NF-IL6 (C/EBPbeta ) vigorously activates il1b gene expression via a Spi-1 (PU.1) protein-protein tether.

Two classes of transcription factors, ETS and bZIP, stand out as key mediators of monocyte commitment and differentiation. The ETS domain factor Spi-1 (also called PU.1) and the bZIP factor NF-IL6 (also called C/EBPbeta) have been shown to be involved in the transcriptional regulation of interleukin-1beta gene (il1b) and other monocyte-specific genes. We now show that these two factors strongly cooperate on the il1b core promoter (-59/+12) in the absence of direct NF-IL6 binding to DNA. Transient transfection assays, using mutated il1b core promoters, showed that the Spi-1, but not the NF-IL6, binding site is absolutely required for functional cooperativity. Furthermore, the NF-IL6 transactivation domain (TAD) is functionally indispensable and more critical than that of Spi-1. Additionally, TAD-deficient NF-IL6 functions as a dominant negative for Spi-1-mediated activation, suggesting the involvement of the bZIP DNA binding domain. This is supported by the demonstration of in vitro interaction between the NF-IL6 bZIP and Spi-1 winged helix-turn-helix (wHTH) DNA binding domains, arguing that NF-IL6 vigorously activates the il1b core promoter via protein-tethered transactivation mediated by Spi-1.

Myeloid lineage differentiation and the expression of activated monocyte/macrophage genes, such as the gene encoding interleukin 1␤ (IL-1␤), 1 depend upon the activity of at least two key transcriptional regulatory factors. One of these is NF-IL6 (also called C/EBP␤, NF-M, AGP/EBP, LAP, IL6-DBP, and CRP2) (1)(2)(3)(4)(5)(6), which is constitutively expressed in resting monocytes and immediately activated by translocation from cytoplasm to nucleus by agents such as lipopolysaccharide, phorbol myristate acetate (PMA), and IL-6. NF-IL6 is a bZIP transcription factor that belongs to the C/EBP family of proteins (1,3,7). Recent studies have shown that NF-IL6 is capable of synergistically cooperating with other transcription factors, including NF-B (8), Sp1 (9), and GATA-1 (10). In these cases, the binding of NF-IL6 and its partners to their recognition sites in the promoters is required for the functional cooperativity. A second factor, Spi-1/PU.1, is a winged helix-turn-helix (wHTH) transcription factor that belongs to the ETS family of proteins. Spi-1 expression is primarily restricted to myeloid cells, whereas NF-IL6 is more broadly expressed. The importance of C/EBP factors and Spi-1 in myeloid cells emphasizes the need to understand the mechanisms regulating their functions.
IL-1␤ is an important inflammatory and immunoregulatory cytokine expressed by primarily activated monocytes/macrophages in response to a variety of stimuli, including lipopolysaccharide, PMA, IL-1␤, and other cytokines (11). Uncovering the mechanisms that drive the expression of il1b will elucidate the events of normal myeloid commitment and differentiation as well as the dysregulation of gene expression that leads to inflammatory diseases. Zhang and Rom (12) reported that the Ϫ131/ϩ12 il1b promoter contained two NF-IL6 binding elements, located at positions Ϫ90/Ϫ82 and Ϫ41/Ϫ33. While the importance of the more upstream NF-IL6 site has been established (13), the function of the Ϫ41/Ϫ33 is poorly defined. This site overlaps the 3Ј-end of a Spi-1 binding site at Ϫ50/Ϫ39 by 3 base pairs. It was shown that in lipopolysaccharide-simulated macrophage cells, Spi-1, but not NF-IL6, is a predominant protein factor bound to this overlapping sequence (13,14) in the Ϫ59/ϩ12 il1b core promoter. These data argue that the Spi-1, but not the overlapping NF-IL6 binding element, is required for maximal activation of the il1b core promoter.
