Novel Combinatorial Interactions of GATA-1, PU.1, and C/EBP
Isoforms Regulate Transcription of the Gene Encoding Eosinophil Granule
Major Basic Protein*
Jian
Du
,
Monika J.
Stankiewicz,
Yang
Liu,
Qing
Xi,
Jonathan E.
Schmitz,
Julie A.
Lekstrom-Himes§, and
Steven J.
Ackerman¶
From the Department of Biochemistry and Molecular
Biology, College of Medicine, University of Illinois,
Chicago, Illinois 60612 and § Laboratory of Host Defenses,
National Institutes of Health, Bethesda, Maryland 20892
Received for publication, May 16, 2002, and in revised form, July 24, 2002
 |
ABSTRACT |
GATA-1 and the ets factor PU.1 have been reported
to functionally antagonize one another in the regulation of erythroid
versus myeloid gene transcription and development. The
CCAAT enhancer binding protein 
(C/EBP) is expressed as
multiple isoforms and has been shown to be essential to myeloid
(granulocyte) terminal differentiation. We have defined a novel
synergistic, as opposed to antagonistic, combinatorial interaction
between GATA-1 and PU.1, and a unique repressor role for certain
C/EBP isoforms in the transcriptional regulation of a model
eosinophil granulocyte gene, the major basic protein (MBP). The
eosinophil-specific P2 promoter of the MBP gene contains GATA-1, C/EBP,
and PU.1 consensus sites that bind these factors in nuclear
extracts of the eosinophil myelocyte cell line, AML14.3D10. The
promoter is transactivated by GATA-1 alone but is synergistically
transactivated by low levels of PU.1 in the context of optimal levels
of GATA-1. The C/EBP27 isoform strongly represses GATA-1
activity and completely blocks GATA-1/PU.1 synergy. In
vitro mutational analyses of the MBP-P2 promoter showed that both
the GATA-1/PU.1 synergy, and repressor activity of
C/EBP27 are mediated via protein-protein interactions
through the C/EBP and/or GATA-binding sites but not the PU.1 sites.
Co-immunoprecipitations using lysates of AML14.3D10 eosinophils show
that both C/EBP32/30 and 27 physically
interact in vivo with PU.1 and GATA-1, demonstrating functional interactions among these factors in eosinophil progenitors. Our findings identify novel combinatorial protein-protein interactions for GATA-1, PU.1, and C/EBP isoforms in eosinophil gene
transcription that include GATA-1/PU.1 synergy and repressor activity
for C/EBP27.
| |
INTRODUCTION |
Hematopoietic development is regulated in part by the
combinatorial actions of transcription factors that coordinate temporal and lineage-specific patterns of gene expression. The transcription factors GATA-1, PU.1, and members of the CCAAT enhancer binding protein
(C/EBP)1 family (C/EBP
and
C/EBP in particular) are essential for the commitment and/or
terminal differentiation of myeloid progenitors to the eosinophil
lineage (1-9). A direct interaction and functional antagonism between
GATA-1 and PU.1 have been reported recently by several groups (10-12).
PU.1 has also been shown to directly or indirectly interact with
multiple transcription factors including C/EBP, C/EBP
(NF-IL6
,
CRP3), c-Jun, c-Myb, AML1, and NF-
B (13-17).
GATA-1, a zinc finger family member, has been shown to be essential for
virtually all aspects of erythroid and megakaryocyte gene transcription
and lineage development (18-20). We have shown that
eosinophil-committed promyelocytic cell lines express GATA-1-3 mRNA in an inducible fashion, and peripheral blood eosinophils of
mixed developmental maturity isolated from patients with the hypereosinophilic syndrome also express GATA-1 mRNA (21). Studies in the avian system using myb-ets-transformed myeloid progenitors have
suggested that an intermediate level of GATA-1 expression in the
context of C/EBP and PU.1 is critical for eosinophil
lineage-specific gene expression and differentiation (1, 2, 7, 12, 22). As well, GATA-1 expression is slowly down-regulated during
eosinophilopoiesis in the avian system (3, 7). However, the GATA-1 null
mutation in the mouse has not been evaluated for defects in eosinophil development.
PU.1, an ets family member, is selectively expressed in B lymphocytes,
granulocytes, and monocytes and is required for the development of
these hematopoietic lineages (23-28). During the proliferation and
differentiation of multipotential lymphoid-myeloid progenitors, PU.1 is
expressed at all stages of granulocyte development, both during the
early onset of expression of the GM-CSF and G-CSF receptors in myeloid progenitors (27-30), and during the more terminal differentiation steps that include expression of the CD11b, CD18, and
Fc
RI receptors (31-35). PU.1 expression increases gradually through
the myelocytic stages of hematopoietic development and then remains
relatively constant through terminal differentiation (26, 36).
Importantly, graded expression levels of PU.1 in the mouse have been
shown to specify distinct cell lineage fates in hematopoietic
development, with low levels of PU.1 specifying lymphocytic and high
levels specifying monocytic differentiation (37). Moreover, disruption
of the PU.1 gene affects multiple hematopoietic lineages, including
defects in granulocyte terminal differentiation, resulting in the
absence of functionally mature neutrophils (24, 25, 28, 38) and
eosinophils2 (39). Thus, PU.1
is a critical regulator of hematopoietic gene transcription that
possesses multiple and varied roles at different stages of both
lymphoid and myeloid development.
The CCAAT/enhancer-binding proteins (C/EBPs), members of leucine zipper
transcription factor family, include six members to date, of which four
(C/EBP, -, -, and -) are expressed in the myeloid lineages
and play key roles in hematopoiesis and/or functional maturation of
these cells, and two (C/EBP and -
) lack transactivation domains
and act as dominant negative repressors of transcription (40, 41).
C/EBP is expressed in a variety of tissues with the highest
expression levels in adipose and liver (42, 43). C/EBP plays a
critical role in the terminal differentiation of hepatocytes, lipid
storage, and early granulocyte development (14, 15, 42, 44-48). The
null mutation of C/EBP in mice results in profound abnormalities in
glucose metabolism and granulocyte (both neutrophil and eosinophil)
development, including impaired G-CSF receptor signaling (15,
16, 49-52). C/EBP, originally identified as a mediator of IL-6
signaling (53, 54), is critical for lipid storage, Th1 immune
responses, and macrophage function (55, 56), but targeted disruption of
the C/EBP gene had no effect on hematopoietic development.
C/EBP, the most recently identified C/EBP family member, is
expressed almost exclusively in the myeloid lineages and is mostly restricted to the later stages of differentiation from the promyelocyte to mature granulocyte (57). Disruption of the C/EBP gene blocks the
terminal differentiation of granulocytes, including impaired production
of eosinophils (4) and decreased, but not loss, expression of a number
of eosinophil secondary (specific) granule protein genes (see Fig.
