Published online before print June 16, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
binds and activates the promoter of the PU.1 gene
Division of Pediatric Oncology, Johns Hopkins University, Baltimore, Maryland
1Correspondence: Johns Hopkins University, Cancer Research Bldg., Rm. 253, 1650 Orleans St., Baltimore, MD 21231. E-mail: afriedm2{at}jhem.jhmi.edu
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and PU.1 are required for myelopoiesis. Examination of the murine PU.1 promoter revealed several potential C/EBP-binding sites. Gel-shift assay demonstrated that C/EBP expressed in 293T cells bound the site centered at –68 most potently. C/EBP from 32D cl3 myeloid cell nuclear extracts also bound this site strongly, and endogenous C/EBPß did so to a lesser extent, whereas these C/EBP isoforms bound the neutrophil elastase promoter with equal affinity. The –68 site in the murine PU.1 promoter is conserved in the human PU.1 promoter. Mutation of the –68 C/EBP-binding site in a -85/+152 promoter segment linked to the luciferase cDNA reduced promoter activity fourfold in 293T cells in the presence of cotransfected C/EBP and twofold in 32D cl3 myeloid cells. Induction of endogenous PU.1 RNA by C/EBP-estradiol receptor (ER) in the presence of cycloheximide is obviated by mutation of the C/EBP DNA-binding domain, and chromosomal immunoprecipitation demonstrated specific interaction of C/EBP and C/EBP-ER with the PU.1 promoter. Finally, PU.1 RNA is reduced several-fold in immortalized C/EBP (-/-) compared with (+/-) cells. Together, these findings indicate that C/EBP binds and activates the endogenous PU.1 gene in myeloid cells. Induction of PU.1 by C/EBP may account for increased levels of PU.1 in myeloid as compared with B lymphoid cells and in this way, may contribute to the specification of myeloid progenitors.
Key Words: hematopoiesis • differentiation • transcription
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
is the founding member of a family of transcription factors [1
] and binds DNA via its basic region after dimerizing via its leucine zipper domain [2
]. Once bound to DNA, C/EBP activates transcription via its N-terminal transactivation domains [3
, 4
]. C/EBP is expressed in several lineages, including hepatocytes, adipocytes, pneumocytes, and mammary and intestinal epithelia [5
6
7
8
]. Within hematopoiesis, C/EBP, C/EBPß, C/EBP
, and C/EBP
are expressed largely in the granulocytic and monocytic (myeloid) lineages, and expression of C/EBP predominates in more immature cells [9
10
11
]. Neonatal C/EBP (-/-) mice lack granulocytes but retain monocytes; older C/EBP (-/-) mice lack granulocytes and monocytes as a result of a defect in the formation of the granulocyte-monocyte progenitor (GMP); mice lacking C/EBPß and C/EBP apparently develop all the hematopoietic lineages but have functional macrophage defects; and mice lacking C/EBP do not develop mature granulocytes [12
13
14
15
16
17
18
]. Functional redundancy amongst C/EBPs in myelopoiesis is suggested by the finding that a dominant inhibitor of the entire family greatly reduces the generation of granulocytic and monocytic but not erythroid progenitor colonies in vitro [19
]. The central role of C/EBP in granulopoiesis is further suggested by the finding that exogenous C/EBP induces the U937, HL-60, or 32D cl3 myeloblastic cell lines to differentiate to neutrophils [10
, 20
].