There are accumulating evidence showing that Spi-1 and C/EBP family factors are important for the regulation of many genes involved in immunity and hematopoiesis, such as macrophage colony-stimulating factor receptor (15) and neutrophil elastase (16), in addition to the IL-1␤ gene. However, the mechanism responsible for the functional cooperativity is still poorly understood. Many ETS target sites are found adjacent to binding sites for other protein factors, which appear to functionally cooperate. The most frequently reported type of composite site involves cooperative interactions between ETS proteins and bZIP factors (17,18), such as the interaction between the Spi-1 wHTH ETS DNA binding domain and the NF-IL6␤ (C/EBP␦) leucine zipper region (17). In transient expression assays, using an artificial promoter containing adjacent Spi-1 and NF-IL6␤ sites, Nagulapalli et al. (17) observed that Spi-1 and NF-IL6␤ could functionally cooperate to activate transcription. However, the combined roles of NF-IL6␤ and Spi-1 in naturally occurring promoters have not yet been reported. In most cases, the association of another factor with Spi-1 results in strong synergistic activation of target gene. A recent example is c-Jun, which acts as a Spi-1 coactivator on the promoters of myeloid genes coding for macrophage scavenger receptor (19), and macrophage colony-stimulating factor receptor (20).
The close proximity of the Spi-1 and the putative NF-IL6 sites does not seem to be fortuitous. Even though the putative NF-IL6 site at positions Ϫ41/Ϫ33 seemed to not be critical as a transcription factor binding site, as suggested by Buras' in vitro data (14), we still investigated the possible functional involvement of NF-IL6 in the regulation of the il1b core promoter. The study described here demonstrates that NF-IL6 strongly cooperates with Spi-1 to activate the il1b core promoter (Ϫ59/ϩ12), in which the integrity of the Spi-1 binding site, but not the putative NF-IL6 binding site, is critical for the synergy. In addition, the functional cooperativity between Spi-1 and NF-IL6 definitely requires the transactivation domain (TAD) of NF-IL6, but not those of Spi-1. Spi-1 seems to act as an anchor, which tethers NF-IL6 to the il1b core promoter to exert activation, without NF-IL6 binding to its cognate binding site. This mechanism, which we have called protein-tethered transactivation (PTT) (20,21), may be more widely used in gene activation than is presently appreciated.
Transfections and Luciferase Assays-HeLa S3 cells were plated in 24-well plates 24 h before transfection. A total of 1.8 g plasmid DNA, including 0.5 g of reporter, 0.5 g of each expression vector, and 0.3 g of pCMV⅐Sport-␤-gal (Life Technologies, Inc.), except as noted, was transfected into the cells using DOTAP transfection reagent (Roche Molecular Biochemicals GmbH) as described previously (21). After incubation for 24 h, cells were stimulated with 50 ng/ml PMA (Sigma) for 20 h. The cells were then harvested and lysed in 150 l of cell culture lysis reagent (Promega). The lysates were assayed for luciferase activity using the Promega luciferase assay kit.
Expression and Purification of GST Fusion Protein-Glutathione S-transferase fusion proteins were prepared by standard procedures as described previously (21). Equivalent amounts of GST fusion proteins (as determined by Bio-Rad and confirmed by Coomassie Blue staining) were bound to 50 l of glutathione-Sepharose beads by incubation in a total volume of 500 l of NETN (20 mM Tris chloride, pH 8.0, 100 mM NaCl, 1 mM EDTA, and 0.5% Nonidet P-40) (21) for 1 h at 4°C. The beads were washed three times in NETN buffer.
GST Fusion Protein Interaction Assays-The Sepharose beads bound with GST fusion protein were incubated with 35 S-labeled in vitro translated protein (TNT T7 coupled reticulocyte lysate system, Promega) at 4°C for 1 h. The beads were washed five times in NETN buffer, and the bound proteins were eluted by boiling for 5 min in SDS-PAGE loading buffer (50 mM Tris⅐Cl, pH 6.8, 30% glycerol, 0.4% SDS, and 0.1% bromphenol blue) containing ␤-mercaptoethanol. Proteins were analyzed by SDS-PAGE (15% polyacrylamide gel), followed by soaking in Amplify fluorographic reagent (Amersham Pharmacia Biotech) and exposed to Kodak X-Omat film.
