associated, differentiation-induced factor (DIF) as a potential molecular link ... gem.univie.ac.at. 25, 26 ... binding sites for proteins rapidly induced by epidermal.
MOLECULAR AND CELLULAR BIoLOGY, Feb. 1994, p. 1364-1373 0270-7306/94/$04.00+0 Copyright C) 1994, American Society for Microbiology
Vol. 14, No. 2
A Factor Induced by Differentiation Signals in Cells of the Macrophage Lineage Binds to the Gamma Interferon Activation Site ANDREAS EILERS,1 MANUELA BACCARINI,2 FRIEDEMANN HORN,3 ROBERT A. HIPSKIND,4 CHRIS SCHINDLER,s AND THOMAS DECKER'* Department of Cell and Microbiology, Institute for Microbiology and Genetics, Vienna Biocenter, A-1030 Vienna, Austria1; Fraunhofer Institute for Toxicology and Molecular Biology, 2 and Institute for Molecular Biology, Hannover Medical School,4 D-30625 Hannover, and Institute for Biochemistry, RJWTHJAachen, D-52057Aachen,3 Germany; and Division of Molecular Medicine, Department of Medicine, College of Physicians and Surgeons of Columbia University, New York, New York 100325 Received 7 July 1993/Returned for modification 30 August 1993/Accepted 9 November 1993
Rapid transcriptional induction of genes in response to gamma interferon (IFN-y) is mediated by the IFN-y activation site (GAS) and its cognate protein, the IFN-y activation factor (GAF). We describe a GASassociated, differentiation-induced factor (DIF) as a potential molecular link between the activities of IFN-y and of growth and differentiation factors. DIF DNA binding was activated by colony-stimulating factor 1 in murine macrophages and also during tetradecanoyl phorbol acetate-induced differentiation or IFN-y treatment in myeloid U937 cells. IFN-y activation of DIF decreased significantly upon monocytic differentiation. DIF binding to DNA was inhibited by antiphosphotyrosine antibodies and could be induced by treatment of U937 cells with vanadate. Unlike GAF, DIF-DNA complexes did not contain the 91-kDa protein (p91) from ISGF-3. DIF bound with high affinity to GAS from the promoters of the IFP 53/tryptophanyl-tRNA synthetase and Fc-yRI genes, intermediate affinity to the Ly6A/E GAS, and low affinity to the guanylate-binding protein GAS. DIF may belong to a family of cytokine- or growth factor-induced factors binding with variable affinities to GAS-related elements: the interleukin-6-responsive acute-phase response factor associated with GAS from different IFN-inducible promoters but with a different preference of binding compared with DEF. The sis-inducible elemeit of the c-fos promoter bound GAF but not DIF. However, the sis-inducible element could be changed by point mutation to compete for GAF and DIF binding. Our data show DIF to be a novel DNA-binding protein which is activated in response to differentiating signals. Moreover, they suggest that a family of cytokine- or growth factor-regulated proteins integrates and coordinates the responses to cytokines and to growth and differentiation factors by binding to GAS-related elements.