10).3 A frameshift mutation
in the C/EBP gene has been implicated recently (58, 59) as the
primary defect specifying neutrophil-specific granule deficiency. Four
isoforms of C/EBP (32, 30, 27, and 14 kDa), generated by alternative
promoter usage, mRNA splicing, and different translation start
sites have been identified (60-62). Structural and functional analyses
have shown that full-length C/EBP32 (32-kDa isoform)
contains both transcriptional activation and repression domains
(61-63). The C/EBP27 isoform contains DNA binding,
dimerization, transactivation, and possible repression domains, whereas
C/EBP14 is the only isoform that lacks a transactivation
domain. Essentially nothing is known about the differential expression
of the C/EBP27 and 14 isoforms during
hematopoietic development nor their activities or specific roles in
myeloid gene expression. Although C/EBP family proteins generally bind
DNA as either homo- or heterodimers (64), dimer formation was not
detected for C/EBP32 expressed recombinantly in either
bacteria or mammalian cells (65).
Functional and physical interactions between C/EBP and GATA-1 were
first identified in vitro through our studies of the
transcriptional regulation of the gene encoding the human eosinophil
granule major basic protein (MBP) (9), in which we demonstrated synergy
between these factors in the transactivation of the MBP promoter.
Regulation of an eosinophil-specific gene (EOS47) by the combinatorial
actions of C/EBP, ets-1, and GATA-1 has also been reported for the
avian system (2, 3). In the present study, we have further
characterized the eosinophil-specific MBP-P2 promoter, identifying
novel combinatorial interactions among GATA-1, PU.1, and the
C/EBP27 isoform that result in either synergistic
activation or transcriptional repression. Human MBP, one of the
principal mediators of inflammation and tissue damage in
eosinophil-associated allergic responses, is expressed by eosinophils
(66), and to a lesser extent by basophils (67), and a proform of MBP is
elaborated by placental trophoblasts during gestation (68, 69). Two
different MBP transcripts are regulated by alternative splicing from
two distinct promoters (P1 and P2) located 34 kb apart (70).
Eosinophils and basophils transcribe principally from the P2 promoter,
whereas the P1 promoter predominates in placental cells (8, 70).
We identified previously a major functional region for the MBP-P2
promoter between bp 
117 and 67 (8, 9). In the present studies, a
more extensive consensus binding site analysis showed that the bp 117
to +1 region of the promoter contains not one but two adjacent
GATA-binding sites, two C/EBP sites, and two previously unidentified
PU.1 sites, suggesting that the combinatorial activities of all three
of these factors are involved in MBP gene transcription. Because PU.1
has been reported to antagonize GATA-1-mediated erythroid gene
expression and development (10-12), and we have shown that GATA-1 is a
key regulator of the MBP gene, we utilized the MBP-P2 promoter as a
model to characterize further the combinatorial interactions and roles
of GATA-1, PU.1, and C/EBP in the eosinophil lineage. Consensus
sites in the functional region of the MBP-P2 promoter were evaluated
for their ability to bind GATA, PU.1, or C/EBP family members using
nuclear extracts from the MBP-expressing eosinophil myelocyte cell
line, AML14.3D10, and mature peripheral blood eosinophils. The
functional relevance of these binding sites and transcription factor
interactions were evaluated using mutagenesis, transactivation analyses
in CV-1 cells, and transient transfection analyses in AML14.3D10
eosinophils. Co-immunoprecipitation (co-IP) experiments were performed
to confirm the in vivo physical interactions of GATA-1,
PU.1, and the various C/EBP isoforms in eosinophil cell lines and
mature peripheral blood eosinophils.
Our results show that GATA-1, PU.1, and the various C/EBP isoforms
are differentially expressed, and physically and functionally interact
during human eosinophil development to activate and then repress
eosinophil gene transcription. In contrast to the functional antagonism
reported for PU.1 and GATA-1 for activation of the M-CSFR promoter and
various erythroid genes, we show that these factors cooperate
synergistically in the activation of the eosinophil MBP-P2 promoter.
Importantly, we demonstrate that PU.1 expression levels relative to
GATA-1 determine whether the GATA-1/PU.1 interaction leads to synergy
for GATA-1-regulated eosinophil target genes such as MBP. Moreover, we
identify a novel role for the C/EBP27 isoform as a
repressor of GATA-1-mediated transactivation and GATA-1/PU.1 synergy.
Finally, co-IP analyses demonstrate for the first time that the
C/EBP32/30 and C/EBP27 isoforms
physically interact with both PU.1 and GATA-1 in vivo, in
eosinophils. Our findings demonstrate key combinatorial interactions of
GATA-1, PU.1, and C/EBP isoforms that may mediate either synergy or
antagonism (repression) of granulocyte (eosinophil) gene transcription during myeloid development and terminal differentiation.
| |
EXPERIMENTAL PROCEDURES |
Primary Cells and Cell Lines--
Human blood eosinophils were
isolated from a unit (475 ml) of blood obtained from normal,
non-allergic, healthy donors. The use of normal human subjects as blood
donors was in full compliance with all federal guidelines and was
approved by the University of Illinois Institutional Review Board.
Eosinophil isolation and purification were performed as described
previously (71) using a Miltenyi Biotec SuperMacs immunomagnetic
sorting kit and apparatus according to manufacturer's specifications.
Eosinophil preparations of >99% purity were routinely obtained and
utilized for these studies. The human AML14.3D10 cell line is a fully
differentiated eosinophil myelocyte line that contains eosinophil
secondary (specific) granules and displays many characteristics of
mature peripheral blood eosinophils. These include expression of the
secondary granule cationic proteins major basic protein, eosinophil
peroxidase, eosinophil-derived neurotoxin, and eosinophil cationic
protein (72, 73). AML14.3D10 cells can be induced to express chemokine receptors (e.g. for eotaxins) (74) and express GM-CSF
which drives their proliferation, differentiation, and survival in
culture (75). The AML14 parental cell line is a human myeloid
leukemia-derived myeloblast committed to the eosinophil lineage. AML14
and AML14.3D10 cells were maintained in RPMI 1640 supplemented with 8%
fetal bovine serum, 2 mM L-glutamine, 1 mM sodium pyruvate, and 5 × 10
5
M -mercaptoethanol without any antibiotics. Cells were
passaged every 3-4 days and maintained at a concentration between
3 × 105 and 1 × 106 cells/ml. CV-1
cells were maintained by passage twice weekly in Dulbecco's modified
Eagle's medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, penicillin (10 units/ml), and streptomycin (10 µg/ml) and were used for transfection at 60-70% confluence.
Whole Cell Lysates and Nuclear Extract--
AML14.3D10
eosinophils were lysed in a lysis buffer containing 0.5-1% Triton
X-100, 150 mM NaCl, 1 mM EDTA, 200 µM sodium orthovanadate, 50 mM Hepes, 10 mM sodium pyrophosphate, 100 mM sodium
fluoride, 1.5 mM magnesium chloride, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, and 10 µg/ml aprotinin
(76), supplemented with protease inhibitor mixture tablets (Roche
Molecular Biochemicals) for 1 h at 4 °C. Whole cell lysates
were collected by centrifugation at 14,000 × g for 20 min at 4 °C to remove cell debris, and lysates were stored at
80 °C. Nuclear extracts were prepared by the method of Dignam
et al. (77) with minor modifications, including the use of
protease inhibitor mixture tablets (Roche Molecular Biochemicals) and
the addition of phenylmethylsulfonyl fluoride (0.5 mM) and
diisopropylfluorophosphate (1 mM) to the resuspension and
lysis buffers (78, 79). The protein concentration of nuclear or whole
cell extracts was determined by the BCA method (Pierce), and extracts
were aliquoted and stored at 80 °C.