GST 1-Hybrid DNA Binding Assay-A sensitive method modified (21) from a technique previously reported by Chittenden et al. (24) was used as to detect weak protein-DNA interaction. Briefly, equivalent amounts of GST fusion proteins were bound to 50 l of glutathione-Sepharose beads as described above. After washing with NETN buffer three times, the beads were resuspended in 200 l of DNA probe binding buffer (20 mM HEPES, pH 8.0, 1 mM EDTA, 50 mM NaCl, 3 mM MgCl 2 , 33 ng/l poly(dI-dC), 1 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride) and incubated for 5 min at 4°C. The beads were then incubated for 20 min at 4°C with 200,000 cpm of 32 P-labeled wild-type probe corresponding to the il1b promoter sequence between Ϫ56 and Ϫ21 or the same probe with the putative NF-IL6 site mutated. The beads were washed twice with the binding buffer containing 10 ng/ml poly(dI-dC) and 0.5% Nonidet P-40. Specific binding was assessed by Cerenkov counting of the protein-DNA complex on the glutathione-Sepharose beads. Electrophoretic Mobility Shift Assays-Double-stranded oligonucleotides spanning the overlapping Spi-1 and NF-IL6 sites (il1b promoter region from Ϫ56 to Ϫ21) were synthesized and labeled by using DNA polymerase Klenow fragment in the presence of [␣-32 P]dATP and [␣-32 P]dGTP. EMSAs were carried out by incubating 0.5 l in vitro translated Spi-1 or NF-IL6 protein (TNT T7 coupled reticulocyte lysate system, Promega) with 10,000 cpm of the wild-type il1b probe, or the probe carrying mutations in either the Spi-1 site or the NF-IL6 site, under binding conditions of 10 mM Tris⅐Cl, pH 7.5, 50 mM NaCl, 3.3 mM MgCl 2 , 1 mM EDTA, 1 mM dithiothreitol, 5% glycerol with 1 g of poly(dI-dC) in a final volume of 15 l. The binding reactions were performed on ice for 20 min and then subjected to electrophoresis on 4% nondenaturing low ionic strength polyacrylamide gels using 0.5 ϫ TBE buffer (TBE, 45 mM Tris borate, pH 8.3, and 1 mM EDTA). The gels were then dried and analyzed by autoradiography.

NF-IL6 Strongly
Cooperates with Spi-1 to Activate the (Ϫ59/ ϩ12) il1b Core Promoter-We and others (12-14) have previously reported that the Ϫ59/ϩ12 il1b core promoter contains a Spi-1 (Ϫ50/Ϫ39) and an NF-IL6 (Ϫ41/Ϫ33) binding site. To verify that Spi-1 and NF-IL6 could transactivate the il1b core promoter, we performed transient cotransfection assays. A luciferase reporter plasmid containing the il1b core promoter was cotransfected into Spi-1-deficient HeLa S3 cells (25) along with plasmids either expressing Spi-1, NF-IL6, or both (Spi-1ϩNF-IL6). The cells were stimulated with PMA 20 h before luciferase assays to activate the NF-IL6 protein. As shown in Fig. 1,   FIG. 1. Effect of Spi-1and NF-IL6 on the transcription of the wild-type or mutant il1b promoter. The wild type il1b (Ϫ59/ϩ12) core promoter-luciferase reporter plasmid, or reporter constructs containing mutations of the Spi-1 and NF-IL6 sites, were transfected into HeLa S3 cells along with vectors expressing either Spi-1 or NF-IL6 as shown. The luciferase reporter vector pGL3-Basic containing no insertion was also cotransfected with various expression vectors as a control. Luciferase activities were normalized to ␤-galactosidase activities expressed by a cotransfected plasmid. The level of the wild-type reporter construct in the presence of empty expression vectors is set to 1. Data represent the mean and the S.E. of three repetitions.
NF-IL6 alone only had a minimal effect, whereas Spi-1 stimulated activity by about 20-fold. However, NF-IL6 together with Spi-1 activated the il1b core promoter by about 380-fold, suggesting a strong cooperativity between these two factors.
Mutation of the Spi-1 Site, but Not the NF-IL6 Site, Abolishes Cooperative Activation-To determine whether the cooperative activation of the il1b core promoter by Spi-1 and NF-IL6 depends on their binding sites, site-directed mutagenesis was used to introduce multiple nucleotide sequence substitutions into the Spi-1 (mut Spi-1), NF-IL6 (mut NF-IL6), or both (mut both) binding sites, without changing the nucleotides in the overlapping region (Fig. 2). In agreement with previous studies (13), mutation of the Spi-1 binding site completely prevented activation by Spi-1, either alone or with NF-IL6 (Fig. 1). Transfection of NF-IL6 alone did not affect the promoter activity, even when the overlapping Spi-1 site was disrupted. A promoter containing an intact Spi-1 site and either an intact or mutated NF-IL6 site (WT or mut NF-IL6) supported activation by Spi-1 alone. Importantly, a promoter containing an intact Spi-1, but a mutated NF-IL6 site (mut NF-IL6) retained a majority of the synergistic activation in the presence of both Spi-1 and NF-IL6 (Fig. 1). Consequently, the Spi-1 binding site, but not the NF-IL6 binding site, is critical for the cooperative transactivation of this promoter by Spi-1 and NF-IL6.