Interferons (IFNs), and in particular gamma IFN (IFN-y), exert pleiotropic biological effects. IFN--y not only influences the host immune response through its antiviral properties and its ability to activate macrophages but also controls cellular growth and promotes differentiation (40, 55). In
25, 26, 29). ISGF-3 is an IFN-a-responsive transcription factor which consists of distinct polypeptides of 113, 91, 84, and 48 kDa (15, 26). The 113-, 91-, and 84-kDa species are activated through IFN-a-induced signals, leading to phosphorylation on tyrosine. Tyrosine phosphorylation, in turn, triggers their association with each other and the 48-kDa protein to form a transcriptionally competent complex (15, 18, 45). In contrast, only the 91-kDa protein is phosphorylated on tyrosine after IFN-y stimulation (48, 49). As an IFN--y activation factor (GAF), p91 displays a distinct DNA sequence specificity by binding to DNA either alone or complexed with p48 (47, 48). It is currently unclear whether p91-containing protein complexes may consist of subunits distinct from those found in ISGF-3. The DNA element recognized by the GAF is known as the IFN--y activation site (GAS) (30). GAS sequences of different IFN-regulated promoters vary considerably (53). Sequences of response elements for other polypeptide ligands suggest that GAS-like elements may be involved in a number of signaling pathways besides that for IFN--y. For example, the binding sites for proteins rapidly induced by epidermal growth factor (EGF) or platelet-derived growth factor (PDGF) (sis-inducible factor [SIF] [42, 57]), IL-6 (acutephase response factor [APRF] [58]) and/or lactogenic hormones (MGF [46]) are potential GAF-binding sites. This raises the intriguing possibility that slight variations of a principal sequence element determine the responsiveness of
myeloid progenitors like the U937 cell line, IFN-y induces growth arrest, produces a partially differentiated phenotype, and synergizes with other stimuli in generating the fully differentiated, monocytic phenotype (20, 21, 36, 54). Therefore, some of the genes regulated by IFN--y are likely to exert control on the cell cycle and to be part of a differentiation program. Such genes may in fact represent common targets of IFN-,y and growth and differentiation factors. Cytokines or growth factors can rapidly influence the DNA-binding or transactivation activities of transcription factors and thereby regulate gene transcription (24). Type I and type II IFNs (IFN-a and -1 and IFN--y, respectively) have been shown to rapidly activate latent, cytoplasmic DNA-binding proteins. Once activated, they translocate to the nucleus and induce expression of specific genes (9, 10, * Corresponding author. Mailing address: Institute for Microbiology and Genetics, Department of Cell and Microbiology, Vienna Biocenter, Dr. Bohr-Gasse 9, A-1030 Vienna, Austria. Phone: (1) 79515-4518. Fax: (1) 79862-24. Electronic mail address: decker@
gem.univie.ac.at. 1364
VOL. 14, 1994
genes to external stimuli, as well as allowing the possibility
of cross-talk between signaling pathways. One of the main biological activities of IFN--y is to cause macrophages to progress into an activated, immunocompetent state. The promonocytic human cell line U937 is an attractive tool with which to study macrophage biology. U937 cells can be induced to differentiate into monocytes, which allows monitoring of changes in cytokine responsiveness that accompany differentiation. We have recently reported that differentiation of U937 cells changes their potential to activate IFN-responsive transcription factors, in particular GAF and ISGF-3 (12). We describe here a DNAbinding factor, differentiation-induced factor (DIF), induced during monocytic differentiation of U937 cells, that binds to a subset of GAS from different IFN-inducible promoters. DIF can also be induced by IFN--y treatment of U937 promonocytes as well as by treatment of mature murine macrophages with the growth and differentiation factor colony-stimulating factor 1 (CSF-1). We have studied signaling events leading to DIF activation and the relationship of DIF to GAF, SIF, and APRF. Our data support a model in which related DNA elements create a regulatory network capable of differentially responding to external stimuli. MATERIALS AND METHODS Cells and reagents. U937 cells (ATCC CRL 1593; kindly provided by K. Resch, Hannover Medical School) were cultured in RPMI medium supplemented with 10% fetal calf serum (FCS). For differentiation, U937 cells were transferred from petri dishes to tissue culture dishes at a concentration of 5 x 105 cells per ml and treated with a final concentration of 5 nM 12-O-tetradecanoylphorbol 13-acetate (TPA; Sigma). The murine macrophage cell line BAC-1.2F5 has been previously described (32). The cells were maintained in Dulbecco's modified Eagle's medium supplemented with 15% FCS and 1.32 nM human recombinant CSF-1. BAC1.2F5 cells were starved by