 and Fli-1 bind these endogenous promoters, binding of either Ets factor is not sufficient for gene transcription in Meg-01 cells. Sequential ChIP analysis demonstrated that Fli-1 and GATA-1 co-occupy the C-MPL and GPIX promoters in vivo. Such cooccupancy was not observed for Ets-1 and GATA-1 (data not shown). Furthermore, although the C-MPL promoter contains several potential EBSs, we did not detect Ets-1 and Fli-1 cooccupancy of the C-MPL promoter as assessed by sequential ChIP experiments (data not shown). Transient transfection with either Ets-1 or Fli-1 with GATA-1 activates transcription of the C-MPL promoter in non-hematopoietic cells (43,44). Our in vitro analysis confirmed these findings in HeLa cells and demonstrated that Fli-1 and GATA-1 synergistically activate C-MPL (data not shown). These results support the model that Fli-1 and GATA-1 interact and mediate synergistic expression of specific megakaryocyte-specific genes (27).
Promoter occupancy by Fli-1 or Ets-1 alone is not correlated with transcription of C-MPL or GPIX. The presence of acetylated histone H3 on the Fli-1 occupied C-MPL and GPIX promoters was directly correlated with their expression status. Relevant to these observations, GATA-1 has been reported to induce histone H3 acetylation at the ␤-globin promoter (71,72). Furthermore, GATA-1 associates with histone acetyltransferases and CREB protein (73), supporting a role in chromatin remodeling necessary to enhance transcription. Also, GATA-1 has been associated with RNA polymerase II recruitment to the ␤-globin promoter (74). Consistent with the expression status of the C-MPL and GPIX genes in CMK11-5 and Meg-01 cells, we further demonstrate that phosphorylated RNA polymerase II is present on these promoters in only CMK11-5 cells. The relative mRNA expression level (81%) of GPIIb in Meg-01 and CMK11-5 cells is correlated with that of Fli-1/acetylated H3 occupied DNA (79%) and Fli-1/phosphorylated RNA polymerase II occupied DNA (70%).
Exogenous expression of GATA-1 in Meg-01 cells resulted in the expression of the C-MPL and GPIX genes (Fig. 5), demonstrating a critical role for GATA-1 for the expression of these genes. This result complements the previous observation that exogenous expression of Fli-1 in K562 cells, which expresses GATA-1 endogenously, results in expression of megakaryocytic genes, GPIIb, GPVI, GPIb␣, and GPIX (25)(26)(27). Furthermore, gpVI mRNA was only detected in megakaryocytic cell lines expressing both Fli-1 and GATA-1 (26). Taken together, it is clear that both Fli-1 and GATA-1 are critical for the expression of many, but not all, megakaryocytic genes. However, other transcription factors, including FOG-1 (Friend of GATA-1) (37), MafB/Kreisler (75), and TEL (19) also regulate expression of these genes.
Models for Ets factor and GATA-1-mediated transcriptional control have recently been proposed (27,37,71). Although Fli-1 and GATA-1 expression is not limited to megakaryocytes, we show that GATA-1 is important for transcriptional activation of specific Fli-1 target genes in megakaryocytes. Our results demonstrate that both Ets-1 and Fli-1 are able to bind to target promoters independently of GATA-1. It has been shown that Fli-1 and GATA-1 proteins directly interact via their Ets and zinc finger domains, respectively, and the in vitro binding of GATA-1 to DNA is markedly increased in the presence of Fli-1 (27). When GATA-1 and Fli-1 are both present on a promoter, Fli-1 converts FOG-1 into a co-activator (37). GATA-1 also recruits other nuclear factor(s) with histone acetyltransferase activity (73). Histone acetylation-dependent chromatin remodeling allows other nuclear factors to bind, enhancing phosphorylated RNA polymerase II-mediated transcription. Further studies will be required to identify the precise molecular mechanisms that regulate each of these events.