Using the wild-type il1b probe, containing both the Spi-1 and NF-IL6 sites (Fig. 3A), in vitro translated Spi-1 can bind avidly to the probe in EMSA (lane 2), as suggested by previous reports (13,14). Mutations in the core Spi-1 recognition sequence (AGAA to CTAA) abolished Spi-1 binding (lane 3). Due to the weak equilibrium binding observed by EMSA between NF-IL6 and the il1b probe (Ϫ56/Ϫ21) (data not shown), we employed a GST-1-hybrid assay, capable of detecting complexes with either low affinity or high decay rates (21). Fig. 3B shows that a GST-bZIP fusion protein containing the NF-IL6 bZIP DNAbinding domain could bind to the wild-type probe about six times better than the binding between GST control protein and the probe. As expected, the substitution of 4 nucleotides in the NF-IL6 consensus region prevented the probe from being recognized by NF-IL6.
The NF-IL6 TAD Is Indispensable for Functional Cooperativity, while Spi-1 TADs Only Partly Contribute-NF-IL6 is a 345-amino acid protein with a COOH-terminal basic leucine zipper structure (bZIP) that binds to DNA. The TAD of NF-IL6 has been reported to bind directly to CBP/p300 (26). The extreme amino terminus of NF-IL6 is capable of recruiting SWI/ SNF chromatin-remodeling complex and activating endogenous target genes in concert with the TAD (27). Between the TAD and the bZIP domains reside regulatory domains, which are involved in intramolecular interactions that inhibit transactivation and DNA binding when NF-IL6 is not activated by phosphorylation (28, 29) (Fig. 4A) 3) were incubated with the wild-type il1b probe (between nucleotides Ϫ56 and Ϫ21) in EMSA. Also, in vitro translated Spi-1 was incubated with the same probe containing mutations in the core Spi-1 recognition site (AGAA to CTAA) in EMSA (lane 3). B, the GST-1-hybrid assay was used to detect weak protein-DNA interaction between NF-IL6 bZIP and the il1b probe. GST and GST-bZIP fusion proteins immobilized on glutathione-Sepharose were tested for DNA binding activity using either wild type (wt) il1b probe (between nucleotides Ϫ56 and Ϫ21) or probe containing mutations in the putative NF-IL6 site (mNF-IL6). The results are the relative binding affinity with the cpm of mNF-IL6 probe bound to GST-bZIP set as 1. Error bars represent the S.E. in results from three repetitions.
To determine the domains of Spi-1 and NF-IL6 required for transcriptional cooperativity, we assayed deletion mutations of both proteins in transient transfection assays. First, vectors coding for full-length NF-IL6 and a truncation (⌬Spl) that can bind to DNA, but lacks the TAD and a portion of the regulatory region, were co-expressed with full-length Spi-1 and assayed for il1b core promoter reporter activity in HeLa cells. Western analysis showed that both the full-length and the truncated NF-IL6 (⌬Spl) were correctly expressed by the transfected plasmids in HeLa cells (data not shown). The truncated NF-IL6 could not synergize with Spi-1 (Fig. 4B). Unlike the observations with NF-IL6, the Spi-1 mutant constructs (Fig. 5, A and  B) with deletions of either the Q domain (⌬PN), or the Q domain together with the NH 2 -terminal TBP binding domain (⌬100, ⌬NN) retained significant ability to cooperate with NF-IL6. Deletion of the PEST region increased Spi-1 activation of the il1b core promoter and its ability to synergize with NF-IL6, arguing for its dispensability. We have previously shown that expression of the various Spi-1 derivatives in HeLa cells with these vectors yielded comparable levels of proteins capable of binding specifically to DNA (13).