incubation in growth factor-deprived medium for 18 to 24 h prior to CSF-1 treatment. Human hepatoma (HepG2) cells were cultured in Dulbecco's modified Eagle's medium and Ham's F12 medium (1:1, vol/vol) supplemented with 10% FCS. For generation of APRF-containing extracts, the cells were treated with 100 U of recombinant interleukin-6 (IL-6) for 15 min. Recombinant human IFN--y was kindly provided by P. von Wussow and J. Roesler (Hannover Medical School) and used at a final concentration of 1 ng/ml. Incubation times are indicated for the individual experiments. Recombinant human CSF-1 was from Chiron Corp., Emeryville, Calif., and was used at a concentration of 10,000 U/ml. This concentration was previously shown to cause a maximal CSF-1 response in murine BAC-1.2F5 cells (3). Antibodies. The antiserum against the C-terminal or N-terminal portion of the 91-kDa protein of ISGF-3 and the antiserum with specificity for the 113-kDa protein have recently been described (16, 44). The antiserum recognizing the SH2 domains of both p91 and p84 was generated by immunization of rabbits with a glutathione S-transferase fusion protein comprising amino acids 504 to 596 of p91. The antiphosphotyrosine monoclonal antibody (MAb) Ab-1 was purchased from Dianova (Hamburg, Germany). As a control for Ab-1, a monoclonal immunoglobulin G2b purified from the supernatant of the murine myeloma MPC11 (Coulter Clone) was used in identical amounts. Oligonucleotides. For gel shift experiments and competitions, the following double-stranded oligonucleotides were
DIFFERENTIATION AND IFN-y RESPONSE
1365
used (the bottom strand is shown for IFP 53, since the GAS element in the IFP 53 promoter has an inverse orientation with respect to all others; the top strands are shown for the other GAS- or GAS-related elements): IFP 53, 5'GATCAATCACCCAGATFCTCAGAAACACTT-3'; IFP 53C, 5'-GATCCAGATTCTCAGAAA-3'; IFP 53A, 5'-GAT CATTCTCAGAAA-3'; Ly6, 5'-GATCAATlTATGCATAT TCCTGTAAGTGAC-3'; guanylate-binding protein (GBP), 5' -GATCCGTCAGTTTCATATTACTCTAAATCCA-3'; Fc-yRI (kindly provided by Phillip Benech, Institut Curie, Paris, France): 5'-GATCTGT'TTCAAGGAT'lTGAGATGT ATFT'ITCCCAGAAAAGGAAC-3'; c-fos sis-inducible element (SIE), 5'-AATTCAGCAGTTCCCGTCAATCCCTC CCG-3'; SIE 67, AATTCAGCA'TTCCCGTAAATCCCTC CCG 3'; SIE m34, AATTCAGCAGTTCACGTCAATC CCTCCCG-3'; human a2-macroglobulin promoter acutephase response element (APRE): 5'-GCTGTACGGTAAA AGTGAGCTC'TTACGGGAATGGGAAT-3'; rat a2-macroglobulin, 5'-GATCC(TCTGGGAATTCCTA-3'; and rat ac2-macroglobulin mutant, 5'-GATCCTITCTCTAGATTCC TA-3'. Cell extracts. The generation of nuclear and cytoplasmic extracts from cells after hypotonic lysis, based on the method of Dignam et al. (11), has been described elsewhere (29). Alternatively to breaking cells in hypotonic buffer by Dounce homogenization, we have used the method of Lee et al. (28), in which the cells are lysed by forcing them repeatedly through a 25-gauge hypodermic needle. If separation into nuclear and cytoplasmic extracts was not required, whole-cell extracts (WCE) were prepared as described by Andrews and Faller (1). Mobility shift assay. To measure the association of DNAbinding proteins with different DNA sequences, the doublestranded oligonucleotides described above were labeled with 32P by filling in the ends with Klenow polymerase and radiolabeled deoxynucleotides. The protein-binding reaction was in a buffer containing 20 mM [N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES; pH 7.9), 40 mM KCl, 1 mM MgCl2, 0.1 mM EGTA, 0.5 mM dithiothreitol, 10% (vol/vol) glycerol, and poly(dI-dC) (250 ng/ml), to which 0.5 ng of 32P-labeled oligonucleotide (about 105 cpm) was added. Protein-DNA complexes formed after 30 min at room temperature were analyzed on a 6% acrylamide gel at 4°C, using 0.2x Tris-borate-EDTA as a running buffer. To prevent complex formation, anti-p91 antiserum or control serum was added to the binding reaction in a final dilution of 1:100 unless otherwise indicated. Antiphosphotyrosine antibody was added in the indicated amounts. Western blotting (immunoblotting). Before the proteinDNA complexes were transferred to nitrocellulose, one control lane was removed from the mobility shift gel and processed for autoradiography. The resulting autoradiograph served to align the antibody-stained bands on Western blots with the positions of protein-DNA complexes in the mobility shift assay. The remaining gel was equilibrated in transfer buffer (40 mM 6-aminocapronic acid, 20% methanol, 0.01% sodium dodecyl sulfate) for 30 min. Subsequently, proteins and protein-DNA complexes were transferred to nitrocellulose by using a semidry blotting procedure. The nitrocellulose was incubated in blocking buffer (1 h at room temperature) and the appropriate dilutions of first antibody (1 h at room temperature). Bound first antibody was detected by using biotinylated second antibody and streptavidinbiotinylated alkaline phosphatase as recently described (12).