DNA Constructs and Mutagenesis--
The bp 117MBP-P2-pXP2
luciferase reporter gene construct was described in detail previously
(8, 9). Mutagenesis PCR was performed to create mutations in the PU.1,
GATA-1, or C/EBP consensus sites that block transcription factor
binding using the QuickChangeTM PCR kit (Stratagene, CA).
Dr. Daniel Tenen kindly provided the M-CSFR promoter construct
(M-CSFR-pXP2). The expression vector PU.1-pECE contains the mouse
cDNA for PU.1 and was a gift from Dr. R. Maki (23). The GATA-1-pXM
expression vector contains the murine GATA-1 cDNA and was kindly
provided by Dr. L. Zon (80, 81). Both C/EBP and -, in the pMSV
expression vector, were kindly provided by Dr. A. Friedman, and the
various C/EBP isoforms in the pcDNA3 expression vector
(pc32, pc27, and pc14) were kindly provided by Dr. K. Xanthopoulos. All DNA constructs used in transfection and
transactivation experiments were prepared by alkaline lysis
maxi-preparation followed by CsCl2 purification as
described previously (79, 82).
Western Blotting and Co-immunoprecipitations--
Western
blotting was performed as described previously (71, 76) using either
whole cell lysates or nuclear extracts from blood eosinophils,
AML14.3D10 eosinophil myelocytes, and undifferentiated AML14 parental
cells. HL-60 and K562 cell lines were used as positive controls as
appropriate to the transcription factors being analyzed. Twenty µg of
each extract was separated by SDS-PAGE and electroblotted onto
Immobilon-P transfer membranes (Millipore, MA) following standard
procedures. Antibodies against GATA-1 (M-20), PU.1 (T-21 and N-19), and
C/EBP (C-20) and negative control antibodies were purchased from
Santa Cruz Biotechnology (Santa Cruz, CA). All antibodies were
non-cross-reactive with other members of their respective transcription
factor families. In particular, the antibodies to PU.1 do not recognize
other ets family proteins. The blots were analyzed using an enhanced
chemiluminescence kit (Pierce) following the manufacturer's
recommended procedure. Immunoprecipitation (IP) and co-IP were carried
out using total cell lysates. Antibodies used for IP and co-IP were the
same as used for Western blotting. IP and co-IP were performed
essentially as described previously (71, 76). Briefly, the whole cell
lysate from 2 to 4 × 107 AML14.3D10 cells was
precleared with 1-2 µg of normal rabbit, goat, or rat
affinity-purified IgG antibodies plus protein A- or G-Sepharose
(Amersham Biosciences) with rotation at 4 °C for 2.5 h.
Antibodies were added to the precleared whole cell lysates for 2 h
of rotation at 4 °C. Non-immune purified rabbit, rat, and goat
immunoglobulin (Santa Cruz Biotechnology) were utilized as negative
controls. Protein A- or protein G-Sepharose was added to the lysates
for another 2 h at 4 °C. Immunoprecipitates were washed 4 times
with 0.1% Triton X-100 lysis buffer and analyzed by Western blotting.
Electrophoretic Mobility Shift Assays
(EMSA)--
Oligonucleotides containing the GATA, PU.1, and C/EBP
consensus sites in the MBP-P2 promoter were synthesized and purified by
Integrated DNA Technologies, Inc. (Coralville, Iowa). Double-stranded probes were end-labeled with [-32P]ATP (PerkinElmer
Life Sciences) and purified on 15% polyacrylamide gels as described
previously (79). Probe sequences were as follows: dual GATA sites
(5'-GTCCTTATCAGCCTTGCTATCTCCCT-3'), PU.1 site 1 (5'-CCCTGGGGGAAGTTCCTCCAAGGCCT-3'), PU.1 site 2 (5'-AAGTCTTTGTGAGAGGAAGCAAAGAA-3'), C/EBP site
(5'-GAAGTGATGAAATGGTCC-3'), C/EBP and GATA-1 sites (5'-GAAGTGATGAAATGGTCCTTATCAGCCT-3').
EMSAs were performed as described previously (71). The same antibodies
used for Western blotting and co-IP were also used for supershift
analyses but were supershift grade (Santa Cruz Biotechnology). The
antibodies to PU.1 (T-21 and N-19) were specifically chosen because
they are non-cross-reactive with other ets family member proteins.
Transactivation and Transient Transfection
Assays--
Transactivation experiments were performed using the
DMRIE-C reagent (Invitrogen) for transfections according to the
manufacturer's protocol. CV-1 cells (60-80% confluent in 6-well
plates) were transfected with 0.5 µg of the MBP-P2-pXP2 wild-type or
mutant promoter constructs in the presence or absence of expression
vectors for PU.1 (0.001-0.1 µg), GATA-1 (0.2 µg), and/or C/EBP
and -, or 32, 27, or 14
isoforms (0.2 µg). Because GATA-1 may be an activator at low levels
or a repressor at high levels of expression as reported for the avian
Eos-47 eosinophil promoter (2, 7), the amount of the GATA-1 expression
vector added (0.2 µg) to co-transactivations with PU.1 was first
optimized by dose-response titration in order to provide maximum
transactivation of the MBP-P2-pXP2 promoter construct. The empty
vectors corresponding to each of the transcription factors were used to
normalize the total DNA content of each reaction. Transfection
efficiency was normalized by the addition of 0.2 µg of the pRL-CMV
(Renilla luciferase) control reporter. Promoter activities
were determined using the Dual-luciferase Reporter Assay System
(Promega) 48 h after transfection. Luciferase activities were
measured as relative light units using an EG & G Berthold Lumat LB 1507 Luminometer as described previously (71). Results are reported as the
mean ± S.D. (fold induction over control) for at least three
independent transactivation experiments. Transient transfections of the
AML14 parental and AML14.3D10 eosinophil cell lines were performed by
electroporation as described previously (71, 79). Briefly, 10 µg of
each DNA construct and 1.5 × 107 cells were used in
each reaction. Promoter activities (relative light units) were measured
6 h after electroporation, and transfection efficiency was
normalized using the pRL-CMV vector as above in the dual luciferase
assay. Results are reported as the mean ± S.D. (fold induction
over control) for at least three independent experiments.
Semi-quantitative Reverse Transcriptase-PCR--
Total RNA was
prepared from fetal liver or bone marrow cells of mice with targeted
disruptions of the C/EBP (50, 52) and C/EBP (4) genes using
standard purification methods as described previously (4, 83). C/EBP
null mice were first stimulated for 4.5 or 24 h with 4%
thioglycolate broth by intraperitoneal injection prior to harvesting
the bone marrow as described previously (4). The RNA was
electrophoresed on 1% formaldehyde agarose gels, and 18 S and 28 S
bands were visualized by ethidium bromide staining to assess the
quality. Three µg of total RNA was reverse-transcribed with random
hexadeoxynucleotide primers using a First-strand cDNA Synthesis kit
(Amersham Biosciences) according to the manufacturer's protocol.