Dominant Negative NF-IL6 Represses Spi-1-dependent Activation of the il1b Core Promoter in HeLa Cells-Previously we have shown that the ⌬Spl truncated NF-IL6 could antagonize NF-IL6-dependent activation (23) of the il1b upstream inducible enhancer (UIS), by competing with the wild-type endogenous NF-IL6 for a specific NF-IL6 binding site. We have transfected this dominant negative NF-IL6 mutant into HeLa cells in either the presence or absence of Spi-1. As shown in Fig. 6, Spi-1 alone activated the il1b core promoter. Strikingly, co-transfection of the NF-IL6 dominant negative repressed the Spi-1-mediated activation in a dose-dependent fashion, presumably by interfering with endogenous NF-IL6 function. Cotransfection of 0.5 or 1 g of ⌬Spl NF-IL6 vector with a constant amount of Spi-1 expression plasmid (0.25 g) revealed a 20 and 60% decrease of Spi-1-induced activity, respectively. However, the small amount of basal il1b core promoter activity was not affected by the NF-IL6 dominant negative when Spi-1 was not present, even though NF-IL6 is known to be endogenously expressed in HeLa cells (35).
Spi-1 and NF-IL6 Directly Interact through Their DNA Binding Regions, the ETS and bZIP Domains-It has been shown by a number of studies that the wHTH domains of Spi-1 and other ETS factors directly interact with various transcription factors, playing a possible role in functional cooperativity (17,18,30,36). To determine whether Spi-1 and NF-IL6 cooperativity in il1b expression is mediated by protein-protein binding, we performed GST pull-down protein-interaction assays. The fulllength Spi-1 cDNA and truncated sequences were subcloned into plasmid pGEX-2T to produce GST fusion constructs. The 35 S-labeled NF-IL6 prepared by in vitro transcription and translation was incubated with similar amounts of either GST or various GST Spi-1 truncations linked to Sepharose beads. After incubation, the beads were intensively washed, and bound proteins were resolved by SDS-PAGE. As shown in  6). Both of these fusion proteins were capable of weakly binding to NF-IL6. It is possible that regions containing both the ␤2/ ␣2/␣3 and ␤3/␤4 (Fig. 5A) structural elements are important in mediating Spi-1 and NF-IL6 interaction.
To identify sequences within NF-IL6 necessary for interaction with Spi-1, we prepared in vitro translated 35 S-labeled NF-IL6 with amino-terminal deletions and tested for interaction with GST- Spi-1 171-272 (Fig. 7B). Similar to the full- . It is noteworthy that deletion of the amino-terminal 268 amino acids of NF-IL6, leaving only the bZIP DNA binding domain, dramatically increased the binding affinity observed in the GST pull-down assay (lane 6). It was reported that, in the absence of activation, one of two regulatory elements (RD2, Fig. 4) from rat NF-IL6 could inhibit DNA binding by intramolecular interaction, whereas the other element (RD1) similarly inhibited transactivation (28,29). Our data suggest that the RD2 domain, specifically the region from aa 206 to 268, may also prevent NF-IL6 from interacting with its cofactors.
In a reciprocal assay, 35 S-labeled Spi-1 ETS domain (from amino acids 171-272) was prepared by in vitro transcription and translation and incubated with GST-bZIP fusion protein containing the bZIP region of NF-IL6 from amino acids 269 -345. As expected, the labeled Spi-1 wHTH domain bound strongly to GST-bZIP, but not to GST alone (Fig. 7C). DISCUSSION The expression of il1b is regulated by two independent elements, an upstream inducible sequence (the UIS enhancer) and a cell type-specific promoter element (37). Although strong enhancer-dependent activity depends upon a long promoter extending from Ϫ131 to ϩ12 (13), weaker enhancer-independent activity can be detected with the shorter Ϫ59/ϩ12 promoter. Moreover, we have recently shown that the Spi-1 binding site located at position Ϫ50/Ϫ39 is responsible for mediating transactivation of il1b expression by cytomegalovi-rus IE2 protein, which eliminates the need for the otherwise essential upstream enhancer (21).