PU.1 is a member of the Ets family of transcription factors expressed specifically in hematopoietic cells [21
]. PU.1 (-/-) mice lack B-lymphoid cells and monocytes and have greatly reduced numbers of neutrophils [22
, 23
]. Myeloid cells have higher levels of PU.1 than lymphoid cells, and introduction of PU.1 into lymphoid progenitors redirects them to the monocytic lineage [24
]. We found that activation of a C/EBP-estradiol receptor (ER) fusion protein with estradiol rapidly induced PU.1 mRNA expression in 32D cl3 myeloid and Ba/F3 lymphoid cells, even in the presence of cycloheximide, suggesting that C/EBP directly activates the PU.1 gene [20
]. We now provide evidence indicating that C/EBP binds and activates the PU.1 promoter and discuss the implications of this finding for the regulation of myelopoiesis.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
cells [25
, 26
] were maintained in phenol red-free Iscove’s modified Dulbecco’s medium (IMDM) with 10% heat-inactivated fetal bovine serum (HI-FBS) and 1 ng/ml murine interleukin-3 (IL-3; Peprotech, Rocky Hill, NJ). For transfer to granulocyte-colony stimulating factor (G-CSF), 32D cl3 cells were washed twice with phosphate-buffered saline (PBS) and were then placed in IMDM with 10% HI-FBS and 20 ng/ml G-CSF (Amgen, Thousand Oaks, CA). Stable 32D cl3 lines were maintained in 2 µg/ml puromycin. Estradiol was added to 1 µM from a 1000x stock in ethanol, and cycloheximide was used at 50 µg/ml. 32DPKC cells were transiently transfected using a diethylaminoethyl-dextran procedure at a ratio of 10 µg DNA per 5 x 106 cells [27
]. 293T cells were transfected using 3 µl Lipofectamine 2000 (Invitrogen, Carlsbad, CA) and 0.8 µg DNA per 8 x 104 cells in a 35-mm dish. Plasmid cytomegalovirus (pCMV)-ß-galactosidase (ßGal) was included in each transfection as an internal control. Forty-eight hours later, luciferase and ßGal activities were assessed in cell extracts as described [27
]. To express C/EBP for gel-shift assay, 5 µg murine sarcoma virus promoter/enhancer (pMSV)-C/EBP was transfected per 5 x 105 cells 293T on a 100-mm dish using 15 µl Lipofectamine. Immortalized C/EBP (-/-) and (+/-) cells, kindly provided by Steven J. Collins (Fred Hutchinson Cancer Research Center, Seattle, WA), were maintained in IMDM with 20% HI-FBS, 10 ng/ml murine IL-3, and 30 ng/ml murine stem-cell factor (SCF; Peprotech).
DNA constructs
A murine PU.1 genomic subclone was kindly provided by Edward Scott (University of Florida, Gainesville, FL). A 487-bp BamHI fragment, comprising -335–+152, was linked to the luciferase (LUC) cDNA in p19Luc. An XhoI site was then inserted between –65 and –78 by site-directed mutagenesis. Oligonucleotides containing the wild-type (WT) C/EBP-binding site at –68, 5'-TAGCGCAAG-3', or a mutant site (Mut) 5'-TAGCCCTGG-3' and having intact PU.1 promoter sequence until –85 followed by a BamHI site were inserted in the latter plasmid to obtain PU(–85/+152)WT-LUC and PU(–85/+152)Mut-LUC.
Northern blotting and gel-shift assay
Total cellar RNA was prepared using the RNAeasy kit (Qiagen, Valencia, CA) and was subjected to Northern blotting as described [19
]. Nuclear extracts prepared from 293T cells transiently transfected with pMSV-C/EBP, from untransfected cells, or from 32D cl3 cells were subjected to gel-shift and super-shift assay as described [27
]. Rabbit immunoglobulin G (IgG), rabbit anti-C/EBP (14AA), and rabbit anti-C/EBP (C22) were obtained commercially (Santa Cruz Biotechnology, Santa Cruz, CA), and C/EBPß and C/EBP antiserum were kindly provided by Steven McKnight (University of Texas, Dallas, TX) and have been described [9
]. The sense strands of the oligonucleotides used, including 4 bp overhangs to allow Klenow labeling, are listed below, and the potential WT or mutant C/EBP-binding sites are underlined: WT-neutrophil elastase (NE)-C/EBP: 5'-TCGAGGCCAGGATGGGGCAATACAACCCG-3'; Mut-NE-C/EBP: 5'-TCGAGGCCAGGACTCGAGGATACAACCCG-3'; PU.1(68/75): 5'-CTAGGCCAGAGACTTCCTGTAGCGCAAGAGATTTATG-3'; WT-PU.1(–68): 5'-CTAGGACGGCCTGTAGCGCAA-GAGATTTATG-3'; Mut-PU.1(–68): 5'-CTAGGACGGCCTGTAGCCCTGGAGATTTATG-3'; PU.1(–75): 5'-CTAGGCCAGAGACTTCCTGTAGCGCTTGAG-3'; PU.1(–174): 5'-GATCAGTGCTAGCCTTTCTCCCTCCCAGCCC-3'; PU.1(–227): 5'-GATCGTGGACTACTTCAGCAAGGCCTAGCGA-3'.