Calculations were based on the assumption that 100% of the RNA was
reverse-transcribed to cDNA. The PCR primer pairs for cDNA
amplification of murine MBP were forward 5'-CCAAGGAAGAGGACACAACAAGTC-3' and reverse 5'-TTGACCCTGGTTGATTCCCC-3'; GAPDH forward
5'-CCATGGAGAAGGCTGGGG-3' and reverse 5'-CAAAGTTGTCATGGATGACC-3'. BLAST
search analysis of these primers indicated they were specific for the
genes of interest. The signal from amplification of the GAPDH cDNA
(mRNA) was used first as a control to normalize the amount of
cDNA input for subsequent PCR amplification of each sample. The PCR
mix contained 10-50 ng of cDNA, 10 pmol of each primer, 200 µM dNTPs, 0.3 µl of 10 mCi/ml
[-32P]dCTP (PerkinElmer Life Sciences), 0.5 µl of
100 mg/µl of bovine serum albumin, 10× PCR buffer, and 0.5 ml (2.5 units) of Taq DNA polymerase from Roche Molecular
Biochemicals. The total reaction volume was 50 µl. The number of PCR
cycles determined to yield a linear, quantitative signal for MBP and
GAPDH were 30 and 21, respectively. Each reaction cycle consisted of
50 s at 94 °C for DNA denaturation, 45 s at the annealing
temperature for the primer pair (55 °C and 53 °C for mMBP1 and
GAPDH, respectively), and 45 s at 72 °C for extension.
Following PCR amplification, 5 µl of PCR buffer was analyzed by
electrophoresis on 6% native polyacrylamide gels.
| |
RESULTS |
PU.1, GATA-1, and C/EBP Isoforms Are Differentially
Expressed in Eosinophilic Cell Lines and Mature Blood Eosinophils and
Bind to the Their Consensus Sites in the Functional Region of the
MBP-P2 Promoter--
We identified a number of additional GATA-1 and
C/EBP consensus sites, as well as two PU.1 consensus sites not
characterized previously in the MBP-P2 promoter, using subsequence
homology searches of the bp 117 to +1 functionally active region
(Fig. 1). This is the region for which we
had shown previously the most 5' GATA-1 and C/EBP sites to be necessary
for MBP promoter activity in eosinophilic cell lines using deletional
and mutagenesis analyses (8, 9). We first determined the expression
patterns for GATA-1, PU.1, and C/EBP in AML14 eosinophilic cell
lines and mature peripheral blood eosinophils by Western blotting (Fig. 2), and their binding activity for the
multiple consensus sites was identified in the MBP-P2 promoter by gel
shifts (Fig. 3). Expression levels for
GATA-1, PU.1, and the various C/EBP isoforms was compared for AML14
parental myeloblasts, AML14.3D10 eosinophil myelocytes, and peripheral
blood eosinophils from normal donors, covering the range of
differentiation stages from undifferentiated progenitors to terminally
differentiated mature eosinophils. GATA-1 was well expressed both for
undifferentiated AML14 parental cells and differentiated AML14.3D10
eosinophilic myelocytes (Fig. 2A, lanes 1 and
2) but was undetectable in normal peripheral blood eosinophils (Fig. 2A, lane 3) from two different
donors (2nd donor not shown). K562 cells, which express abundant
endogenous GATA-1 protein, were used as the positive control (Fig.
2A, lane 4). This finding suggests that GATA-1
expression is significantly down-regulated during eosinophil terminal
differentiation, supporting prior results in the avian system (2, 7).
PU.1 was also expressed by both AML14 parental myeloblasts and
AML14.3D10 eosinophil myelocytes with AML14 < AML14.3D10 (Fig.
2B, lanes 1 and 2) but not detectable
in mature blood eosinophils and K562 cells (Fig. 2B,
lanes 3 and 4). However, much longer exposure times
revealed weak expression in both blood eosinophils and K562 cells (not shown). This finding indicates a significant reduction in PU.1 expression as eosinophils terminally differentiate. All four isoforms of C/EBP including C/EBP32/30,
C/EBP27, and C/EBP14 were well expressed
in AML14.3D10 eosinophilic myelocytes (Fig. 2C, lane
2), comparable with HL-60 promyelocytes (lane 4), but expression was lower in the AML14 myeloblast line (Fig. 2C,
lane 1). Importantly, only the C/EBP14
isoform is abundantly expressed in terminally differentiated blood
eosinophils (Fig. 2C, lane 3), a finding
confirmed using two different blood donors (Fig. 2C,
lanes 5 and 6). The high level expression of
C/EBP14 in blood eosinophils was initially detected as a
broad smear on Western blots using equal protein loading for all
nuclear extracts (Fig. 2C, lane 3). We therefore
reduced protein loading by a factor of 10 to show the
C/EBP14 signal as a single 14-kDa band in the blood
eosinophils from two different normal subjects (Fig. 2C,
lanes 5 and 6). Further reductions in protein
loading eliminated all isoforms of C/EBP expressed by AML14.3D10
cells but still showed a sharp single band for the
C/EBP14 isoform expressed by authentic blood eosinophils
(not shown), indicating very high level expression of
C/EBP14 in the mature eosinophil. These findings
demonstrate that GATA-1, PU.1, and C/EBP are all well expressed in
the AML14.3D10 eosinophil myelocyte cell line, which is therefore an
ideal model for functional studies of the roles of GATA-1, PU.1, and
C/EBP isoforms in vivo and in vitro. The
differential expression of the C/EBP isoforms in developing
eosinophil progenitors versus the mature, terminally differentiated cell may reflect their specific roles as activators and
then repressors of eosinophil gene expression during granulocyte development (see below and Fig. 11).
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 1.
Schematic of the functional 5'-upstream
region of the MBP-P2 promoter. Binding sites for three principal
transcription factors have been identified in the most functional bp
117 to +1 region of the MBP-P2 promoter by nucleic acid subsequence
analysis. These include two GATA, two C/EBP, and two PU.1 consensus
sites as indicated (top schematic). Mutations in the GATA-1,
C/EBP, and PU.1 sites were generated as indicated in the pXP2
luciferase reporter vector for transactivation (Fig. 7A) and
transfection (Fig. 7B) experiments.
|
|
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 2.
Differential expression of PU.1, GATA-1, and
C/EBP isoforms in eosinophil progenitor and
myelocytic cell lines compared with mature peripheral blood
eosinophils. Western blot analysis of the levels of PU.1, GATA-1,
and C/EBP isoforms using nuclear extracts from AML14 (eosinophil
myeloblast) and AML4.3D10 (eosinophil myelocyte) cell lines and mature
peripheral blood eosinophils (>99% purity) from two different normal
subjects. Western blots were probed with GATA-1 antibody
(A), PU.1 antibody (B), and C/EBP antibody
(C). K562 nuclear extract was used as a positive control for
GATA-1 and PU.1. HL-60 nuclear extract was used as a positive control
for C/EBP expression.
|
|
View larger version (40K):
[in this window]
[in a new window]
|
Fig. 3.
