The Small RNA Gene Activator Protein, SphI Postoctamer Homology-binding Factor/Selenocysteine tRNA Gene Transcription Activating Factor, Stimulates Transcription of the Human Interferon Regulatory Factor-3 Gene*

Many small nuclear RNA gene promoters are activated by SphI postoctamer homology (SPH)-binding factor/selenocysteine tRNA gene transcription activating factor (SBF/Staf). Whereas this transcription factor was initially identified by its ability to bind to SPH elements in such promoters, it was more recently shown to have the capacity to activate transcription of a synthetic mRNA gene promoter through a distinct activation domain. Here, we show that the human interferon regulatory factor-3 (IRF-3) gene promoter contains a functional SPH element that is bound by SBF/Staf in vitro and in transfected cells.

Eukaryotic transcriptional activator proteins are usually multifunctional. For example, activator proteins that bind a specific DNA sequence are organized in a modular fashion, with a DNA-binding domain (DBD) 1 and one or more activation domains (1). In addition, activator proteins are likely to be used in the transcription of more than one gene, each of which contains a similar DNA-binding site in the promoter or enhancer. Furthermore, a single activation domain can target multiple general transcription factors or coactivator complexes while effecting transcriptional stimulation (2,3).
Small nuclear RNA (snRNA)-type promoters appear to be relatively simple models in which to investigate transcriptional activation (recently reviewed in Ref. 4). SnRNA genes are transcribed by either RNA polymerase II (pol II) (e.g. U1 and U2) or RNA polymerase III (pol III) (e.g. U6) (5,6). In addition, a number of other small RNA genes, such as those encoding 7SK, MRP, and selenocysteine tRNA, contain snRNA-type pol III promoters (7)(8)(9). All vertebrate snRNA-type promoters are organized into a proximal region, containing a proximal sequence element (PSE), and a distal, enhancer-like region. For pol III snRNA-type promoters a TATA box is juxtaposed between the PSE and transcription start site, and serves as a specificity element for pol III recognition of the promoter (10,11). The PSE is bound by a multisubunit complex known in human cells as SNAP C or PTF (12,13). The distal region is almost universally composed of an octamer motif (OCT) and, often, a closely spaced SPH element (14 -17). OCT is bound by the Oct-1 activator, and part of the mechanism for Oct-1 activation is accounted by direct contact between its DNA-binding POU domain and the SNAP-190 subunit that facilitates ternary complex formation with DNA (18). Furthermore, this proteinprotein contact is mediated by a positioned nucleosome located between the distal region and PSE on the human U6 promoter (19,20).
The snRNA SPH motif was first recognized in chicken U1, U2, and U4 distal regions (21,22), but it is now apparent that the SPH motif is present in many vertebrate snRNA-type promoters (17). For the human U6 snRNA gene promoter, the OCT and SPH motifs contribute approximately equivalent stimulatory effects (23). The cDNA for SPH-binding factor has been cloned from Xenopus, mouse, and humans, and is called SBF, Staf (selenocysteine tRNA gene transcription activating factor), or ZNF143 (24 -27) (referred to as SBF/ Staf here). SBF/Staf contains a 7-zinc finger DBD and an apparently unique 15-amino acid repeat that is present in four copies in the amino-terminal region (24). For Xenopus Staf, the snRNA activation region, functional for both pol II and pol III snRNA-like promoters, has been localized to a segment of 18 amino acids positioned in the primary structure between the 15-amino acid repeats and the DBD (28). Unexpectedly, SBF/Staf also stimulated transcription from an mRNA-type pol II promoter. The mRNA activation activity was demonstrated using a synthetic tk/CAT promoter and localized to the 15-amino acid repeat region of SBF/Staf (28). Hence, SBF/Staf has the unusual property of distinct activation domains that target either snRNA-type or mRNA-type promoters, and therefore illustrates another example of multifunctionality among activators. However, a genuine mRNA promoter activated by SBF/Staf was not identified initially. In the course of this work, a report was published that described the presence of SPH elements in the promoter of the mouse cytosolic chaperonin containing t-complex polypeptide 1␣ subunit gene, and demonstrated transcriptional stimulation by SBF/Staf (29). Here we show a second example of a bona fide mRNA promoter that employs SBF/Staf, the human interferon regulatory factor-3 (IRF-3) gene.