Chromosomal immunoprecipitation (ChIP) assay
Cells (10x106) per sample were exposed to 1% formaldehyde at room temperature for 10 min, followed by addition of glycine to 0.125 M. Cells were then washed at 4°C with PBS containing 1 mM phenylmethylsulfonyl fluoride, 1 µg/ml aprotinin, and 1 µg/ml pepstatin A (protease inhibitors) and were resuspended in 200 µl 1% sodium dodecyl sulfate (SDS), 10 mM EDTA, 50 mM Tris, pH 8.1, and left on ice 10 min. The cells were then sonicated on ice six times at 15 s each, with 1-min intervals in between, using a Branson Sonifier 250. Sheared DNA was between 200 and 1000 bp. The extract was spun in a microfuge at 13,000 rpm for 10 min; the supernatant was diluted to 2 ml with 0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris, pH 8.1, and 167 mM NaCl containing protease inhibitors; and 20 µl was saved as input DNA. The extract was precleared using 120 µl salmon sperm DNA–protein A agarose 50% slurry (Upstate Biotechnology, Lake Placid, NY) for 30 min on ice. C/EBP antiserum (14AA; 2 µg), acute myeloid leukemia (AML)1 antiserum (kindly provided by Harry Drabkin, University of Colorado, Denver, CO), ER antiserum (HC-ER; Santa Cruz Biotechnology), rabbit IgG, or no antibody was added overnight with rocking at 4°C. DNA–agarose (60 µl) was then added for 1 h, followed by successive washing with low-salt, high-salt, and LiCl buffers (Upstate Biotechnology) and two washes with 10 mM Tris, 1 mM EDTA, pH 8.0. Chromatin fragments were eluted from the beads using 1% SDS, 0.1 M NaHCO3, at room temperature for 15 min, combining two 250 µl fractions. The input DNA was brought to 500 µl using water. NaCl was added to 0.2 M and RNase A, to 20 µg/ml to all three samples, followed by incubation at 65°C for 4 h to reverse the cross-links. EDTA was brought to 10 mM Tris, pH 6.5, to 40 mM and proteinase K, to 40 µg/ml, and the samples were incubated at 45°C for 1 h. After phenol:chloroform and chloroform extractions, 20 µg tRNA and glycogen was added followed by 2.5 vol ethanol. Precipitates were washed with 70% ethanol and resuspended in 30 µl water. Each sample or water (3 µl) was subjected to 30 cycles of polymerase chain reaction (PCR) at 95°C x 1 min, 50°C x 1 min, and 72°C x 1 min. The PCR products were electrophoresed and subjected to Southern blotting using the corresponding promoter segments as probe. The oligonucleotides used for PCR were: PU.1(–215): 5'-CGACCGGAGCAGCAGAAG-3'; PU.1(+25): 5'-CAAGTTCCTGATTTTATCGAAG-3'; D3(–1383): 5'-CTAGCCACCCAGGAAAGAAC-3'; D3(–1180): 5'-GAGATCTTTGCAGCTACTATTG-3'; NE(–157): 5'-ATGGATGATGCTGAAATGGAG-3'; NE(+92): 5'-CTCACCACCCAGGA-ACAATG-3'.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-binding site in the PU.1 promoter
 View larger version (36K): [in a new window] |
Figure 1. Diagram of the PU.1 promoter region and potential C/EBP-binding sites. The murine PU.1 RNA has a 204-bp 5'-untranslated leader followed by the PU.1 open reading frame (ORF). Segments of the promoter linked to the luciferase cDNA in reporter plasmids are shown. An expanded segment depicts the binding sites for factors previously shown to activate the promoter and for the functional C/EBP-binding site identified herein. The sequences of four potential C/EBP-binding sites within the PU.1 promoter are compared with the C/EBP consensus. The –174 and –75 sites are shown in the antisense (as) direction. Flanking sequences are also shown. Bases that differ from the consensus are underlined. A mutant –68 site (–68m) used in this study is also shown. Sp1, Selective promoter factor 1; Oct, Octamer; Ets family = binding site.
