Aug 19, 1999 - Gerondakis and was prepared as previously described (17). .... Nuclear extracts prepared form Jurkat T cells 6 h following treatment with ...
International Immunology, Vol. 11, No. 12, pp. 1945–1955
© 1999 The Japanese Society for Immunology
Nuclear factor of activated T cells contributes to the function of the CD28 response region of the granulocyte macrophage-colony stimulating factor promoter Catherine Shang1, Joanne Attema2, Dimitrios Cakouros2, Peter N. Cockerill and M. Frances Shannon2 Division of Human Immunology, Hanson Centre for Cancer Research, Institute of Medical and Veterinary Science, Adelaide, South Australia 5001, Australia 1Present
address: Department of Physiology and Pharmacology, University of Queensland, Brisbane, Queensland 4072, Australia 2Present address: Division of Biochemistry and Molecular Biology, John Curtin School of Medical Research, Australian National University, Canberra, ACT 2601, Australia Keywords: co-stimulatory molecules, cytokines, gene regulation, T cell, transcription factors
Abstract The granulocyte macrophage colony stimulating factor (GM-CSF) promoter contains a 10 bp element known as CK-1 or CD28RE that specifically responds to the co-stimulatory signal delivered to T cells via the CD28 surface receptor. This element is a variant NFκB site that does not function alone but requires an adjacent promoter region that includes a classical NFκB element, an Sp-1 site and a putative activator protein-1 (AP-1)-like binding site. The entire region is referred to as the CD28 response region (CD28RR). The GM-CSF CK-1 element has been shown to bind NFκB proteins, in particular c-Rel, whose binding and function is dependent on the architectural transcription factor HMGI(Y). It has been previously suggested that the nuclear factor of activated T cells (NFAT) family of proteins also plays a role in the activity of this region. We show here that recombinant NFATp but not AP-1 can bind to the GM-CSF CD28RR. NFATp present in activated Jurkat T cell extracts can also interact with the CD28RR. The binding of NFATp and Rel proteins requires the same core CK-1 sequences, and appears to be mutually exclusive. We investigated the functional significance of NFATp binding to CK-1 by overexpressing the protein in Jurkat T cells and found that NFATp cannot activate the CD28RR alone but can cooperate with signals generated by phorbol 12-myristate 13-acetate/calcium ionophore. The CD28RR is therefore a complex region that can bind and respond to a combination of transcription factors and signals. Introduction The efficient activation of naive T cells requires not only the activation of the T cell receptor (TCR) by MHC–antigen complexes on antigen presenting cells (APC) but also the presence of co-stimulatory signals provided by the APC. The most important of these co-stimulatory signals is provided by the interaction of the B7 molecules on the APC with CD28 on the T cell surface (1,2). TCR stimulation alone leads to anergy
or apoptosis (3), and the co-activation of the CD28 receptor appears to overcome the anergic or apoptotic response and activates a proliferative response in the T cells (1–3). TCR and CD28 generated signals are highly cooperative, and result in high-level production of cytokines such as IL-2, IL-4 and granulocyte macrophage colony stimulating factor (GM-CSF) (4). While IL-2 and IL-4 are involved in T cell
Correspondence to: M. F. Shannon, Division of Biochemistry and Molecular Biology, John Curtin School of Medical Research, Australian National University, Canberra, ACT 2601, Australia Transmitting editor: A. Kelso
Received 21 June 1999, accepted 19 August 1999
1946 NFAT and the GM-CSF CD28 response region proliferation and function, cytokines such as GM-CSF affect other arms of the immune system and help to coordinate the immune response. GM-CSF affects the differentiation, function and survival of cells of the myeloid lineage (5). These cells perform important functions in the inflammatory response to foreign antigen as well as being involved in antigen presentation (5). The coordinate expression, by activated T cells, of cytokines that control distinct arms of the immune system is clearly important in a correctly controlled immune response. Much of the cooperative activation of cytokine expression by CD28 and TCR activation has been attributed to increased transcriptional activity from the cytokine gene promoters (4,6,7). There have been many detailed studies on the promoter regions and the transcription factors involved in the transcription of the genes encoding IL-2 and other T cell cytokines (8–10). A consensus has emerged showing that the inducible transcription factor families, nuclear factor of activated T cells (NFAT), activator protein-1 (AP-1) and NFκB, play major roles in the activation of transcription from the promoters of these cytokine genes. Activation of the TCR with anti-CD3 antibody or phytohemagglutinin or mimicking this activation with phorbol 12-myristate 13-acetate (PMA) and calcium ionophore (I) leads to the induction of certain members of these families of transcription factors but this in turn drives only low-level promoter function (4,6,7,11). CD28 costimulation leads to an augmentation of the activity or level of some of these transcription factors and hence increased transcription (4,6,7,11). NFκB family members, especially cRel, are activated by CD28 (6,12,13) as are members of the AP-1 family via the activation of JNK (14,15). Whether CD28 co-stimulation leads to a qualitative or simply a quantitative change in these transcription factors has yet to be determined. Specific regions that are required for response to CD28 have been identified in the promoters of the IL-2 and GMCSF genes, and are referred to as the CD28 response regions (CD28RR) (4,6,7,11). These CD28RR contain an element known as either CK-1 in the GM-CSF promoter (6,16) or CD28RE in the IL-2 promoter (4,7) that is essential for a CD28 response. The proteins that bind to and activate the CD28RR have been difficult to define. It has been shown that the CK-1 or CD28RE is a variant NFκB site that binds RelA homodimers as well as c-Rel-containing complexes (6,17). NFκB proteins are activated in a biphasic fashion in T cells (13,18,19). Within 10 min of activation there is an accumulation of RelA and p50 in the nucleus that peaks at around 30 min but returns to near basal levels by 1 h. The second wave of NFκB activation is characterized by a shift in the NFκB family members that are found in the nucleus, with c-Rel and RelB dominating the later activation that peaks at ~6 h (18,19). CD28 activation appears to have its major impact on the late activation of cRel and thus not only increases the level of activation but also prolongs the signal (6,13,19). The CD28RR does not respond to signals that only activate the early induction of RelA and p50 such as PMA or tumor necrosis factor (TNF) (6,17). The late induction of c-Rel leads to a large increase in protein binding to the CD28RR (6,20) which is thought to be essential for the activity of this element (6). There is evidence both from in vitro studies and from c-Rel deficient mice that c-Rel plays a vital role in CD28 activation of cytokine genes (6,21).