Transfections, Primer Extension, and Luciferase Assays-Human 293 cells were transfected in 100-mm dishes with 10 g of U6/CFREE reporter plasmid plus 10 g of pGEM/c␤3 using the calcium phosphate method, and total RNA was isolated and analyzed by primer extension as described previously, except that a longer primer was used to detect ␤-tubulin transcripts (5Ј-AGGTGCACGATCTCCCTCATG-3Ј) (34). Relative band intensities were quantitated using a STORM PhosphorImager (Molecular Dynamics). Transfection experiments shown in Fig. 4 used 60-mm dishes of human 293 cells containing 1.0 ϫ 10 6 cells seeded 24 h prior to transfection. One g of pIRF3 reporter DNA and 0.5 g of pSV-␤gal (Promega) were introduced using FuGENE 6 reagent (Roche Molecular Biochemicals) following the manufacturers protocol. After 48 h, cells were harvested by scraping, washed with phosphate-buffered saline, and total lysates prepared in 0.1 M potassium phosphate (pH 7.8), 0.2% Triton X-100. Luciferase and ␤-galactosidase (Galacto-Light Plus assay system (Tropix)) activities of cell lysates were measured with a microplate luminometer (Packard Lumicount). Transfection experiments shown in Fig. 5 used 24-well microtiter plates containing 1.0 ϫ 10 5 human 293 cells per well seeded 24 h prior to transfection. Using the calcium phosphate protocol, each sample was transfected with 100 ng of pIRF3 reporter plasmid, 50 ng of pSV-␤gal, and 1 ng of pCI/GAL4/ SBF expression plasmid. Cells were harvested after 36 h, and lysates assayed for luciferase and ␤-galactosidase.
Chromatin Immunoprecipitation-Stably transfected 293 cells containing either pIRF3(Ϫ394)(wt) or pIRF3(Ϫ394)/SPHMUT were prepared using the calcium phosphate protocol with a combination of the appropriate IRF-3 plasmid plus pCI-neo (Promega) as a neomycin resistance marker. After selection in Dulbecco's modified Eagle's medium (Invitrogen), 10% fetal bovine serum (HyClone), 0.4 mg/ml G418 sulfate (Invitrogen), individual clones were isolated and maintained in the same medium. 100-mm dishes of cells were treated with formaldehyde, harvested, lysed, and sonicated as described by the Farnham laboratory (35). Per 50 l of sonicated chromatin were added 50 l of 2% Triton X-100, 200 mM NaCl, and 20 l of Ultralink-Protein A beads (Pierce), with incubation at 4°C for 4 h. After pelleting the beads, the superna-tant was immunoselected with the addition of 20 g of sonicated, salmon sperm DNA (Sigma), 2 g of affinity-selected anti-SBF antibodies (27), and 20 l of Ultralink-Protein A beads, followed by incubation at 4°C overnight. Beads were washed and immunoprecipitates eluted as described in Ref. 36. Cross-links were reversed in the immunoprecipitated samples by incubation at 65°C for 6 h, and DNA samples were purified using the Wizard PCR Preps System (Promega). IRF-3 promoter DNA from exogenous genes was detected by PCR, using the IRF-3-206 oligonucleotide and GLprimer2 (Promega), an oligonucleotide that is complementary to the luciferase coding region. Endogenous U6-1 genomic sequences were detected using two primer sets. For the near upstream region (Ϫ307 to Ϫ190), HU6-307 (5Ј-GCAAAACGCAC-CACGTGACG-3Ј) and CHU6-190 (5Ј-CTCTCTAACAGCCTTGTATC-3Ј) were employed. For the far upstream region (Ϫ1880 to Ϫ1760), HU61-1880 (5Ј-ACCTCTGGCTCAAAGTAACACATG-3Ј) and CHU61-1760 (5Ј-TTAGAACAGGTAGGGCTGGTGGTT-3Ј) were used for PCR amplification. The first round of amplication used 20 cycles, and was followed with a second round of 10 cycles in the presence of 5 Ci of [ 32 P]dGTP. The PCR signal was not saturated using these conditions. PCR products were electrophoresed on a 20% nondenaturing polyacrylamide gel.