PU.1, GATA-1, and C/EBP
in AML14.3D10 eosinophil nuclear extracts bind to their consensus
sites in the MBP-P2 promoter. A, GATA-1 binds to the dual
GATA sites in the MBP-P2 promoter. The gel shift assay used the dual
GATA consensus site in the MBP-P2 promoter as probe. Two specific
complexes formed with nuclear extract from AML14.3D10 eosinophilic
myelocytes (lane 2). These complexes were completely
inhibited by the addition of unlabeled self-competitors (lane
3) and partially inhibited by an oligonucleotide containing only a
single GATA site (lanes 4 and 5) but not by a
dual GATA site mutation (lane 6). The protein-DNA complexes
were completely supershifted by the addition of antibody to GATA-1
(lane 7) but not by antibodies to PU.1 or the C/EBPs (,
, and ) (lanes 8-11). B, PU.1 binds to
both PU.1 consensus sites in the MBP-P2 promoter. The gel shift was
performed using the second PU.1 consensus site of the MBP-P2 promoter
as probe. Two protein-DNA complexes formed with nuclear extract of
AML14.3D10 cells as indicated by the arrows (lane
2). Although both complexes were completely inhibited by the
addition of unlabeled probe (the 2nd PU.1 site) (lane 3),
only the higher mobility complex (labeled PU.1) was specifically
eliminated by the addition of competitor containing the first PU.1 site
(lane 4) but not by other transcription factor-binding sites
(lanes 5-7). This PU.1 complex was also eliminated by the
addition of two different antibodies to PU.1 (lanes 8 and
9) but not by antibodies to GATA-1, C/EBP, -, or -
(lanes 10-13). C, C/EBP and C/EBP both
bind to the first (most 5') C/EBP consensus site in the MBP-P2
promoter. The gel shift was performed using both C/EBP consensus sites
of the MBP-P2 promoter as probes. However, no binding was detected
using the 2nd C/EBP site. Two specific complexes formed
with the first C/EBP consensus site using nuclear extract of AML14.3D10
eosinophils (lane 1, bands and ). Complex
formation was completely inhibited by the addition of unlabeled
competitors including self or the C/EBP consensus site of the M-CSFR
promoter (lanes 2 and 4, respectively) but not by
a C/EBP binding site mutation (lane 3). Antibody to C/EBP
completely removed and supershifted complex and partially removed
complex (lane 8). The addition of antibody to
C/EBP partially removed complex (lane 6).
Supershifted bands (S and S ) were
detected by addition of both C/EBP and C/EBP antibodies
(lanes 6 and 8) but not by antibodies to C/EBP
and C/EBP (lanes 5 and 7).
|
|
The bp 117 to +1 functional region of the MBP-P2 promoter possesses
two consensus sites each for GATA-1, PU.1, and the C/EBPs (Fig. 1). To
assess their ability to bind these factors, we performed gel and
antibody supershift analyses using nuclear extracts from the AML14.3D10
eosinophil myelocyte cell line and
[-32P]ATP-end-labeled oligonucleotides containing the
GATA-1, PU.1, and C/EBP consensus sites of interest (Fig. 3). GATA-1
binds to the two tandem GATA sites in the promoter (Fig.
3A). Two specific protein-DNA-binding complexes formed with
the AML14.3D10 nuclear extract using the dual GATA site probe (Fig.
3A, lane 2). Complex formation was completely
inhibited by the addition of a 100-fold molar excess of the unlabeled
dual GATA site oligonucleotide (lane 3), partially inhibited
by oligonucleotides containing only the first or second GATA site
(lanes 4 and 5) but not inhibited by oligonucleotides containing a dual GATA site mutation (lane
6). The GATA-binding complexes were completely supershifted by an antibody to GATA-1 (lane 7) but not by antibodies to PU.1 or
the various C/EBPs (lanes 8-11). These results demonstrate
that GATA-1 binds to both GATA consensus sites comprising the dual
GATA-1 site of the MBP-P2 promoter, with maximum binding to the dual site as compared with the single sites (data not shown).
For the PU.1 consensus sites, two major protein-DNA complexes formed
with nuclear extract from AML14.3D10 cells using the second PU.1
consensus site as a probe (Fig. 3B, lane 2). Formation of these complexes was completely inhibited by the addition of a
100-fold molar excess of the unlabeled probe (lane 3) but
only the higher mobility complex (PU.1) was removed by competition with
both the first (not shown) and second PU.1 sites (lane 4), indicating the specificity of this complex. This PU.1 complex was
selectively blocked by the addition of two different antibodies to PU.1
(lanes 8 and 9) but not by antibodies against
GATA-1 or the C/EBPs (lanes 10-13). These results
demonstrate that both PU.1 consensus sites in the MBP promoter have the
capacity to bind PU.1 expressed by AML14.3D10 eosinophils. For the
C/EBP consensus sites in the MBP promoter, both C/EBP and C/EBP
in AML14.3D10 eosinophil nuclear extracts formed specific protein-DNA
complexes with the first C/EBP site (Fig. 3C). Complex
formation was completely inhibited by the addition of a 100-fold molar
excess of an unlabeled C/EBP consensus site oligonucleotide matching
the MBP-P2 or M-CSFR promoters (lanes 2 and
4) but not by an oligonucleotide containing a C/EBP site
mutation (lane 3). One of the DNA-binding complexes was
completely supershifted by the addition of antibody to C/EBP (lane 8). A second, higher mobility binding complex was
partially removed by the addition of antibody to C/EBP or C/EBP.
No C/EBP or C/EBP complexes were detected using AML14.3D10
eosinophil nuclear extracts (lanes 5 and
7), whereas in equivalent gel shifts using AML14 parental
cell nuclear extracts, only C/EBP complexes could be identified by
antibody supershift (not shown). The second C/EBP consensus site in the
promoter did not form any complexes, nor did it inhibit complex
formation with the first C/EBP site (not shown). These results indicate
that C/EBP and C/EBP homodimers, as well as C/EBP-C/EBP
heterodimers, present in AML14.3D10 eosinophil nuclear extracts bind to
the first C/EBP site in the MBP-P2 promoter; this is the C/EBP site we
previously showed to be functionally transactivated by C/EBP (9) and
in the present study by C/EBP as well (see Fig. 9A).
Synergy Versus Antagonism between GATA-1 and PU.1--
As noted
above, a functional antagonism between GATA-1 and PU.1 has been
reported for the transcriptional regulation of myeloid genes such as
the M-CSFR (11), for which the promoter possesses a functional PU.1
site but no GATA sites, and GATA-1 antagonizes PU.1 transactivating
activity. Because the MBP-P2 promoter possesses both GATA-1- and
PU.1-binding sites, and the MBP-P2 promoter is strongly transactivated
by GATA-1, we sought to determine whether PU.1 would similarly
antagonize GATA-1 activity in our system. Transactivation experiments
were performed in CV-1 cells by transfection of the bp 117MBP-P2
luciferase reporter construct with GATA-1 and/or PU.1 expression
vectors (Fig. 4A). Because
GATA-1 has been shown previously (2, 7) to function as an activator at
low concentrations and repressor at high concentrations for an avian eosinophil promoter EOS47, we first performed dose-response titrations of the GATA-1 expression vector to determine the optimal concentration for maximal GATA-1 transactivation of the MBP-P2-pXP2 reporter construct (data not shown); this optimal amount was used for all subsequent co-transactivation assays. A PU.1-regulated M-CSFR promoter
construct, used previously to demonstrate GATA-1 antagonism of PU.1
transactivation (11), was also transfected into CV-1 cells along with
the same GATA-1 and/or PU.1 expression vectors in these experiments for
comparison (Fig. 4B). Contrary to the anticipated PU.1
antagonism of GATA-1, our results clearly show that MBP-P2 promoter
activity was induced ~20-fold by GATA-1 alone and ~60-fold by
GATA-1 plus PU.1, with PU.1 alone being essentially inactive (Fig.