1366
EILERS ET AL. TPA (h) 0 3 - _
MOL. CELL. BIOL. 6
12
_
-
F 18 24 48 72 -
-
- O
IFN--
+-
CSF-1 (min)I.
I
-* 1
5 .
.
10 15 30 360 *
-
.
.
DIF 4
v- DIF GAF v
_
u
k:
.4
'jo -
*,,#
:0.
..
FIG. 1. DIF binds to the GAS of the gene encoding IFP 53/ tryptophanyl-tRNA synthetase. WCE of U937 cells after various times of TPA treatment were used in mobility shift assays with an IFP 53 GAS oligonucleotide.
RESULTS Activation of a GAS-binding factor during differentiation of U937 cells. The IFP 53 gene encodes a tryptophanyl-tRNA synthetase and is transcriptionally activated by IFN--y via a GAS in its promoter. In HeLa cell extracts, an IFP 53 GAS oligonucleotide can interact with one IFN--y-regulated nuclear protein that we have identified as GAF/p91 (53). In U937 cell extracts, binding of a factor to the IFP 53 GAS was induced during TPA-mediated differentiation (Fig. 1). This factor, DIF, was at its highest levels during the differentiation process and decreased once the cells had completed progression into the monocytic stage (48 h). Induction of DIF was blocked by cycloheximide and thus required active protein synthesis (data not shown). The factor was not detected in extracts from TPA-treated HeLa cells (data not shown). To determine whether DIF induction was a specific TPA response in macrophages or whether TPA might induce a response typical of physiological differentiation factors, we conducted experiments with the CSF-1-dependent mouse macrophage cell line BAC-1.2 F5. CSF-1 not only regulates growth and survival of cells committed to the macrophage lineage but also stimulates their differentiation (4, 52, 56). U937 cells produce CSF-1 during differentiation, and this factor may in fact serve as an autocrine stimulator of differentiation (31). Treatment of growth factor-starved macrophages with CSF-1 induced a DNA-binding factor indistinguishable from DIF (Fig. 2). In contrast, GAF/p91 activity was not observed in extracts from CSF-1-treated cells (Fig. 2). U937 cell DIF and the activity derived from CSF-1treated BAC-1.2F5 cells had identical mobilities when run side by side (data not shown) and showed the same relative mobility toward GAF (Fig. 2 and 3). They also behaved identically in competition experiments (see Fig. 6a and b). In contrast to TPA-treated U937 cells, DIF was induced rapidly and transiently in BAC-1.2F5 cells by CSF-1. DIF is activated by IFN-y. The ability of DIF to bind an IFN-y response element prompted us to investigate whether DIF activation occurred in cells treated with this cytokine. IFN--y treatment of U937 cells produced three distinct activities binding specifically to the IFP 53 GAS (Fig. 3). Two of these were identified as DIF and GAF, while the third one, y--y, with a slightly faster mobility than DIF, appeared to be
FIG. 2. Induction of DIF in murine BAC-1.2F5 macrophages by CSF-1. CSF-1-starved cells were treated with the growth/differentiation factor for the indicated periods. WCE were prepared and used in a mobility shift assay with the IFP 53 GAS as a probe. Control extracts were from untreated cells or cells after a 15-min incubation with murine IFN-,y. a novel activity. Y-y is not a major object of this study and will be described in more detail in future reports. Interestingly, the relative amount of the three GAS-associated proteins after IFN--y treatment depended on the differentiation status of the cells: y--y was maximally induced before the onset of differentiation, DIF was maximally induced during differentiation, and GAF was maximally induced after terminal differentiation had occurred (Fig. 4) (12). The rapid induction of DIF by IFN-,y suggests that the factor is preformed in U937 cells. Thus, the delayed type of response observed after TPA treatment might be due either to the TPA-induced synthesis of an endogenous ligand with DIFinducing activity (possibly CSF-1; see below) or to the synthesis of signaling components which are required for DIF activation in response to differentiation. IFN--y treatment induced very little DIF DNA-binding activity in BAC1.2F5 cells, probably because these cells are in a mature stage of differentiation (Fig. 2). This would preclude the
IFN-y
-
TPA
+
i
0
+
+
-
s
_
m
DIF I. yi,
I
*
GAF
01
4
FIG. 3. IFN-,y induces three distinct IFP 53 GAS-binding activities. U937 cells before or after differentiation (48 h of TPA treatment) were either left untreated or stimulated with IFN-y for 30 min. Nuclear extracts were tested in a mobility shift assay with the IFP 53 GAS oligonucleotide as a probe.