|
The core sequences of the potential C/EBP-binding sites at –68 and –75 partially overlap within the sequence 5'-CTTCCTGTAGCGCAAG-3'. We first assessed the ability of an oligonucleotide containing both of these sites to bind C/EBP. Nuclear extracts from 293T cells programmed to express C/EBP were incubated with a radiolabeled oligonucleotide containing a perfect C/EBP consensus site derived from the murine NE promoter. As described previously [27
], ten- or 50-fold excess of unlabeled WT-NE-C/EBP oligonucleotide readily competed for binding, whereas even 50-fold excess of Mut-NE-C/EBP was ineffective (Fig. 2A
, lanes 1–5). PU.1(68/75) competed at least as well as WT-NE-C/EBP (lanes 6 and 7). To determine whether C/EBP was interacting at the –68 or –75 site within PU.1(68/75), we performed the same competition using oligonucleotides containing only one or the other site (Fig. 2A
, lanes 8–11). WT-PU.1(–68) competed effectively, whereas PU.1(–75) did not compete, even at 50-fold excess. Lack of binding to the –75 site may reflect differences from the –68 site within the consensus segment, differences in the flanking sequences, or both. To demonstrate that C/EBP interacts with the predicted binding site within WT-PU.1(–68), we used Mut-PU.1(–68), carrying clustered point mutations in the C/EBP consensus site (Fig. 2B
, lanes 1–9). WT-PU.1(–68) again competed as successfully as WT-NE-C/EBP, whereas Mut-PU.1(–68) was ineffective. To confirm that C/EBP binds WT-PU.1(–68) but not Mut-PU.1(–68), these oligonucleotides were radiolabeled and incubated with nuclear extracts prepared from 293T cells or from 293T cells expressing exogenous C/EBP (Fig. 2C)
. Mutation of the C/EBP consensus at –68 again prevented binding to C/EBP.
 View larger version (51K): [in a new window] |
Figure 2. C/EBP binds the PU.1 promoter strongly at –68. (A) 293T nuclear extracts (10 µg) expressing C/EBP were subjected to gel-shift analysis using 1 ng radiolabeled WT-NE-C/EBP oligonucleotide in the absence of competitor (comp; –) or in the presence of 10 ng or 50 ng unlabeled WT-NE-C/EBP, Mut-NE-C/EBP, PU.1(68/75), WT-PU.1(–68), or PU.1(–75) oligonucleotides. The position of the major C/EBP gel-shift species is indicated (C/EBP). (B) 293T extracts expressing C/EBP were subjected to gel-shift analysis using 1 ng radiolabeled WT-NE-C/EBP in the absence of competitor or in the presence of 10 ng or 50 ng WT-NE-C/EBP, Mut-NE-C/EBP, WT-PU.1(–68), Mut-PU.1(–68), PU.1(–174), or PU.1(–227). (C) Radiolabeled WT-PU.1(–68) or Mut-PU.1(–68) oligonucleotides (–68, m–68; 1 ng) were incubated with 10 µg 293T (–) or C/EBP-transfected 293T (+) nuclear extracts and subjected to gel-shift analysis.
|
with potential binding sites centered at –174 and –227 was also assessed by using oligonucleotides containing these sites as competitors in a gel-shift assay (Fig. 2B) . Both competed with WT-NE-C/EBP at 50-fold excess but not at tenfold excess, indicating a moderate affinity. Notably, these sites are not conserved in the human PU.1 promoter, whereas the human –68 site along with two 5' and four 3' flanking bases, 5'-TG-TAGCGCAAG-AGAT-3', match the murine –68 C/EBP-binding site exactly.