4A).
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 4.
Synergy versus antagonism of
GATA-1 and PU.1. A, GATA-1 and PU.1 synergistically
transactivate the MBP-P2 promoter. CV-1 cells were transfected with 0.5 µg of the MBP-P2-pXP2 luciferase construct, 0.2 µg of GATA-1-pXM,
and 0.01 µg of PU.1-pECE expression vectors. Promoter activity was
measured 48 h after transfection and analyzed for fold induction
over the level of MBP-P2 base-line activity in the absence of
expression vectors. B, GATA-1 represses PU.1 transactivation
of the M-CSFR promoter. CV-1 cells were transfected using the same
transfection method as above with 0.5 µg of an M-CSFR promoter
luciferase construct and expression vectors for GATA-1 and PU.1 as
above. Promoter activity was determined 48 h after transfection
and analyzed for fold induction over the M-CSFR promoter base-line
activity. Means (± S.D.) are shown for four independent
co-transactivation experiments. C and D, the
synergy versus antagonism of GATA-1 and PU.1 is dependent
upon the expression level of PU.1 relative to GATA-1, an increased
amount of PU.1 reduces GATA-1/PU.1 synergistic transactivation of the
MBP-P2 promoter (C). A 10-fold increase in the amount of
PU.1 expression vector (0.1 µg) was compared with the amount (0.01 µg) used in transactivation experiments performed as in A.
The synergy between GATA-1 and PU.1 for the MBP-P2 promoter was
eliminated by the addition of an increased amount of the PU.1
expression vector (0.1 µg). D, an increased amount of PU.1
reverses GATA-1/PU.1 antagonism of the M-CSFR. A 10-fold increase in
the amount of PU.1 vector (0.1 µg) was compared with the amount (0.01 µg) used in the transactivation experiments performed as described in
B. The GATA-1 inhibition of PU.1-mediated transactivation of
the M-CSFR promoter was overcome by the addition of an increased amount
of the PU.1 expression vector. +, 0.01 µg; ++, 0.1 µg of
PU.1-pECE).
|
|
These results demonstrate a significant synergy of PU.1 for the
GATA-1-mediated transactivation of the MBP-P2 promoter, in contrast to
the antagonism of GATA-1 for the PU.1-mediated transactivation of the
M-CSFR promoter as reported previously (11) and confirmed in the
current "control" experiments (Fig. 4B). The findings
provide evidence for important promoter-specific and possible
lineage-specific differences in the combinatorial outcome of
GATA-1/PU.1 interactions in myeloid gene regulation in general and
eosinophil gene regulation in particular during eosinophil development.
To further evaluate the synergistic activities of GATA-1 and PU.1 for
transactivation of the MBP-P2 promoter, we performed dose-response
analyses using different amounts of PU.1 expression vector at an
optimal concentration of the GATA-1 expression vector. The M-CSFR
promoter was again evaluated in these experiments as a control for a
PU.1-regulated gene that is antagonized by GATA-1. Surprisingly, we
found that the synergistic transactivation of the MBP-P2 promoter by
GATA-1 and PU.1 occurred in the presence of low levels of PU.1 but was completely abrogated by high levels of PU.1 (Fig. 4C). In
contrast, PU.1-mediated transactivation of the M-CSFR promoter was
antagonized by GATA-1 in the context of a low level of PU.1 but
overcome by a high level of PU.1 (Fig. 4D). In addition, a
positive interaction between PU.1 and GATA-1 was also obtained in
transactivation experiments performed in the K562 erythroleukemia cell
line that expresses high levels of endogenous GATA-1, such that low
levels of the PU.1 expression vector doubled MBP-P2 promoter activity
due to endogenous GATA-1, whereas high levels of PU.1 had no effect
(Fig. 5). These findings suggest that the
level of PU.1 expression is critical in determining whether GATA-1 and
PU.1 functionally antagonize or synergize in the regulation of PU.1
versus GATA-1 target genes in the myeloid lineages.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 5.
PU.1 augments MBP-P2 promoter activity in the
GATA-1-expressing K562 cell line. K562 cells were transfected by
electroporation with the MBP-P2 promoter construct along with a low
(0.1 µg) or high (1.0 µg) dose of the PU.1-pECE expression vector
and the pRL-CMV (Renilla luciferase) control vector. Mean
promoter activity (± S.D.) relative to the promoterless pXP2 control
luciferase vector is shown for three independent experiments.
Luciferase activity was determined 6 h after transfection and
normalized using the dual luciferase method.
|
|
C/EBP27 Represses GATA-1 Transactivation
and Blocks GATA-1/PU.1 Synergy in the Activation of the
MBP-P2 Promoter--
We have shown that all four C/EBP isoforms are
expressed by AML14.3D10 eosinophilic myelocytes (Fig. 2) and that
C/EBP in nuclear extracts of these cells binds the C/EBP consensus
site of the MBP-P2 promoter (Fig. 3C). As well, we reported
previously (9) that C/EBP functionally synergizes with and
physically interacts with GATA-1. We therefore sought to characterize
the functional activities of the different C/EBP isoforms in terms of both direct transactivating activity and combinatorial effects with
GATA-1 and PU.1. Transactivation experiments were carried out in CV-1
cells by co-transfection of the bp 117-MBP-P2 promoter construct with
expression vectors for C/EBP32, C/EBP27,
and C/EBP14, and/or GATA-1, and/or PU.1 as shown in Fig.
6. The MBP-P2 promoter was transactivated
by GATA-1 alone but not by any of the individual C/EBP isoforms
(Fig. 6, A and B). In contrast, GATA-1-induced transactivation was significantly inhibited by the addition of C/EBP27 but not by the C/EBP32 or
C/EBP14 isoforms (Fig. 6A). Furthermore, the
synergistic activity of GATA-1 and PU.1 (~75-fold induction) was
completely suppressed by the addition of the C/EBP27
isoform, but not by the other C/EBP isoforms (Fig. 6B).
Importantly, the repressor activity of the C/EBP27
isoform was not unique to the individual or combinatorial activities of
GATA-1 and GATA-1/PU.1 on MBP-P2 promoter activity, as PU.1-mediated transactivation of the M-CSFR promoter was likewise blocked by C/EBP27 in these experiments (Fig. 6C). In
addition, the C/EBP27 isoform completely blocked PU.1
synergy with endogenous GATA-1 in the K562 cell line in which the
MBP-P2 promoter is active (Fig. 6D). The
C/EBP27 isoform contains both transactivation and
putative repression domains (40, 61, 63). Its negative regulatory
activity in our system may be mediated by the repression domain through
either binding to the MBP promoter C/EBP site directly or by
protein-protein interactions that interfere with GATA-1 DNA binding,
transcriptional activity, or the physical interactions of GATA-1 with
PU.1.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 6.