Identification of SPH and OCT Elements in Human IRF-3
Gene Promoter-The 5Ј-flanking sequence of the human IRF-3 gene contains a putative SPH element centered at position Ϫ163 and a consensus octamer motif centered at position Ϫ181 (Fig. 1). Close linkage of OCT and SPH motifs is a hallmark of the enhancer-like regions of many vertebrate snRNA gene promoters, whether they are transcribed by RNA polymerase II or III (16,17). The presence of snRNA gene SPH elements in mRNA gene promoters has been described recently in only a single example, the mouse cytosolic chaperonin containing tcomplex subunit a gene (29). To examine whether the IRF-3 SPH and OCT sites could be bound by factors in vitro, electrophoretic mobility gel shift, and DNase I footprinting assays were performed using a radiolabeled DNA fragment containing the sequence from Ϫ206 to Ϫ129. Several complexes were formed on the IRF-3 fragment using a HeLa cell S100 extract ( Fig. 2A). The major complexes were composed of octamerbinding factor, presumably Oct-1, and SBF/Staf, since they were specifically competed by the addition of excess unlabeled consensus octamer oligonucleotide ( Fig. 2A, lanes 2-4) or human U6 SPH oligonucleotide ( Fig. 2A, lanes 5-7). Recombinant SBF/Staf expressed in Escherichia coli bound the same IRF-3 promoter fragment, and binding specificity was demonstrated by competition with U6 SPH, but not OCT, oligonucleotide (Fig.  2B). Furthermore, sequence-specific binding to the IRF-3 gene promoter by either recombinant Oct-1 POU domain, or recombinant SBF/Staf-(76 -626) was shown by DNase I footprinting (Fig. 2C). The downstream portion of the IRF-3 SPH element was not protected in the footprint (Fig. 2C, lanes 4 and 5), similar to the footprint of recombinant SBF/Staf on the human U6 SPH element (27). This lack of protection may reflect the poor fit of the 3Ј-end of the IRF-3 SPH element to the consensus (Fig. 1), and is consistent with the documented flexibility of DNA binding by Staf zinc fingers, notably zinc finger 1 (37).

FIG. 1. Presence of SPH and OCT elements in human IRF-3 gene promoter region. DNA sequences in the human IRF-3 and U6
snRNA 5Ј-flanking regions are compared with a consensus SPH element determined by binding site selection with the Xenopus Staf protein (37). The uppercase type for the SPH consensus represents nucleotide bases that were present in Ͼ70% of the selected fragments, whereas the positions shown in lowercase were selected at lower frequencies. The underlined sequence highlights a consensus octamer motif. The nucleotide at position Ϫ155 of the IRF-3 promoter (denoted in bold) was reported as a T in the original sequence (30), but is a G in all plasmid constructs used in this work.
Enhancer Activity of the IRF-3 SPH ϩ OCT Region for the Human U6 Promoter-After demonstrating that the IRF-3 SPH and OCT sites were potential targets for factor binding in vitro, their capacity for transcriptional activation was tested. Initially, the IRF-3 SPH ϩ OCT segment (Ϫ206/Ϫ129) was fused to the basal human U6 snRNA gene promoter (dl-84/U6/ CFREE) in both orientations, and tested for activity in transfected human 293 cells. In either orientation the IRF-3 SPH ϩ OCT segment stimulated snRNA promoter transcription approximately 6 -9-fold (Fig. 3, lanes 2 and 3). This stimulatory activity was comparable to that exerted by the U6 SPH motif (Fig. 3, lane 4), but ϳ20% of that provided by the U6 SPH ϩ OCT distal region (Fig. 3, lane 5). Previously we have shown that the U6 SPH and OCT elements were approximately equally effective in a such a transfection assay, and their combination was only slightly greater than additive (23).