Table 1. Sequences of oligonucleotides used in EMSA –111
–67
GM CD28RR TGATAAGGGCCAGGAGATTCCACAGTTCAGGTAGTTCCCCCGCCTCa GMCK-1 TGATAAGGGCCAGGAGATTCCACAGTTCAGGTAGTTb,c GMCK-1M -----------------gg-t---------------d GMm19 ----------gtct---------------------GMm21 ---------------------------gtct----GMm23 ----------gtct-------------gtct----GMm25 -----------------cgg-----cgg-------GM430 gatcTCACACATCTTTCTCATGGAAAGATGA IL2CD28RR TGGGGTTTAAAGAAATTCCAGAGAGTCATCAGAA I-CAMCK-1 GATCCCGATTGCTTTAGCTTGGAAATTCCGGAc IL-2 NFAT gattCGAAAGGAGGAAAAACTGTTTCATACAGAAG SP-1 ATTCGATCGGGGCGGGGCGAGC aThe sequence of the GM-CSF CD28RR with the CK-1 element underlined and bold. The NFκB site is bold and the Sp-1 site is in italics. The region was used to generate the pGMCK-1(2) construct. bThe CK-1 elements in the GMCD28RR, IL-2CD28RR and I-CAMCK1 oligonucleotides, and the AP-1 site in the IL-2 CD28RR are underlined. cThe GGA/TCC palindrome in the GMCK-1 and I-CAMCK-1 is shown in bold. dBases that differ from the wild-type sequences are shown in lower case.
The CD28RE (CK-1 elements) do not function as isolated elements but require adjacent transcription factor binding sites in the CD28RR for activity (6,22). The GM-CSF CD28RR also contains a consensus NFκB site that binds RelA/p50 heterodimers and, although not itself responsive to CD28 activation, appears to play a important role in the activity of this region (6). Recently it has been shown that an AP-1-like site in the CD28RR of IL-2 is required for activity together with the CD28RE (22), although there is some controversy concerning the proteins that operate on this AP-1 site (22– 25). The GM-CSF CD28RR contains a very weak match to a consensus AP-1 site but mutation of this site does not affect the ability of the HTLV-1 transactivator, Tax, to activate this region (26). The DNA binding domains of NFκB and NFAT proteins share a low degree of sequence similarity and three-dimensional structure (27,28). The similarity is also manifested in some sequence similarity in their binding sites on DNA (28). It is not surprising, therefore, that there has been a report that NFAT-like proteins can bind to the GM-CSF CK-1 element (29) but their function is unknown. Here, we have characterized the binding of NFATp to the GM-CSF CD28RR and investigated the functional consequences of this binding. Methods Plasmids and oligonucleotides The pGMCK-1(2) reporter plasmid contains two copies of the GM-CSF CD28RR (Table 1) cloned upstream of the thymidine kinase basal promoter and the luciferase reporter gene, and has been previously described (26). The expression vector pRc/CMV was obtained from Invitrogen (San Diego, CA) and pCMVp65(⫹) contains the Rel homology domain of RelA cloned downstream of the CMV promoter and has been
NFAT and the GM-CSF CD28 response region 1947 described previously (17). The NFATp expression plasmid, pLGPNFAT1c, contains the ‘c’ isoform of NFATp (NFAT1) cloned into the plasmid pLGP3 (30) and was a gift from Dr Anjana Rao. All oligonucleotides were synthesized at Bresatec (Adelaide, Australia) except Sp-1 which was purchased from Promega (Madison, WI). The sequences of the oligonucleotides used for gel-shift analysis are shown in Table 1. Double-stranded oligonucleotide probes for gel shift were prepared by end-labeling with [γ-P32]ATP using T4 polynucleotide kinase as previously described (6). Antibodies The antibody against NFATp (R59) was a gift from Dr A. Rao and is specific for NFATp (31). Polyclonal p50 and c-Rel antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Polyclonal RelA antibody (32) was provided by Dr Steve Gerondakis. Where indicated, 0.5 µl of anti-NFAT (R59), anti-c-Rel and anti-RelA was used in DNA binding assays.
in 0.5⫻TBE, dried and autoradiographed. Conditions for EMSA using recombinant HMGI(Y) were as above but with the addition of 2 mM MgCl2 to the binding reactions. Cell culture and transfection Jurkat T cells were maintained in RPMI containing 10% FCS, supplemented with L-glutamine, penicillin and gentamicin antibiotics (6). Cells were electroporated at 250 V with a capacitance of 960 µF using a BioRad Gene Pulser. Approximately 4.5⫻106 cells were electroporated in 300 µl of RPMI containing 20% FCS per transfection. Routinely, 10 µg of pGMCK-1(2) reporter plasmid was used in transfections. For the overexpression study, 10 and 20 µg of expression constructs pCMVp65(⫹), pLGPNFAT1c and pRc/CMV were co-transfected with reporter plasmid. Cells were allowed to recover from transfection for 24 h, and then stimulated for 8 h with 20 ng/ml PMA, 1.0 µM calcium ionophore (A23187) and a 1:10,000 dilution of anti-CD28 ascites (WA93165; Bristol Myers Squibb, Seattle, WA). Reporter assays
Recombinant proteins Recombinant RelA protein was a generous gift from Dr Steve Gerondakis and was prepared as previously described (17). Purified recombinant NFATp protein and truncated c-fos and c-jun proteins were prepared and purified as described (33). The expression vector p∆M1.1 which contains the Rel homology domain of human c-Rel truncated at amino acid 288 was a gift from Dr Charles Kunsch and was used to produce recombinant c-Rel protein as described (34). The cRel protein was used as a crude Escherichia coli lysate in gel-shift assays. Recombinant p50 protein was purchased from Promega. Purified recombinant high-mobility group protein I(Y) [HMGI(Y)] was a gift from Dr Raymond Reeves and has been previously described (6). The amount of recombinant proteins used in gel-shift analysis was determined empirically by titration experiments. Purified recombinant c-fos and c-jun were provided by Dr David Tremethick, and were used as previously described (35). Electrophoretic mobility shift assay (EMSA) Nuclear extracts from Jurkat T cells were prepared as described by Schreiber et al. (36). An aliquot of 10 µg of nuclear extracts was incubated in the presence of 0.2 ng of radiolabeled oligonucleotide probe with 25 mM HEPES, pH 7.9, 100 mM NaCl, 2.5 µg poly(dI:dC), 1 mM Tris, 0.5 mM EDTA, 0.5 mM EGTA, 0.5 mM DTT and 0.5 mM PMSF in a final volume of 15 µl (37). Binding was carried out for 15 min at room temperature. The binding reactions were separated on a 5% non-denaturing acrylamide gel in 0.5⫻TBE (0.5 mM Tris, 42 mM boric acid and 1 mM EDTA, pH 8.3) at 200 V for ~2 h, dried and autoradiographed. For assays using recombinant proteins, 5 µl of protein diluted in 0.1 mM EDTA, 100 mM KCl, 20 mM HEPES pH 7.8, 5 µg/ml BSA and 20% glycerol was incubated for 15 min at room temperature with 0.2 ng probe, 200 ng dI:dC in a binding buffer containing 30 mM KCl, 16.6 mM NaCl, 20 mM HEPES (pH 7.9), 10% glycerol, 1 mg/ml leupeptin, 1 mg/ml aprotinin, 1.5 mM DTT and 0.15 mM PMSF in a final volume of 15 µl. This reaction was then separated on a 5% non-denaturing acrlyamide gel