C/EBP27 inhibits GATA-1 and
GATA-1/PU.1 synergistic transactivation of the MBP-P2 promoter and PU.1
transactivation of the M-CSFR promoter. A,
C/EBP27 inhibits GATA-1 transactivation of the MBP-P2
promoter. CV-1 cells were transfected with 0.5 µg of the MBP-P2-pXP2
luciferase reporter and also co-transfected with or without 0.2 µg of
GATA-1 and 0.5 µg of C/EBP32, C/EBP27,
or C/EBP14 expression vectors as indicated. Promoter
activity is shown as the fold induction over the MBP-P2-pXP2 reporter
alone. GATA-1 transactivation was repressed by the addition of the
C/EBP27 but not by C/EBP32 or
C/EBP14 isoforms. B, C/EBP27
isoform inhibits GATA-1/PU.1 synergistic transactivation of the MBP-P2
promoter. Transactivation of the MBP-P2 promoter by GATA-1, PU.1, and
the C/EBP isoforms was performed as described above. Synergistic
transactivation of the MBP-P2 promoter by GATA-1/PU.1 was also blocked
by the addition of the C/EBP27 but not
C/EBP32 or C/EBP14 isoforms.
C, C/EBP27 inhibits PU.1-mediated
transactivation of the M-CSFR promoter. Transactivation of a
PU.1-responsive M-CSFR-pXP2 reporter construct was performed in CV-1
cells as described in Fig. 4B. 0.01 µg of PU.1 expression
vector and/or 0.5 µg of C/EBP27 expression vector were
used in the co-transactivation. PU.1 transactivation of the M-CSFR
promoter was fully repressed by the expression of the
C/EBP27 isoform. D, the
C/EBP27 isoform blocks PU.1-mediated augmentation of
MBP-P2 promoter activity in the K562 cell line. K562 cells were
transfected by electroporation with the MBP-P2 promoter construct, a
low dose of PU.1 (0.1 µg), and the C/EBP27 expression
vector as indicated, along with the pRL-CMV (Renilla
luciferase) control vector. Mean (± S.D.) promoter activity is
expressed as the fold activity relative to the promoterless pXP2
luciferase control vector. Luciferase activity was determined 6 h
after transfection and normalized using the dual luciferase
method.
|
|
We showed previously that the synergistic interactions of GATA-1 and
C/EBP likely occur through protein/protein interaction in that they
required either a functional GATA- or C/EBP-binding site but not both
(9). As well, GATA-1 and PU.1 have been shown to antagonize
functionally one another's functions via protein/protein interactions,
whether the target genes contain GATA-1- or PU.1-binding sites (11,
84). Because GATA-1, PU.1, and C/EBP can all bind to their
respective sites in the MBP-P2 promoter (Fig. 3), we sought to
determine the functional relevance of these binding sites in the
combinatorial interactions of these transcription factors on this
eosinophil promoter. Constructs containing mutations in the dual GATA
site, the two C/EBP sites, or both PU.1 sites were generated by PCR
mutagenesis using the wild type MBP-P2 promoter construct as template
and tested in both CV-1 cells (Fig.
7A) and the AML14 eosinophil
cell lines (Fig. 7B). Transactivation analyses were
performed in CV-1 cells as described above by co-transfection of the
MBP-P2 mutant and wild type constructs with GATA-1, PU.1, and
C/EBP27 expression vectors (Fig. 7A).
Activity of the MBP-P2 promoter was essentially eliminated by the
mutation of the C/EBP sites or the GATA sites, indicating a requirement
for both sites in this promoter. No GATA-1-induced transactivation or
GATA-1/PU.1 synergistic transactivation was seen for either the C/EBP
or GATA site mutants. In contrast, mutation of the two PU.1 sites in
the MBP-P2 promoter was essentially without effect, and both
GATA-1-induced transactivation and GATA-1/PU.1 synergy was equivalent
to that obtained for the wild type promoter. To confirm this finding
for the endogenously expressed factors in eosinophils, the same
MBP-P2 promoter wild type and mutant constructs were transiently
transfected by electroporation into the AML14.3D10 eosinophil and AML14
parental cell lines (Fig. 7B). The results confirmed that
mutation of the PU.1-binding sites was essentially without effect on
MBP-P2 promoter activity compared with the wild type promoter in both
cell lines, whereas mutation of either the GATA sites or C/EBP sites
resulted in a significant loss of activity. We conclude from these
analyses that the synergistic activity of PU.1 on GATA-1-mediated
transactivation occurs independently of direct PU.1 binding
to either of the two PU.1-binding sites in the MBP-P2 promoter. These
results imply that the synergy of PU.1 and GATA-1 in this system is
mediated via protein/protein interactions between these factors and
binding of the complex through the dual GATA-binding site in the
promoter. These results also confirm our prior findings that the C/EBP
and GATA sites are both critical for transcriptional activation of this
promoter by transactivation in either heterologous cell lines or by the
endogenous factors in eosinophil progenitor lines (8, 9).
View larger version (33K):
[in this window]
[in a new window]
|
Fig. 7.
Activity of the MBP-P2 promoter is eliminated
by mutation of the GATA or C/EBP sites but not the PU.1 sites; analysis
by transactivation in CV-1 cells and transfection of AML14.3D10
eosinophilic myelocytes. A, CV-1 cells were transfected with
0.5 µg of the MBP-P2 wild type or mutant constructs including a dual
C/EBP site mutation, dual GATA-1 site mutation, or dual PU.1 site
mutation as indicated and shown schematically in Fig. 2.
Transactivation experiments were performed with or without the addition
of 0.2 µg of GATA-1, 0.01 µg of PU.1, or 0.5 µg of
C/EBP27 expression vectors as indicated. Promoter
activity is shown as fold induction over the base-line MBP-P2-pXP2
construct alone. Transactivation of the MBP-P2 promoter by GATA-1 or
GATA-1/PU.1 synergy and its repression by C/EBP27 was
blocked by mutation of the C/EBP or GATA-binding sites but not the PU.1
sites. B, AML14 and AML14.3D10 cells were transiently
transfected by electroporation using 10 µg of promoter constructs
including the MBP-P2 wild type or dual GATA-site, C/EBP site, or PU.1
site mutation constructs as above. Luciferase activity was measured
6 h after transfection, normalized, and analyzed as fold
differences compared with the promoterless pXP2 luciferase plasmid. The
mean (± S.D.) for three independent transactivation or transfection
experiments is shown.
|
|
Endogenous GATA-1, PU.1, and C/EBP Isoforms Expressed
by Eosinophils Physically Interact; Analysis by
Co-immunopre-cipitation--
To evaluate further the physical
interactions that may occur among GATA-1, PU.1. and C/EBP in
vivo in the eosinophil lineage, we performed co-IP experiments
using whole cell lysates of the AML14.3D10 eosinophil myelocyte cell
line that expresses all three transcription factors. Whole cell lysates
were first immunoprecipitated with antibodies to C/EBP, GATA-1, or
PU.1 and two control antibodies including non-immune rabbit and goat
IgGs. Immunoprecipitates were analyzed by Western blotting using
antibodies against C/EBP and GATA-1 as shown in Fig.