DIFFERENTIATION AND IFN-y RESPONSE
VOL. 14, 1994 a
IFN-TPA (h)
1367
0 M-
3
6 1 2 20 48
anti p-tyr (1g) control mAb(g)-
-
1 -
2 -
-
1
2
3
4
5
____ s&f af
-
3
1
2
3
-
1 -
2 -
3 -
6
7
8
9
10
11
12
DIF
- Y-Y
DIF
Y-7
--%%GAF
FIG. 4. The activation of IFN--y-regulated factors in U937 cells depends on the state of differentiation. U937 cells were treated with TPA for the indicated periods and then stimulated for 30 min with IFN--y. Nuclear extracts were tested in a mobility shift assay with the IFP 53 GAS oligonucleotide as a probe.
PoV
~
b
CIP IFP53
~~ d,
--%%GAF
..
_..t
C
_
VanadateI FN-'Y . IFP53 FcR -I
+
+ .
+
+
-
'
+
+
. +
+ + .
.
.
Xt+~ "
_
DIF
DIF
activation of DIF by IFN--y, according to the results presented in Fig. 4. The DNA-binding form of DIF is phosphorylated on tyrosine. To study the signaling events leading to the activation of DIF, we first tested whether the DNA-binding form of DIF is, like GAF, modified by tyrosine phosphorylation. Antiphosphotyrosine MAb Ab-1, but not a control MAb, efficiently prevented binding of DIF to the IFP 53 GAS. This indicates that the binding of this MAb to DIF interferes with its DNA association (Fig. 5a). The effect of Ab-1 on DIF DNA binding was stronger than that observed with tyrosinephosphorylated GAF (most evident in lanes 10 to 12). Phosphorylation appears to be required for DIF binding to DNA. Consistently, treatment with alkaline phosphatase decreased the level of DIF binding (Fig. Sb). Moreover, both DIF and GAF could be induced by treatment of U937 with vanadate (Fig. 5c). This finding raised the possibility of their regulation by a constitutively active tyrosine phosphatase. As might be expected, vanadate and IFN--y acted synergistically in activating both DIF and GAF (Fig. Sc). Extracts from IFN-ylvanadate-treated cells contained a yet different complex with slower mobility than DIF. Low amounts of the slowly migrating complex could also be detected, upon longer exposure of the autoradiograms, in extracts from cells treated with IFN-y alone (data not shown). DIF preferentially binds to selected GAS and does not contain p91. To preliminarily characterize DIF and assess its relationship to GAF, we tested the sensitivity of complexes containing either factor to competition by GAS oligonucleotides from promoters of different IFN--y-inducible genes, namely, the GBP (30), Fc-yRI (34, 35, 59), and Ly6A/E (8) genes, and shortened versions of the IFP 53 binding site, IFP 53A and IFP 53C (53). To further establish the identity of the U937- and BAC-1.2F5-derived factors, these experiments were performed with DIF-containing extracts from either TPA/IFN--y-treated U937 cells or CSF-1-treated BAC-1.2F5 cells (Fig. 6a and b). GAF binding to the IFP 53 GAS was competed for by all tested GAS, albeit with different efficiencies. In contrast, DIF binding was competed for by the IFP 53 and Fc-yRI GAS but not by the GAS from the GBP gene promoter. The Ly6A/E oligonucleotide showed an intermediate level of competition for DIF. Using the Ly6A/E sequence directly as a probe, we failed to detect the binding
. l_ GAF U-
-W
--GAF mWmw