Transfected cells were lysed by three cycles of freeze–thaw. The protein concentration of cell lysates was determined by the Bradford assay as previously described (6). Luciferase reporter assays were carried out as described (26). Cell extract (10 µg) was assayed in triplicate for each transfection. Results The proteins that bind to the CK-1 element depend on the stimulation conditions There has been a previous report showing that a NFAT-like activity bound to the GM-CSF CK-1 element (29). To further characterize this activity, we initially activated Jurkat T cells with I alone to avoid activation of NFκB proteins that could bind to the CK-1 element. Jurkat T cells treated with I alone contained an inducible complex that bound to the GM-CSF CD28RR (Fig. 1a). This complex co-migrated with a previously characterized NFATp complex that bound to a NFAT binding site in the GM-CSF enhancer (GM430) (Fig. 1a). The GM430 element is the NFAT portion of a NFAT/AP-1 composite site that can act as a strong NFAT binding site in the absence of AP-1 (33,38). The complex was present at a range of times examined from 30 min to 6 h (data not shown). The behavior of this complex was identical to that of the NFATp complex binding to the GM430 element in that it was supershifted by an anti-NFATp antibody (Fig. 1a, lanes 4 and 8) but not by pre-immune serum or an antibody to RelA (Fig. 1a, lanes 2/3 and 6/7). The majority of the NFAT complex binding to the GMCK-1 probe was supershifted with the NFATp antibody as described above. An antibody to NFATc did not supershift the complex (data not shown). Therefore, all subsequent experiments concentrated on the analysis of NFATp. We have previously shown that NFκB/Rel proteins can bind to the GMCSF CD28RR in extracts prepared from cells stimulated with PMA/I, with or without CD28 activation (6). In order to determine whether NFATp also binds to the CD28RR under these stimulation conditions, antibody supershift experiments were performed on extracts made at 30 min or 6 h following
1948 NFAT and the GM-CSF CD28 response region
Fig. 1. NFATp but not AP-1 can bind to the GM-CSF CK-1 element. (a) Extracts from Jurkat T cells stimulated for 30 min with I were used in EMSA with radiolabeled GMCK-1 or GM430 oligonucleotides under conditions optimized for NFAT binding. Antibodies to RelA, NFATp or a pre-immune serum were included in the DNA binding reactions. The NFATp complex and the antibody supershifted complex (Ns) are indicated. (b) Jurkat T cells were stimulated with PMA/I/CD28 for either 30 min or 6 h and the extracts used in EMSA with the GMCK-1 radiolabeled oligonucleotide. Either a pre-immune serum or antibodies to p50, RelA, c-Rel or NFATp were included in the reactions. The supershifted complexes are indicated as Ns and Rs. (c) Complexes binding to the GM-CSF CD28RR from Jurkat T cell nuclear extracts do not contain cjun and c-fos proteins. Nuclear extracts prepared form Jurkat T cells 6 h following treatment with PMA/I/CD28 were used in binding reactions with the IL-2 NFAT (lanes 2–7) or the GMCK-1 (lanes 9–14) probes. Lanes 1 and 8 are unstimulated cell extracts. The inducible NFAT/AP-1 and NFκB complexes are marked. Antibodies to c-Rel, NFATp, c-jun, c-fos or pre-immune control serum were added to the binding reactions. Supershifted complexes in the c-Rel, NFATp and c-jun lanes are indicated by Rs, Ns and Js respectively. The EMSA shown here are representative of at least three individual experiments.
PMA/I/CD28 activation with binding conditions optimized for NFAT (37). As expected supershifts were observed with antibodies to RelA or c-Rel with c-Rel being dominant at 6 h (Fig. 1b, lanes 4/5 and 10/11). The p50 antibody did not generate a specific supershift as previously described but did result in non-specific smearing when a large amount of NFκB binding was observed (Fig. 1b, lane 9). The NFATp antibody generated a weak supershifted complex at 30 min and a stronger supershifted complex at 6 h implying that NFATp could bind to the CD28RR under these conditions (Fig. 1b, lanes 6 and 12). The disappearance of the NFATp band is more obvious in Fig. 1(a) because there was no NFκB protein in the I-treated cells. When NFκB proteins are present they migrate at the same position as the NFAT complexes making it difficult to distinguish a reduction in NFATp binding (Fig. 1b). In many of the cytokine promoter and enhancer regions NFATp functions as a complex with AP-1 bound to an adjacent site (40), and cooperative binding of NFATp and AP-1 is observed when the NFAT and AP-1 binding sites are correctly spaced (33). Immediately downstream from the CK-1 element is a very weak match (CAGTTCA) to a consensus AP-1 site
whose binding capacity is unknown. The presence of c-jun and c-fos was examined in complexes formed on the GMCSF CD28RR from nuclear extracts prepared from cells stimulated with PMA/I/CD28 for 6 h. Antibodies to c-Rel and NFATp generated supershifts when added to these binding reactions as described above, but the c-fos or c-jun antibodies did not (Fig. 1c, lanes 10–13). On the other hand, the c-jun and NFATp antibodies supershifted complexes binding to the NFAT/AP-1 IL-2 probe, which forms a cooperative NFAT/AP-1 complex, but c-Rel antibody did not (Fig. 1c, lanes 3–5). The c-fos antibody has been shown to inhibit recombinant c-fos/ c-jun binding to the IL-2 CD28RR (data not shown). These results show that NFATp but not AP-1 can bind to the GM-CSF CD28RR in extracts from cells stimulated either with I alone or under conditions such as PMA/I/CD28 where NFκB and also AP-1 are activated. Recombinant NFATp but not AP-1 binds to the GM-CSF CD28RR To further analyze the binding of NFATp and c-fos/c-jun (AP-1) to the CD28RR we used recombinant proteins to perform
NFAT and the GM-CSF CD28 response region 1949
Fig. 2. (a) AP-1 binds to the IL-2 but not the GM-CSF CD28RR. EMSA were carried out using recombinant c-fos and c-jun (AP-1) and NFATp with the IL-2 CD28RR oligonucleotide (lanes 1–6) or the GMCK-1 oligonucleotide (lanes 7–12). Each probe was bound to AP-1 (A) or NFATp (N) alone or together (A ⫹ N). Antibodies to c-jun, NFATp or a control serum were added to the binding reactions. The positions of NFATp and AP-1 are indicated. (b) NFATp and AP-1 do not form higher-order complexes on the GM-CSF CD28RR. Recombinant NFATp and truncated c-fos and c-jun proteins were bound either alone [NFATp (N) or AP-1(A)] or together (N ⫹ A) to the IL-2 NFAT (lanes 1–3), the IL-2CD28RR (lanes 4–6) or the GMCK-1 (lanes 7–9) probes. The positions of the NFATp, AP-1 and higher-order NFAT/AP-1 complexes is indicated. (c) A mutation in the CK-1 site eliminates NFATp binding. EMSA was carried out with recombinant RelA (lanes 2 and 4) and NFATp (lanes 1 and 3), and the GMCK-1 oligonucleotide (lanes 1 and 2) or the same oligonucleotide containing a mutation in the CK-1 element (CK-1M, Table 1) (lanes 3 and 4). The band in lane 4 is a non-specific band that is generated with this mutant from the E. coli extract that was used for RelA binding. The more slowly migrating band in lane 1 is most likely a NFATp dimer. (d) EMSA was performed using the GMCK-1 oligonucleotide and recombinant NFATp. Five, 10 and 20 ng of the wild-type competitor (lanes 14–16) as well as various mutant oligonucleotides (Table 1), m19 (lanes 2–4), m21 (lanes 5–7), m23 (lanes 8–10) and m25 (lanes 11–13) were included in the binding reactions. Lanes 1 and 17 have no competitor added. The EMSA is representative of at least three individual experiments.