8. As shown on the C/EBP Western blot
(Fig. 8A), antibody to C/EBP strongly immunoprecipitated both the C/EBP32/30 and C/EBP27 isoforms
(lane 2). Antibodies to PU.1 and GATA-1 both
co-immunoprecipitated the C/EBP32/30 and
C/EBP27 isoforms (Fig. 8A, lanes 3 and 5) but not the C/EBP14 isoform (not
shown). As shown on the GATA-1 Western blot (Fig. 8B),
GATA-1 was efficiently immunoprecipitated by antibody to GATA-1
(lane 5), and antibodies to C/EBP and PU.1 both
co-immunoprecipitated GATA-1 (lanes 2 and 3,
respectively). We were unable to co-immunoprecipitate PU.1 with
antibodies to either C/EBP or GATA-1 (data not shown); the low level
expression of endogenous PU.1 in AML14.3D10 cells likely precludes its
visualization in these co-immunoprecipitates. Alternatively, antibodies
to GATA-1 and C/EBP may block their interactions with PU.1. Our
co-immunoprecipitation results demonstrate for the first time that
C/EBP physically interacts with both PU.1 and GATA-1 in
vivo. Our results indicate that C/EBP, GATA-1. and PU.1
physically interact with one another in vivo in AML14.3D10 eosinophil myelocytes.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 8.
Co-immunoprecipitation of
C/EBP, PU.1, and GATA-1 from lysates of
AML14.3D10 eosinophilic myelocytes. Whole cell lysates from
AML14.3D10 cells were immunoprecipitated or co-immunoprecipitated with
anti-C/EBP, PU.1, and GATA-1 antibodies. Non-immune, normal rabbit
IgG (lane 4), or goat IgG (lane 6) was used as
the negative control. Twenty micrograms of lysate from AML14.3D10 cells
was loaded on each gel to show input (lane 1). A,
Western blot analysis was performed using anti-C/EBP antibody.
C/EBP32/30 and C/EBP27 were
immunoprecipitated by rabbit polyclonal C/EBP antibody (lane
2) but not by normal rabbit IgG (lane 4). These
C/EBP isoforms were also co-immunoprecipitated with antibodies to
both PU.1 (lane 3) and GATA-1 (lane 5).
B, Western blot analysis was performed using GATA-1
antibody. GATA-1 was immunoprecipitated by goat polyclonal antibody to
GATA-1 (lane 5) but not by non-immune, normal goat IgG
(lane 6). GATA-1 was co-immunoprecipitated with antibodies
to both C/EBP and PU.1 (lanes 2 and 3).
|
|
Activation Versus Repression of Eosinophil Gene Transcription by
C/EBP Family Members--
Previous studies (4, 52) have
shown that granulopoiesis is impaired in C/EBP- or
C/EBP-deficient mice. C/EBP has been shown to play a critical
role in early aspects of myeloid lineage commitment and gene
transcription, whereas C/EBP has been shown to be critical for the
more terminal stages of lineage differentiation and maturation (4, 40).
To define further the transcriptional regulation of the MBP gene during
the various stages of eosinophil development by the members of the
C/EBP family, we sought to characterize their combinatorial
interactions in the activation of the MBP promoter. Co-transactivation
analyses were performed in CV-1 cells by co-transfection of the MBP-P2
promoter construct with expression vectors for C/EBP, C/EBP, and
the three C/EBP isoforms (32, 27, and
14). C/EBP or C/EBP equivalently transactivated
the MBP-P2 promoter with 4-8-fold inductions, and these activities
were competitively inhibited by the addition of the
C/EBP14, C/EBP27, and
C/EBP32 isoforms as shown in Fig.
9, A and B,
respectively. The C/EBP14 isoform in particular, which
has no transactivating activity in this system (Fig. 6), inhibited both
C/EBP- and C/EBP-induced transactivation in a
dose-dependent manner, possibly by forming inactive
heterodimers with these C/EBPs through their shared leucine zipper
dimerization domains. The C/EBP32 and
C/EBP27 isoforms, which likewise have no inherent
transactivating activity of their own for the MBP-P2 promoter (Fig.
6.), may also form inactive heterodimers or homodimers that compete for
occupation of the functional C/EBP site in the MBP-P2 promoter,
decreasing transactivation by C/EBP or -.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 9.
C/EBP- and
C/EBP-mediated transactivation of the MBP-P2
promoter is inhibited by the C/EBP
isoforms. Transactivation of the MBP-P2 promoter was
performed in CV-1 cells by the addition of C/EBP (A) or
C/EBP (B) and increasing amounts (0.5, 1.0, and 2.0 µg
of plasmid DNA) of the C/EBP14, 27, and
32 isoform expression vectors. Luciferase activity was
assayed 72 h after transfection. The mean (± S.D.) for three
independent co-transactivation experiments is shown. All C/EBP
isoforms repressed transactivation of the MBP-P2 promoter by C/EBP
and C/EBP to a greater or lesser extent. Only the
C/EBP14 isoform inhibited the activity of both C/EBP
and - in a dose-dependent fashion.
|
|
Targeted Disruption of the C/EBP Gene Results in Only
a Partial Decrease in MBP Gene Expression--
Mice with a targeted
disruption of the C/EBP gene fail to produce normal mature
granulocytes in their bone marrow or peripheral blood, but they instead
produce neutrophils that have atypical nuclear morphology, lack
expression of secondary and tertiary granule proteins, and have
functional abnormalities. As well, mature eosinophils are not
detectable in these mice using standard histochemical staining with
Wrights'-Giemsa (4). By comparison, mice with a targeted disruption of
the C/EBP gene have a profound absence of both neutrophils and
eosinophils in the fetal liver (52). In C/EBP null mice, we were
unable to detect any expression of MBP whatsoever, consistent with a
complete block in granulocyte development (Fig.
10). In addition, there was no
expression of the gene encoding murine eosinophil peroxidase and
decreased expression of the gene encoding the IL-5R subunit (not
shown).3 In contrast, in the
C/EBP null mice, there was only a partial (~50%) decrease in MBP
mRNA expression (Fig. 10), with only a partial decrease in murine
eosinophil peroxidase, and no effect whatsoever on IL-5R subunit
expression (not shown). Thus, although C/EBP knockout mice do not
produce any terminally differentiated mature eosinophils, they must
still produce eosinophil progenitors capable of expressing secondary
granule protein genes such as MBP and eosinophil peroxidase, albeit at
somewhat reduced levels.
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 10.
Effects of targeted disruption of the
C/EBP and C/EBP genes
on eosinophil MBP gene expression. Semi-quantitative reverse
transcriptase-PCR analysis of MBP mRNA expression in C/EBP- and
C/EBP-deficient mice is shown. The GAPDH gene was used as a control
for semi-quantitation of the mRNA (cDNA) sample input. The
mRNA was prepared from the fetal liver (for C/EBP) or bone
marrow (for C/EBP) of wild-type (+/+), heterozygote (+/), and
knockout mice (/), respectively. For the wild-type and C/EBP
knockout analysis (right panels), mice were first stimulated
intraperitoneal with thioglycolate broth for 4.5 or 24 h prior to
harvesting bone marrow cells as describe |
|
TOP
ABSTRACT
INTRODUCTION