C/EBP from myeloid nuclear extracts preferentially binds the PU.1 promoter
To assess interaction of C/EBP proteins from myeloid cells with the PU.1 promoter, radiolabeled WT-PU.1(–68) or WT-NE-C/EBP was subjected to gel-shift assay with nuclear extracts from 32D cl3 myeloblasts (Fig. 3A
). Each probe bound three predominant species, which were competed by WT-PU.1(–68) but not by Mut-PU.1(–68). To identify the C/EBP isoforms bound to each probe, they were subjected to super-shift assay using a panel of C/EBP antisera (Fig. 3B
, left panel). Compared with no antibody or rabbit IgG control, the top band bound to each probe super-shifted with C/EBP antiserum, the bottom band with C/EBPß antiserum, and the central band with both antisera, identifying these bands as C/EBP homodimers (), C/EBPß homodimers (ßß), or C/EBP:C/EBPß heterodimers (ß). Neither band super-shifted with C/EBP or C/EBP antisera. As C/EBPß is smaller than C/EBP, it is not surprising that its homodimer has greater mobility. Of note, the WT-NE-C/EBP probe bound C/EBP and C/EBPß homodimers with similar affinity, whereas the WT-PU.1(–68) probe bound C/EBP more strongly. Finally, we compared the affinity of WT and mutant PU.1(–68) probes for endogenous proteins in 32D cl3 nuclear extracts (Fig. 3B
, right panel). Introduction of clustered point mutations into the WT-PU.1(–68) oligonucleotide reduced endogenous C/EBP binding several-fold and allowed increased binding by two more slowly migrating species (*) but did not lead to binding novel factors. Perhaps the bands indicated by an asterisk represent binding to the previously defined Ets site at –76.
 View larger version (68K): [in a new window] |
Figure 3. C/EBP from myeloid cells binds the PU.1 promoter. (A) 32D cl3 nuclear extracts (10 µg) were subjected to gel-shift analysis using 1 ng radiolabeled WT-PU.1(–68) or WT-NE-C/EBP oligonucleotides in the presence of 10 ng or 50 ng unlabeled WT-PU.1(–68) or Mut-PU.1(–68; PU-WT, PU-M) or in the absence of competitor (comp; –). (B) 32D cl3 nuclear extracts (10 µg) were incubated with these same probes in the presence of no antibody (–), rabbit IgG (Ig), or antisera specific for C/EBP, C/EBPß, C/EBP, or C/EBP (, ß, , , left panel). Similar extracts were incubated with radiolabeled WT-PU.1(–68) or Mut-PU.1(–68; –68, m–68, right panel). The locations of C/EBP homodimers (), C/EBPß homodimers (ßß), and C/EBP:C/EBPß heterodimers (ß) are indicated.
|
trans-activates the PU.1 promoter via its binding site at –68, pMSV-C/EBPß, or pMSV-C/EBP leucine 1,2-valine (L12V) and pCMV-ßGal (Fig. 4A
). C/EBPL12V carries point mutations in the leucine zipper domain, which obviates DNA binding. Relative to pMSV, pMSV-C/EBP activated the WT reporter twofold on average in four experiments, whereas the mutant reporter was repressed approximately twofold. ßGal activity generated from pCMV-ßGal was not affected by C/EBP, and a threefold lower dose of pMSV-C/EBP neither activated nor repressed the reporters (not shown). In contrast to C/EBP, neither C/EBPß nor C/EBPL12V activated the PU.1 promoter. Lack of activation by C/EBPß is consistent with the greater affinity of C/EBP for the C/EBP-binding site at –68. The human PU.1 promoter was also recently found to be induced twofold by C/EBP in 293T cells, but the relevant cis element was not identified [31
].