EMSA experiments with the GMCK-1 and IL-2CD28RR oligonucleotides. Both the GMCK-1 and IL-2CD28RR oligonucleotides could bind recombinant NFATp, but only the IL-2 and not the GMCSF CD28RR bound AP-1 (Fig. 2a, lanes 1 and 7). There was no evidence of higher-order complex formation when NFAT and AP-1 were present on either probe (Fig. 2a, lanes 3 and 9), but an apparent increase in NFATp binding strength was observed on both probes in the presence of AP-1 (Fig. 2a, lanes 3 and 9) (see below). The addition of an anti-c-jun antibody to the binding reactions reduced the AP-1 complex on the IL-2 probe but had no effect on the NFATp complexes binding to either probe (Fig. 2a, lanes 4 and 10), whereas the addition of a NFATp antibody inhibited NFATp binding to both probes (Fig. 2a, lanes 5 and 11). The presence of the NFAT antibody or pre-immune serum leads to an apparent increase in AP-1 binding to the IL-2 CD28RR and a faint AP1 band appearing with the GMCK-1 probe which could be a non-specific effect due to an increase in the protein concentrations in the binding reactions.
We also used truncated c-fos/c-jun proteins that we have previously shown can form cooperative complexes on composite NFAT/AP-1 sites (33) to confirm the lack of cooperation between NFAT and AP-1 on either the IL-2 or GM-CSF CD28RR. As expected NFATp and the truncated c-fos and cjun proteins bound cooperatively to a previously described composite NFAT/AP-1 site from the IL-2 promoter (33) (Fig. 2b, lanes 1–3). Each of the proteins bound independently to the IL-2 CD28RR (Fig. 2b, lanes 4–6), but only NFAT bound to the GMCK-1 (Fig. 2b, lanes 7–9). The binding of NFATp to the GMCK-1 probe and AP-1 to the IL-2 CD28RR was very weak here due to the low concentrations used in these assays in order to detect cooperative binding. An apparent increase in NFATp binding to the GMCK-1 probe in the presence of AP-1 and AP-1 binding to the IL-2 probe in the presence of NFATp was observed at the low concentrations of protein used here. The reason for such increases, that vary depending on the binding conditions, is not clear. These results verify the results observed with nuclear extracts showing that the GM-CSF CD28RR binds NFATp but not AP-1 proteins.
1950 NFAT and the GM-CSF CD28 response region We next defined the sequence requirements for NFATp binding to the GM-CSF CD28RR. NFATp could not bind to a CK-1 mutant that we have previously characterized as a nonfunctional mutant that cannot bind RelA (Fig. 2c). The more slowly migrating band in the NFATp lane most likely represents a NFATp dimer which forms at high concentrations of protein (see below). Using a series of mutations across the GMCK-1 probe (Table 1) as competitors in gel-shift assays we found that mutations in the 5⬘ end of the CK-1 element (Fig. 2d, m19 and m23) or outside the CK-1 element (Fig. 2d, m21) did not greatly affect NFATp binding but that mutations that included the 3⬘ end of CK-1 were weak competitors of NFATp binding (Fig. 2d, m25). These results show that NFATp binds to the core CK-1 element as has previously been determined for NFκB proteins (6,20,26). Mutually exclusive binding of NFATp and NFκB proteins to the GM-CSF CK-1 element Because NFATp and NFκB proteins both bound to the CK-1 element in the CD28RR, we tested whether they could bind simultaneously or form heteromeric complexes on the DNA. When NFATp and RelA were both added to GMCK-1 binding reactions, we observed that each protein appeared to bind independently and there was no evidence for the formation of heteromeric or higher-order complexes (Fig. 3a). In fact there appears to be an inhibition of NFATp binding by the addition of RelA even in the presence of excess probe (Fig. 3a, cf. lanes 2 and 9/10). When increasing amounts of RelA (Fig. 3b, lanes 2–6), c-Rel (Fig. 3c, lanes 2–7) or p50 (Fig. 3b, lanes 8–12) were titrated with a fixed amount of NFATp in DNA binding assays again no higher-order or heteromeric complexes appeared. Both RelA and to a lesser extent c-Rel but not p50 appeared to inhibit NFATp binding in a dosedependent manner (Fig. 3b and c). This was not due to competition for limiting probe as the probe concentration was always in excess in the binding reactions. p50 binding to the GMCK-1 probe was observed here because of the large amounts of protein used. Altering the relative levels of NFATp and the Rel proteins showed that under the appropriate conditions NFATp also inhibits RelA or c-Rel binding to the CK-1 element (data not shown). We next tested whether the addition of recombinant truncated NFκB/Rel proteins to nuclear extracts containing only NFATp (extracts from cells stimulated with only I) lead to the generation of additional complexes. As expected recombinant RelA and c-Rel but not p50 bound to the GMCK-1 probe but none of these proteins bound to the GM430 probe (Fig. 4a and b). The truncated NFκB proteins generated complexes that migrate lower on the gel than full-length NFAT proteins from cells. As was observed for recombinant NFATp, no additional complexes were detected in these experiments. Some inhibition of binding of NFATp was observed to both the GMCK-1 and GM430 probes with the addition of RelA but not c-Rel or p50 (Fig. 4a and b, lanes 5–7, 8–10 and 11–13). Taken together these results imply that NFATp and NFκB bind to the GMCK-1 element independently and that binding is mutually exclusive. HMGI(Y) enhances NFATp binding to the CK-1 elements Since we had previously shown that the binding of c-Rel but not RelA to the GM-CSF and IL-2 CD28RR was dependent
Fig. 3. RelA and NFATp show mutually exclusive binding to the GMCSF CK-1 element. (a) Recombinant NFATp or RelA was used in EMSA reactions with the radiolabeled GMCK-1 oligonucleotides either alone (NFATp lanes 1 and 2; RelA lanes 3 and 4) or together at various concentrations (lanes 5–10). The quantities of protein extract are indicated above the lanes and the RelA, NFATp and free DNA bands are indicated. (b) Titrations of recombinant RelA (lanes 2–6) or p50 (lanes 8–12) were added to EMSA containing a fixed amount of NFATp and the GMCK-1 radiolabeled oligonucleotide. Lanes 1 and 7 had NFATp alone. (c) A titration of recombinant c-Rel was carried out as described in (b) for RelA and p50. Binding reactions and EMSA were performed as described in Fig. 3. The EMSA gels shown here were repeated with identical results at least 3 times.