FIG. 5. Regulation of DIF DNA-binding activity by phosphorylation. (a) Sensitivity of DIF binding to the IFP 53 GAS to antiphosphotyrosine antibody. Cytoplasmic extracts were generated from U937 cells after treatment with vanadate (lanes 1 to 8) or treatment with vanadate and IFN--y (lanes 9 to 12; details of vanadate/IFN--y treatment are given below). Antiphosphotyrosine (anti p-tyr) antibody Ab-1 was added to the extracts in the indicated amounts. Control monoclonal immunoglobulin G2b (mAb; lanes 6 to 8) was purified from the supernatant of the murine myeloma MPC-11. (b) Sensitivity of DIF DNA-binding activity to treatment with alkaline phosphatase. WCE from vanadate/IFN-y-treated cells were either left untreated or treated in vitro with 40 U of calf intestinal phosphatase (CIP) per ml for 30 min at room temperature. Lane 2, competition with a 50-fold molar excess of cold IFP 53 GAS oligonucleotide. (c) Induction of IFP 53 GAS-binding factors by treatment of U937 cells with vanadate or a combination of vanadate and IFN-,y. Cytoplasmic extracts were obtained from cells after a 3.5-h incubation with vanadate (0.1 mg/ml) with or without addition of IFN-y for the last 30 min. Lanes 3 and 4, competition with a 50-fold molar excess of the indicated oligonucleotides.
of a differentiation-induced protein (data not shown). The shortened IFP 53 oligonucleotides did not differentially compete for GAF and DIF binding. In both instances, the IFP 53A oligonucleotide competed less well. This finding indicates that the binding of DIF to a certain GAS is determined by the base composition rather than the extension of the binding site. A control oligonucleotide containing the IFN stimulation response element from the ISG15 gene was ineffective in binding either GAF or DIF. We tested whether the DIF-DNA complex contained p91 or an immunologically related protein. Antibodies specific for the p91 C terminus (44) were added to the binding reaction and tested in a gel shift experiment (Fig. 6a and b). The antibody abolished the binding of GAF and of the slow-mobility U937 complex but did not affect the associa-
EILERS ET AL.
1368
MOL. CELL. BIOL.
a
D
a4If
(0
u
01h
Competitor IFN-y &TPA
* -
_--W
01
01
X
L
_
-J
+
c
a
C a
a
+ + + + + + + .4 _
+
+
wkDIF 4
4
A
+
. 4 1*4
-
GAF 4
bw_
b
D
0.CSF-1 - + + + v IFN-.-
a+
-
+
+
-
+
+
+
+
DIF-*
so
m -
b-S-
J 4i
a
g~~~~~" i44
-GA F
a
c
anti-SH 1 2
anti-N 1_ _2 ,"
_
t.-,
anti-C 1 2
S F-
i_.-;
p91 -
_
I
-
I
I
-DIF -GAF
j
FIG. 6. Sensitivity of DIF and GAF binding to the IFP 53 GAS to competition with IFN-responsive regions of other promoters or to antibodies against the 91-kDa ISGF-3 protein. (a) WCE from U937 cells treated with TPA and IFN-y (48 h of TPA followed by 30 min of IFN--y) were used in a mobility shift assay with the IFP 53 GAS oligonucleotide. The indicated oligonucleotides corresponding to other GAS sequences (Ly6ANE, GBP, and FcyRI), shortened versions of the IFP 53 GAS (IFP 53A and IFP 53C), or an oligonucleotide corresponding to the IFN stimulation response element of ISG 15 (015 [38]) was added in a 50-fold molar excess. Antibodies to the carboxy terminus of p91 or to pl13 were added in a dilution of 1:100. (b) BAC-1.2F5 DIF in WCE from cells after a 15-min CSF-1 treatment was characterized as described above for U937 DIF. GAF was characterized by using antibodies to the p91 carboxy terminus in a dilution of 1:50. (c) IFP 53 GAS-protein complexes separated by nondenaturing PAGE were transferred to nitrocellulose and stained with antibodies to the carboxy terminus (dilution, 1:2,000), the amino terminus (1:250), or the SH2 domain (1:500) of p91. Lanes: 1,