In this report, we have demonstrated that NF-IL6 dramatically cooperates with Spi-1 to activate the il1b core promoter, where the Spi-1 binding site, but not the putative NF-IL6 site, is critical. Although the Spi-1 recognition site is sufficiently important that mutations leading to a complete loss of Spi-1 binding result in a total loss of promoter activity in the presence of cotransfected Spi-1 and NF-IL6, deletion of the transcription activation domains of Spi-1 results in only a partial loss of its ability to functionally cooperate with NF-IL6. In contrast, the deletion of the NF-IL6 transactivation domain (aa 41-205) completely abolishes its ability to synergize with Spi-1 on the il1b core promoter. Physical interaction between the Spi-1 wHTH and the NF-IL6 bZIP DNA binding domains provides the basis for our model. In this model, the Spi-1 wHTH domain functions to recognize a specific site in the il1b core promoter and tether NF-IL6, which contains an efficient transcription activation domain, which, unlike those of Spi-1, is able to strongly activate il1b expression (Fig. 8). This is distinct from an earlier report of NF-IL6 cooperativity with glucocorticoid receptor, in which the NF-IL6 played a TAD-independent and indirect role (38). It should be noted that we could not detect a reproducible ternary complex (data not shown), involving NF-IL6, Spi-1, and DNA using either EMSA or a more sensitive GST-based two-hybrid approach (21), suggesting a tenuous interaction. Others have also attempted to detect ternary complexes involving NF-IL6 and have failed (38,39), supporting this conclusion.
Our model is supported by two facts. First, cotransfection of NF-IL6 expression vector significantly increased the ability of both Spi-1⌬100 (lacking both TBP and Q TADs) and Spi-1⌬PEST (lacking the PEST region) to activate the il1b core promoter (Fig. 5). Second, although the dominant negative NF-IL6 was not able to repress the Spi-1-independent activity of the il1b core promoter in NF-IL6-expressing HeLa cells (22), it antagonized the Spi-1-mediated activation of the same promoter in a dose-dependent manner. This is consistent with the notion that Spi-1 activation of the il1b core promoter is mediated by NF-IL6 and that the putative overlapping NF-IL6 site is not functional. GST pull-down assays demonstrate that both NF-IL6⌬Spl and the full-length protein interact with Spi-1 at a similar level (Fig. 7), implying that NF-IL6⌬Spl may compete with endogenous NF-IL6 for interaction with Spi-1. It has been reported that Spi-1 without the TBP and Q TADs can activate transcription by playing an architectural role in interaction with NF-EM5/Pip/IRF-4 mediated by the PEST region (40).  -1 and NF-IL6. Spi-1 mainly functions as a DNA-binding protein, tethering NF-IL6 to the proximity of the transcriptional initiation machinery. Although the Spi-1 TADs contribute to the cooperativity, they are not as critical as that of NF-IL6. Also shown is a speculated role for Spi-1 as both a DNA-and protein-binding factor, which may integrate the il1b far-upstream inducible enhancer (UIS) containing two NF-IL6 binding sites to the core promoter through NF-IL6 interaction.
Recently, we have shown that functional cooperativity between Spi-1 and the cytomegalovirus IE2 transcription factor does not require any Spi-1 TADs (including PEST) for activity (21). Our data now provide a new example in which the Spi-1 wHTH functions both to bind DNA and to tether a nonviral transcription factor containing a more potent TAD.
The mechanism by which the il1b UIS is integrated into the core promoter has always been a puzzle. We have reported that the UIS sequence between Ϫ3134 and Ϫ2729 contains two NF-IL6 binding sites (11,23). Also we have shown that the Ϫ131/Ϫ59, which contains an additional Spi-1 binding site, is critical for enhancer activity (13). Our results now suggest the possibility that factors bound to the UIS, including LIL-Stat, CREB, and NF-IL6 (23), may be tethered to the proximity of the transcriptional initiation machinery through NF-IL6-Spi-1 interactions (Fig. 8). However, carefully designed experiments are needed to confirm this speculation.
In this report, we have shown that NF-IL6, which is abundant in myeloid cells (41), strongly synergizes with Spi-1 on the il1b core promoter via PTT (21) in transient transfection assays using Spi-1-deficient HeLa cells. This suggests that PTT also functions in IL-1␤-producing monocyte/macrophage cells. However, other bZIP family factors, such as NF-IL6␤ (C/EBP␦) (17), c-Jun (20), as well as other cellular or viral transcription factors (21,42,43), have been shown to physically interact with Spi-1. It would be interesting to determine whether other factors, which are also expressed at various levels over the course of myeloid differentiation, particularly c-Jun (44 -46), C/EBP␦ (17), C/EBP␣ (41), and C/EBP⑀ (47), can also synergize with Spi-1 on the il1b core promoter. Consistent with this prediction, we have recently demonstrated that the viral protein HCMV IE2 strongly activates the il1b promoter via PTT using a similar Spi-1 tether (21).