on the architectural protein HMGI(Y), and that HMGI(Y) was required for CD28RR function (6), we determined whether HMGI(Y) affected NFATp binding to the CD28RR. When increasing amounts of HMGI(Y) was mixed with NFATp prior to DNA binding, a dose-dependent increase in NFATp binding was observed to both the GMCK-1 and IL-2CD28RR oligonucleotides (Fig. 5a and b). An inhibition of binding was observed at the highest concentrations used. We also observed a similar increase in NFATp binding to the GM430 site, a site that does not bind HMGI(Y) alone (Fig. 5c). The decrease in NFATp binding at high concentrations of HMGI(Y)
NFAT and the GM-CSF CD28 response region 1951
Fig. 4. NFκB and NFAT protein bind independently to the GM-CSF CD28RR. Nuclear extracts prepared from Jurkat T cells stimulated with I were used in EMSA with the GMCK-1 (a) or the GM430 (b) radiolabeled oligonucleotides. The identity of the NFATp band was determined by a competition experiment where the IL-2 NFAT/AP-1 oligonucleotide (lane 2) but not an SP-1 binding site (lane 1) nor a random oligonucleotide (lane 3) competed for binding. Recombinant RelA (lanes 5–7), c-Rel (lanes 8–10), p50 (lanes 11–13) or crude E. coli extract (lanes 14–16 in A) were titrated into the binding reactions at concentrations where RelA and c-Rel but not p50 bound to the GMCK-1 element. Lane 4 had no protein added. The section of the gel incorporating the NFATp and Rel bands is shown, and positions of these bands are indicated. The EMSA gels are representative of at least three repeated experiments.
Fig. 5. HMGI(Y) promotes NFATp binding to CK-1 elements. EMSA were carried out using radiolabeled GMCK-1 (A), IL-2CD28RR (B), GM430 (C) or ICAMCK-1 (D) oligonucleotides and recombinant purified NFATp. No HMGI(Y) (lane 1) or increasing amounts of HMGI(Y) (1, 5, 10 and 20 ng in lanes 2–5) were added to the binding reactions. The HMGI(Y), NFATp and probable NFATp dimer [(NFATp)2 ] bands are indicated. The EMSA shown here was repeated 3 times.
that was observed with the GMCK-1 and IL-2 sites was not, however, observed here (Fig. 5c). The addition of HMGI(Y) also appeared to promote the formation of a more slowly migrating complex, that could represent a NFATp dimer, on the GMCK-1 probe but not the IL-2CD28RR nor the GM430 probes (Fig. 5a). An examination
of the sequence of these elements showed that there is a GGA/TCC palindromic sequence across the GM-CSF CK-1 element but not the IL-2 CD28RE or the GM430. The core of a NFAT site is a GGAA/G element and therefore it is possible that two molecules of NFAT could bind to each arm of the palindrome. We examined the sequences of other CK-1-like elements, and found that the ICAM-1 and the IL-3 CK-1-like elements also contained GGA/TCC palindromes. The ICAM1 site but not the IL-3 CK-1 element can bind HMGI(Y) (data not shown). In the case of ICAM-1 a possible dimeric NFATp band was detected together with the increased NFATp binding seen with the addition of HMGI(Y) (Fig. 5d). In the case of IL-3, NFATp can bind efficiently to the site alone and a strong dimeric complex is also observed (data not shown). The formation of these complexes is inhibited by the addition of antibodies to NFATp or HMGI(Y) (data not shown). In summary, HMGI(Y) has the capacity to promote the formation of several different complex types containing either NFκB or NFAT proteins on the CD28RR and therefore may be an integral component of the function of this site. NFATp can potentiate the function of the GM-CSF CD28RR To investigate whether NFATp contributes to the activity of the GM-CSF CD28RR, we initially transfected the pGMCK-1(2) reporter plasmid, containing two copies of the GM-CSF CD28RR, together with the constructs expressing either NFATp or RelA into Jurkat T cells. The expression of RelA activated the CD28RR reporter but NFATp expression did not (Fig. 6a). A combination of NFATp and RelA had no increased effect over RelA alone showing that the proteins could not synergize with each other (Fig. 6a). Titration experiments using different amounts of each plasmid gave the same result (data not shown). It has been shown that NFATp translocation to the nucleus requires its dephosphorylation by calcineurin phosphatase, which is activated by increased levels of intracellular Ca2⫹ (41,42). It is possible that none of the NFATp expressed in the experiments above can translocate to the nucleus. We therefore examined whether co-expressing a constitutively active calcineurin phosphatase would lead to NFATp activation. No increased activity for NFATp alone or synergy with RelA were observed (data not shown). In the same experiments the pGMCK-1(2) construct responded to PMA/I activation and showed an increased activity in response to CD28 co-stimulation (Fig. 6a) as previously described (6). We next examined whether activation of the NFATp-transfected cells could unmask a role for NFATp in CD28RR function since it is possible that NFATp could function with other proteins, which bound to this region. Cells were transfected as above followed by treatment with P and I alone or PMA/I combined. The CD28RR did not respond to either P or I alone but was activated by PMA/I (Fig. 6b). Transfection of the NFATp expression plasmid did not overcome the lack of response to P or I alone but with a combined PMA/I signal a highly cooperative effect was observed (Fig. 6b). These results show that activating the nuclear translocation of NFAT with I treatment alone does not lead to NFAT function on this region, as we had observed with active calcineurin expression, but that signals activated by P treatment are also required for cooperative activity on the CD28RR. The activity of the CD28RR can be increased by stimulation
1952 NFAT and the GM-CSF CD28 response region of the CD28 signaling pathway (6). We therefore investigated whether NFATp overexpression could also cooperate with all three signals or was substituting for the CD28 signal in the cooperation with PMA/I observed here. The results showed that NFATp expression cooperated with PMA/I in the presence of CD28 signals to lead to increased levels of activity from the CD28RR (Fig. 6c). While NFATp expression enhanced both PMA/I and PMA/I/CD28 activity, the ratio of the PMA/I/ CD28 to PMA/I response was not affected by the presence of NFATp. We also examined whether NFATp expression could cooperate with the combinations of PMA ⫹ CD28 or I ⫹ CD28 to activate the CD28RR. These combinations of signals alone could not activate the CD28RR and expressing NFATp did not compensate for this lack of activity (Fig. 6c). These results imply that a combination of signals activated by PMA/I and CD28 are required for the function of the CD28RR, and that NFATp cannot substitute for but can cooperate with these signals. Discussion The CD28RR of the GM-CSF gene is a complex binding site for many different transcription factors and requires the activation of multiple signaling pathways for function. This is illustrated here by the finding that the GM-CSF CD28RR does not respond to PMA or I signals alone but requires both signals for activity. CD28 activation together with the above signals optimally activates the CD28RR, but CD28 activation cannot combine with PMA or I individually to activate the CD28RR. The need for multiple signals is borne out at the level of transcription factor binding to the CD28RR. We have previously shown that NFκB family proteins play an important role in the function of the GM-CSF CD28RR (6). Experiments described here show that NFATp may also play a role in GMCSF CD28RR activity. There was an earlier report that NFATlike proteins bound to the GM-CSF CK-1 element but the exact identity and functional consequences were not investigated (29). We confirmed this interaction using recombinant NFATp and showed that NFATp binding required the TTCC/
Fig. 6. Activation of the GM-CSF CD28RR by NFATp in Jurkat T cells. (a) Jurkat T cells were transfected by electroporation with the pGMCK1(2) reporter construct (10 µg) and expression constructs for RelA and NFATp either alone or together at the indicated amounts. Cells without the expression constructs or those with lower amounts were co-transfected with a blank CMV expression plasmid to ensure equal DNA levels in each transfection. Transfected cells were either left unstimulated or stimulated with PMA (20 ng/ml) and I (1.0 µM) (PMA/ I) in the presence or absence of a CD28 activating antibody (1:10,000 dilution of ascites WA93165 from Bristol Myers Squibb) (PMA/I/CD28) 24 h later. Cells were harvested 8 h after stimulation and cell extracts prepared for luciferase assays. The data is presented as mean luciferase activity from three independent transfection with duplicates within each experiment. (b) Jurkat T cells were transfected with a NFATp reporter plasmid (10 µg) or a blank CMV plasmid (10 µg) and the pGMCK-1(2) reporter plasmid and stimulated 24 h later with PMA (P), I alone or together (PMA/I) for 8 h before harvesting for luciferase activity. The data is presented as in (a). (c) Jurkat T cells were transfected as described in (b) and were stimulated with combinations of PMA (P), I and the activating CD28 antibody (C) at the same concentrations as described for (a).