The SPH Element Is Functional in the IRF-3 Promoter-Next, the roles of the SPH and OCT elements were tested in the IRF-3 promoter. Clustered point mutations were introduced to disrupt each element separately (SPHMUT or OCTMUT), and together in the same promoter (SPHMUT/OCTMUT). IRF-3 Approximately 3 fmol of radiolabeled IRF-3 DNA fragment containing the sequence from Ϫ206 to Ϫ129 were incubated with HeLa S100 extract and electrophoresed on a 4% native, polyacrylamide gel. Complexes delineated as "OCT" or "SBF/Staf" were identified by competition with excess double-stranded oligonucleotides added with the radiolabeled probe in the following amounts: lanes 2-4, 30-, 300-and 3000-fold molar excess, respectively, of consensus octamer motif from the human U6 distal region (OCTCON); lanes 5-7, 30-, 300-, and 3000-fold molar excess, respectively, of human U6 SPH motif. Human 293 cells were transfected by the calcium phosphate technique with plasmid DNAs containing the basal human U6 promoter (dl-84) or various constructs in which the IRF-3 (SPH ϩ OCT) region (IRF3FORW or IRF3REV), or human U6 SPH or U6 (SPH ϩ OCT) elements were ligated to the basal promoter. Expression from exogenous promoters was detected by primer extension. "CFREE" represents transcription from the U6 promoter, and "c␤3" represents transcription from a co-transfected chicken ␤-tubulin plasmid used as a control to normalize for variable transfection efficiency. The fold-activation noted at the bottom was determined after quantitation of band intensities by phosphorimaging. After background subtraction, the CFREE/c␤3 ratio was calculated and compared with the value from the dl-84 (basal) promoter (lane 1). promoter activities were determined by quantitation of luciferase reporter gene levels in transfected human 293 cells. Disruption of the SPH element resulted in a decrease of promoter activity to ϳ25% of wt (Fig. 4, SPHMUT). In contrast, mutation of OCT caused a small, but reproducible increase in activity (Fig. 4, OCTMUT). When both elements were mutated, IRF-3 promoter activity was reduced to the same extent as for the SPHMUT construct. The data shown in Fig. 4 are from transfection experiments using 1 g of luciferase reporter plasmid (in 60-mm dishes), but similar results were obtained when either 0.5 or 2 g of reporter plasmid were employed (data not shown). Furthermore, IRF-3/SPHMUT promoter activity was similarly reduced to ϳ25% of wt in transfected HeLa cells (data not shown).
Binding of SBF/Staf to the IRF-3 Promoter in Transfected Cells-Reduced transcription from the SPHMUT promoter demonstrated the importance of the DNA sequence from Ϫ170 to Ϫ161, but did not prove a role for SBF/Staf in the IRF-3 promoter in vivo. In order to examine the activation of this promoter by SBF/Staf, cells were co-transfected with expression plasmids encoding GAL4 DBD/SBF fusion proteins and an IRF-3 promoter/luciferase reporter plasmid containing a GAL4-binding site. The SPHMUT/IRF-3 reporter plasmid had been designed to convert the SPH element to a GAL4 element. Two GAL4 DBD/SBF fusion plasmids were constructed. One, pCI/GAL4/SBF- (1-140), contained the amino-terminal se-quence of SBF/Staf, a region that includes 4 imperfect 15amino acid repeats previously shown to activate a synthetic thymidine kinase reporter gene in injected Xenopus oocytes or transfected Drosophila cells (28). A second fusion protein, expressed from pCI/GAL4/SBF-(136 -223), contained a region of SBF/Staf reported to activate snRNA gene promoters (28). 2 Expression of GAL4/SBF-(1-140) stimulated transcription from the SPHMUT/IRF-3 promoter ϳ2-fold, whereas the GAL4/SBF-(136 -223) fusion protein had no effect (Fig. 5). Furthermore, expression of the GAL4 DBD-(1-94) alone did not significantly stimulate expression from the SPHMUT promoter (results not shown). In these experiments the effect of the SPH mutation was somewhat more severe than shown previously (13% of wt in Fig. 5 versus 27% of wt in Fig. 4), possibly because smaller numbers of cells and plasmid DNAs were used for the transfection assays. Both GAL4 fusion proteins and the GAL4 DBD protein were expressed at similar levels in transfected 293 cells as determined on Western blots using antibody directed against the GAL4 DBD (data not shown). The GAL4/ SBF-(1-140) fusion protein was unable to reconstitute full, wt IRF-3 activity to the SPHMUT promoter, but the stimulatory activity was significant compared with the control GAL4/SBF-(136 -223) and GAL4 DBD proteins.
To further investigate the binding of SBF/Staf to the IRF-3 promoter, chromatin immunoprecipitation was employed with 2 G. R. Kunkel, unpublished results.

FIG. 4. Disruption of the SPH element in the IRF-3 promoter reduces expression in transfected cells.
Human 293 cells in 60-mm dishes were transfected using FuGENE 6 reagent (Roche Molecular Biochemicals) with 1 g of various luciferase expression constructs containing the wt human IRF-3 promoter, or mutants that were disrupted singly in the SPH element (SPHMUT) or OCT element (OCT-MUT), or both elements disrupted in the same promoter (SPHMUT/ OCTMUT). Cells in each dish were co-transfected with 0.5 g of pSV-␤gal plasmid DNA to control for variable transfection efficiency. After background subtraction, luciferase/␤Ϫgalactosidase values were compared with the wt promoter. The height of each bar represents the average value from three (SPHMUT, OCTMUT) or two (SPHMUT/ OCTMUT) separate transfections, and the height from the midpoint of the error bar shows one standard deviation from the mean.