tion of DIF with the IFP 53 GAS. The carboxy terminus of p91 is unique, whereas other domains of p91 are related in sequence to other ISGF-3 proteins (16, 44) and may be common to other members of this family of proteins. Thus, the same experiments were performed with antibodies directed against the SH2 domains or the amino terminus of p91. Neither antiserum significantly affected the complex corresponding to DIF (data not shown). To further verify our findings, we performed a mobility shift assay with extracts containing abundant amounts of DIF and GAF activity. After separation by polyacrylamide gel electrophoresis (PAGE), the protein-DNA complexes were transferred to a nitrocellulose membrane. In this experiment, none of the p91 antisera stained the DIF complex while reacting with p91 contained in GAF (Fig. 6c). Control antibodies directed against a related protein, p113, had no effect on either factor. Relationship between DIF, GAF, SIF, and APRF. The sequences of the binding sites for the DNA-binding proteins SIF and APRF, termed SIE and APRE, display some homology to the GAS consensus (53, 57, 58). SIF is rapidly activated in response to PDGF and EGF and binds to the SIE in the promoter of the c-fos proto-oncogene (42, 57). APRF is rapidly induced by IL-6 and binds to the APRE of several genes contributing to the acute-phase response in hepatocytes (reference 58 and references therein). We used SIE and APRE oligonucleotides to compete for the binding of DIF and GAF to the IFP 53 GAS. The SIE competed for GAF binding in U937 cell extracts but had much less effect on the binding of DIF (Fig. 7a). A mutant SIE oligonucleotide, SIE 67 (57), that shows higher affinity for SIF competed for the binding of both GAF and DIF. SIE 67 has two nucleotides altered with respect to the wild-type fos SIE, and comparison with the IFP 53- and FclyRI sequences shows that only one of these changes (C-12-.A-12) should confer higher affinity for DIF (Table 1). Thus, this single nucleotide appears to be a crucial determinant for the association of DIF. A second mutant SIE oligonucleotide, SIE m34, with decreased affinity for SIF (57) did not compete for GAF or DIF binding. This finding demonstrates that a pyrimidine nucleotide is essential in position 8 for binding by both GAF and DIF. Identical results were obtained with extracts from CSF-1-treated BAC-1.2F5 cells, further confirming the identity of the BAC-1.2F5 and U937
proteins. The IL-6-responsive region of the human a2-macroglobulin promoter contains two APRE consensus sequences (Table 1) (58) and competed very efficiently for GAF binding (Fig. 7b). In contrast, the oligonucleotide competed inefficiently for DIF binding to the IFP 53 GAS. Thus, GAF/p91 binds all GAS, the SIE, and APRE elements, while DIF preferentially binds to the IFP 53 and FcyRI GAS. This was confirmed by direct binding experiments using SIE and APRE oligonucleotides as probes. Cotreatment with IFN--y and vanadate, which activates DIF and GAF DNA binding (Fig. 5c), induced only one factor binding to the SIE probe (Fig. 7c). This factor was identical to GAF/p91, since it was absent from extracts containing DIF activity exclusively and could be blocked with antibodies to p91 (data not shown). An activity induced by vanadate treatment weakly bound the APRE. This factor was competed for by the APRE oligonu-
cytoplasmic extracts from cells treated with vanadate and IFN--y (see above); 2, cytoplasmic extracts from untreated cells; S, autoradiogram showing the complexes corresponding to lanes 1.
DIFFERENTIATION AND IFN-y RESPONSE
VOL. 14, 1994 Probe: IFP 53
b
a probe:IFP53 : 1
E
cn
2L 3 4
:~Lc
:
2
S E
3
4
!