NFAT and the GM-CSF CD28 response region 1953 GGAA bases of the CK-1 element. This sequence has been shown to be the core consensus required for NFAT binding to DNA (42) and is part of the already defined binding site for Rel proteins across CK-1 (6,26). The Rel and NFAT families of proteins share some similarity across their DNA binding domains, and display similar DNA contacts across the GGAA region of their binding sites (42). However, NFAT generally binds as a monomer whereas Rel proteins bind as obligate dimers (reviewed in 42,43). Interestingly, it appears that two molecules of NFATp may bind to some of the CK-1 sites examined here and this appears to correspond to the presence of a GGA/TCC palindromic sequence across the region. The ability to form such a dimer of NFATp may alter the functional significance of NFAT binding to the CK-1 elements. Similar dimeric binding of NFATp has previously been reported on the TNF-α promoter, but the functional significance is not clear (44). The architectural protein HMGI(Y) appears to promote NFATp binding to the CD28RR. We have previously shown that HMGI(Y) promotes c-Rel binding to the CD28RR, and is essential for the function of the GM-CSF and IL-2 CD28RR (6). The fact that it also supports NFATp binding suggests that it plays a central role in the formation and architecture of many complexes on the CD28RR. NFATp present in nuclear extracts from activated T cells can also bind to the GM-CSF CD28RR. This binding was observed both in extracts where NFκB/Rel proteins were absent (I treated cells) or where there were high levels of RelA and c-Rel (PMA/I and CD28 treated cells). Similar complexes of proteins including NFATp and NFκB/Rel proteins have been observed binding to the IL-2 CD28RR (45). It is clear that the Rel and NFATp proteins do not form heteromeric complexes on the DNA and appear to bind independently. In fact an apparent cross-inhibition of binding by these proteins was observed but the significance of this inhibition is unclear since NFATp binding can be observed in gel-shift experiments under stimulation conditions in cells (PMA/I/CD28) that generate large amounts of NFκB proteins. Attempts to detect NFAT/ RelA interactions in the absence of DNA, that could account for such cross-inhibition, have shown that there may be a weak but unstable interaction (data not shown). It is possible that the NFAT and NFκB proteins, binding independently, play a separate time- or stimulus-dependent role in the activation of this region. Proteins such as NFATp and RelA are activated within minutes of cell stimulation (18,42,43), whereas c-Rel requires several hours of activation to reach high levels in the nucleus (6,13,18,19). It has been shown that RelA and NFATp can interact with the histone acetylases p300/CBP (46,47), and it is possible that they play an early role in recruitment of the acetylases to the promoter and reorganization of the chromatin. On the other hand, c-Rel may play a role in prolonging the activation of the promoter via a different mechanism. Our analysis of the GM-CSF CD28RR has shown that cfos/c-jun does not bind to or form complexes with NFATp on this site. This is in contrast to the IL-2 CD28RR where, as shown here and by others, AP-1 family members can bind and are required for the activity of this region (22–25,48). There has been recent evidence presented that AP-1 (24,25) or activating transcription factor/cAMP response element binding protein (23), proteins binding to the AP-1 site of the
IL-2 CD28RR, functionally cooperate with Rel proteins binding to the CD28RE. This cooperation is not at the level of DNA binding since the combination of AP-1 and c-Rel appear to bind independently to the IL-2 CD28RR (J. Attema and M. F. Shannon, unpublished). AP-1 and NFATp also bind independently to the IL-2 CD28RR and not in a cooperative fashion as has been shown for many other NFAT/AP-1 sites. The cooperative binding of these proteins depends on the appropriate spacing of the binding sites (33), and the NFAT and AP-1 sites of the IL-2 CD28RR are not appropriately spaced to fit this model. Mutation of the IL-2 AP-1 site abolishes the response of the region to HTLV-1 Tax and CD28 activation (24,25). On the other hand, in agreement with the lack of AP1 binding, we have previously shown that mutating the putative AP-1 region of the GM-CSF CD28RR has no effect on its response to HTLV-1 Tax nor to PMA and I treatment (20 and data not shown). These data suggest that there is a distinction between the proteins that bind to and activate the IL-2 and GM-CSF CD28RR, and implies that different physiological signals may act on each of the genes. NFATp activity on the CD28RR is only manifested when the transfected cells are stimulated with both PMA and I. Stimulation of the cells with I alone or co-expression of a constitutively active calcineurin, both of which should lead to NFATp translocation to the nucleus, were not sufficient for NFATp function on the CD28RR. NFATp has been shown to activate other cytokine promoter elements under these conditions (45). While NFATp expression enhanced PMA/I or PMA/I/CD28 activation it did not appear to alter the ratio of the PMA/I/CD28 to PMA/I responses. This suggests that the CD28 signaling pathway does not cooperate with or influence NFATp function, whereas the PMA/I signals do. Under the conditions where NFATp generates increased activity, NFκB/ Rel proteins are also activated and bind to the CD28RR. We can find no evidence either at the functional level or DNA binding level that NFATp and NFκB proteins can cooperate on the GM-CSF CD28RR and thus it is unlikely that the increased activity is a result of NFAT/NFκB cooperation at this level. There are several other possible explanations for the finding that NFATp cooperates with both I and PMA generated signals. Firstly, it is possible that while NFATp is translocated to the nucleus by the I signal it needs to cooperate functionally with PMA-activated transcription factors that have yet to be identified. Our experiments imply that these factors are not NFκB/Rel or AP-1-like proteins. A second possible explanation for the requirement for both PMA and I signals to cooperate with NFATp relates to the fact that NFATp can be translocated to the nucleus by an increase in intracellular calcium but may require a second signal for activity. There is evidence that PMA-activated signals can affect NFAT function but the mechanism is not understood (42). The GM-CSF CD28RR appears to be capable of binding and reacting to different combinations of transcription factors [NFAT, c-Rel, RelA and HMGI(Y)] all of which are important in the regulation of T cell activation. Each factor may play an individual time- or stimulus-dependent role in the activity of this region. Whether any of these activities are redundant or exclusive remains to be determined. The ability to bind these different factors may confer on the CD28RR a unique role in sensing the relative levels of each of these important factors
1954 NFAT and the GM-CSF CD28 response region in the nucleus and acting as a type of rheostat for the activation state of the cells. An analysis of transcription driven by the CD28RR in cells lacking individual factors will be required to ultimately define the importance of each protein. Acknowledgements We are indebted to Donna Woltring for expert technical assistance, and to Andrew Bert and Roy Himes for technical advice. The following people kindly donated reagents: Drs Steve Gerondakis, Anjana Rao, Charles Kunsch, David Tremethick and Raymond Reeves. We thank Dr Adele Holloway for comments on the manuscript, Ms Jan Lee for secretarial skills and the photography section at the JCSMR for help in preparation of the figures. The work was supported by National Health and Medical Research Council, Australia funding to M. F. S. and P. N. C.