FIG. 5. Expression of the amino-terminal human SBF/Staf mRNA activation domain can enhance transcription from the IRF-3/SPHMUT promoter in transfected cells. Human 293 cells in
24-well microtiter plates were transfected by the calcium phosphate protocol with a pIRF3(Ϫ394) or pIRF3/SPHMUT reporter plasmid DNA, a pCI/GAL4/SBF expression plasmid DNA, and pSV-␤gal plasmid DNA. 36 h post-transfection, cell lysates were prepared, and luciferase and ␤-galactosidase activities were determined using chemiluminescent assays. After background subtraction, luciferase/␤-galactosidase ratios were compared with the wt promoter (pIRF3(Ϫ394)). The height of each bar represents the average value from four separate transfections, and the height from the midpoint of the error bar shows one standard deviation from the mean. stably transfected cell lines containing either the wt IRF-3 gene promoter or the SPHMUT promoter. Individual clones were selected for resistance to G418 due to co-transfection with pCI-neo plasmid DNA, and two that showed relatively high luciferase activities were investigated further. Luciferase expression from the SPHMUT cell line was ϳ9% of that from the wt IRF-3 promoter cells (data not shown). Furthermore, using genomic DNA and quantitative PCR conditions, it was estimated that an approximately equal number of copies of exogenous wt or SPHMUT promoter DNAs were incorporated into the stably transformed cell lines (data not shown). Sheared chromatin from formaldehyde-treated cells was immunoselected with affinity purified anti-SBF/Staf antibodies or preimmune antiserum. IRF-3 promoter sequence was detected at significantly higher levels by PCR in the DNA from anti-SBF/ Staf immunoselected chromatin compared with that selected by the preimmune antibodies (Fig. 6A, compare lanes 2 and 3), using primers spanning the region from Ϫ206 (IRF-3-206 primer) to just downstream of the luciferase translation start site (GLprimer2; ϩ97 from the putative IRF-3 transcription start site). In contrast, the PCR signals from anti-SBF/Staf and preimmune antibodies were the same background level from SPHMUT chromatin samples (Fig. 6A, compare lanes 5 and 6). The average of three independent PCR assays resulted in a 9.8-fold higher band intensity for wt versus SPHMUT anti-SBF/Staf selected DNA when normalized to the signals from total, unselected DNA (i.e. Fig. 6A, compare lanes 2 and 5). As a control, the selection of human U6-1 gene promoter DNA by the anti-SBF/Staf antibodies was examined. Primers spanning the SPH element (Ϫ307 to Ϫ190) amplified DNA in the anti-SBF/Staf immunoselected chromatin, but not after preimmune antibody selection (Fig. 6B, compare lanes 4 and 5). However, no DNA from a far upstream region of the human U6-1 gene (Ϫ1880 to Ϫ1760) was selected by either anti-SBF/Staf or preimmune antibodies (Fig. 6B, lanes 9 and 10). Therefore, similar to the endogenous U6-1 gene, SBF/Staf is bound to the IRF-3 gene promoter in transfected cells, and SBF/Staf promoter occupancy is correlated with increased transcription activity of the wt compared with SPHMUT templates in transfected cells. DISCUSSION In this report, we demonstrate that the transcriptional activator, SBF/Staf, first characterized for its role at snRNA gene promoters, is an important stimulatory factor for the human IRF-3 gene promoter. The SPH element at Ϫ160 is bound by recombinant SBF/Staf or the endogenous protein in HeLa cell extracts (Fig. 2). Disruption of the SPH element causes decreased promoter activity that is partially restored by co-expression of the SBF/Staf mRNA activation domain which is directed to the mutant promoter (Fig. 5). Furthermore, chromatin immunoprecipitation results show that SBF/Staf binds to the wt IRF-3 promoter in transfected cells (Fig. 6).