 View larger version (18K): [in a new window] |
Figure 4. C/EBP activates the PU.1 promoter in 293T cells and in myeloid cells. (A) PU(–85/+152)WT-LUC or PU(–85/+152)Mut-LUC (750 ng) was cotransfected with 50 ng pMSV, pMSV-C/EBP, pMSV-C/EBPß, or pMSV-C/EBPL12V and 0.25 ng pCMV-ßGal. Luciferase and ßGal assays were performed on cell extracts 48 h later. Luciferase activity was normalized for transfection efficiency using ßGal activity. Shown is the mean activity of each reporter in the presence of pMSV-C/EBP, pMSV-C/EBPß, or pMSV-C/EBPL12V divided by their activities with pMSV (mean and SE from two to four determinations). (B) 32DPKC cells were transfected with 10 µg of the indicated luciferase reporters and with 0.5 µg pCMV-ßGal. Luciferase activities were determined 48 h later. Shown is the mean activity of each reporter relative to pCMV-ßGal (mean and SE from three determinations). The activity of PU(–85/+152)Mut-LUC was set at one in each experiment.
|
cells, a myeloid cell line. These cells were used in lieu of the related 32D cl3 line, as they are more readily transfected. In this paradigm, we rely on endogenous C/EBPs to activate the reporters. Mutation of the C/EBP-binding site at –68 in PU(–85/+152)WT reduced promoter activity 1.7-fold on average in three experiments (Fig. 4B)
. As will be discussed, this degree of reduction is similar to the effect of mutating several other factor binding sites in the promoter. Perhaps several transcription factors each contribute modest activation.
C/EBP binds the endogenous PU.1 promoter
To determine whether induction of endogenous PU.1 RNA by C/EBP-ER requires DNA binding, we compared 32D-C/EBP-ER and 32D-C/EBPL12V-ER cells, which express equivalent amounts of exogenous protein [32
]. As reported previously [20
], activation of C/EBP-ER with estradiol induced PU.1 RNA within 8 h, even in the presence of cycloheximide, whereas C/EBPL12V-ER was ineffective (Fig. 5
).
 View larger version (105K): [in a new window] |
Figure 5. C/EBP must bind DNA to induce the endogenous PU.1 gene. 32D-C/EBP-ER or 32D-C/EBPL12V-ER cells in IL-3 were cultured in the absence or presence of cycloheximide (CHX) for 30 min, followed by addition of estradiol (Est.) for 8 h. Total cellular RNAs, 10 µg per sample, were then prepared and subjected to Northern blotting for PU.1 (top panel). An equivalent aliquot of RNA was stained for ethidium bromide as a control for loading and RNA integrity (bottom panel).
|
can bind the endogenous PU.1 promoter, we performed ChIP analysis. Parental 32D cl3 cells exposed to G-CSF for 24 h, to induce endogenous C/EBP levels [9
], were cross-linked with formaldehyde and were sonicated. Chromosomal fragments were immunoprecipitated with or without C/EBP antiserum and with an AML1 antiserum as a control. The presence of the PU.1 and cyclin D3 promoters in DNA purified from these precipitates was assessed by PCR (Fig. 6A
). The input (I) lane represents the signal derived from 1% of the cell equivalents used for the immunoprecipations. The PU.1 promoter was detected with the PU.1 but not the AML1 antiserum, and the cyclin D3 promoter was only minimally precipitated with the PU.1 antiserum. Precipitation of the NE promoter was also minimal (data not shown), perhaps as a result of its greater affinity for other C/EBP isoforms.
 View larger version (25K): [in a new window] |
Figure 6. C/EBP-ER binds the endogenous PU.1 promoter. (A) 32D cl3 cells transferred from IL-3 to G-CSF for 24 h were subjected to ChIP analysis using C/EBP or AML1 antiserum (Ab) and PCR primers designed to amplify -215/+25 of the PU.1 promoter (top panels) or -1383/-1180 of the cyclin D3 promoter as a negative control (bottom panel). - or +, The cell extracts incubated in the absence or presence of antiserum; I, 1% of the input DNA. (B) 32D-C/EBP-ER cells exposed to estradiol for 24 h (ER+Est.) were subjected to ChIP analysis using ER antiserum and PCR primers designed to amplify the PU.1 promoter, the cyclin D3 promoter, and -157/+92 of the NE promoter as a positive control (left panels). 32D cl3 cells exposed to estradiol for 24 h (32D+Est.), 32D-C/EBP-ER cells not exposed to estradiol (ER-Est.), and 32D-C/EBPL12V-ER cells exposed to estradiol (L12VER+Est.) were also subjected to ChIP analysis for binding to the PU.1 promoter using ER antiserum (right panels). H, Water alone used for PCR; Ig, cell extract incubated in the presence of rabbit IgG.