Abbreviations AP-1 APC CD28RR CK-1 CD28RE EMSA GM-CSF HMGI(Y) I NFAT PMA TNF
activator protein-1 antigen-presenting cell region containing multiple elements required for CD28 response 10 bp CD28 response element essential for CD28 response in the GM-CSF CD28RR 10 bp CD28 response element in the IL-2 CD28RR electrophoretic mobility shift assay granulocyte macrophage colony stimulating factor high-mobility group protein I(Y) calcium ionophore nuclear factor of activated T cells phorbol 12-myristate 13-acetate tumor necrosis factor
References 1 Bluestone, J. A. 1995. New perspectives of CD28-B7-mediated T cell co-stimulation. Immunity 2:555. 2 Rudd, C. E. 1996. Upstream–downstream: CD28 co-signaling pathways and T cell function. Immunity 4:527. 3 Jenkins, M. K. 1994. The ups and downs of T cell co-stimulation. Immunity 1:443. 4 Fraser, J. D., Irving, B. A., Crabtree, G. R. and Weiss, A. 1991. Regulation of Interleukin-2 gene enhancer activity by the T cell accessory molecule CD28. Science 251:313. 5 Metcalf, D. 1991. Control of granulocytes and macrophages: molecular, cellular and clinical aspects. Science 254:529. 6 Himes, S. R., Coles, L. S., Reeves, R. and Shannon, M. F. 1996. High mobility group protein I(Y) is required for function and for c-Rel binding to CD28 response elements within the GM-CSF and IL-2 promoters. Immunity 5:479. 7 Fraser, J. D. and Weiss, A. 1992. Regulation of T-cell lymphokine gene transcription by the accessory molecule CD28. Mol. Cell. Biol. 12:4357. 8 Serfling, E., Avots, A. and Neumann, M. 1995. The architecture of the Interleukin-2 promoter: a reflection of T lymphocyte activation. Biochim. Biophys. Acta 1263:181. 9 Jain, J., Loh, C. and Rao, A. 1995. Transcriptional regulation of the IL-2 gene. Curr. Opin. Immunol. 7:333. 10 Shannon, M. F., Coles, L. S., Vadas, M. A. and Cockerill, P. N. 1997. Signals for activation of the GM-CSF promoter and enhancer in T cells. Crit. Rev. Immunol. 17:301. 11 Verweij, C. L., Geerts, M. and Aarden, L. A. 1991. Activation of Interleukin-2 gene transcription via the T cell surface molecule CD28 is mediated through an NFκB-like response element. J. Biol. Chem. 266:14179. 12 Ghosh, P., Tan, T.-H., Rice, N. R., Sica, A. and Young, H. A. 1993. The interleukin 2 CD28-responsive complex contains at least three members of the NFκB family: c-Rel, p50, and p65. Proc. Natl Acad. Sci. USA 90:1696.
13 Bryan, R. G., Li, Y., Lai, J.-H., Van, M., Rice, N. R., Rich, R. R. and Tan, T.-H. 1994. Effect of CD28 signal transduction on c-Rel in human peripheral blood T cells. Mol. Cell. Biol. 14:7933. 14 Rincon, M. and Flavell, R. A. 1994. AP-1 transcriptional activity requires both T-cell receptor-mediated and co-stimulatory signals in primary T lymphocytes. EMBO J. 13:4370. 15 Su, B., Jacinto, E., Hibi, M., Kallunki, T., Karin, M. and BenNeriah, Y. 1994. JNK is involved in signal integration during costimulation of T lymphocytes. Cell 77:727. 16 Shannon, M. F., Gamble, J. R. and Vadas, M. A. 1988. Nuclear proteins interacting with the promoter region of the human granulocyte/macrophage colony-stimulating factor gene. Proc. Natl Acad. Sci. USA 85:674. 17 Dunn, S. M., Coles, L. S., Lang, R. K., Gerondakis, S., Vadas, M. A. and Shannon, M. F. 1994. Requirement for nuclear factor (NF)-kappa B p65 and NF–interleukin-6 binding elements in the tumor necrosis factor response region of the granulocyte colonystimulating factor promoter. Blood 83:2469. 18 Molitor, J. A., Walker, W. H., Doerre, S., Ballard, D. W. and Greene, W. C. 1990. NF kappa B: a family of inducible and differentially expressed enhancer-binding proteins in human T cells. Proc. Natl Acad. Sci. USA 87:10028. 19 Kalli K., Huntoon, C., Bell, M. and McKean, D. J. 1998. Mechanism responsible for T-cell antigen receptor- and CD28- or interleukin 1 (IL-1) receptor-initiated regulation of IL-2 gene expression by NFκB. Mol. Cell. Biol. 18:3140. 20 Shannon, M. F., Himes, S. R. and Coles, L. S. 1995. GM-CSF and IL-2 share common control mechanisms in response to costimulatory signals in T cells. J. Leuk. Biol. 57:767. 21 Gerondakis, S., Strasser, A., Metcalf, D., Grigoriadis, G., Scheerlinck, J. Y. and Grumont, R. J. 1996. Rel-deficient T cells exhibit defects in production of interleukin 3 and granulocytemacrophage colony-stimulating factor. Proc. Natl Acad. Sci. USA 93:3405. 22 Shapiro, V. S., Truitt, K. E., Imboden, J. B. and Weiss, A. 1997. CD28 mediates transcriptional upregulation of the interleukin-2 (IL-2) promoter through a composite element containing the CD28RE and NFIL-2B AP-1 sites. Mol. Cell. Biol. 17:4051. 23 Butscher, W. G., Powers, C., Olive, M., Vinson, C. and Gardner, K. 1998. Coordinate transactivation of the interleukin-2 CD28 response element by c-Rel and ATF/CREB2. J. Biol. Chem. 273:552. 24 Parra, E., McGuire, K., Hedlund, G. and Dohlsten, M. 1998. Overexpression of p65 and c-Jun substitutes for B7-1 costimulation by targeting the CD28RE within the IL-2 promoter. J. Immunol. 