The potential for SBF/Staf activation at mRNA gene promoters was demonstrated first using a synthetic tk/CAT promoter containing multiple activator-binding sites (28). With this report, at least two bona fide mammalian mRNA promoters have been shown to contain functional SPH elements that are binding sites for this activator. Recently, the mouse chaperonin containing t-complex polypeptide 1 ␣-subunit gene promoter was found to contain two SPH elements at positions Ϫ70 and Ϫ20 (29). At present, the limited dataset of mRNA promoters containing SPH elements makes it impossible to determine whether correlations exist between other features of these promoters. Whereas the mouse chaperonin containing t-complex polypeptide 1 ␣-subunit promoter is apparently TATA-less, the tk promoter contains a documented TATA box (38). The human IRF-3 promoter does not contain a canonical TATA box at the normal location, but it is not known whether a functional TATA element is present (30). Both the aforementioned synthetic tk/CAT promoter and mouse chaperonin containing t-complex polypeptide 1 ␣-subunit promoter contain multiple SBF/Stafbinding sites. We found only one SPH site in the Ϫ206/Ϫ129 region of the human IRF-3 promoter, but cannot rule out the possibility of other sites outside this region. However, our chromatin immunoprecipitation results with the SPHMUT promoter do not indicate any strong SBF/Staf-binding sites in a more proximal location of this promoter.
It is not known what other transcription factors operate at the human IRF-3 promoter. We noticed a consensus octamer motif just upstream of the SPH motif at a position of approximately Ϫ180 (Fig. 1), and showed that this site was bound by Oct protein or recombinant Oct-1 POU domain in vitro (Fig. 2,  A and C). However, in contrast to the SPHMUT mutation, disruption of the octamer motif (OCTMUT) resulted in only a minimal effect on the IRF-3 promoter (Fig. 4). In fact, expression from the OCTMUT promoter was reproducibly higher, suggesting that a negative regulatory element was disrupted. Presently, it is not known whether Oct protein binds to this site in vivo, although this could be examined using chromatin immunoprecipitation. Indeed, a possible USF site overlaps the consensus octamer element, and could occlude Oct binding (30). Other possible transcription factor-binding sites have been identified within the proximal 210 bp of the IRF-3 promoter, FIG. 6. SBF/Staf interacts with wt but not SPHMUT IRF-3 promoter in stably transfected cells. A, sheared chromatin from stably transfected pIRF-3(Ϫ394) or pIRF-3/SPHMUT human 293 cells that had been cross-linked with formaldehyde was immunoselected with anti-SBF antibodies (lanes 2 and 5) or preimmune antibodies (lanes 3 and 6), and DNA was purified following reversal of cross-links. The relative amounts of IRF-3 promoter sequence in the unselected or antibody-selected DNA samples were determined by electrophoresis of radiolabeled PCR products on a 20% gel. The "total" lanes (1 and 4) show PCR of approximately double the amount of sample used for antibody selection. B, binding of SBF/Staf to the endogenous human U6-1 gene was investigated by chromatin immunoprecipitation using primer sets to amplify either the near upstream region (Ϫ307 to Ϫ190), encompassing the SPH element (lanes 2-5), or a far upstream region (Ϫ1880 to Ϫ1760) as a control (lanes 7-10). Lanes marked "Buffer" contained no added DNA, and "5% Total" contained unselected DNA at 5% of the amount of immunoselected sample assayed by PCR. but none have been verified experimentally (30).
A previous report did not identify a difference in IRF-3 promoter activity when the region containing the SPH element (Ϫ195 to Ϫ127) was deleted (30). It is not clear why our results are apparently in contradiction. However, it is possible that this region contains both positive (e.g. SPH) and negative regulatory elements that counteract each other. Alternatively, the previous investigators used a different cell line for their transfection experiments (2FTGH cells), that could express a different repertoire of factors that bind to the IRF-3 promoter. We found similar results with both human 293 and HeLa cells.
In co-transfection experiments, expression of a GAL4 DBD/ SBF-(1-140) fusion protein partially restored IRF-3 promoter activity to the SPHMUT template that contained a single GAL4-binding site (Fig. 5). However, even when larger amounts of this expression plasmid were added, wt promoter activity was not attained (results not shown). Possibly, to regain wt activity, multiple GAL4-binding sites would be necessary in this synthetic system. Alternatively, a larger segment of SBF/Staf containing multiple activation domains might be necessary for efficient IRF-3 promoter transcription complex assembly. Nevertheless, our results indicate that the amino-terminal portion of SBF/Staf, containing the 15-amino acid repeats, can function as an activation domain for the IRF-3 gene promoter, and corroborate previous results using Xenopus Staf and the synthetic tk/CAT promoter (28). Future work will delineate the mechanisms by which SBF/Staf uses two separate domains to activate mRNA versus snRNA gene promoters.