|
bound to the PU.1 promoter was modest, perhaps as a result of low-level C/EBP expression or weak antiserum affinity. We therefore sought to determine whether C/EBP-ER also bound the PU.1 promoter. 32D-C/EBP-ER cells exposed to estradiol for 24 h were cross-linked and sonicated. Chromosomal fragments were immunoprecipitated with or without human ER antiserum (which does not bind murine ER), and the presence of the PU.1, cyclin D3, and NE promoters in DNA purified from these precipitates was assessed by PCR (Fig. 6B
, left panels). The PU.1 and NE promoters were specifically detected using the ER antiserum. As controls, ChIP analysis of the PU.1 promoter was also performed on parental 32D cl3 cells exposed to estradiol, on 32D-C/EBP-ER cells in the absence of estradiol, and on 32D-C/EBPL12V-ER cells in the presence of estradiol (Fig. 6B
, center and right panels). Significant immunoprecipation of the PU.1 promoter by ER antibody was not detected in these samples. In addition, rabbit IgG did not immunoprecipitate the PU.1 promoter in cells expressing C/EBPL12V-ER (Fig. 6)
or C/EBP-ER (data not shown).
PU.1 RNA is reduced in C/EBP (-/-) cells
Total cellular RNAs were isolated from C/EBP (-/-) and C/EBP (+/-) murine cell lines, which had been generated by transducing a dominant-negative retinoic acid receptor into marrow isolated from C/EBP (-/-) and (+/-) littermates [33
]. These lines were apparently isolated from a related myeloid progenitor, as each requires IL-3 and SCF to survive and proliferate. These RNAs were subjected to Northern blotting for PU.1 and ß-actin (Fig. 7
). PU.1 RNA was detected in the C/EBP (-/-) cells but at a level several-fold lower than in the C/EBP (+/-) line, relative to ß-actin.
 View larger version (39K): [in a new window] |
Figure 7. C/EBP (-/-) cells have reduced PU.1 RNA. Total cellular RNA from immortalized C/EBP (+/-) and C/EBP (-/-) myeloid cells (10 µg), each maintained in IL-3 and SCF, were subjected to Northern blotting for PU.1 and ß-actin.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
binds and activates the PU.1 promoter in myeloid cells, via an evolutionarily conserved binding site centered at –68 and situated in the midst of previously identified, functional Ets-, Oct-1/2-, Sp1-, and PU.1-binding sites. Modest activation of the murine PU.1 promoter via this site is consistent with its ability to activate the human promoter to a similar extent [31
] and with the effect of mutation of other functional sites in the promoter. Mutation of the Ets site at –76 decreased promoter activity twofold in B-lineage and macrophage cell lines [28
]. Mutation of the Oct site at –50 decreased promoter activity two- to fourfold in B-lymphoid cells and twofold in U937 myeloblasts but had no effect in the RAW264.7 macrophage line [28
29
30
]. The Oct site is likely more important in B-lymphoid cells than in myeloid cells as a result of the presence of OBF-1, a B cell-specific coactivator [30
]. Mutation of the Sp1 site at –37 diminished PU.1 promoter activity twofold in myeloid and lymphoid cells, and mutations of the PU.1 site at +18 reduced activity threefold in lymphoid cells and 40-fold in myeloid cells [29
, 30
]. As the murine and human PU.1 promoters are highly conserved between -173 and +147, they may also be regulated by additional factors yet to be identified [30
]. In addition, C/EBP and other factors may activate PU.1 transcription via distal regulatory elements [34
].