160:5374. 25 McGuire, K. L. and Iacobelli, M. 1997. Involvement of Rel, Fos and Jun proteins in binding activity to the IL-2 promoter CD28 response element/AP-1 sequence in human T cells. J. Immunol. 159:1319. 26 Himes, S. R., Katsikeros, R. and Shannon, M. F. 1996. Costimulation of cytokine gene expression in T cells by the human T leukemia/lymphotropic virus type 1 trans activation Tax. J. Virol. 70:4001. 27 Wolfe, S. A., Zhou, P., Dotsch, V., Chen, L., You, A., Ho, S. N., Crabtree, G. R., Wagner, G. and Verdine, G. L. 1997. Unusual Rel-like architecture in the DNA-binding domain of the transcription factor NFATc. Nature 385:172. 28 Chen, L., Glover, J. N., Hogan, P. G., Rao, A. and Harrison, S. C. 1998. Structure of the DNA-binding domains of NFAT, Fos and Jun bound specifically to DNA. Nature 392:42. 29 McCaffrey, P. G., Jain, J., Jamieson, C., Sen, R. and Rao, A. 1992. A T cell nuclear factor resembling NFAT binds to an NFκB site and to the conserved lymphokine promoter ‘cytokine-1’. J. Biol. Chem. 267:1864. 30 Luo, C., Burgeon, E., Carew, J. A., McCaffrey, P. G., Badalian, T. M., Lane, W. S., Hogan, P. G. and Rao, A. 1996. Recombinant NFATI (NFATp) is regulated by calcineurin in T cells and mediates transcription of several cytokine genes. Mol. Cell. Biol. 16:3955. 31 McCaffrey, P. G., Luo, C., Kerppola, T. K., Jain, J., Badalian, T. M., Ho, A. M., Burgeon, E., Lane, W. S., Lambert, J. N., Curran, T., et al. 1993. Isolation of the cyclosporin-sensitive T cell transcription factor NFATp. Science 262:750.
NFAT and the GM-CSF CD28 response region 1955 32 Grumont, R. J. and Gerondakis, S. 1994. The subunit composition of NFκB complexes changes during B-cell development. Cell Growth Different. 5:1321. 33 Cockerill, P. N., Bert, A. G., Jenkins, F., Ryan, G. R., Shannon, M. F. and Vadas, M. A. 1995. Human granulocyte-macrophage colony-stimulating factor enhancer function is associated with cooperative interactions between AP-1 and NFATp/c. Mol. Cell. Biol. 15:2071. 34 Kunsch, C., Ruben S. M. and Rosen, C. A. 1992. Selection of optimal kappa B/Rel DNA-binding motifs: interaction of both subunits of NF kappa B with DNA is required for transcriptional activation. Mol. Cell. Biol. 12:4412. 35 Ng. K. W., Ridgway, P., Cohen, D. R. and Tremethick, D. J. 1997. The binding of a Fos/Jun heterodimer can completely disrupt the structure of a nucleosome. EMBO J. 16:2072. 36 Schreiber, E., Matthias, P., Muller, M. M. and Schaffner, W. 1989. Rapid detection of octamer binding proteins with ‘mini-extracts’, prepared from a small number of cells. Nucleic Acids Res. 17:6419. 37 Jenkins, F., Cockerill, P. N., Bohmann, D. and Shannon, M. F. 1995. Multiple signals are required for function of the human granulocyte-macrophage colony-stimulating factor gene promoter in T cells. J. Immunol. 155:1240. 38 Cockerill, P. N., Shannon, M. F., Bert, A. G., Ryan, G. R. and Vadas, M. A 1993. The granulocyte-macrophage colonystimulating factor/interleukin 3 locus is regulated by an inducible cyclosporin A-sensitive enhancer. Proc. Natl Acad. Sci USA 90:2466. 39 Cockerill, G. W, Bert, A. G., Ryan, G. R., Gamble, J. R., Vadas, M. A. and Cockerill, P. N. 1995. Regulation of GM-CSF and E-
40 41
42 43 44 45 46 47
48
selectin expression in endothelial cells by cyclosporin A and the T cell transcription factor NFAT. Blood 86:2689. Rao, A. 1994. NFATp: a transcription factor required for the coordinate induction of several cytokine genes. Immunol. Today 15:274. Shaw, K. T., Ho, A. M., Raghavan, A., Kim, J., Jain, J., Park, J., Sharma, S., Rao, A. and Hogan, P. G. 1995. Immunosuppressive drugs prevent a rapid dephosphorylation of transcription factor NFAT1 in stimulated immune cells. Proc. Natl Acad. Sci. USA 92:11205. Rao, A., Luo, C. and Hogan, P. G. 1997. Transcription factors of the NFAT family: regulation and function. Annu. Rev. Immunol. 15:707. Siebenlist, U., Franzoso, G. and Brown, K. 1994. Structure, regulation and function of NFκB. Annu. Rev. Cell Biol. 10:405. McCaffrey, P. G., Goldfeld, A. E. and Rao, A. 1994. The role of NFATp in cyclosporin A-sensitive TNFα gene transcription. J. Biol. Chem. 269:30445. Maggirwar, S. B., Harhaj, E. W. and Sun, S.-C. 1997. Regulation of the interleukin-2 CD28-responsive element by NFATp and various NFκB/Rel transcription factors. Mol. Cell. Biol. 17:2605. Gerritsen, M. E., Williams, A. J., Neish, A. S., Moore, S., Shi Y. and Collins, T. 1997. CREB-binding protein/p300 are transcriptional coactivators of p65. Proc. Natl Acad. Sci. USA 94:2927. Garcia-Rodriguez C. and Rao, A. 1998. Nuclear factor of activated T cells (NFAT)-dependent transactivation regulated by the coactivators p300/CREB-binding protein (CBP). J. Exp. Med. 187:2031. Rooney, J. W., Sun, Y. L., Glimcher, L. H. and Hoey, T. 1995. Novel NFAT sites that mediate activation of the interleukin-2 promoter in response to T-cell receptor stimulation. Mol. Cell. Biol. 15:6299.