PU.1 and C/EBP are each key regulators of hematopoiesis. PU.1 (-/-) mice have a more severe defect than C/EBP (-/-) mice, suggesting that PU.1 acts earlier than C/EBP in the hierarchy of myeloid commitment. Yet, myeloid cells have higher levels of PU.1 than B-lymphoid cells [24
], and we reproducibly find that exogenous C/EBP increases endogenous PU.1 mRNA levels in 32D cl3, 32DPKC, and Ba/F3 hematopoietic cells [20
, 35
]. These findings support a model we have described previously [36
] in which generation of GMPs occurs by C/EBP- or myeloid cytokine-mediated induction of PU.1 levels (Fig. 8
). The ability of PU.1 to autoregulate its own promoter would serve to amplify and solidify developmental decisions triggered by small increases in PU.1 or C/EBP in less committed progenitors. The existence of two pathways for increasing PU.1 levels in hematopoietic stem cells is supported by the finding that G-CSF receptor signals elevate PU.1 RNA in the presence of a dominant inhibitor of C/EBP [19
]. Perhaps related to this observation is the finding that a C/EBP (-/-) cell line immortalized using homeobox (HOX)-11 differentiates to neutrophils in response to exogenous C/EBP or when cultured in IL-3 and GM-CSF [37
]. Of note, in these HOX-11-expressing C/EBP (-/-) cells, IL-3 and GM-CSF or activation of C/EBP-ER increase endogenous PU.1 expression (Daniel Tenen, personal communication).
 View larger version (20K): [in a new window] |
Figure 8. Model for the role of PU.1 induction in myeloid differentiation. In GMPs, C/EBP or cytokine signals elevate PU.l levels. PU.1 then cooperates with C/EBP or other C/EBPs to specify and play a role in the maturation of a commited granulocytic progenitor (G) and contributes, potentially even in the absence of C/EBPs, to the specification and terminal differentiation of a commited monocytic progenitor (M). GM-CSF, G macrophage-CSF.
|
or PU.1 protein-coding regions were recently identified in subsets of AML [38
39
40
]. Our findings suggest that leukemias with C/EBP mutations also express reduced levels of PU.1.
In summary, activation of the PU.1 promoter by C/EBP may help direct myeloid as compared with lymphoid commitment of multipotent hematopoietic progenitors.
L12V-ER cells, K. Ravid for the cyclin D3 promoter probe, S. J. Collins for C/EBP (-/-) and (+/-) cells, and W. Wang for technical assistance. Received December 23, 2002; revised April 3, 2003; accepted April 23, 2003.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
deficiency in hematopoiesis induces accumulation of non-malignant myeloblasts mimicking acute myelogenous leukemia Blood 100,61a by-passes G-CSF signals to rapidly induce PU.1 gene expression, stimulate granulocytic differentiation, and limit proliferation in 32D cl3 myeloblasts Blood 94,560-571 inactivates the myeloid master regulator PU.1: possible role in lineage commitment decisions Blood 100,483-490 basic region is required but not sufficient for granulopoiesis Oncogene 22,2548-2557[CrossRef][Medline](-/-) knockout mice display granulocyte macrophage-colony stimulating factor, granulocyte-colony stimulating factor, and retinoic acid-induced granulocytic differentiation Blood 98,2382-2388 and PKC on JunB expression and monocytic differentiation depend upon the C/EBP basic region Blood 101,3885-3892This article has been cited by other articles:
 |
|
 |
|
  X. Liu, B. Wu, J. Szary, E. M. Kofoed, and F. Schaufele Functional Sequestration of Transcription Factor Activity by Repetitive DNA J. Biol. Chem., July 20, 2007; 282(29): 20868 - 20876. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Zheng, A. D. Friedman, M. Levis, L. Li, E. G. Weir, and D. Small Internal tandem duplication mutation of FLT3 blocks myeloid differentiation through suppression of C/EBP{alpha} expression Blood, March 1, 2004; 103(5): 1883 - 1890. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
